avrdude/stk500v2.c

4724 lines
136 KiB
C
Raw Permalink Normal View History

Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* avrdude - A Downloader/Uploader for AVR device programmers
* Copyright (C) 2005 Erik Walthinsen
* Copyright (C) 2002-2004 Brian S. Dean <bsd@bsdhome.com>
* Copyright (C) 2006 David Moore
* Copyright (C) 2006,2007,2010 Joerg Wunsch <j@uriah.heep.sax.de>
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
*/
/* $Id$ */
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/* Based on Id: stk500.c,v 1.46 2004/12/22 01:52:45 bdean Exp */
/*
* avrdude interface for Atmel STK500V2 programmer
*
* As the AVRISP mkII device is basically an STK500v2 one that can
* only talk across USB, and that misses any kind of framing protocol,
* this is handled here as well.
*
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
* Note: most commands use the "universal command" feature of the
* programmer in a "pass through" mode, exceptions are "program
* enable", "paged read", and "paged write".
*
*/
#include "ac_cfg.h"
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
#include <math.h>
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <errno.h>
#include <limits.h>
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#include <unistd.h>
#include <sys/time.h>
#include <time.h>
#include "avrdude.h"
#include "libavrdude.h"
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#include "stk500_private.h" // temp until all code converted
#include "stk500v2.h"
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#include "stk500v2_private.h"
#include "usbdevs.h"
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* We need to import enough from the JTAG ICE mkII definitions to be
* able to talk to the ICE, query some parameters etc. The macro
* JTAGMKII_PRIVATE_EXPORTED limits the amount of definitions that
* jtagmkII_private.h will export, so to avoid conflicts with those
* names that are identical to the STK500v2 ones.
*/
#include "jtagmkII.h" // public interfaces from jtagmkII.c
#define JTAGMKII_PRIVATE_EXPORTED
#include "jtagmkII_private.h"
#include "jtag3.h" // public interfaces from jtagmkII.c
#define JTAG3_PRIVATE_EXPORTED
#include "jtag3_private.h"
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#define STK500V2_XTAL 7372800U
// Timeout (in seconds) for waiting for serial response
#define SERIAL_TIMEOUT 2
// Retry count
#define RETRIES 5
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#if 0
#define DEBUG(...) avrdude_message(MSG_INFO, __VA_ARGS__)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#else
#define DEBUG(...)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#endif
#if 0
#define DEBUGRECV(...) avrdude_message(MSG_INFO, __VA_ARGS__)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#else
#define DEBUGRECV(...)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#endif
enum hvmode
{
PPMODE, HVSPMODE
};
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
#define PDATA(pgm) ((struct pdata *)(pgm->cookie))
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
/*
* Data structure for displaying STK600 routing and socket cards.
*/
struct carddata
{
int id;
const char *name;
};
static const char *pgmname[] =
{
"unknown",
"STK500",
"AVRISP",
"AVRISP mkII",
"JTAG ICE mkII",
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
"STK600",
"JTAGICE3",
};
struct jtagispentry
{
unsigned char cmd;
unsigned short size;
#define SZ_READ_FLASH_EE USHRT_MAX
#define SZ_SPI_MULTI (USHRT_MAX - 1)
};
static const struct jtagispentry jtagispcmds[] = {
/* generic */
{ CMD_SET_PARAMETER, 2 },
{ CMD_GET_PARAMETER, 3 },
{ CMD_OSCCAL, 2 },
{ CMD_LOAD_ADDRESS, 2 },
/* ISP mode */
{ CMD_ENTER_PROGMODE_ISP, 2 },
{ CMD_LEAVE_PROGMODE_ISP, 2 },
{ CMD_CHIP_ERASE_ISP, 2 },
{ CMD_PROGRAM_FLASH_ISP, 2 },
{ CMD_READ_FLASH_ISP, SZ_READ_FLASH_EE },
{ CMD_PROGRAM_EEPROM_ISP, 2 },
{ CMD_READ_EEPROM_ISP, SZ_READ_FLASH_EE },
{ CMD_PROGRAM_FUSE_ISP, 3 },
{ CMD_READ_FUSE_ISP, 4 },
{ CMD_PROGRAM_LOCK_ISP, 3 },
{ CMD_READ_LOCK_ISP, 4 },
{ CMD_READ_SIGNATURE_ISP, 4 },
{ CMD_READ_OSCCAL_ISP, 4 },
{ CMD_SPI_MULTI, SZ_SPI_MULTI },
/* all HV modes */
{ CMD_SET_CONTROL_STACK, 2 },
/* HVSP mode */
{ CMD_ENTER_PROGMODE_HVSP, 2 },
{ CMD_LEAVE_PROGMODE_HVSP, 2 },
{ CMD_CHIP_ERASE_HVSP, 2 },
{ CMD_PROGRAM_FLASH_HVSP, 2 },
{ CMD_READ_FLASH_HVSP, SZ_READ_FLASH_EE },
{ CMD_PROGRAM_EEPROM_HVSP, 2 },
{ CMD_READ_EEPROM_HVSP, SZ_READ_FLASH_EE },
{ CMD_PROGRAM_FUSE_HVSP, 2 },
{ CMD_READ_FUSE_HVSP, 3 },
{ CMD_PROGRAM_LOCK_HVSP, 2 },
{ CMD_READ_LOCK_HVSP, 3 },
{ CMD_READ_SIGNATURE_HVSP, 3 },
{ CMD_READ_OSCCAL_HVSP, 3 },
/* PP mode */
{ CMD_ENTER_PROGMODE_PP, 2 },
{ CMD_LEAVE_PROGMODE_PP, 2 },
{ CMD_CHIP_ERASE_PP, 2 },
{ CMD_PROGRAM_FLASH_PP, 2 },
{ CMD_READ_FLASH_PP, SZ_READ_FLASH_EE },
{ CMD_PROGRAM_EEPROM_PP, 2 },
{ CMD_READ_EEPROM_PP, SZ_READ_FLASH_EE },
{ CMD_PROGRAM_FUSE_PP, 2 },
{ CMD_READ_FUSE_PP, 3 },
{ CMD_PROGRAM_LOCK_PP, 2 },
{ CMD_READ_LOCK_PP, 3 },
{ CMD_READ_SIGNATURE_PP, 3 },
{ CMD_READ_OSCCAL_PP, 3 },
};
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* From XML file:
<REVISION>
<RC_ID_MAJOR>0</RC_ID_MAJOR>
<RC_ID_MINOR>56</RC_ID_MINOR>
<EC_ID_MAJOR>0</EC_ID_MAJOR>
<EC_ID_MINOR>1</EC_ID_MINOR>
</REVISION>
*/
/*
* These two tables can be semi-automatically updated from
* targetboards.xml using tools/get-stk600-cards.xsl.
*/
static const struct carddata routing_cards[] =
{
{ 0x01, "STK600-RC020T-1" },
{ 0x03, "STK600-RC028T-3" },
{ 0x05, "STK600-RC040M-5" },
{ 0x08, "STK600-RC020T-8" },
{ 0x0A, "STK600-RC040M-4" },
{ 0x0C, "STK600-RC008T-2" },
{ 0x0D, "STK600-RC028M-6" },
{ 0x10, "STK600-RC064M-10" },
{ 0x11, "STK600-RC100M-11" },
{ 0x13, "STK600-RC100X-13" },
{ 0x15, "STK600-RC044X-15" },
{ 0x18, "STK600-RC100M-18" },
{ 0x19, "STK600-RCPWM-19" },
{ 0x1A, "STK600-RC064X-14" },
{ 0x1B, "STK600-RC032U-20" },
{ 0x1C, "STK600-RC014T-12" },
{ 0x1E, "STK600-RC064U-17" },
{ 0x1F, "STK600-RCuC3B0-21" },
{ 0x20, "STK600-RCPWM-22" },
{ 0x21, "STK600-RC020T-23" },
{ 0x22, "STK600-RC044M-24" },
{ 0x23, "STK600-RC044U-25" },
{ 0x24, "STK600-RCPWM-26" },
{ 0x25, "STK600-RCuC3B48-27" },
{ 0x27, "STK600-RC032M-29" },
{ 0x28, "STK600-RC044M-30" },
{ 0x29, "STK600-RC044M-31" },
{ 0x2A, "STK600-RC014T-42" },
{ 0x2B, "STK600-RC020T-43" },
{ 0x30, "STK600-RCUC3A144-32" },
{ 0x34, "STK600-RCUC3L0-34" },
{ 0x38, "STK600-RCUC3C0-36" },
{ 0x3B, "STK600-RCUC3C0-37" },
{ 0x3E, "STK600-RCUC3A144-33" },
{ 0x46, "STK600-RCuC3A100-28" },
{ 0x55, "STK600-RC064M-9" },
{ 0x88, "STK600-RCUC3C1-38" },
{ 0x8B, "STK600-RCUC3C1-39" },
{ 0xA0, "STK600-RC008T-7" },
{ 0xB8, "STK600-RCUC3C2-40" },
{ 0xBB, "STK600-RCUC3C2-41" },
};
static const struct carddata socket_cards[] =
{
{ 0x01, "STK600-TQFP48" },
{ 0x02, "STK600-TQFP32" },
{ 0x03, "STK600-TQFP100" },
{ 0x04, "STK600-SOIC" },
{ 0x06, "STK600-TQFP144" },
{ 0x09, "STK600-TinyX3U" },
{ 0x0C, "STK600-TSSOP44" },
{ 0x0D, "STK600-TQFP44" },
{ 0x0E, "STK600-TQFP64-2" },
{ 0x0F, "STK600-ATMEGA2560" },
{ 0x15, "STK600-MLF64" },
{ 0x16, "STK600-ATXMEGAT0" },
{ 0x18, "QT600-ATMEGA324-QM64" },
{ 0x19, "STK600-ATMEGA128RFA1" },
{ 0x1A, "QT600-ATTINY88-QT8" },
{ 0x1B, "QT600-ATXMEGA128A1-QT16" },
{ 0x1C, "QT600-AT32UC3L-QM64" },
{ 0x1D, "STK600-HVE2" },
{ 0x1E, "STK600-ATTINY10" },
{ 0x55, "STK600-TQFP64" },
{ 0x69, "STK600-uC3-144" },
{ 0xF0, "STK600-ATXMEGA1281A1" },
{ 0xF1, "STK600-DIP" },
};
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static int stk500v2_getparm(PROGRAMMER * pgm, unsigned char parm, unsigned char * value);
static int stk500v2_setparm(PROGRAMMER * pgm, unsigned char parm, unsigned char value);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
static int stk500v2_getparm2(PROGRAMMER * pgm, unsigned char parm, unsigned int * value);
static int stk500v2_setparm2(PROGRAMMER * pgm, unsigned char parm, unsigned int value);
static int stk500v2_setparm_real(PROGRAMMER * pgm, unsigned char parm, unsigned char value);
static void stk500v2_print_parms1(PROGRAMMER * pgm, const char * p);
static int stk500v2_paged_load(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes);
static int stk500v2_paged_write(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static unsigned int stk500v2_mode_for_pagesize(unsigned int pagesize);
static double stk500v2_sck_to_us(PROGRAMMER * pgm, unsigned char dur);
static int stk500v2_set_sck_period_mk2(PROGRAMMER * pgm, double v);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
static int stk600_set_sck_period(PROGRAMMER * pgm, double v);
static void stk600_setup_xprog(PROGRAMMER * pgm);
static void stk600_setup_isp(PROGRAMMER * pgm);
static int stk600_xprog_program_enable(PROGRAMMER * pgm, AVRPART * p);
void stk500v2_setup(PROGRAMMER * pgm)
{
if ((pgm->cookie = malloc(sizeof(struct pdata))) == 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_setup(): Out of memory allocating private data\n",
progname);
exit(1);
}
memset(pgm->cookie, 0, sizeof(struct pdata));
PDATA(pgm)->command_sequence = 1;
PDATA(pgm)->boot_start = ULONG_MAX;
}
static void stk500v2_jtagmkII_setup(PROGRAMMER * pgm)
{
void *mycookie, *theircookie;
if ((pgm->cookie = malloc(sizeof(struct pdata))) == 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_setup(): Out of memory allocating private data\n",
progname);
exit(1);
}
memset(pgm->cookie, 0, sizeof(struct pdata));
PDATA(pgm)->command_sequence = 1;
/*
* Now, have the JTAG ICE mkII backend allocate its own private
* data. Store our own cookie in a safe place for the time being.
*/
mycookie = pgm->cookie;
jtagmkII_setup(pgm);
theircookie = pgm->cookie;
pgm->cookie = mycookie;
PDATA(pgm)->chained_pdata = theircookie;
}
static void stk500v2_jtag3_setup(PROGRAMMER * pgm)
{
void *mycookie, *theircookie;
if ((pgm->cookie = malloc(sizeof(struct pdata))) == 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_setup(): Out of memory allocating private data\n",
progname);
exit(1);
}
memset(pgm->cookie, 0, sizeof(struct pdata));
PDATA(pgm)->command_sequence = 1;
/*
* Now, have the JTAGICE3 backend allocate its own private
* data. Store our own cookie in a safe place for the time being.
*/
mycookie = pgm->cookie;
jtag3_setup(pgm);
theircookie = pgm->cookie;
pgm->cookie = mycookie;
PDATA(pgm)->chained_pdata = theircookie;
}
void stk500v2_teardown(PROGRAMMER * pgm)
{
free(pgm->cookie);
}
static void stk500v2_jtagmkII_teardown(PROGRAMMER * pgm)
{
void *mycookie;
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
jtagmkII_teardown(pgm);
free(mycookie);
}
static void stk500v2_jtag3_teardown(PROGRAMMER * pgm)
{
void *mycookie;
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
jtag3_teardown(pgm);
free(mycookie);
}
static unsigned short
b2_to_u16(unsigned char *b)
{
unsigned short l;
l = b[0];
l += (unsigned)b[1] << 8;
return l;
}
static int stk500v2_send_mk2(PROGRAMMER * pgm, unsigned char * data, size_t len)
{
if (serial_send(&pgm->fd, data, len) != 0) {
avrdude_message(MSG_INFO, "%s: stk500_send_mk2(): failed to send command to serial port\n",progname);
return -1;
}
return 0;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static unsigned short get_jtagisp_return_size(unsigned char cmd)
{
int i;
for (i = 0; i < sizeof jtagispcmds / sizeof jtagispcmds[0]; i++)
if (jtagispcmds[i].cmd == cmd)
return jtagispcmds[i].size;
return 0;
}
/*
* Send the data as a JTAG ICE mkII encapsulated ISP packet.
* Unlike what AVR067 says, the packet gets a length of our
* response buffer prepended, and replies with RSP_SPI_DATA
* if successful.
*/
static int stk500v2_jtagmkII_send(PROGRAMMER * pgm, unsigned char * data, size_t len)
{
unsigned char *cmdbuf;
int rv;
unsigned short sz;
void *mycookie;
sz = get_jtagisp_return_size(data[0]);
if (sz == 0) {
avrdude_message(MSG_INFO, "%s: unsupported encapsulated ISP command: %#x\n",
progname, data[0]);
return -1;
}
if (sz == SZ_READ_FLASH_EE) {
/*
* For CMND_READ_FLASH_ISP and CMND_READ_EEPROM_ISP, extract the
* size of the return data from the request. Note that the
* request itself has the size in big endian format, while we are
* supposed to deliver it in little endian.
*/
sz = 3 + (data[1] << 8) + data[2];
} else if (sz == SZ_SPI_MULTI) {
/*
* CMND_SPI_MULTI has the Rx size encoded in its 3rd byte.
*/
sz = 3 + data[2];
}
if ((cmdbuf = malloc(len + 3)) == NULL) {
avrdude_message(MSG_INFO, "%s: out of memory for command packet\n",
progname);
exit(1);
}
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
cmdbuf[0] = CMND_ISP_PACKET;
cmdbuf[1] = sz & 0xff;
cmdbuf[2] = (sz >> 8) & 0xff;
memcpy(cmdbuf + 3, data, len);
rv = jtagmkII_send(pgm, cmdbuf, len + 3);
free(cmdbuf);
pgm->cookie = mycookie;
return rv;
}
/*
* Send the data as a JTAGICE3 encapsulated ISP packet.
*/
static int stk500v2_jtag3_send(PROGRAMMER * pgm, unsigned char * data, size_t len)
{
unsigned char *cmdbuf;
int rv;
void *mycookie;
if ((cmdbuf = malloc(len + 1)) == NULL) {
avrdude_message(MSG_INFO, "%s: out of memory for command packet\n",
progname);
exit(1);
}
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
cmdbuf[0] = SCOPE_AVR_ISP;
memcpy(cmdbuf + 1, data, len);
rv = jtag3_send(pgm, cmdbuf, len + 1);
free(cmdbuf);
pgm->cookie = mycookie;
return rv;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static int stk500v2_send(PROGRAMMER * pgm, unsigned char * data, size_t len)
{
unsigned char buf[275 + 6]; // max MESSAGE_BODY of 275 bytes, 6 bytes overhead
int i;
if (PDATA(pgm)->pgmtype == PGMTYPE_AVRISP_MKII ||
PDATA(pgm)->pgmtype == PGMTYPE_STK600)
return stk500v2_send_mk2(pgm, data, len);
else if (PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE_MKII)
return stk500v2_jtagmkII_send(pgm, data, len);
else if (PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE3)
return stk500v2_jtag3_send(pgm, data, len);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
buf[0] = MESSAGE_START;
buf[1] = PDATA(pgm)->command_sequence;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
buf[2] = len / 256;
buf[3] = len % 256;
buf[4] = TOKEN;
memcpy(buf+5, data, len);
// calculate the XOR checksum
buf[5+len] = 0;
for (i=0;i<5+len;i++)
buf[5+len] ^= buf[i];
DEBUG("STK500V2: stk500v2_send(");
for (i=0;i<len+6;i++) DEBUG("0x%02x ",buf[i]);
DEBUG(", %d)\n",len+6);
if (serial_send(&pgm->fd, buf, len+6) != 0) {
avrdude_message(MSG_INFO, "%s: stk500_send(): failed to send command to serial port\n",progname);
return -1;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
return 0;
}
int stk500v2_drain(PROGRAMMER * pgm, int display)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
return serial_drain(&pgm->fd, display);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
static int stk500v2_recv_mk2(PROGRAMMER * pgm, unsigned char *msg,
size_t maxsize)
{
int rv;
rv = serial_recv(&pgm->fd, msg, maxsize);
if (rv < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_recv_mk2: error in USB receive\n", progname);
return -1;
}
return rv;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static int stk500v2_jtagmkII_recv(PROGRAMMER * pgm, unsigned char *msg,
size_t maxsize)
{
int rv;
unsigned char *jtagmsg;
void *mycookie;
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
rv = jtagmkII_recv(pgm, &jtagmsg);
pgm->cookie = mycookie;
if (rv <= 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_jtagmkII_recv(): error in jtagmkII_recv()\n",
progname);
return -1;
}
if (rv - 1 > maxsize) {
avrdude_message(MSG_INFO, "%s: stk500v2_jtagmkII_recv(): got %u bytes, have only room for %u bytes\n",
progname, (unsigned)rv - 1, (unsigned)maxsize);
rv = maxsize;
}
switch (jtagmsg[0]) {
case RSP_SPI_DATA:
break;
case RSP_FAILED:
avrdude_message(MSG_INFO, "%s: stk500v2_jtagmkII_recv(): failed\n",
progname);
return -1;
case RSP_ILLEGAL_MCU_STATE:
avrdude_message(MSG_INFO, "%s: stk500v2_jtagmkII_recv(): illegal MCU state\n",
progname);
return -1;
default:
avrdude_message(MSG_INFO, "%s: stk500v2_jtagmkII_recv(): unknown status %d\n",
progname, jtagmsg[0]);
return -1;
}
memcpy(msg, jtagmsg + 1, rv - 1);
return rv;
}
static int stk500v2_jtag3_recv(PROGRAMMER * pgm, unsigned char *msg,
size_t maxsize)
{
int rv;
unsigned char *jtagmsg;
void *mycookie;
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
rv = jtag3_recv(pgm, &jtagmsg);
pgm->cookie = mycookie;
if (rv <= 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_jtag3_recv(): error in jtagmkII_recv()\n",
progname);
return -1;
}
/* Getting more data than expected is a normal case for the EDBG
implementation of JTAGICE3, as they always request a full 512
octets from the ICE. Thus, only complain at high verbose
levels. */
if (rv - 1 > maxsize) {
avrdude_message(MSG_DEBUG, "%s: stk500v2_jtag3_recv(): got %u bytes, have only room for %u bytes\n",
progname, (unsigned)rv - 1, (unsigned)maxsize);
rv = maxsize;
}
if (jtagmsg[0] != SCOPE_AVR_ISP) {
avrdude_message(MSG_INFO, "%s: stk500v2_jtag3_recv(): message is not AVR ISP: 0x%02x\n",
progname, jtagmsg[0]);
free(jtagmsg);
return -1;
}
memcpy(msg, jtagmsg + 1, rv - 1);
free(jtagmsg);
return rv;
}
static int stk500v2_recv(PROGRAMMER * pgm, unsigned char *msg, size_t maxsize) {
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
enum states { sINIT, sSTART, sSEQNUM, sSIZE1, sSIZE2, sTOKEN, sDATA, sCSUM, sDONE } state = sSTART;
unsigned int msglen = 0;
unsigned int curlen = 0;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
int timeout = 0;
unsigned char c, checksum = 0;
/*
* The entire timeout handling here is not very consistent, see
*
* https://savannah.nongnu.org/bugs/index.php?43626
*/
long timeoutval = SERIAL_TIMEOUT; // seconds
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
struct timeval tv;
double tstart, tnow;
if (PDATA(pgm)->pgmtype == PGMTYPE_AVRISP_MKII ||
PDATA(pgm)->pgmtype == PGMTYPE_STK600)
return stk500v2_recv_mk2(pgm, msg, maxsize);
else if (PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE_MKII)
return stk500v2_jtagmkII_recv(pgm, msg, maxsize);
else if (PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE3)
return stk500v2_jtag3_recv(pgm, msg, maxsize);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
DEBUG("STK500V2: stk500v2_recv(): ");
gettimeofday(&tv, NULL);
tstart = tv.tv_sec;
while ( (state != sDONE ) && (!timeout) ) {
if (serial_recv(&pgm->fd, &c, 1) < 0)
goto timedout;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
DEBUG("0x%02x ",c);
checksum ^= c;
switch (state) {
case sSTART:
DEBUGRECV("hoping for start token...");
if (c == MESSAGE_START) {
DEBUGRECV("got it\n");
checksum = MESSAGE_START;
state = sSEQNUM;
} else
DEBUGRECV("sorry\n");
break;
case sSEQNUM:
DEBUGRECV("hoping for sequence...\n");
if (c == PDATA(pgm)->command_sequence) {
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
DEBUGRECV("got it, incrementing\n");
state = sSIZE1;
PDATA(pgm)->command_sequence++;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
} else {
DEBUGRECV("sorry\n");
state = sSTART;
}
break;
case sSIZE1:
DEBUGRECV("hoping for size LSB\n");
msglen = (unsigned)c * 256;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
state = sSIZE2;
break;
case sSIZE2:
DEBUGRECV("hoping for size MSB...");
msglen += (unsigned)c;
DEBUG(" msg is %u bytes\n",msglen);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
state = sTOKEN;
break;
case sTOKEN:
if (c == TOKEN) state = sDATA;
else state = sSTART;
break;
case sDATA:
if (curlen < maxsize) {
msg[curlen] = c;
} else {
avrdude_message(MSG_INFO, "%s: stk500v2_recv(): buffer too small, received %d byte into %u byte buffer\n",
progname,curlen,(unsigned int)maxsize);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -2;
}
if ((curlen == 0) && (msg[0] == ANSWER_CKSUM_ERROR)) {
avrdude_message(MSG_INFO, "%s: stk500v2_recv(): previous packet sent with wrong checksum\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname);
return -3;
}
curlen++;
if (curlen == msglen) state = sCSUM;
break;
case sCSUM:
if (checksum == 0) {
state = sDONE;
} else {
state = sSTART;
avrdude_message(MSG_INFO, "%s: stk500v2_recv(): checksum error\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname);
return -4;
}
break;
default:
avrdude_message(MSG_INFO, "%s: stk500v2_recv(): unknown state\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname);
return -5;
} /* switch */
gettimeofday(&tv, NULL);
tnow = tv.tv_sec;
if (tnow-tstart > timeoutval) { // wuff - signed/unsigned/overflow
timedout:
avrdude_message(MSG_INFO, "%s: stk500v2_ReceiveMessage(): timeout\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname);
return -1;
}
} /* while */
DEBUG("\n");
return (int)(msglen+6);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
int stk500v2_getsync(PROGRAMMER * pgm) {
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
int tries = 0;
unsigned char buf[1], resp[32];
int status;
DEBUG("STK500V2: stk500v2_getsync()\n");
if (PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE_MKII ||
PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE3)
return 0;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
retry:
tries++;
// send the sync command and see if we can get there
buf[0] = CMD_SIGN_ON;
stk500v2_send(pgm, buf, 1);
// try to get the response back and see where we got
status = stk500v2_recv(pgm, resp, sizeof(resp));
// if we got bytes returned, check to see what came back
if (status > 0) {
if ((resp[0] == CMD_SIGN_ON) && (resp[1] == STATUS_CMD_OK) &&
(status > 3)) {
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
// success!
unsigned int siglen = resp[2];
if (siglen >= strlen("STK500_2") &&
memcmp(resp + 3, "STK500_2", strlen("STK500_2")) == 0) {
PDATA(pgm)->pgmtype = PGMTYPE_STK500;
} else if (siglen >= strlen("AVRISP_2") &&
memcmp(resp + 3, "AVRISP_2", strlen("AVRISP_2")) == 0) {
PDATA(pgm)->pgmtype = PGMTYPE_AVRISP;
} else if (siglen >= strlen("AVRISP_MK2") &&
memcmp(resp + 3, "AVRISP_MK2", strlen("AVRISP_MK2")) == 0) {
PDATA(pgm)->pgmtype = PGMTYPE_AVRISP_MKII;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
} else if (siglen >= strlen("STK600") &&
memcmp(resp + 3, "STK600", strlen("STK600")) == 0) {
PDATA(pgm)->pgmtype = PGMTYPE_STK600;
} else {
resp[siglen + 3] = 0;
avrdude_message(MSG_NOTICE, "%s: stk500v2_getsync(): got response from unknown "
"programmer %s, assuming STK500\n",
progname, resp + 3);
PDATA(pgm)->pgmtype = PGMTYPE_STK500;
}
avrdude_message(MSG_DEBUG, "%s: stk500v2_getsync(): found %s programmer\n",
progname, pgmname[PDATA(pgm)->pgmtype]);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return 0;
} else {
if (tries > RETRIES) {
avrdude_message(MSG_INFO, "%s: stk500v2_getsync(): can't communicate with device: resp=0x%02x\n",
progname, resp[0]);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -6;
} else
goto retry;
}
// or if we got a timeout
} else if (status == -1) {
if (tries > RETRIES) {
avrdude_message(MSG_INFO, "%s: stk500v2_getsync(): timeout communicating with programmer\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname);
return -1;
} else
goto retry;
// or any other error
} else {
if (tries > RETRIES) {
avrdude_message(MSG_INFO, "%s: stk500v2_getsync(): error communicating with programmer: (%d)\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname,status);
} else
goto retry;
}
return 0;
}
static int stk500v2_command(PROGRAMMER * pgm, unsigned char * buf,
size_t len, size_t maxlen) {
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
int i;
int tries = 0;
int status;
DEBUG("STK500V2: stk500v2_command(");
for (i=0;i<len;i++) DEBUG("0x%02x ",buf[i]);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
DEBUG(", %d)\n",len);
retry:
tries++;
// send the command to the programmer
stk500v2_send(pgm,buf,len);
// attempt to read the status back
status = stk500v2_recv(pgm,buf,maxlen);
// if we got a successful readback, return
if (status > 0) {
DEBUG(" = %d\n",status);
if (status < 2) {
avrdude_message(MSG_INFO, "%s: stk500v2_command(): short reply\n", progname);
return -1;
}
if (buf[0] == CMD_XPROG_SETMODE || buf[0] == CMD_XPROG) {
/*
* Decode XPROG wrapper errors.
*/
const char *msg;
int i;
/*
* For CMD_XPROG_SETMODE, the status is returned in buf[1].
* For CMD_XPROG, buf[1] contains the XPRG_CMD_* command, and
* buf[2] contains the status.
*/
i = buf[0] == CMD_XPROG_SETMODE? 1: 2;
if (buf[i] != XPRG_ERR_OK) {
switch (buf[i]) {
case XPRG_ERR_FAILED: msg = "Failed"; break;
case XPRG_ERR_COLLISION: msg = "Collision"; break;
case XPRG_ERR_TIMEOUT: msg = "Timeout"; break;
default: msg = "Unknown"; break;
}
avrdude_message(MSG_INFO, "%s: stk500v2_command(): error in %s: %s\n",
progname,
(buf[0] == CMD_XPROG_SETMODE? "CMD_XPROG_SETMODE": "CMD_XPROG"),
msg);
return -1;
}
return 0;
} else {
/*
* Decode STK500v2 errors.
*/
if (buf[1] >= STATUS_CMD_TOUT && buf[1] < 0xa0) {
const char *msg;
char msgbuf[30];
switch (buf[1]) {
case STATUS_CMD_TOUT:
msg = "Command timed out";
break;
case STATUS_RDY_BSY_TOUT:
msg = "Sampling of the RDY/nBSY pin timed out";
break;
case STATUS_SET_PARAM_MISSING:
msg = "The `Set Device Parameters' have not been "
"executed in advance of this command";
default:
sprintf(msgbuf, "unknown, code 0x%02x", buf[1]);
msg = msgbuf;
break;
}
if (quell_progress < 2) {
avrdude_message(MSG_INFO, "%s: stk500v2_command(): warning: %s\n",
progname, msg);
}
} else if (buf[1] == STATUS_CMD_OK) {
return status;
} else if (buf[1] == STATUS_CMD_FAILED) {
avrdude_message(MSG_INFO, "%s: stk500v2_command(): command failed\n",
progname);
} else if (buf[1] == STATUS_CMD_UNKNOWN) {
avrdude_message(MSG_INFO, "%s: stk500v2_command(): unknown command\n",
progname);
} else {
avrdude_message(MSG_INFO, "%s: stk500v2_command(): unknown status 0x%02x\n",
progname, buf[1]);
}
return -1;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
// otherwise try to sync up again
status = stk500v2_getsync(pgm);
if (status != 0) {
if (tries > RETRIES) {
avrdude_message(MSG_INFO, "%s: stk500v2_command(): failed miserably to execute command 0x%02x\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname,buf[0]);
return -1;
} else
goto retry;
}
DEBUG(" = 0\n");
return 0;
}
static int stk500v2_cmd(PROGRAMMER * pgm, const unsigned char *cmd,
unsigned char *res)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
unsigned char buf[8];
int result;
DEBUG("STK500V2: stk500v2_cmd(%02x,%02x,%02x,%02x)\n",cmd[0],cmd[1],cmd[2],cmd[3]);
buf[0] = CMD_SPI_MULTI;
buf[1] = 4;
buf[2] = 4;
buf[3] = 0;
buf[4] = cmd[0];
buf[5] = cmd[1];
buf[6] = cmd[2];
buf[7] = cmd[3];
result = stk500v2_command(pgm, buf, 8, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_cmd(): failed to send command\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname);
return -1;
} else if (result < 6) {
avrdude_message(MSG_INFO, "%s: stk500v2_cmd(): short reply, len = %d\n",
progname, result);
return -1;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
res[0] = buf[2];
res[1] = buf[3];
res[2] = buf[4];
res[3] = buf[5];
return 0;
}
static int stk500v2_jtag3_cmd(PROGRAMMER * pgm, const unsigned char *cmd,
unsigned char *res)
{
avrdude_message(MSG_INFO, "%s: stk500v2_jtag3_cmd(): Not available in JTAGICE3\n",
progname);
return -1;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* issue the 'chip erase' command to the AVR device
*/
static int stk500v2_chip_erase(PROGRAMMER * pgm, AVRPART * p)
{
int result;
unsigned char buf[16];
if (p->op[AVR_OP_CHIP_ERASE] == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_chip_erase: chip erase instruction not defined for part \"%s\"\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, p->desc);
return -1;
}
pgm->pgm_led(pgm, ON);
buf[0] = CMD_CHIP_ERASE_ISP;
buf[1] = p->chip_erase_delay / 1000;
buf[2] = 0; // use delay (?)
avr_set_bits(p->op[AVR_OP_CHIP_ERASE], buf+3);
result = stk500v2_command(pgm, buf, 7, sizeof(buf));
usleep(p->chip_erase_delay);
pgm->initialize(pgm, p);
pgm->pgm_led(pgm, OFF);
return result >= 0? 0: -1;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
/*
* issue the 'chip erase' command to the AVR device, generic HV mode
*/
static int stk500hv_chip_erase(PROGRAMMER * pgm, AVRPART * p, enum hvmode mode)
{
int result;
unsigned char buf[3];
pgm->pgm_led(pgm, ON);
if (mode == PPMODE) {
buf[0] = CMD_CHIP_ERASE_PP;
buf[1] = p->chiperasepulsewidth;
buf[2] = p->chiperasepolltimeout;
} else {
buf[0] = CMD_CHIP_ERASE_HVSP;
buf[1] = p->chiperasepolltimeout;
buf[2] = p->chiperasetime;
}
result = stk500v2_command(pgm, buf, 3, sizeof(buf));
usleep(p->chip_erase_delay);
pgm->initialize(pgm, p);
pgm->pgm_led(pgm, OFF);
return result >= 0? 0: -1;
}
/*
* issue the 'chip erase' command to the AVR device, parallel mode
*/
static int stk500pp_chip_erase(PROGRAMMER * pgm, AVRPART * p)
{
return stk500hv_chip_erase(pgm, p, PPMODE);
}
/*
* issue the 'chip erase' command to the AVR device, HVSP mode
*/
static int stk500hvsp_chip_erase(PROGRAMMER * pgm, AVRPART * p)
{
return stk500hv_chip_erase(pgm, p, HVSPMODE);
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
static struct
{
unsigned int state;
const char *description;
} connection_status[] =
{
{ STATUS_CONN_FAIL_MOSI, "MOSI fail" },
{ STATUS_CONN_FAIL_RST, "RST fail" },
{ STATUS_CONN_FAIL_SCK, "SCK fail" },
{ STATUS_TGT_NOT_DETECTED, "Target not detected" },
{ STATUS_TGT_REVERSE_INSERTED, "Target reverse inserted" },
};
/*
* Max length of returned message is the sum of all the description
* strings in the table above, plus 2 characters for separation.
* Currently, this is 76 chars.
*/
static void
stk500v2_translate_conn_status(unsigned char status, char *msg)
{
size_t i;
int need_comma;
*msg = 0;
need_comma = 0;
for (i = 0;
i < sizeof connection_status / sizeof connection_status[0];
i++)
{
if ((status & connection_status[i].state) != 0)
{
if (need_comma)
strcat(msg, ", ");
strcat(msg, connection_status[i].description);
need_comma = 1;
}
}
if (*msg == 0)
sprintf(msg, "Unknown status 0x%02x", status);
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* issue the 'program enable' command to the AVR device
*/
static int stk500v2_program_enable(PROGRAMMER * pgm, AVRPART * p)
{
unsigned char buf[16];
char msg[100]; /* see remarks above about size needed */
int rv, tries;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
PDATA(pgm)->lastpart = p;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
if (p->op[AVR_OP_PGM_ENABLE] == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_program_enable(): program enable instruction not defined for part \"%s\"\n",
progname, p->desc);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -1;
}
if (PDATA(pgm)->pgmtype == PGMTYPE_STK500 ||
PDATA(pgm)->pgmtype == PGMTYPE_STK600)
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
/* Activate AVR-style (low active) RESET */
stk500v2_setparm_real(pgm, PARAM_RESET_POLARITY, 0x01);
tries = 0;
retry:
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
buf[0] = CMD_ENTER_PROGMODE_ISP;
buf[1] = p->timeout;
buf[2] = p->stabdelay;
buf[3] = p->cmdexedelay;
buf[4] = p->synchloops;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
buf[5] = p->bytedelay;
buf[6] = p->pollvalue;
buf[7] = p->pollindex;
avr_set_bits(p->op[AVR_OP_PGM_ENABLE], buf+8);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
buf[10] = buf[11] = 0;
rv = stk500v2_command(pgm, buf, 12, sizeof(buf));
if (rv < 0) {
switch (PDATA(pgm)->pgmtype)
{
case PGMTYPE_STK600:
case PGMTYPE_AVRISP_MKII:
if (stk500v2_getparm(pgm, PARAM_STATUS_TGT_CONN, &buf[0]) != 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_program_enable(): cannot get connection status\n",
progname);
} else {
stk500v2_translate_conn_status(buf[0], msg);
avrdude_message(MSG_INFO, "%s: stk500v2_program_enable():"
" bad AVRISPmkII connection status: %s\n",
progname, msg);
}
break;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
case PGMTYPE_JTAGICE3:
if (buf[1] == STATUS_CMD_FAILED &&
(p->flags & AVRPART_HAS_DW) != 0) {
void *mycookie;
unsigned char cmd[4], *resp;
/* Try debugWIRE, and MONCON_DISABLE */
avrdude_message(MSG_NOTICE2, "%s: No response in ISP mode, trying debugWIRE\n",
progname);
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
cmd[0] = PARM3_CONN_DW;
if (jtag3_setparm(pgm, SCOPE_AVR, 1, PARM3_CONNECTION, cmd, 1) < 0) {
pgm->cookie = mycookie;
break;
}
cmd[0] = SCOPE_AVR;
cmd[1] = CMD3_SIGN_ON;
cmd[2] = cmd[3] = 0;
if (jtag3_command(pgm, cmd, 4, &resp, "AVR sign-on") >= 0) {
free(resp);
cmd[1] = CMD3_START_DW_DEBUG;
if (jtag3_command(pgm, cmd, 4, &resp, "start DW debug") >= 0) {
free(resp);
cmd[1] = CMD3_MONCON_DISABLE;
if (jtag3_command(pgm, cmd, 3, &resp, "MonCon disable") >= 0)
free(resp);
}
}
pgm->cookie = mycookie;
if (tries++ > 3) {
avrdude_message(MSG_INFO, "%s: Failed to return from debugWIRE to ISP.\n",
progname);
break;
}
avrdude_message(MSG_INFO, "%s: Target prepared for ISP, signed off.\n"
"%s: Now retrying without power-cycling the target.\n",
progname, progname);
goto retry;
}
break;
default:
/* cannot report anything for other pgmtypes */
break;
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
}
return rv;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
/*
* issue the 'program enable' command to the AVR device, parallel mode
*/
static int stk500pp_program_enable(PROGRAMMER * pgm, AVRPART * p)
{
unsigned char buf[16];
PDATA(pgm)->lastpart = p;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
buf[0] = CMD_ENTER_PROGMODE_PP;
buf[1] = p->hventerstabdelay;
buf[2] = p->progmodedelay;
buf[3] = p->latchcycles;
buf[4] = p->togglevtg;
buf[5] = p->poweroffdelay;
buf[6] = p->resetdelayms;
buf[7] = p->resetdelayus;
return stk500v2_command(pgm, buf, 8, sizeof(buf));
}
/*
* issue the 'program enable' command to the AVR device, HVSP mode
*/
static int stk500hvsp_program_enable(PROGRAMMER * pgm, AVRPART * p)
{
unsigned char buf[16];
PDATA(pgm)->lastpart = p;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
buf[0] = PDATA(pgm)->pgmtype == PGMTYPE_STK600?
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
CMD_ENTER_PROGMODE_HVSP_STK600:
CMD_ENTER_PROGMODE_HVSP;
buf[1] = p->hventerstabdelay;
buf[2] = p->hvspcmdexedelay;
buf[3] = p->synchcycles;
buf[4] = p->latchcycles;
buf[5] = p->togglevtg;
buf[6] = p->poweroffdelay;
buf[7] = p->resetdelayms;
buf[8] = p->resetdelayus;
return stk500v2_command(pgm, buf, 9, sizeof(buf));
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* initialize the AVR device and prepare it to accept commands
*/
static int stk500v2_initialize(PROGRAMMER * pgm, AVRPART * p)
{
LNODEID ln;
AVRMEM * m;
if ((PDATA(pgm)->pgmtype == PGMTYPE_STK600 ||
PDATA(pgm)->pgmtype == PGMTYPE_AVRISP_MKII ||
PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE_MKII) != 0
&& (p->flags & (AVRPART_HAS_PDI | AVRPART_HAS_TPI)) != 0) {
/*
* This is an ATxmega device, must use XPROG protocol for the
* remaining actions.
*/
if ((p->flags & AVRPART_HAS_PDI) != 0) {
/*
* Find out where the border between application and boot area
* is.
*/
AVRMEM *bootmem = avr_locate_mem(p, "boot");
AVRMEM *flashmem = avr_locate_mem(p, "flash");
if (bootmem == NULL || flashmem == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_initialize(): Cannot locate \"flash\" and \"boot\" memories in description\n",
progname);
} else {
PDATA(pgm)->boot_start = bootmem->offset - flashmem->offset;
}
}
stk600_setup_xprog(pgm);
} else {
stk600_setup_isp(pgm);
}
/*
* Examine the avrpart's memory definitions, and initialize the page
* caches. For devices/memory that are not page oriented, treat
* them as page size 1 for EEPROM, and 2 for flash.
*/
PDATA(pgm)->flash_pagesize = 2;
PDATA(pgm)->eeprom_pagesize = 1;
for (ln = lfirst(p->mem); ln; ln = lnext(ln)) {
m = ldata(ln);
if (strcmp(m->desc, "flash") == 0) {
if (m->page_size > 0) {
if (m->page_size > 256)
PDATA(pgm)->flash_pagesize = 256;
else
PDATA(pgm)->flash_pagesize = m->page_size;
}
} else if (strcmp(m->desc, "eeprom") == 0) {
if (m->page_size > 0)
PDATA(pgm)->eeprom_pagesize = m->page_size;
}
}
free(PDATA(pgm)->flash_pagecache);
free(PDATA(pgm)->eeprom_pagecache);
if ((PDATA(pgm)->flash_pagecache = malloc(PDATA(pgm)->flash_pagesize)) == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_initialize(): Out of memory\n",
progname);
return -1;
}
if ((PDATA(pgm)->eeprom_pagecache = malloc(PDATA(pgm)->eeprom_pagesize)) == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_initialize(): Out of memory\n",
progname);
free(PDATA(pgm)->flash_pagecache);
return -1;
}
PDATA(pgm)->flash_pageaddr = PDATA(pgm)->eeprom_pageaddr = (unsigned long)-1L;
if (p->flags & AVRPART_IS_AT90S1200) {
/*
* AT90S1200 needs a positive reset pulse after a chip erase.
*/
pgm->disable(pgm);
usleep(10000);
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return pgm->program_enable(pgm, p);
}
/*
* initialize the AVR device and prepare it to accept commands
*/
static int stk500v2_jtag3_initialize(PROGRAMMER * pgm, AVRPART * p)
{
unsigned char parm[4], *resp;
LNODEID ln;
AVRMEM * m;
void *mycookie;
if ((p->flags & AVRPART_HAS_PDI) ||
(p->flags & AVRPART_HAS_TPI)) {
avrdude_message(MSG_INFO, "%s: jtag3_initialize(): part %s has no ISP interface\n",
progname, p->desc);
return -1;
}
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
if (p->flags & AVRPART_HAS_DW)
parm[0] = PARM3_ARCH_TINY;
else
parm[0] = PARM3_ARCH_MEGA;
if (jtag3_setparm(pgm, SCOPE_AVR, 0, PARM3_ARCH, parm, 1) < 0) {
pgm->cookie = mycookie;
return -1;
}
parm[0] = PARM3_SESS_PROGRAMMING;
if (jtag3_setparm(pgm, SCOPE_AVR, 0, PARM3_SESS_PURPOSE, parm, 1) < 0) {
pgm->cookie = mycookie;
return -1;
}
parm[0] = PARM3_CONN_ISP;
if (jtag3_setparm(pgm, SCOPE_AVR, 1, PARM3_CONNECTION, parm, 1) < 0) {
pgm->cookie = mycookie;
return -1;
}
parm[0] = SCOPE_AVR_ISP;
parm[1] = 0x1e;
jtag3_send(pgm, parm, 2);
if (jtag3_recv(pgm, &resp) > 0)
free(resp);
pgm->cookie = mycookie;
/*
* Examine the avrpart's memory definitions, and initialize the page
* caches. For devices/memory that are not page oriented, treat
* them as page size 1 for EEPROM, and 2 for flash.
*/
PDATA(pgm)->flash_pagesize = 2;
PDATA(pgm)->eeprom_pagesize = 1;
for (ln = lfirst(p->mem); ln; ln = lnext(ln)) {
m = ldata(ln);
if (strcmp(m->desc, "flash") == 0) {
if (m->page_size > 0) {
if (m->page_size > 256)
PDATA(pgm)->flash_pagesize = 256;
else
PDATA(pgm)->flash_pagesize = m->page_size;
}
} else if (strcmp(m->desc, "eeprom") == 0) {
if (m->page_size > 0)
PDATA(pgm)->eeprom_pagesize = m->page_size;
}
}
free(PDATA(pgm)->flash_pagecache);
free(PDATA(pgm)->eeprom_pagecache);
if ((PDATA(pgm)->flash_pagecache = malloc(PDATA(pgm)->flash_pagesize)) == NULL) {
avrdude_message(MSG_INFO, "%s: stk500hv_initialize(): Out of memory\n",
progname);
return -1;
}
if ((PDATA(pgm)->eeprom_pagecache = malloc(PDATA(pgm)->eeprom_pagesize)) == NULL) {
avrdude_message(MSG_INFO, "%s: stk500hv_initialize(): Out of memory\n",
progname);
free(PDATA(pgm)->flash_pagecache);
return -1;
}
PDATA(pgm)->flash_pageaddr = PDATA(pgm)->eeprom_pageaddr = (unsigned long)-1L;
return pgm->program_enable(pgm, p);
}
/*
* initialize the AVR device and prepare it to accept commands, generic HV mode
*/
static int stk500hv_initialize(PROGRAMMER * pgm, AVRPART * p, enum hvmode mode)
{
unsigned char buf[CTL_STACK_SIZE + 1];
int result;
LNODEID ln;
AVRMEM * m;
if (p->ctl_stack_type != (mode == PPMODE? CTL_STACK_PP: CTL_STACK_HVSP)) {
avrdude_message(MSG_INFO, "%s: stk500hv_initialize(): "
"%s programming control stack not defined for part \"%s\"\n",
progname,
(mode == PPMODE? "parallel": "high-voltage serial"),
p->desc);
return -1;
}
buf[0] = CMD_SET_CONTROL_STACK;
memcpy(buf + 1, p->controlstack, CTL_STACK_SIZE);
result = stk500v2_command(pgm, buf, CTL_STACK_SIZE + 1, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500hv_initalize(): "
"failed to set control stack\n",
progname);
return -1;
}
/*
* Examine the avrpart's memory definitions, and initialize the page
* caches. For devices/memory that are not page oriented, treat
* them as page size 1 for EEPROM, and 2 for flash.
*/
PDATA(pgm)->flash_pagesize = 2;
PDATA(pgm)->eeprom_pagesize = 1;
for (ln = lfirst(p->mem); ln; ln = lnext(ln)) {
m = ldata(ln);
if (strcmp(m->desc, "flash") == 0) {
if (m->page_size > 0) {
if (m->page_size > 256)
PDATA(pgm)->flash_pagesize = 256;
else
PDATA(pgm)->flash_pagesize = m->page_size;
}
} else if (strcmp(m->desc, "eeprom") == 0) {
if (m->page_size > 0)
PDATA(pgm)->eeprom_pagesize = m->page_size;
}
}
free(PDATA(pgm)->flash_pagecache);
free(PDATA(pgm)->eeprom_pagecache);
if ((PDATA(pgm)->flash_pagecache = malloc(PDATA(pgm)->flash_pagesize)) == NULL) {
avrdude_message(MSG_INFO, "%s: stk500hv_initialize(): Out of memory\n",
progname);
return -1;
}
if ((PDATA(pgm)->eeprom_pagecache = malloc(PDATA(pgm)->eeprom_pagesize)) == NULL) {
avrdude_message(MSG_INFO, "%s: stk500hv_initialize(): Out of memory\n",
progname);
free(PDATA(pgm)->flash_pagecache);
return -1;
}
PDATA(pgm)->flash_pageaddr = PDATA(pgm)->eeprom_pageaddr = (unsigned long)-1L;
return pgm->program_enable(pgm, p);
}
/*
* initialize the AVR device and prepare it to accept commands, PP mode
*/
static int stk500pp_initialize(PROGRAMMER * pgm, AVRPART * p)
{
return stk500hv_initialize(pgm, p, PPMODE);
}
/*
* initialize the AVR device and prepare it to accept commands, HVSP mode
*/
static int stk500hvsp_initialize(PROGRAMMER * pgm, AVRPART * p)
{
return stk500hv_initialize(pgm, p, HVSPMODE);
}
static void stk500v2_jtag3_disable(PROGRAMMER * pgm)
{
unsigned char buf[16];
int result;
free(PDATA(pgm)->flash_pagecache);
PDATA(pgm)->flash_pagecache = NULL;
free(PDATA(pgm)->eeprom_pagecache);
PDATA(pgm)->eeprom_pagecache = NULL;
buf[0] = CMD_LEAVE_PROGMODE_ISP;
buf[1] = 1; // preDelay;
buf[2] = 1; // postDelay;
result = stk500v2_command(pgm, buf, 3, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_disable(): failed to leave programming mode\n",
progname);
}
return;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static void stk500v2_disable(PROGRAMMER * pgm)
{
unsigned char buf[16];
int result;
buf[0] = CMD_LEAVE_PROGMODE_ISP;
buf[1] = 1; // preDelay;
buf[2] = 1; // postDelay;
result = stk500v2_command(pgm, buf, 3, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_disable(): failed to leave programming mode\n",
progname);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
return;
}
/*
* Leave programming mode, generic HV mode
*/
static void stk500hv_disable(PROGRAMMER * pgm, enum hvmode mode)
{
unsigned char buf[16];
int result;
free(PDATA(pgm)->flash_pagecache);
PDATA(pgm)->flash_pagecache = NULL;
free(PDATA(pgm)->eeprom_pagecache);
PDATA(pgm)->eeprom_pagecache = NULL;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
buf[0] = mode == PPMODE? CMD_LEAVE_PROGMODE_PP:
(PDATA(pgm)->pgmtype == PGMTYPE_STK600?
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
CMD_LEAVE_PROGMODE_HVSP_STK600:
CMD_LEAVE_PROGMODE_HVSP);
buf[1] = 15; // p->hvleavestabdelay;
buf[2] = 15; // p->resetdelay;
result = stk500v2_command(pgm, buf, 3, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500hv_disable(): "
"failed to leave programming mode\n",
progname);
}
return;
}
/*
* Leave programming mode, PP mode
*/
static void stk500pp_disable(PROGRAMMER * pgm)
{
stk500hv_disable(pgm, PPMODE);
}
/*
* Leave programming mode, HVSP mode
*/
static void stk500hvsp_disable(PROGRAMMER * pgm)
{
stk500hv_disable(pgm, HVSPMODE);
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static void stk500v2_enable(PROGRAMMER * pgm)
{
return;
}
static int stk500v2_open(PROGRAMMER * pgm, char * port)
{
union pinfo pinfo = { .baud = 115200 };
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
DEBUG("STK500V2: stk500v2_open()\n");
if (pgm->baudrate)
pinfo.baud = pgm->baudrate;
PDATA(pgm)->pgmtype = PGMTYPE_UNKNOWN;
if(strcasecmp(port, "avrdoper") == 0){
#if defined(HAVE_LIBUSB) || (defined(WIN32NATIVE) && defined(HAVE_LIBHID))
serdev = &avrdoper_serdev;
PDATA(pgm)->pgmtype = PGMTYPE_STK500;
#else
avrdude_message(MSG_INFO, "avrdude was compiled without usb support.\n");
return -1;
#endif
}
/*
* If the port name starts with "usb", divert the serial routines
* to the USB ones. The serial_open() function for USB overrides
* the meaning of the "baud" parameter to be the USB device ID to
* search for.
*/
if (strncmp(port, "usb", 3) == 0) {
#if defined(HAVE_LIBUSB)
serdev = &usb_serdev_frame;
pinfo.usbinfo.vid = USB_VENDOR_ATMEL;
pinfo.usbinfo.flags = 0;
pinfo.usbinfo.pid = USB_DEVICE_AVRISPMKII;
PDATA(pgm)->pgmtype = PGMTYPE_AVRISP_MKII;
pgm->set_sck_period = stk500v2_set_sck_period_mk2;
pgm->fd.usb.max_xfer = USBDEV_MAX_XFER_MKII;
pgm->fd.usb.rep = USBDEV_BULK_EP_READ_MKII;
pgm->fd.usb.wep = USBDEV_BULK_EP_WRITE_MKII;
pgm->fd.usb.eep = 0; /* no seperate EP for events */
#else
avrdude_message(MSG_INFO, "avrdude was compiled without usb support.\n");
return -1;
#endif
}
strcpy(pgm->port, port);
if (serial_open(port, pinfo, &pgm->fd)==-1) {
return -1;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* drain any extraneous input
*/
stk500v2_drain(pgm, 0);
stk500v2_getsync(pgm);
stk500v2_drain(pgm, 0);
if (pgm->bitclock != 0.0) {
if (pgm->set_sck_period(pgm, pgm->bitclock) != 0)
return -1;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return 0;
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
static int stk600_open(PROGRAMMER * pgm, char * port)
{
union pinfo pinfo = { .baud = 115200 };
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
DEBUG("STK500V2: stk600_open()\n");
if (pgm->baudrate)
pinfo.baud = pgm->baudrate;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
PDATA(pgm)->pgmtype = PGMTYPE_UNKNOWN;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
/*
* If the port name starts with "usb", divert the serial routines
* to the USB ones. The serial_open() function for USB overrides
* the meaning of the "baud" parameter to be the USB device ID to
* search for.
*/
if (strncmp(port, "usb", 3) == 0) {
#if defined(HAVE_LIBUSB)
serdev = &usb_serdev_frame;
pinfo.usbinfo.vid = USB_VENDOR_ATMEL;
pinfo.usbinfo.flags = 0;
pinfo.usbinfo.pid = USB_DEVICE_STK600;
PDATA(pgm)->pgmtype = PGMTYPE_STK600;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
pgm->set_sck_period = stk600_set_sck_period;
pgm->fd.usb.max_xfer = USBDEV_MAX_XFER_MKII;
pgm->fd.usb.rep = USBDEV_BULK_EP_READ_STK600;
pgm->fd.usb.wep = USBDEV_BULK_EP_WRITE_STK600;
pgm->fd.usb.eep = 0; /* no seperate EP for events */
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
#else
avrdude_message(MSG_INFO, "avrdude was compiled without usb support.\n");
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
return -1;
#endif
}
strcpy(pgm->port, port);
if (serial_open(port, pinfo, &pgm->fd)==-1) {
return -1;
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
/*
* drain any extraneous input
*/
stk500v2_drain(pgm, 0);
stk500v2_getsync(pgm);
stk500v2_drain(pgm, 0);
if (pgm->bitclock != 0.0) {
if (pgm->set_sck_period(pgm, pgm->bitclock) != 0)
return -1;
}
return 0;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static void stk500v2_close(PROGRAMMER * pgm)
{
DEBUG("STK500V2: stk500v2_close()\n");
serial_close(&pgm->fd);
pgm->fd.ifd = -1;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
static int stk500v2_loadaddr(PROGRAMMER * pgm, unsigned int addr)
{
unsigned char buf[16];
int result;
DEBUG("STK500V2: stk500v2_loadaddr(%d)\n",addr);
buf[0] = CMD_LOAD_ADDRESS;
buf[1] = (addr >> 24) & 0xff;
buf[2] = (addr >> 16) & 0xff;
buf[3] = (addr >> 8) & 0xff;
buf[4] = addr & 0xff;
result = stk500v2_command(pgm, buf, 5, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_loadaddr(): failed to set load address\n",
progname);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -1;
}
return 0;
}
/*
* Read a single byte, generic HV mode
*/
static int stk500hv_read_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char * value,
enum hvmode mode)
{
int result, cmdlen = 2;
unsigned char buf[266];
unsigned long paddr = 0UL, *paddr_ptr = NULL;
unsigned int pagesize = 0, use_ext_addr = 0, addrshift = 0;
unsigned char *cache_ptr = NULL;
avrdude_message(MSG_NOTICE2, "%s: stk500hv_read_byte(.., %s, 0x%lx, ...)\n",
progname, mem->desc, addr);
if (strcmp(mem->desc, "flash") == 0) {
buf[0] = mode == PPMODE? CMD_READ_FLASH_PP: CMD_READ_FLASH_HVSP;
cmdlen = 3;
pagesize = PDATA(pgm)->flash_pagesize;
paddr = addr & ~(pagesize - 1);
paddr_ptr = &PDATA(pgm)->flash_pageaddr;
cache_ptr = PDATA(pgm)->flash_pagecache;
addrshift = 1;
/*
* If bit 31 is set, this indicates that the following read/write
* operation will be performed on a memory that is larger than
* 64KBytes. This is an indication to STK500 that a load extended
* address must be executed.
*/
if (mem->op[AVR_OP_LOAD_EXT_ADDR] != NULL) {
use_ext_addr = (1U << 31);
}
} else if (strcmp(mem->desc, "eeprom") == 0) {
buf[0] = mode == PPMODE? CMD_READ_EEPROM_PP: CMD_READ_EEPROM_HVSP;
cmdlen = 3;
pagesize = mem->page_size;
if (pagesize == 0)
pagesize = 1;
paddr = addr & ~(pagesize - 1);
paddr_ptr = &PDATA(pgm)->eeprom_pageaddr;
cache_ptr = PDATA(pgm)->eeprom_pagecache;
} else if (strcmp(mem->desc, "lfuse") == 0 ||
strcmp(mem->desc, "fuse") == 0) {
buf[0] = mode == PPMODE? CMD_READ_FUSE_PP: CMD_READ_FUSE_HVSP;
addr = 0;
} else if (strcmp(mem->desc, "hfuse") == 0) {
buf[0] = mode == PPMODE? CMD_READ_FUSE_PP: CMD_READ_FUSE_HVSP;
addr = 1;
} else if (strcmp(mem->desc, "efuse") == 0) {
buf[0] = mode == PPMODE? CMD_READ_FUSE_PP: CMD_READ_FUSE_HVSP;
addr = 2;
} else if (strcmp(mem->desc, "lock") == 0) {
buf[0] = mode == PPMODE? CMD_READ_LOCK_PP: CMD_READ_LOCK_HVSP;
} else if (strcmp(mem->desc, "calibration") == 0) {
buf[0] = mode == PPMODE? CMD_READ_OSCCAL_PP: CMD_READ_OSCCAL_HVSP;
} else if (strcmp(mem->desc, "signature") == 0) {
buf[0] = mode == PPMODE? CMD_READ_SIGNATURE_PP: CMD_READ_SIGNATURE_HVSP;
}
/*
* In HV mode, we have to use paged reads for flash and
* EEPROM, and cache the results in a page cache.
*
* Page cache validation is based on "{flash,eeprom}_pageaddr"
* (holding the base address of the most recent cache fill
* operation). This variable is set to (unsigned long)-1L when the
* cache needs to be invalidated.
*/
if (pagesize && paddr == *paddr_ptr) {
*value = cache_ptr[addr & (pagesize - 1)];
return 0;
}
if (cmdlen == 3) {
/* long command, fill in # of bytes */
buf[1] = (pagesize >> 8) & 0xff;
buf[2] = pagesize & 0xff;
/* flash and EEPROM reads require the load address command */
if (stk500v2_loadaddr(pgm, use_ext_addr | (paddr >> addrshift)) < 0)
return -1;
} else {
buf[1] = addr;
}
avrdude_message(MSG_NOTICE2, "%s: stk500hv_read_byte(): Sending read memory command: ",
progname);
result = stk500v2_command(pgm, buf, cmdlen, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500hv_read_byte(): "
"timeout/error communicating with programmer\n",
progname);
return -1;
}
if (pagesize) {
*paddr_ptr = paddr;
memcpy(cache_ptr, buf + 2, pagesize);
*value = cache_ptr[addr & (pagesize - 1)];
} else {
*value = buf[2];
}
return 0;
}
/*
* Read a single byte, PP mode
*/
static int stk500pp_read_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char * value)
{
return stk500hv_read_byte(pgm, p, mem, addr, value, PPMODE);
}
/*
* Read a single byte, HVSP mode
*/
static int stk500hvsp_read_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char * value)
{
return stk500hv_read_byte(pgm, p, mem, addr, value, HVSPMODE);
}
/*
* Read a single byte, ISP mode
*
* By now, only used on the JTAGICE3 which does not implement the
* CMD_SPI_MULTI SPI passthrough command.
*/
static int stk500isp_read_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char * value)
{
int result, pollidx;
unsigned char buf[6];
unsigned long paddr = 0UL, *paddr_ptr = NULL;
unsigned int pagesize = 0;
unsigned char *cache_ptr = NULL;
OPCODE *op;
avrdude_message(MSG_NOTICE2, "%s: stk500isp_read_byte(.., %s, 0x%lx, ...)\n",
progname, mem->desc, addr);
if (strcmp(mem->desc, "flash") == 0 ||
strcmp(mem->desc, "eeprom") == 0) {
// use paged access, and cache result
if (strcmp(mem->desc, "flash") == 0) {
pagesize = PDATA(pgm)->flash_pagesize;
paddr = addr & ~(pagesize - 1);
paddr_ptr = &PDATA(pgm)->flash_pageaddr;
cache_ptr = PDATA(pgm)->flash_pagecache;
} else {
pagesize = mem->page_size;
if (pagesize == 0)
pagesize = 1;
paddr = addr & ~(pagesize - 1);
paddr_ptr = &PDATA(pgm)->eeprom_pageaddr;
cache_ptr = PDATA(pgm)->eeprom_pagecache;
}
if (paddr == *paddr_ptr) {
*value = cache_ptr[addr & (pagesize - 1)];
return 0;
}
if (stk500v2_paged_load(pgm, p, mem, pagesize, paddr, pagesize) < 0)
return -1;
*paddr_ptr = paddr;
memcpy(cache_ptr, &mem->buf[paddr], pagesize);
*value = cache_ptr[addr & (pagesize - 1)];
return 0;
}
if (strcmp(mem->desc, "lfuse") == 0 ||
strcmp(mem->desc, "fuse") == 0) {
buf[0] = CMD_READ_FUSE_ISP;
addr = 0;
} else if (strcmp(mem->desc, "hfuse") == 0) {
buf[0] = CMD_READ_FUSE_ISP;
addr = 1;
} else if (strcmp(mem->desc, "efuse") == 0) {
buf[0] = CMD_READ_FUSE_ISP;
addr = 2;
} else if (strcmp(mem->desc, "lock") == 0) {
buf[0] = CMD_READ_LOCK_ISP;
} else if (strcmp(mem->desc, "calibration") == 0) {
buf[0] = CMD_READ_OSCCAL_ISP;
} else if (strcmp(mem->desc, "signature") == 0) {
buf[0] = CMD_READ_SIGNATURE_ISP;
}
memset(buf + 1, 0, 5);
if ((op = mem->op[AVR_OP_READ]) == NULL) {
avrdude_message(MSG_INFO, "%s: stk500isp_read_byte(): invalid operation AVR_OP_READ on %s memory\n",
progname, mem->desc);
return -1;
}
avr_set_bits(op, buf + 2);
if ((pollidx = avr_get_output_index(op)) == -1) {
avrdude_message(MSG_INFO, "%s: stk500isp_read_byte(): cannot determine pollidx to read %s memory\n",
progname, mem->desc);
pollidx = 3;
}
buf[1] = pollidx + 1;
avr_set_addr(op, buf + 2, addr);
avrdude_message(MSG_NOTICE2, "%s: stk500isp_read_byte(): Sending read memory command: ",
progname);
result = stk500v2_command(pgm, buf, 6, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500isp_read_byte(): "
"timeout/error communicating with programmer\n",
progname);
return -1;
}
*value = buf[2];
return 0;
}
/*
* Write one byte, generic HV mode
*/
static int stk500hv_write_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char data,
enum hvmode mode)
{
int result, cmdlen, timeout = 0, pulsewidth = 0;
unsigned char buf[266];
unsigned long paddr = 0UL, *paddr_ptr = NULL;
unsigned int pagesize = 0, use_ext_addr = 0, addrshift = 0;
unsigned char *cache_ptr = NULL;
avrdude_message(MSG_NOTICE2, "%s: stk500hv_write_byte(.., %s, 0x%lx, ...)\n",
progname, mem->desc, addr);
if (strcmp(mem->desc, "flash") == 0) {
buf[0] = mode == PPMODE? CMD_PROGRAM_FLASH_PP: CMD_PROGRAM_FLASH_HVSP;
pagesize = PDATA(pgm)->flash_pagesize;
paddr = addr & ~(pagesize - 1);
paddr_ptr = &PDATA(pgm)->flash_pageaddr;
cache_ptr = PDATA(pgm)->flash_pagecache;
addrshift = 1;
/*
* If bit 31 is set, this indicates that the following read/write
* operation will be performed on a memory that is larger than
* 64KBytes. This is an indication to STK500 that a load extended
* address must be executed.
*/
if (mem->op[AVR_OP_LOAD_EXT_ADDR] != NULL) {
use_ext_addr = (1U << 31);
}
} else if (strcmp(mem->desc, "eeprom") == 0) {
buf[0] = mode == PPMODE? CMD_PROGRAM_EEPROM_PP: CMD_PROGRAM_EEPROM_HVSP;
pagesize = mem->page_size;
if (pagesize == 0)
pagesize = 1;
paddr = addr & ~(pagesize - 1);
paddr_ptr = &PDATA(pgm)->eeprom_pageaddr;
cache_ptr = PDATA(pgm)->eeprom_pagecache;
} else if (strcmp(mem->desc, "lfuse") == 0 ||
strcmp(mem->desc, "fuse") == 0) {
buf[0] = mode == PPMODE? CMD_PROGRAM_FUSE_PP: CMD_PROGRAM_FUSE_HVSP;
addr = 0;
pulsewidth = p->programfusepulsewidth;
timeout = p->programfusepolltimeout;
} else if (strcmp(mem->desc, "hfuse") == 0) {
buf[0] = mode == PPMODE? CMD_PROGRAM_FUSE_PP: CMD_PROGRAM_FUSE_HVSP;
addr = 1;
pulsewidth = p->programfusepulsewidth;
timeout = p->programfusepolltimeout;
} else if (strcmp(mem->desc, "efuse") == 0) {
buf[0] = mode == PPMODE? CMD_PROGRAM_FUSE_PP: CMD_PROGRAM_FUSE_HVSP;
addr = 2;
pulsewidth = p->programfusepulsewidth;
timeout = p->programfusepolltimeout;
} else if (strcmp(mem->desc, "lock") == 0) {
buf[0] = mode == PPMODE? CMD_PROGRAM_LOCK_PP: CMD_PROGRAM_LOCK_HVSP;
pulsewidth = p->programlockpulsewidth;
timeout = p->programlockpolltimeout;
} else {
avrdude_message(MSG_INFO, "%s: stk500hv_write_byte(): "
"unsupported memory type: %s\n",
progname, mem->desc);
return -1;
}
cmdlen = 5 + pagesize;
/*
* In HV mode, we have to use paged writes for flash and
* EEPROM. As both, flash and EEPROM cells can only be programmed
* from `1' to `0' bits (even EEPROM does not support auto-erase in
* parallel mode), we just pre-fill the page cache with 0xff, so all
* those cells that are outside our current address will remain
* unaffected.
*/
if (pagesize) {
memset(cache_ptr, 0xff, pagesize);
cache_ptr[addr & (pagesize - 1)] = data;
/* long command, fill in # of bytes */
buf[1] = (pagesize >> 8) & 0xff;
buf[2] = pagesize & 0xff;
/*
* Synthesize the mode byte. This is simpler than adding yet
* another parameter to the avrdude.conf file. We calculate the
* bits corresponding to the page size, as explained in AVR068.
* We set bit 7, to indicate this is to actually write the page to
* the target device. We set bit 6 to indicate this is the very
* last page to be programmed, whatever this means -- we just
* pretend we don't know any better. ;-) Bit 0 is set if this is
* a paged memory, which means it has a page size of more than 2.
*/
buf[3] = 0x80 | 0x40;
if (pagesize > 2) {
unsigned int rv = stk500v2_mode_for_pagesize(pagesize);
if (rv == 0)
return -1;
buf[3] |= rv;
buf[3] |= 0x01;
}
buf[4] = mem->delay;
memcpy(buf + 5, cache_ptr, pagesize);
/* flash and EEPROM reads require the load address command */
if (stk500v2_loadaddr(pgm, use_ext_addr | (paddr >> addrshift)) < 0)
return -1;
} else {
buf[1] = addr;
buf[2] = data;
if (mode == PPMODE) {
buf[3] = pulsewidth;
buf[4] = timeout;
} else {
buf[3] = timeout;
cmdlen--;
}
}
avrdude_message(MSG_NOTICE2, "%s: stk500hv_write_byte(): Sending write memory command: ",
progname);
result = stk500v2_command(pgm, buf, cmdlen, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500hv_write_byte(): "
"timeout/error communicating with programmer\n",
progname);
return -1;
}
if (pagesize) {
/* Invalidate the page cache. */
*paddr_ptr = (unsigned long)-1L;
}
return 0;
}
/*
* Write one byte, PP mode
*/
static int stk500pp_write_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char data)
{
return stk500hv_write_byte(pgm, p, mem, addr, data, PPMODE);
}
/*
* Write one byte, HVSP mode
*/
static int stk500hvsp_write_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char data)
{
return stk500hv_write_byte(pgm, p, mem, addr, data, HVSPMODE);
}
/*
* Write one byte, ISP mode
*/
static int stk500isp_write_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char data)
{
int result;
unsigned char buf[5];
unsigned long paddr = 0UL, *paddr_ptr = NULL;
unsigned int pagesize = 0;
unsigned char *cache_ptr = NULL;
OPCODE *op;
avrdude_message(MSG_NOTICE2, "%s: stk500isp_write_byte(.., %s, 0x%lx, ...)\n",
progname, mem->desc, addr);
if (strcmp(mem->desc, "flash") == 0 ||
strcmp(mem->desc, "eeprom") == 0) {
if (strcmp(mem->desc, "flash") == 0) {
pagesize = PDATA(pgm)->flash_pagesize;
paddr = addr & ~(pagesize - 1);
paddr_ptr = &PDATA(pgm)->flash_pageaddr;
cache_ptr = PDATA(pgm)->flash_pagecache;
} else {
pagesize = mem->page_size;
if (pagesize == 0)
pagesize = 1;
paddr = addr & ~(pagesize - 1);
paddr_ptr = &PDATA(pgm)->eeprom_pageaddr;
cache_ptr = PDATA(pgm)->eeprom_pagecache;
}
/*
* We use paged writes for flash and EEPROM, reading back the
* current page first, modify the byte to write, and write out the
* entire page.
*/
if (stk500v2_paged_load(pgm, p, mem, pagesize, paddr, pagesize) < 0)
return -1;
memcpy(cache_ptr, mem->buf + paddr, pagesize);
*paddr_ptr = paddr;
cache_ptr[addr & (pagesize - 1)] = data;
memcpy(mem->buf + paddr, cache_ptr, pagesize);
stk500v2_paged_write(pgm, p, mem, pagesize, paddr, pagesize);
return 0;
}
memset(buf, 0, sizeof buf);
if (strcmp(mem->desc, "lfuse") == 0 ||
strcmp(mem->desc, "fuse") == 0) {
buf[0] = CMD_PROGRAM_FUSE_ISP;
addr = 0;
} else if (strcmp(mem->desc, "hfuse") == 0) {
buf[0] = CMD_PROGRAM_FUSE_ISP;
addr = 1;
} else if (strcmp(mem->desc, "efuse") == 0) {
buf[0] = CMD_PROGRAM_FUSE_ISP;
addr = 2;
} else if (strcmp(mem->desc, "lock") == 0) {
buf[0] = CMD_PROGRAM_LOCK_ISP;
} else {
avrdude_message(MSG_INFO, "%s: stk500isp_write_byte(): "
"unsupported memory type: %s\n",
progname, mem->desc);
return -1;
}
if ((op = mem->op[AVR_OP_WRITE]) == NULL) {
avrdude_message(MSG_INFO, "%s: stk500isp_write_byte(): "
"no AVR_OP_WRITE for %s memory\n",
progname, mem->desc);
return -1;
}
avr_set_bits(op, buf + 1);
avr_set_addr(op, buf + 1, addr);
avr_set_input(op, buf + 1, data);
avrdude_message(MSG_NOTICE2, "%s: stk500isp_write_byte(): Sending write memory command: ",
progname);
result = stk500v2_command(pgm, buf, 5, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500isp_write_byte(): "
"timeout/error communicating with programmer\n",
progname);
return -1;
}
/*
* Prevent verification readback to be too fast, see
* https://savannah.nongnu.org/bugs/index.php?42267
*
* After all, this is just an ugly hack working around some
* brokeness in the Atmel firmware starting with the AVRISPmkII (the
* old JTAGICEmkII isn't affected). Let's hope 10 ms of additional
* delay are good enough for everyone.
*/
usleep(10000);
return 0;
}
static int stk500v2_paged_write(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
unsigned int block_size, last_addr, addrshift, use_ext_addr;
unsigned int maxaddr = addr + n_bytes;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
unsigned char commandbuf[10];
unsigned char buf[266];
unsigned char cmds[4];
int result;
OPCODE * rop, * wop;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
DEBUG("STK500V2: stk500v2_paged_write(..,%s,%u,%u,%u)\n",
m->desc, page_size, addr, n_bytes);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
if (page_size == 0) page_size = 256;
addrshift = 0;
use_ext_addr = 0;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
// determine which command is to be used
if (strcmp(m->desc, "flash") == 0) {
addrshift = 1;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
commandbuf[0] = CMD_PROGRAM_FLASH_ISP;
/*
* If bit 31 is set, this indicates that the following read/write
* operation will be performed on a memory that is larger than
* 64KBytes. This is an indication to STK500 that a load extended
* address must be executed.
*/
if (m->op[AVR_OP_LOAD_EXT_ADDR] != NULL) {
use_ext_addr = (1U << 31);
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
} else if (strcmp(m->desc, "eeprom") == 0) {
commandbuf[0] = CMD_PROGRAM_EEPROM_ISP;
}
commandbuf[4] = m->delay;
if (addrshift == 0) {
wop = m->op[AVR_OP_WRITE];
rop = m->op[AVR_OP_READ];
}
else {
wop = m->op[AVR_OP_WRITE_LO];
rop = m->op[AVR_OP_READ_LO];
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
// if the memory is paged, load the appropriate commands into the buffer
if (m->mode & 0x01) {
commandbuf[3] = m->mode | 0x80; // yes, write the page to flash
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
if (m->op[AVR_OP_LOADPAGE_LO] == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_paged_write: loadpage instruction not defined for part \"%s\"\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, p->desc);
return -1;
}
avr_set_bits(m->op[AVR_OP_LOADPAGE_LO], cmds);
commandbuf[5] = cmds[0];
if (m->op[AVR_OP_WRITEPAGE] == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_paged_write: write page instruction not defined for part \"%s\"\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, p->desc);
return -1;
}
avr_set_bits(m->op[AVR_OP_WRITEPAGE], cmds);
commandbuf[6] = cmds[0];
// otherwise, we need to load different commands in
}
else {
commandbuf[3] = m->mode | 0x80; // yes, write the words to flash
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
if (wop == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_paged_write: write instruction not defined for part \"%s\"\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, p->desc);
return -1;
}
avr_set_bits(wop, cmds);
commandbuf[5] = cmds[0];
commandbuf[6] = 0;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
// the read command is common to both methods
if (rop == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_paged_write: read instruction not defined for part \"%s\"\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, p->desc);
return -1;
}
avr_set_bits(rop, cmds);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
commandbuf[7] = cmds[0];
commandbuf[8] = m->readback[0];
commandbuf[9] = m->readback[1];
last_addr=UINT_MAX; /* impossible address */
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
for (; addr < maxaddr; addr += page_size) {
if ((maxaddr - addr) < page_size)
block_size = maxaddr - addr;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
else
block_size = page_size;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
DEBUG("block_size at addr %d is %d\n",addr,block_size);
memcpy(buf,commandbuf,sizeof(commandbuf));
buf[1] = block_size >> 8;
buf[2] = block_size & 0xff;
if((last_addr==UINT_MAX)||(last_addr+block_size != addr)){
if (stk500v2_loadaddr(pgm, use_ext_addr | (addr >> addrshift)) < 0)
return -1;
}
last_addr=addr;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
memcpy(buf+10,m->buf+addr, block_size);
result = stk500v2_command(pgm,buf,block_size+10, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_paged_write: write command failed\n",
progname);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -1;
}
}
return n_bytes;
}
/*
* Write pages of flash/EEPROM, generic HV mode
*/
static int stk500hv_paged_write(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes,
enum hvmode mode)
{
unsigned int block_size, last_addr, addrshift, use_ext_addr;
unsigned int maxaddr = addr + n_bytes;
unsigned char commandbuf[5], buf[266];
int result;
DEBUG("STK500V2: stk500hv_paged_write(..,%s,%u,%u)\n",
m->desc, page_size, addr, n_bytes);
addrshift = 0;
use_ext_addr = 0;
// determine which command is to be used
if (strcmp(m->desc, "flash") == 0) {
addrshift = 1;
PDATA(pgm)->flash_pageaddr = (unsigned long)-1L;
commandbuf[0] = mode == PPMODE? CMD_PROGRAM_FLASH_PP: CMD_PROGRAM_FLASH_HVSP;
/*
* If bit 31 is set, this indicates that the following read/write
* operation will be performed on a memory that is larger than
* 64KBytes. This is an indication to STK500 that a load extended
* address must be executed.
*/
if (m->op[AVR_OP_LOAD_EXT_ADDR] != NULL) {
use_ext_addr = (1U << 31);
}
} else if (strcmp(m->desc, "eeprom") == 0) {
PDATA(pgm)->eeprom_pageaddr = (unsigned long)-1L;
commandbuf[0] = mode == PPMODE? CMD_PROGRAM_EEPROM_PP: CMD_PROGRAM_EEPROM_HVSP;
}
/*
* Synthesize the mode byte. This is simpler than adding yet
* another parameter to the avrdude.conf file. We calculate the
* bits corresponding to the page size, as explained in AVR068. We
* set bit 7, to indicate this is to actually write the page to the
* target device. We set bit 6 to indicate this is the very last
* page to be programmed, whatever this means -- we just pretend we
* don't know any better. ;-) Finally, we set bit 0 to say this is
* a paged memory, after all, that's why we got here at all.
*/
commandbuf[3] = 0x80 | 0x40;
if (page_size > 2) {
unsigned int rv = stk500v2_mode_for_pagesize(page_size);
if (rv == 0)
return -1;
commandbuf[3] |= rv;
commandbuf[3] |= 0x01;
}
commandbuf[4] = m->delay;
if (page_size == 0) page_size = 256;
last_addr = UINT_MAX; /* impossible address */
for (; addr < maxaddr; addr += page_size) {
if ((maxaddr - addr) < page_size)
block_size = maxaddr - addr;
else
block_size = page_size;
DEBUG("block_size at addr %d is %d\n",addr,block_size);
memcpy(buf, commandbuf, sizeof(commandbuf));
buf[1] = page_size >> 8;
buf[2] = page_size & 0xff;
if ((last_addr == UINT_MAX) || (last_addr + block_size != addr)) {
if (stk500v2_loadaddr(pgm, use_ext_addr | (addr >> addrshift)) < 0)
return -1;
}
last_addr=addr;
memcpy(buf + 5, m->buf + addr, block_size);
if (block_size != page_size)
memset(buf + 5 + block_size, 0xff, page_size - block_size);
result = stk500v2_command(pgm, buf, page_size + 5, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500hv_paged_write: write command failed\n",
progname);
return -1;
}
}
return n_bytes;
}
/*
* Write pages of flash/EEPROM, PP mode
*/
static int stk500pp_paged_write(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes)
{
return stk500hv_paged_write(pgm, p, m, page_size, addr, n_bytes, PPMODE);
}
/*
* Write pages of flash/EEPROM, HVSP mode
*/
static int stk500hvsp_paged_write(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
return stk500hv_paged_write(pgm, p, m, page_size, addr, n_bytes, HVSPMODE);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
static int stk500v2_paged_load(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
unsigned int block_size, hiaddr, addrshift, use_ext_addr;
unsigned int maxaddr = addr + n_bytes;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
unsigned char commandbuf[4];
unsigned char buf[275]; // max buffer size for stk500v2 at this point
unsigned char cmds[4];
int result;
OPCODE * rop;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
DEBUG("STK500V2: stk500v2_paged_load(..,%s,%u,%u,%u)\n",
m->desc, page_size, addr, n_bytes);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
page_size = m->readsize;
rop = m->op[AVR_OP_READ];
hiaddr = UINT_MAX;
addrshift = 0;
use_ext_addr = 0;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
// determine which command is to be used
if (strcmp(m->desc, "flash") == 0) {
commandbuf[0] = CMD_READ_FLASH_ISP;
rop = m->op[AVR_OP_READ_LO];
addrshift = 1;
/*
* If bit 31 is set, this indicates that the following read/write
* operation will be performed on a memory that is larger than
* 64KBytes. This is an indication to STK500 that a load extended
* address must be executed.
*/
if (m->op[AVR_OP_LOAD_EXT_ADDR] != NULL) {
use_ext_addr = (1U << 31);
}
}
else if (strcmp(m->desc, "eeprom") == 0) {
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
commandbuf[0] = CMD_READ_EEPROM_ISP;
}
// the read command is common to both methods
if (rop == NULL) {
avrdude_message(MSG_INFO, "%s: stk500v2_paged_load: read instruction not defined for part \"%s\"\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, p->desc);
return -1;
}
avr_set_bits(rop, cmds);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
commandbuf[3] = cmds[0];
for (; addr < maxaddr; addr += page_size) {
if ((maxaddr - addr) < page_size)
block_size = maxaddr - addr;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
else
block_size = page_size;
DEBUG("block_size at addr %d is %d\n",addr,block_size);
memcpy(buf,commandbuf,sizeof(commandbuf));
buf[1] = block_size >> 8;
buf[2] = block_size & 0xff;
// Ensure a new "load extended address" will be issued
// when crossing a 64 KB boundary in flash.
if (hiaddr != (addr & ~0xFFFF)) {
hiaddr = addr & ~0xFFFF;
if (stk500v2_loadaddr(pgm, use_ext_addr | (addr >> addrshift)) < 0)
return -1;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
result = stk500v2_command(pgm,buf,4,sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_paged_load: read command failed\n",
progname);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -1;
}
#if 0
for (i=0;i<page_size;i++) {
avrdude_message(MSG_INFO, "%02X",buf[2+i]);
if (i%16 == 15) avrdude_message(MSG_INFO, "\n");
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
#endif
memcpy(&m->buf[addr], &buf[2], block_size);
}
return n_bytes;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
/*
* Read pages of flash/EEPROM, generic HV mode
*/
static int stk500hv_paged_load(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes,
enum hvmode mode)
{
unsigned int block_size, hiaddr, addrshift, use_ext_addr;
unsigned int maxaddr = addr + n_bytes;
unsigned char commandbuf[3], buf[266];
int result;
DEBUG("STK500V2: stk500hv_paged_load(..,%s,%u,%u,%u)\n",
m->desc, page_size, addr, n_bytes);
page_size = m->readsize;
hiaddr = UINT_MAX;
addrshift = 0;
use_ext_addr = 0;
// determine which command is to be used
if (strcmp(m->desc, "flash") == 0) {
commandbuf[0] = mode == PPMODE? CMD_READ_FLASH_PP: CMD_READ_FLASH_HVSP;
addrshift = 1;
/*
* If bit 31 is set, this indicates that the following read/write
* operation will be performed on a memory that is larger than
* 64KBytes. This is an indication to STK500 that a load extended
* address must be executed.
*/
if (m->op[AVR_OP_LOAD_EXT_ADDR] != NULL) {
use_ext_addr = (1U << 31);
}
}
else if (strcmp(m->desc, "eeprom") == 0) {
commandbuf[0] = mode == PPMODE? CMD_READ_EEPROM_PP: CMD_READ_EEPROM_HVSP;
}
for (; addr < maxaddr; addr += page_size) {
if ((maxaddr - addr) < page_size)
block_size = maxaddr - addr;
else
block_size = page_size;
DEBUG("block_size at addr %d is %d\n",addr,block_size);
memcpy(buf, commandbuf, sizeof(commandbuf));
buf[1] = block_size >> 8;
buf[2] = block_size & 0xff;
// Ensure a new "load extended address" will be issued
// when crossing a 64 KB boundary in flash.
if (hiaddr != (addr & ~0xFFFF)) {
hiaddr = addr & ~0xFFFF;
if (stk500v2_loadaddr(pgm, use_ext_addr | (addr >> addrshift)) < 0)
return -1;
}
result = stk500v2_command(pgm, buf, 3, sizeof(buf));
if (result < 0) {
avrdude_message(MSG_INFO, "%s: stk500hv_paged_load: read command failed\n",
progname);
return -1;
}
#if 0
for (i = 0; i < page_size; i++) {
avrdude_message(MSG_INFO, "%02X", buf[2 + i]);
if (i % 16 == 15) avrdude_message(MSG_INFO, "\n");
}
#endif
memcpy(&m->buf[addr], &buf[2], block_size);
}
return n_bytes;
}
/*
* Read pages of flash/EEPROM, PP mode
*/
static int stk500pp_paged_load(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes)
{
return stk500hv_paged_load(pgm, p, m, page_size, addr, n_bytes, PPMODE);
}
/*
* Read pages of flash/EEPROM, HVSP mode
*/
static int stk500hvsp_paged_load(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes)
{
return stk500hv_paged_load(pgm, p, m, page_size, addr, n_bytes, HVSPMODE);
}
static int stk500v2_page_erase(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int addr)
{
avrdude_message(MSG_INFO, "%s: stk500v2_page_erase(): this function must never be called\n",
progname);
return -1;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static int stk500v2_set_vtarget(PROGRAMMER * pgm, double v)
{
unsigned char uaref, utarg;
utarg = (unsigned)((v + 0.049) * 10);
if (stk500v2_getparm(pgm, PARAM_VADJUST, &uaref) != 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_vtarget(): cannot obtain V[aref]\n",
progname);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -1;
}
if (uaref > utarg) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_vtarget(): reducing V[aref] from %.1f to %.1f\n",
progname, uaref / 10.0, v);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
if (stk500v2_setparm(pgm, PARAM_VADJUST, utarg)
!= 0)
return -1;
}
return stk500v2_setparm(pgm, PARAM_VTARGET, utarg);
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
static int stk500v2_set_varef(PROGRAMMER * pgm, unsigned int chan /* unused */,
double v)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
unsigned char uaref, utarg;
uaref = (unsigned)((v + 0.049) * 10);
if (stk500v2_getparm(pgm, PARAM_VTARGET, &utarg) != 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_varef(): cannot obtain V[target]\n",
progname);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -1;
}
if (uaref > utarg) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_varef(): V[aref] must not be greater than "
"V[target] = %.1f\n",
progname, utarg / 10.0);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return -1;
}
return stk500v2_setparm(pgm, PARAM_VADJUST, uaref);
}
static int stk500v2_set_fosc(PROGRAMMER * pgm, double v)
{
int fosc;
unsigned char prescale, cmatch;
static unsigned ps[] = {
1, 8, 32, 64, 128, 256, 1024
};
int idx, rc;
prescale = cmatch = 0;
if (v > 0.0) {
if (v > STK500V2_XTAL / 2) {
const char *unit;
if (v > 1e6) {
v /= 1e6;
unit = "MHz";
} else if (v > 1e3) {
v /= 1e3;
unit = "kHz";
} else
unit = "Hz";
avrdude_message(MSG_INFO, "%s: stk500v2_set_fosc(): f = %.3f %s too high, using %.3f MHz\n",
progname, v, unit, STK500V2_XTAL / 2e6);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
fosc = STK500V2_XTAL / 2;
} else
fosc = (unsigned)v;
for (idx = 0; idx < sizeof(ps) / sizeof(ps[0]); idx++) {
if (fosc >= STK500V2_XTAL / (256 * ps[idx] * 2)) {
/* this prescaler value can handle our frequency */
prescale = idx + 1;
cmatch = (unsigned)(STK500V2_XTAL / (2 * fosc * ps[idx])) - 1;
break;
}
}
if (idx == sizeof(ps) / sizeof(ps[0])) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_fosc(): f = %u Hz too low, %u Hz min\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, fosc, STK500V2_XTAL / (256 * 1024 * 2));
return -1;
}
}
if ((rc = stk500v2_setparm(pgm, PARAM_OSC_PSCALE, prescale)) != 0
|| (rc = stk500v2_setparm(pgm, PARAM_OSC_CMATCH, cmatch)) != 0)
return rc;
return 0;
}
/* The list of SCK frequencies supported by the AVRISP mkII, as listed
* in AVR069 */
static double avrispmkIIfreqs[] = {
8000000, 4000000, 2000000, 1000000, 500000, 250000, 125000,
96386, 89888, 84211, 79208, 74767, 70797, 67227, 64000,
61069, 58395, 55945, 51613, 49690, 47905, 46243, 43244,
41885, 39409, 38278, 36200, 34335, 32654, 31129, 29740,
28470, 27304, 25724, 24768, 23461, 22285, 21221, 20254,
19371, 18562, 17583, 16914, 16097, 15356, 14520, 13914,
13224, 12599, 12031, 11511, 10944, 10431, 9963, 9468,
9081, 8612, 8239, 7851, 7498, 7137, 6809, 6478, 6178,
5879, 5607, 5359, 5093, 4870, 4633, 4418, 4209, 4019,
3823, 3645, 3474, 3310, 3161, 3011, 2869, 2734, 2611,
2484, 2369, 2257, 2152, 2052, 1956, 1866, 1779, 1695,
1615, 1539, 1468, 1398, 1333, 1271, 1212, 1155, 1101,
1049, 1000, 953, 909, 866, 826, 787, 750, 715, 682,
650, 619, 590, 563, 536, 511, 487, 465, 443, 422,
402, 384, 366, 349, 332, 317, 302, 288, 274, 261,
249, 238, 226, 216, 206, 196, 187, 178, 170, 162,
154, 147, 140, 134, 128, 122, 116, 111, 105, 100,
95.4, 90.9, 86.6, 82.6, 78.7, 75.0, 71.5, 68.2,
65.0, 61.9, 59.0, 56.3, 53.6, 51.1
};
static int stk500v2_set_sck_period_mk2(PROGRAMMER * pgm, double v)
{
int i;
for (i = 0; i < sizeof(avrispmkIIfreqs); i++) {
if (1 / avrispmkIIfreqs[i] >= v)
break;
}
avrdude_message(MSG_NOTICE2, "Using p = %.2f us for SCK (param = %d)\n",
1000000 / avrispmkIIfreqs[i], i);
return stk500v2_setparm(pgm, PARAM_SCK_DURATION, i);
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* Return the "mode" value for the parallel and HVSP modes that
* corresponds to the pagesize.
*/
static unsigned int stk500v2_mode_for_pagesize(unsigned int pagesize)
{
switch (pagesize)
{
case 256: return 0u << 1;
case 2: return 1u << 1;
case 4: return 2u << 1;
case 8: return 3u << 1;
case 16: return 4u << 1;
case 32: return 5u << 1;
case 64: return 6u << 1;
case 128: return 7u << 1;
}
avrdude_message(MSG_INFO, "%s: stk500v2_mode_for_pagesize(): invalid pagesize: %u\n",
progname, pagesize);
return 0;
}
/*
* See pseudo-code in AVR068
*
* This algorithm only fits for the STK500 itself. For the (old)
* AVRISP, the resulting ISP clock is only half. While this would be
* easy to fix in the algorithm, we'd need to add another
* configuration flag for this to the config file. Given the old
* AVRISP devices are virtually no longer around (and the AVRISPmkII
* uses a different algorithm below), it's probably not worth the
* hassle.
*/
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static int stk500v2_set_sck_period(PROGRAMMER * pgm, double v)
{
unsigned int d;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
unsigned char dur;
double f = 1 / v;
if (f >= 1.8432E6)
d = 0;
else if (f > 460.8E3)
d = 1;
else if (f > 115.2E3)
d = 2;
else if (f > 57.6E3)
d = 3;
else
d = (unsigned int)ceil(1 / (24 * f / (double)STK500V2_XTAL) - 10.0 / 12.0);
if (d >= 255)
d = 254;
dur = d;
return stk500v2_setparm(pgm, PARAM_SCK_DURATION, dur);
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static double stk500v2_sck_to_us(PROGRAMMER * pgm, unsigned char dur)
{
double x;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
if (dur == 0)
return 0.5425;
if (dur == 1)
return 2.17;
if (dur == 2)
return 8.68;
if (dur == 3)
return 17.36;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
x = (double)dur + 10.0 / 12.0;
x = 1.0 / x;
x /= 24.0;
x *= (double)STK500V2_XTAL;
return 1E6 / x;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
static int stk600_set_vtarget(PROGRAMMER * pgm, double v)
{
unsigned char utarg;
unsigned int uaref;
int rv;
utarg = (unsigned)((v + 0.049) * 10);
if (stk500v2_getparm2(pgm, PARAM2_AREF0, &uaref) != 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_vtarget(): cannot obtain V[aref][0]\n",
progname);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
return -1;
}
if (uaref > (unsigned)utarg * 10) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_vtarget(): reducing V[aref][0] from %.2f to %.1f\n",
progname, uaref / 100.0, v);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
uaref = 10 * (unsigned)utarg;
if (stk500v2_setparm2(pgm, PARAM2_AREF0, uaref)
!= 0)
return -1;
}
if (stk500v2_getparm2(pgm, PARAM2_AREF1, &uaref) != 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_vtarget(): cannot obtain V[aref][1]\n",
progname);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
return -1;
}
if (uaref > (unsigned)utarg * 10) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_vtarget(): reducing V[aref][1] from %.2f to %.1f\n",
progname, uaref / 100.0, v);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
uaref = 10 * (unsigned)utarg;
if (stk500v2_setparm2(pgm, PARAM2_AREF1, uaref)
!= 0)
return -1;
}
/*
* Vtarget on the STK600 can only be adjusted while not being in
* programming mode.
*/
if (PDATA(pgm)->lastpart)
pgm->disable(pgm);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
rv = stk500v2_setparm(pgm, PARAM_VTARGET, utarg);
if (PDATA(pgm)->lastpart)
pgm->program_enable(pgm, PDATA(pgm)->lastpart);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
return rv;
}
static int stk600_set_varef(PROGRAMMER * pgm, unsigned int chan, double v)
{
unsigned char utarg;
unsigned int uaref;
uaref = (unsigned)((v + 0.0049) * 100);
if (stk500v2_getparm(pgm, PARAM_VTARGET, &utarg) != 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_varef(): cannot obtain V[target]\n",
progname);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
return -1;
}
if (uaref > (unsigned)utarg * 10) {
avrdude_message(MSG_INFO, "%s: stk500v2_set_varef(): V[aref] must not be greater than "
"V[target] = %.1f\n",
progname, utarg / 10.0);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
return -1;
}
switch (chan)
{
case 0:
return stk500v2_setparm2(pgm, PARAM2_AREF0, uaref);
case 1:
return stk500v2_setparm2(pgm, PARAM2_AREF1, uaref);
default:
avrdude_message(MSG_INFO, "%s: stk500v2_set_varef(): invalid channel %d\n",
progname, chan);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
return -1;
}
}
static int stk600_set_fosc(PROGRAMMER * pgm, double v)
{
unsigned int oct, dac;
oct = 1.443 * log(v / 1039.0);
dac = 2048 - (2078.0 * pow(2, (double)(10 + oct))) / v;
return stk500v2_setparm2(pgm, PARAM2_CLOCK_CONF, (oct << 12) | (dac << 2));
}
static int stk600_set_sck_period(PROGRAMMER * pgm, double v)
{
unsigned int sck;
sck = ceil((16e6 / (2 * 1.0 / v)) - 1);
if (sck >= 4096)
sck = 4095;
return stk500v2_setparm2(pgm, PARAM2_SCK_DURATION, sck);
}
static int stk500v2_jtag3_set_sck_period(PROGRAMMER * pgm, double v)
{
unsigned char value[3];
unsigned int sck;
if (v < 1E-6)
sck = 0x400;
else if (v > 1E-3)
sck = 1;
else
sck = 1.0 / (1000.0 * v);
value[0] = CMD_SET_SCK;
value[1] = sck & 0xff;
value[2] = (sck >> 8) & 0xff;
if (stk500v2_jtag3_send(pgm, value, 3) < 0)
return -1;
if (stk500v2_jtag3_recv(pgm, value, 3) < 0)
return -1;
return 0;
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static int stk500v2_getparm(PROGRAMMER * pgm, unsigned char parm, unsigned char * value)
{
unsigned char buf[32];
buf[0] = CMD_GET_PARAMETER;
buf[1] = parm;
if (stk500v2_command(pgm, buf, 2, sizeof(buf)) < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_getparm(): failed to get parameter 0x%02x\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, parm);
return -1;
}
*value = buf[2];
return 0;
}
static int stk500v2_setparm_real(PROGRAMMER * pgm, unsigned char parm, unsigned char value)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
unsigned char buf[32];
buf[0] = CMD_SET_PARAMETER;
buf[1] = parm;
buf[2] = value;
if (stk500v2_command(pgm, buf, 3, sizeof(buf)) < 0) {
avrdude_message(MSG_INFO, "\n%s: stk500v2_setparm(): failed to set parameter 0x%02x\n",
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
progname, parm);
return -1;
}
return 0;
}
static int stk500v2_setparm(PROGRAMMER * pgm, unsigned char parm, unsigned char value)
{
unsigned char current_value;
int res;
res = stk500v2_getparm(pgm, parm, &current_value);
if (res < 0)
avrdude_message(MSG_INFO, "%s: Unable to get parameter 0x%02x\n", progname, parm);
// don't issue a write if the correct value is already set.
if (value == current_value) {
avrdude_message(MSG_NOTICE2, "%s: Skipping parameter write; parameter value already set.\n", progname);
return 0;
}
return stk500v2_setparm_real(pgm, parm, value);
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
static int stk500v2_getparm2(PROGRAMMER * pgm, unsigned char parm, unsigned int * value)
{
unsigned char buf[32];
buf[0] = CMD_GET_PARAMETER;
buf[1] = parm;
if (stk500v2_command(pgm, buf, 2, sizeof(buf)) < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_getparm2(): failed to get parameter 0x%02x\n",
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
progname, parm);
return -1;
}
*value = ((unsigned)buf[2] << 8) | buf[3];
return 0;
}
static int stk500v2_setparm2(PROGRAMMER * pgm, unsigned char parm, unsigned int value)
{
unsigned char buf[32];
buf[0] = CMD_SET_PARAMETER;
buf[1] = parm;
buf[2] = value >> 8;
buf[3] = value;
if (stk500v2_command(pgm, buf, 4, sizeof(buf)) < 0) {
avrdude_message(MSG_INFO, "\n%s: stk500v2_setparm2(): failed to set parameter 0x%02x\n",
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
progname, parm);
return -1;
}
return 0;
}
static const char *stk600_get_cardname(const struct carddata *table,
size_t nele, int id)
{
const struct carddata *cdp;
if (id == 0xFF)
/* 0xFF means this card is not present at all. */
return "Not present";
for (cdp = table; nele > 0; cdp++, nele--)
if (cdp->id == id)
return cdp->name;
return "Unknown";
}
static void stk500v2_display(PROGRAMMER * pgm, const char * p)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
unsigned char maj, min, hdw, topcard, maj_s1, min_s1, maj_s2, min_s2;
unsigned int rev;
const char *topcard_name, *pgmname;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
switch (PDATA(pgm)->pgmtype) {
case PGMTYPE_UNKNOWN: pgmname = "Unknown"; break;
case PGMTYPE_STK500: pgmname = "STK500"; break;
case PGMTYPE_AVRISP: pgmname = "AVRISP"; break;
case PGMTYPE_AVRISP_MKII: pgmname = "AVRISP mkII"; break;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
case PGMTYPE_STK600: pgmname = "STK600"; break;
default: pgmname = "None";
}
if (PDATA(pgm)->pgmtype != PGMTYPE_JTAGICE_MKII &&
PDATA(pgm)->pgmtype != PGMTYPE_JTAGICE3) {
avrdude_message(MSG_INFO, "%sProgrammer Model: %s\n", p, pgmname);
stk500v2_getparm(pgm, PARAM_HW_VER, &hdw);
stk500v2_getparm(pgm, PARAM_SW_MAJOR, &maj);
stk500v2_getparm(pgm, PARAM_SW_MINOR, &min);
avrdude_message(MSG_INFO, "%sHardware Version: %d\n", p, hdw);
avrdude_message(MSG_INFO, "%sFirmware Version Master : %d.%02d\n", p, maj, min);
if (PDATA(pgm)->pgmtype == PGMTYPE_STK600) {
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
stk500v2_getparm(pgm, PARAM_SW_MAJOR_SLAVE1, &maj_s1);
stk500v2_getparm(pgm, PARAM_SW_MINOR_SLAVE1, &min_s1);
stk500v2_getparm(pgm, PARAM_SW_MAJOR_SLAVE2, &maj_s2);
stk500v2_getparm(pgm, PARAM_SW_MINOR_SLAVE2, &min_s2);
avrdude_message(MSG_INFO, "%sFirmware Version Slave 1: %d.%02d\n", p, maj_s1, min_s1);
avrdude_message(MSG_INFO, "%sFirmware Version Slave 2: %d.%02d\n", p, maj_s2, min_s2);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
}
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
if (PDATA(pgm)->pgmtype == PGMTYPE_STK500) {
stk500v2_getparm(pgm, PARAM_TOPCARD_DETECT, &topcard);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
switch (topcard) {
case 0xAA: topcard_name = "STK501"; break;
case 0x55: topcard_name = "STK502"; break;
case 0xFA: topcard_name = "STK503"; break;
case 0xEE: topcard_name = "STK504"; break;
case 0xE4: topcard_name = "STK505"; break;
case 0xDD: topcard_name = "STK520"; break;
default: topcard_name = "Unknown"; break;
}
avrdude_message(MSG_INFO, "%sTopcard : %s\n", p, topcard_name);
} else if (PDATA(pgm)->pgmtype == PGMTYPE_STK600) {
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
stk500v2_getparm(pgm, PARAM_ROUTINGCARD_ID, &topcard);
avrdude_message(MSG_INFO, "%sRouting card : %s\n", p,
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
stk600_get_cardname(routing_cards,
sizeof routing_cards / sizeof routing_cards[0],
topcard));
stk500v2_getparm(pgm, PARAM_SOCKETCARD_ID, &topcard);
avrdude_message(MSG_INFO, "%sSocket card : %s\n", p,
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
stk600_get_cardname(socket_cards,
sizeof socket_cards / sizeof socket_cards[0],
topcard));
stk500v2_getparm2(pgm, PARAM2_RC_ID_TABLE_REV, &rev);
avrdude_message(MSG_INFO, "%sRC_ID table rev : %d\n", p, rev);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
stk500v2_getparm2(pgm, PARAM2_EC_ID_TABLE_REV, &rev);
avrdude_message(MSG_INFO, "%sEC_ID table rev : %d\n", p, rev);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
stk500v2_print_parms1(pgm, p);
return;
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
static double
f_to_kHz_MHz(double f, const char **unit)
{
if (f > 1e6) {
f /= 1e6;
*unit = "MHz";
} else if (f > 1e3) {
f /= 1000;
*unit = "kHz";
} else
*unit = "Hz";
return f;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
static void stk500v2_print_parms1(PROGRAMMER * pgm, const char * p)
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
{
unsigned char vtarget, vadjust, osc_pscale, osc_cmatch, sck_duration =0; //XXX 0 is not correct, check caller
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
unsigned int sck_stk600, clock_conf, dac, oct, varef;
unsigned char vtarget_jtag[4];
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
int prescale;
double f;
const char *unit;
void *mycookie;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
if (PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE_MKII) {
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
jtagmkII_getparm(pgm, PAR_OCD_VTARGET, vtarget_jtag);
pgm->cookie = mycookie;
avrdude_message(MSG_INFO, "%sVtarget : %.1f V\n", p,
b2_to_u16(vtarget_jtag) / 1000.0);
} else if (PDATA(pgm)->pgmtype == PGMTYPE_JTAGICE3) {
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
jtag3_getparm(pgm, SCOPE_GENERAL, 1, PARM3_VTARGET, vtarget_jtag, 2);
pgm->cookie = mycookie;
avrdude_message(MSG_INFO, "%sVtarget : %.1f V\n", p,
b2_to_u16(vtarget_jtag) / 1000.0);
} else {
stk500v2_getparm(pgm, PARAM_VTARGET, &vtarget);
avrdude_message(MSG_INFO, "%sVtarget : %.1f V\n", p, vtarget / 10.0);
}
switch (PDATA(pgm)->pgmtype) {
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
case PGMTYPE_STK500:
stk500v2_getparm(pgm, PARAM_SCK_DURATION, &sck_duration);
stk500v2_getparm(pgm, PARAM_VADJUST, &vadjust);
stk500v2_getparm(pgm, PARAM_OSC_PSCALE, &osc_pscale);
stk500v2_getparm(pgm, PARAM_OSC_CMATCH, &osc_cmatch);
avrdude_message(MSG_INFO, "%sSCK period : %.1f us\n", p,
stk500v2_sck_to_us(pgm, sck_duration));
avrdude_message(MSG_INFO, "%sVaref : %.1f V\n", p, vadjust / 10.0);
avrdude_message(MSG_INFO, "%sOscillator : ", p);
if (osc_pscale == 0)
avrdude_message(MSG_INFO, "Off\n");
else {
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
prescale = 1;
f = STK500V2_XTAL / 2;
switch (osc_pscale) {
case 2: prescale = 8; break;
case 3: prescale = 32; break;
case 4: prescale = 64; break;
case 5: prescale = 128; break;
case 6: prescale = 256; break;
case 7: prescale = 1024; break;
}
f /= prescale;
f /= (osc_cmatch + 1);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
f = f_to_kHz_MHz(f, &unit);
avrdude_message(MSG_INFO, "%.3f %s\n", f, unit);
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
break;
case PGMTYPE_AVRISP_MKII:
case PGMTYPE_JTAGICE_MKII:
stk500v2_getparm(pgm, PARAM_SCK_DURATION, &sck_duration);
avrdude_message(MSG_INFO, "%sSCK period : %.2f us\n", p,
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
(float) 1000000 / avrispmkIIfreqs[sck_duration]);
break;
case PGMTYPE_JTAGICE3:
{
unsigned char cmd[4];
cmd[0] = CMD_GET_SCK;
if (stk500v2_jtag3_send(pgm, cmd, 1) >= 0 &&
stk500v2_jtag3_recv(pgm, cmd, 4) >= 2) {
unsigned int sck = cmd[1] | (cmd[2] << 8);
avrdude_message(MSG_INFO, "%sSCK period : %.2f us\n", p,
(float)(1E6 / (1000.0 * sck)));
}
}
break;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
case PGMTYPE_STK600:
stk500v2_getparm2(pgm, PARAM2_AREF0, &varef);
avrdude_message(MSG_INFO, "%sVaref 0 : %.2f V\n", p, varef / 100.0);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
stk500v2_getparm2(pgm, PARAM2_AREF1, &varef);
avrdude_message(MSG_INFO, "%sVaref 1 : %.2f V\n", p, varef / 100.0);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
stk500v2_getparm2(pgm, PARAM2_SCK_DURATION, &sck_stk600);
avrdude_message(MSG_INFO, "%sSCK period : %.2f us\n", p,
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
(float) (sck_stk600 + 1) / 8.0);
stk500v2_getparm2(pgm, PARAM2_CLOCK_CONF, &clock_conf);
oct = (clock_conf & 0xf000) >> 12u;
dac = (clock_conf & 0x0ffc) >> 2u;
f = pow(2, (double)oct) * 2078.0 / (2 - (double)dac / 1024.0);
f = f_to_kHz_MHz(f, &unit);
avrdude_message(MSG_INFO, "%sOscillator : %.3f %s\n",
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
p, f, unit);
break;
default:
avrdude_message(MSG_INFO, "%sSCK period : %.1f us\n", p,
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
sck_duration * 8.0e6 / STK500V2_XTAL + 0.05);
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
break;
}
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
return;
}
static void stk500v2_print_parms(PROGRAMMER * pgm)
{
stk500v2_print_parms1(pgm, "");
}
static int stk500v2_perform_osccal(PROGRAMMER * pgm)
{
unsigned char buf[32];
int rv;
buf[0] = CMD_OSCCAL;
rv = stk500v2_command(pgm, buf, 1, sizeof(buf));
if (rv < 0) {
avrdude_message(MSG_INFO, "%s: stk500v2_perform_osccal(): failed\n",
progname);
return -1;
}
return 0;
}
/*
* Wrapper functions for the JTAG ICE mkII in ISP mode. This mode
* uses the normal JTAG ICE mkII packet stream to communicate with the
* ICE, but then encapsulates AVRISP mkII commands using
* CMND_ISP_PACKET.
*/
/*
* Open a JTAG ICE mkII in ISP mode.
*/
static int stk500v2_jtagmkII_open(PROGRAMMER * pgm, char * port)
{
union pinfo pinfo;
void *mycookie;
int rv;
avrdude_message(MSG_NOTICE2, "%s: stk500v2_jtagmkII_open()\n", progname);
/*
* The JTAG ICE mkII always starts with a baud rate of 19200 Bd upon
* attaching. If the config file or command-line parameters specify
* a higher baud rate, we switch to it later on, after establishing
* the connection with the ICE.
*/
pinfo.baud = 19200;
/*
* If the port name starts with "usb", divert the serial routines
* to the USB ones. The serial_open() function for USB overrides
* the meaning of the "baud" parameter to be the USB device ID to
* search for.
*/
if (strncmp(port, "usb", 3) == 0) {
#if defined(HAVE_LIBUSB)
serdev = &usb_serdev;
pinfo.usbinfo.vid = USB_VENDOR_ATMEL;
pinfo.usbinfo.flags = 0;
pinfo.usbinfo.pid = USB_DEVICE_JTAGICEMKII;
pgm->fd.usb.max_xfer = USBDEV_MAX_XFER_MKII;
pgm->fd.usb.rep = USBDEV_BULK_EP_READ_MKII;
pgm->fd.usb.wep = USBDEV_BULK_EP_WRITE_MKII;
pgm->fd.usb.eep = 0; /* no seperate EP for events */
#else
avrdude_message(MSG_INFO, "avrdude was compiled without usb support.\n");
return -1;
#endif
}
strcpy(pgm->port, port);
if (serial_open(port, pinfo, &pgm->fd)==-1) {
return -1;
}
/*
* drain any extraneous input
*/
stk500v2_drain(pgm, 0);
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
if ((rv = jtagmkII_getsync(pgm, EMULATOR_MODE_SPI)) != 0) {
if (rv != JTAGII_GETSYNC_FAIL_GRACEFUL)
avrdude_message(MSG_INFO, "%s: failed to sync with the JTAG ICE mkII in ISP mode\n",
progname);
pgm->cookie = mycookie;
return -1;
}
pgm->cookie = mycookie;
PDATA(pgm)->pgmtype = PGMTYPE_JTAGICE_MKII;
if (pgm->bitclock != 0.0) {
if (pgm->set_sck_period(pgm, pgm->bitclock) != 0)
return -1;
}
return 0;
}
/*
* Close an AVR Dragon or JTAG ICE mkII in ISP/HVSP/PP mode.
*/
static void stk500v2_jtagmkII_close(PROGRAMMER * pgm)
{
void *mycookie;
avrdude_message(MSG_NOTICE2, "%s: stk500v2_jtagmkII_close()\n", progname);
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
jtagmkII_close(pgm);
pgm->cookie = mycookie;
}
/*
* Close JTAGICE3.
*/
static void stk500v2_jtag3_close(PROGRAMMER * pgm)
{
void *mycookie;
avrdude_message(MSG_NOTICE2, "%s: stk500v2_jtag3_close()\n", progname);
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
jtag3_close(pgm);
pgm->cookie = mycookie;
}
/*
* Wrapper functions for the AVR Dragon in ISP mode. This mode
* uses the normal JTAG ICE mkII packet stream to communicate with the
* ICE, but then encapsulates AVRISP mkII commands using
* CMND_ISP_PACKET.
*/
/*
* Open an AVR Dragon in ISP mode.
*/
static int stk500v2_dragon_isp_open(PROGRAMMER * pgm, char * port)
{
union pinfo pinfo;
void *mycookie;
avrdude_message(MSG_NOTICE2, "%s: stk500v2_dragon_isp_open()\n", progname);
/*
* The JTAG ICE mkII always starts with a baud rate of 19200 Bd upon
* attaching. If the config file or command-line parameters specify
* a higher baud rate, we switch to it later on, after establishing
* the connection with the ICE.
*/
pinfo.baud = 19200;
/*
* If the port name starts with "usb", divert the serial routines
* to the USB ones. The serial_open() function for USB overrides
* the meaning of the "baud" parameter to be the USB device ID to
* search for.
*/
if (strncmp(port, "usb", 3) == 0) {
#if defined(HAVE_LIBUSB)
serdev = &usb_serdev;
pinfo.usbinfo.vid = USB_VENDOR_ATMEL;
pinfo.usbinfo.flags = 0;
pinfo.usbinfo.pid = USB_DEVICE_AVRDRAGON;
pgm->fd.usb.max_xfer = USBDEV_MAX_XFER_MKII;
pgm->fd.usb.rep = USBDEV_BULK_EP_READ_MKII;
pgm->fd.usb.wep = USBDEV_BULK_EP_WRITE_MKII;
pgm->fd.usb.eep = 0; /* no seperate EP for events */
#else
avrdude_message(MSG_INFO, "avrdude was compiled without usb support.\n");
return -1;
#endif
}
strcpy(pgm->port, port);
if (serial_open(port, pinfo, &pgm->fd)==-1) {
return -1;
}
/*
* drain any extraneous input
*/
stk500v2_drain(pgm, 0);
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
if (jtagmkII_getsync(pgm, EMULATOR_MODE_SPI) != 0) {
avrdude_message(MSG_INFO, "%s: failed to sync with the AVR Dragon in ISP mode\n",
progname);
pgm->cookie = mycookie;
return -1;
}
pgm->cookie = mycookie;
PDATA(pgm)->pgmtype = PGMTYPE_JTAGICE_MKII;
if (pgm->bitclock != 0.0) {
if (pgm->set_sck_period(pgm, pgm->bitclock) != 0)
return -1;
}
return 0;
}
/*
* Wrapper functions for the AVR Dragon in HV mode. This mode
* uses the normal JTAG ICE mkII packet stream to communicate with the
* ICE, but then encapsulates AVRISP mkII commands using
* CMND_ISP_PACKET.
*/
/*
* Open an AVR Dragon in HV mode (HVSP or parallel).
*/
static int stk500v2_dragon_hv_open(PROGRAMMER * pgm, char * port)
{
union pinfo pinfo;
void *mycookie;
avrdude_message(MSG_NOTICE2, "%s: stk500v2_dragon_hv_open()\n", progname);
/*
* The JTAG ICE mkII always starts with a baud rate of 19200 Bd upon
* attaching. If the config file or command-line parameters specify
* a higher baud rate, we switch to it later on, after establishing
* the connection with the ICE.
*/
pinfo.baud = 19200;
/*
* If the port name starts with "usb", divert the serial routines
* to the USB ones. The serial_open() function for USB overrides
* the meaning of the "baud" parameter to be the USB device ID to
* search for.
*/
if (strncmp(port, "usb", 3) == 0) {
#if defined(HAVE_LIBUSB)
serdev = &usb_serdev;
pinfo.usbinfo.vid = USB_VENDOR_ATMEL;
pinfo.usbinfo.flags = 0;
pinfo.usbinfo.pid = USB_DEVICE_AVRDRAGON;
pgm->fd.usb.max_xfer = USBDEV_MAX_XFER_MKII;
pgm->fd.usb.rep = USBDEV_BULK_EP_READ_MKII;
pgm->fd.usb.wep = USBDEV_BULK_EP_WRITE_MKII;
pgm->fd.usb.eep = 0; /* no seperate EP for events */
#else
avrdude_message(MSG_INFO, "avrdude was compiled without usb support.\n");
return -1;
#endif
}
strcpy(pgm->port, port);
if (serial_open(port, pinfo, &pgm->fd)==-1) {
return -1;
}
/*
* drain any extraneous input
*/
stk500v2_drain(pgm, 0);
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
if (jtagmkII_getsync(pgm, EMULATOR_MODE_HV) != 0) {
avrdude_message(MSG_INFO, "%s: failed to sync with the AVR Dragon in HV mode\n",
progname);
pgm->cookie = mycookie;
return -1;
}
pgm->cookie = mycookie;
PDATA(pgm)->pgmtype = PGMTYPE_JTAGICE_MKII;
if (pgm->bitclock != 0.0) {
if (pgm->set_sck_period(pgm, pgm->bitclock) != 0)
return -1;
}
return 0;
}
/*
* Wrapper functions for the JTAGICE3 in ISP mode. This mode
* uses the normal JTAGICE3 packet stream to communicate with the
* ICE, but then encapsulates AVRISP mkII commands using
* scope AVRISP.
*/
/*
* Open a JTAGICE3 in ISP mode.
*/
static int stk500v2_jtag3_open(PROGRAMMER * pgm, char * port)
{
void *mycookie;
int rv;
avrdude_message(MSG_NOTICE2, "%s: stk500v2_jtag3_open()\n", progname);
if (jtag3_open_common(pgm, port) < 0)
return -1;
mycookie = pgm->cookie;
pgm->cookie = PDATA(pgm)->chained_pdata;
if ((rv = jtag3_getsync(pgm, 42)) != 0) {
if (rv != JTAGII_GETSYNC_FAIL_GRACEFUL)
avrdude_message(MSG_INFO, "%s: failed to sync with the JTAGICE3 in ISP mode\n",
progname);
pgm->cookie = mycookie;
return -1;
}
pgm->cookie = mycookie;
PDATA(pgm)->pgmtype = PGMTYPE_JTAGICE3;
if (pgm->bitclock != 0.0) {
if (pgm->set_sck_period(pgm, pgm->bitclock) != 0)
return -1;
}
return 0;
}
/*
* XPROG wrapper
*/
static int stk600_xprog_command(PROGRAMMER * pgm, unsigned char *b,
unsigned int cmdsize, unsigned int responsesize)
{
unsigned char *newb;
unsigned int s;
int rv;
if (cmdsize < responsesize)
s = responsesize;
else
s = cmdsize;
if ((newb = malloc(s + 1)) == 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_cmd(): out of memory\n",
progname);
return -1;
}
newb[0] = CMD_XPROG;
memcpy(newb + 1, b, cmdsize);
rv = stk500v2_command(pgm, newb, cmdsize + 1, responsesize + 1);
if (rv == 0) {
memcpy(b, newb + 1, responsesize);
}
free(newb);
return rv;
}
/*
* issue the 'program enable' command to the AVR device, XPROG version
*/
static int stk600_xprog_program_enable(PROGRAMMER * pgm, AVRPART * p)
{
unsigned char buf[16];
unsigned int eepagesize = 42;
unsigned int nvm_base;
AVRMEM *mem = NULL;
int use_tpi;
use_tpi = (p->flags & AVRPART_HAS_TPI) != 0;
if (!use_tpi) {
if (p->nvm_base == 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_enable(): no nvm_base parameter for PDI device\n",
progname);
return -1;
}
if ((mem = avr_locate_mem(p, "eeprom")) != NULL) {
if (mem->page_size == 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_enable(): no EEPROM page_size parameter for PDI device\n",
progname);
return -1;
}
eepagesize = mem->page_size;
}
}
buf[0] = CMD_XPROG_SETMODE;
buf[1] = use_tpi? XPRG_MODE_TPI: XPRG_MODE_PDI;
if (stk500v2_command(pgm, buf, 2, sizeof(buf)) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_enable(): CMD_XPROG_SETMODE(XPRG_MODE_%s) failed\n",
progname, use_tpi? "TPI": "PDI");
return -1;
}
buf[0] = XPRG_CMD_ENTER_PROGMODE;
if (stk600_xprog_command(pgm, buf, 1, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_enable(): XPRG_CMD_ENTER_PROGMODE failed\n",
progname);
return -1;
}
if (use_tpi) {
/*
* Whatever all that might mean, it matches what AVR Studio
* does.
*/
if (stk500v2_setparm_real(pgm, PARAM_DISCHARGEDELAY, 232) < 0)
return -1;
buf[0] = XPRG_CMD_SET_PARAM;
buf[1] = XPRG_PARAM_TPI_3;
buf[2] = 51;
if (stk600_xprog_command(pgm, buf, 3, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_enable(): XPRG_CMD_SET_PARAM(XPRG_PARAM_TPI_3) failed\n",
progname);
return -1;
}
buf[0] = XPRG_CMD_SET_PARAM;
buf[1] = XPRG_PARAM_TPI_4;
buf[2] = 50;
if (stk600_xprog_command(pgm, buf, 3, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_enable(): XPRG_CMD_SET_PARAM(XPRG_PARAM_TPI_4) failed\n",
progname);
return -1;
}
} else {
buf[0] = XPRG_CMD_SET_PARAM;
buf[1] = XPRG_PARAM_NVMBASE;
nvm_base = p->nvm_base;
/*
* The 0x01000000 appears to be an indication to the programmer
* that the respective address is located in IO (i.e., SRAM)
* memory address space rather than flash. This is not documented
* anywhere in AVR079 but matches what AVR Studio does.
*/
nvm_base |= 0x01000000;
buf[2] = nvm_base >> 24;
buf[3] = nvm_base >> 16;
buf[4] = nvm_base >> 8;
buf[5] = nvm_base;
if (stk600_xprog_command(pgm, buf, 6, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_enable(): XPRG_CMD_SET_PARAM(XPRG_PARAM_NVMBASE) failed\n",
progname);
return -1;
}
if (mem != NULL) {
buf[0] = XPRG_CMD_SET_PARAM;
buf[1] = XPRG_PARAM_EEPPAGESIZE;
buf[2] = eepagesize >> 8;
buf[3] = eepagesize;
if (stk600_xprog_command(pgm, buf, 4, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_enable(): XPRG_CMD_SET_PARAM(XPRG_PARAM_EEPPAGESIZE) failed\n",
progname);
return -1;
}
}
}
return 0;
}
static unsigned char stk600_xprog_memtype(PROGRAMMER * pgm, unsigned long addr)
{
if (addr >= PDATA(pgm)->boot_start)
return XPRG_MEM_TYPE_BOOT;
else
return XPRG_MEM_TYPE_APPL;
}
static void stk600_xprog_disable(PROGRAMMER * pgm)
{
unsigned char buf[2];
buf[0] = XPRG_CMD_LEAVE_PROGMODE;
if (stk600_xprog_command(pgm, buf, 1, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_program_disable(): XPRG_CMD_LEAVE_PROGMODE failed\n",
progname);
}
}
static int stk600_xprog_write_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char data)
{
unsigned char b[9 + 256];
int need_erase = 0;
unsigned char write_size = 1;
unsigned char memcode;
memset(b, 0, sizeof(b));
if (strcmp(mem->desc, "flash") == 0) {
memcode = stk600_xprog_memtype(pgm, addr);
} else if (strcmp(mem->desc, "application") == 0 ||
strcmp(mem->desc, "apptable") == 0) {
memcode = XPRG_MEM_TYPE_APPL;
} else if (strcmp(mem->desc, "boot") == 0) {
memcode = XPRG_MEM_TYPE_BOOT;
} else if (strcmp(mem->desc, "eeprom") == 0) {
memcode = XPRG_MEM_TYPE_EEPROM;
} else if (strncmp(mem->desc, "lock", strlen("lock")) == 0) {
memcode = XPRG_MEM_TYPE_LOCKBITS;
} else if (strncmp(mem->desc, "fuse", strlen("fuse")) == 0) {
memcode = XPRG_MEM_TYPE_FUSE;
if (p->flags & AVRPART_HAS_TPI)
/*
* TPI devices need a mystic erase prior to writing their
* fuses.
*/
need_erase = 1;
} else if (strcmp(mem->desc, "usersig") == 0) {
memcode = XPRG_MEM_TYPE_USERSIG;
} else {
avrdude_message(MSG_INFO, "%s: stk600_xprog_write_byte(): unknown memory \"%s\"\n",
progname, mem->desc);
return -1;
}
addr += mem->offset;
if (need_erase) {
b[0] = XPRG_CMD_ERASE;
b[1] = XPRG_ERASE_CONFIG;
b[2] = mem->offset >> 24;
b[3] = mem->offset >> 16;
b[4] = mem->offset >> 8;
b[5] = mem->offset + 1;
if (stk600_xprog_command(pgm, b, 6, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_chip_erase(): XPRG_CMD_ERASE(XPRG_ERASE_CONFIG) failed\n",
progname);
return -1;
}
}
if (p->flags & AVRPART_HAS_TPI) {
/*
* Some TPI memories (configuration aka. fuse) require a
* larger write block size. We record that as a blocksize in
* avrdude.conf.
*/
if (mem->blocksize != 0)
write_size = mem->blocksize;
}
b[0] = XPRG_CMD_WRITE_MEM;
b[1] = memcode;
b[2] = 0; /* pagemode: non-paged write */
b[3] = addr >> 24;
b[4] = addr >> 16;
b[5] = addr >> 8;
b[6] = addr;
b[7] = 0;
b[8] = write_size;
b[9] = data;
if (stk600_xprog_command(pgm, b, 9 + write_size, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_write_byte(): XPRG_CMD_WRITE_MEM failed\n",
progname);
return -1;
}
return 0;
}
static int stk600_xprog_read_byte(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned long addr, unsigned char * value)
{
unsigned char b[8];
if (strcmp(mem->desc, "flash") == 0) {
b[1] = stk600_xprog_memtype(pgm, addr);
} else if (strcmp(mem->desc, "application") == 0 ||
strcmp(mem->desc, "apptable") == 0) {
b[1] = XPRG_MEM_TYPE_APPL;
} else if (strcmp(mem->desc, "boot") == 0) {
b[1] = XPRG_MEM_TYPE_BOOT;
} else if (strcmp(mem->desc, "eeprom") == 0) {
b[1] = XPRG_MEM_TYPE_EEPROM;
} else if (strcmp(mem->desc, "signature") == 0) {
b[1] = XPRG_MEM_TYPE_APPL;
} else if (strncmp(mem->desc, "fuse", strlen("fuse")) == 0) {
b[1] = XPRG_MEM_TYPE_FUSE;
} else if (strncmp(mem->desc, "lock", strlen("lock")) == 0) {
b[1] = XPRG_MEM_TYPE_LOCKBITS;
} else if (strcmp(mem->desc, "calibration") == 0) {
b[1] = XPRG_MEM_TYPE_FACTORY_CALIBRATION;
} else if (strcmp(mem->desc, "usersig") == 0) {
b[1] = XPRG_MEM_TYPE_USERSIG;
} else {
avrdude_message(MSG_INFO, "%s: stk600_xprog_read_byte(): unknown memory \"%s\"\n",
progname, mem->desc);
return -1;
}
addr += mem->offset;
b[0] = XPRG_CMD_READ_MEM;
b[2] = addr >> 24;
b[3] = addr >> 16;
b[4] = addr >> 8;
b[5] = addr;
b[6] = 0;
b[7] = 1;
if (stk600_xprog_command(pgm, b, 8, 3) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_read_byte(): XPRG_CMD_READ_MEM failed\n",
progname);
return -1;
}
*value = b[2];
return 0;
}
static int stk600_xprog_paged_load(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes)
{
unsigned char *b;
unsigned int offset;
unsigned char memtype;
int n_bytes_orig = n_bytes, dynamic_memtype = 0;
unsigned long use_ext_addr = 0;
/*
* The XPROG read command supports at most 256 bytes in one
* transfer.
*/
if (page_size > 256)
page_size = 256; /* not really a page size anymore */
/*
* Fancy offsets everywhere.
* This is probably what AVR079 means when writing about the
* "TIF address space".
*/
if (strcmp(mem->desc, "flash") == 0) {
memtype = 0;
dynamic_memtype = 1;
if (mem->size > 64 * 1024)
use_ext_addr = (1UL << 31);
} else if (strcmp(mem->desc, "application") == 0 ||
strcmp(mem->desc, "apptable") == 0) {
memtype = XPRG_MEM_TYPE_APPL;
if (mem->size > 64 * 1024)
use_ext_addr = (1UL << 31);
} else if (strcmp(mem->desc, "boot") == 0) {
memtype = XPRG_MEM_TYPE_BOOT;
// Do we have to consider the total amount of flash
// instead to decide whether to use extended addressing?
if (mem->size > 64 * 1024)
use_ext_addr = (1UL << 31);
} else if (strcmp(mem->desc, "eeprom") == 0) {
memtype = XPRG_MEM_TYPE_EEPROM;
} else if (strcmp(mem->desc, "signature") == 0) {
memtype = XPRG_MEM_TYPE_APPL;
} else if (strncmp(mem->desc, "fuse", strlen("fuse")) == 0) {
memtype = XPRG_MEM_TYPE_FUSE;
} else if (strncmp(mem->desc, "lock", strlen("lock")) == 0) {
memtype = XPRG_MEM_TYPE_LOCKBITS;
} else if (strcmp(mem->desc, "calibration") == 0) {
memtype = XPRG_MEM_TYPE_FACTORY_CALIBRATION;
} else if (strcmp(mem->desc, "usersig") == 0) {
memtype = XPRG_MEM_TYPE_USERSIG;
} else {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_load(): unknown paged memory \"%s\"\n",
progname, mem->desc);
return -1;
}
offset = addr;
addr += mem->offset;
if ((b = malloc(page_size + 2)) == NULL) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_load(): out of memory\n",
progname);
return -1;
}
if (stk500v2_loadaddr(pgm, use_ext_addr) < 0) {
free(b);
return -1;
}
while (n_bytes != 0) {
if (dynamic_memtype)
memtype = stk600_xprog_memtype(pgm, addr - mem->offset);
b[0] = XPRG_CMD_READ_MEM;
b[1] = memtype;
b[2] = addr >> 24;
b[3] = addr >> 16;
b[4] = addr >> 8;
b[5] = addr;
b[6] = page_size >> 8;
b[7] = page_size;
if (stk600_xprog_command(pgm, b, 8, page_size + 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_load(): XPRG_CMD_READ_MEM failed\n",
progname);
free(b);
return -1;
}
memcpy(mem->buf + offset, b + 2, page_size);
if (n_bytes < page_size) {
n_bytes = page_size;
}
offset += page_size;
addr += page_size;
n_bytes -= page_size;
}
free(b);
return n_bytes_orig;
}
static int stk600_xprog_paged_write(PROGRAMMER * pgm, AVRPART * p, AVRMEM * mem,
unsigned int page_size,
unsigned int addr, unsigned int n_bytes)
{
unsigned char *b;
unsigned int offset;
unsigned char memtype;
int n_bytes_orig = n_bytes, dynamic_memtype = 0;
size_t writesize;
unsigned long use_ext_addr = 0;
unsigned char writemode;
/*
* The XPROG read command supports at most 256 bytes in one
* transfer.
*/
if (page_size > 512) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_write(): cannot handle page size > 512\n",
progname);
return -1;
}
/*
* Fancy offsets everywhere.
* This is probably what AVR079 means when writing about the
* "TIF address space".
*/
if (strcmp(mem->desc, "flash") == 0) {
memtype = 0;
dynamic_memtype = 1;
writemode = (1 << XPRG_MEM_WRITE_WRITE);
if (mem->size > 64 * 1024)
use_ext_addr = (1UL << 31);
} else if (strcmp(mem->desc, "application") == 0 ||
strcmp(mem->desc, "apptable") == 0) {
memtype = XPRG_MEM_TYPE_APPL;
writemode = (1 << XPRG_MEM_WRITE_WRITE);
if (mem->size > 64 * 1024)
use_ext_addr = (1UL << 31);
} else if (strcmp(mem->desc, "boot") == 0) {
memtype = XPRG_MEM_TYPE_BOOT;
writemode = (1 << XPRG_MEM_WRITE_WRITE);
// Do we have to consider the total amount of flash
// instead to decide whether to use extended addressing?
if (mem->size > 64 * 1024)
use_ext_addr = (1UL << 31);
} else if (strcmp(mem->desc, "eeprom") == 0) {
memtype = XPRG_MEM_TYPE_EEPROM;
writemode = (1 << XPRG_MEM_WRITE_WRITE) | (1 << XPRG_MEM_WRITE_ERASE);
} else if (strcmp(mem->desc, "signature") == 0) {
memtype = XPRG_MEM_TYPE_APPL;
writemode = (1 << XPRG_MEM_WRITE_WRITE);
} else if (strncmp(mem->desc, "fuse", strlen("fuse")) == 0) {
memtype = XPRG_MEM_TYPE_FUSE;
writemode = (1 << XPRG_MEM_WRITE_WRITE);
} else if (strncmp(mem->desc, "lock", strlen("lock")) == 0) {
memtype = XPRG_MEM_TYPE_LOCKBITS;
writemode = (1 << XPRG_MEM_WRITE_WRITE);
} else if (strcmp(mem->desc, "calibration") == 0) {
memtype = XPRG_MEM_TYPE_FACTORY_CALIBRATION;
writemode = (1 << XPRG_MEM_WRITE_WRITE);
} else if (strcmp(mem->desc, "usersig") == 0) {
memtype = XPRG_MEM_TYPE_USERSIG;
writemode = (1 << XPRG_MEM_WRITE_WRITE);
} else {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_write(): unknown paged memory \"%s\"\n",
progname, mem->desc);
return -1;
}
offset = addr;
addr += mem->offset;
if ((b = malloc(page_size + 9)) == NULL) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_write(): out of memory\n",
progname);
return -1;
}
if (stk500v2_loadaddr(pgm, use_ext_addr) < 0) {
free(b);
return -1;
}
while (n_bytes != 0) {
if (dynamic_memtype)
memtype = stk600_xprog_memtype(pgm, addr - mem->offset);
if (page_size > 256) {
/*
* AVR079 is not quite clear. While it suggests that
* downloading up to 512 bytes (256 words) were OK, it
* obviously isn't -- 512-byte pages on the ATxmega128A1
* are getting corrupted when written as a single piece.
* It writes random junk somewhere beyond byte 256.
* Splitting it into 256 byte chunks, and only setting the
* erase page / write page bits in the final chunk helps.
*/
if (page_size % 256 != 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_write(): page size not multiple of 256\n",
progname);
free(b);
return -1;
}
unsigned int chunk;
for (chunk = 0; chunk < page_size; chunk += 256) {
if (n_bytes < 256) {
memset(b + 9 + n_bytes, 0xff, 256 - n_bytes);
writesize = n_bytes;
} else {
writesize = 256;
}
b[0] = XPRG_CMD_WRITE_MEM;
b[1] = memtype;
b[2] = writemode;
b[3] = addr >> 24;
b[4] = addr >> 16;
b[5] = addr >> 8;
b[6] = addr;
b[7] = 1;
b[8] = 0;
memcpy(b + 9, mem->buf + offset, writesize);
if (stk600_xprog_command(pgm, b, 256 + 9, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_write(): XPRG_CMD_WRITE_MEM failed\n",
progname);
free(b);
return -1;
}
if (n_bytes < 256)
n_bytes = 256;
offset += 256;
addr += 256;
n_bytes -= 256;
}
} else {
if (n_bytes < page_size) {
/*
* This can easily happen if the input file was not a
* multiple of the page size.
*/
memset(b + 9 + n_bytes, 0xff, page_size - n_bytes);
writesize = n_bytes;
} else {
writesize = page_size;
}
b[0] = XPRG_CMD_WRITE_MEM;
b[1] = memtype;
b[2] = writemode;
b[3] = addr >> 24;
b[4] = addr >> 16;
b[5] = addr >> 8;
b[6] = addr;
b[7] = page_size >> 8;
b[8] = page_size;
memcpy(b + 9, mem->buf + offset, writesize);
if (stk600_xprog_command(pgm, b, page_size + 9, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_paged_write(): XPRG_CMD_WRITE_MEM failed\n",
progname);
free(b);
return -1;
}
if (n_bytes < page_size)
n_bytes = page_size;
offset += page_size;
addr += page_size;
n_bytes -= page_size;
}
}
free(b);
return n_bytes_orig;
}
static int stk600_xprog_chip_erase(PROGRAMMER * pgm, AVRPART * p)
{
unsigned char b[6];
AVRMEM *mem;
unsigned int addr = 0;
if (p->flags & AVRPART_HAS_TPI) {
if ((mem = avr_locate_mem(p, "flash")) == NULL) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_chip_erase(): no FLASH definition found for TPI device\n",
progname);
return -1;
}
addr = mem->offset + 1;
}
b[0] = XPRG_CMD_ERASE;
b[1] = XPRG_ERASE_CHIP;
b[2] = addr >> 24;
b[3] = addr >> 16;
b[4] = addr >> 8;
b[5] = addr;
if (stk600_xprog_command(pgm, b, 6, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_chip_erase(): XPRG_CMD_ERASE(XPRG_ERASE_CHIP) failed\n",
progname);
return -1;
}
return 0;
}
static int stk600_xprog_page_erase(PROGRAMMER * pgm, AVRPART * p, AVRMEM * m,
unsigned int addr)
{
unsigned char b[6];
if (strcmp(m->desc, "flash") == 0) {
b[1] = stk600_xprog_memtype(pgm, addr) == XPRG_MEM_TYPE_APPL?
XPRG_ERASE_APP_PAGE: XPRG_ERASE_BOOT_PAGE;
} else if (strcmp(m->desc, "application") == 0 ||
strcmp(m->desc, "apptable") == 0) {
b[1] = XPRG_ERASE_APP_PAGE;
} else if (strcmp(m->desc, "boot") == 0) {
b[1] = XPRG_ERASE_BOOT_PAGE;
} else if (strcmp(m->desc, "eeprom") == 0) {
b[1] = XPRG_ERASE_EEPROM_PAGE;
} else if (strcmp(m->desc, "usersig") == 0) {
b[1] = XPRG_ERASE_USERSIG;
} else {
avrdude_message(MSG_INFO, "%s: stk600_xprog_page_erase(): unknown paged memory \"%s\"\n",
progname, m->desc);
return -1;
}
addr += m->offset;
b[0] = XPRG_CMD_ERASE;
b[2] = addr >> 24;
b[3] = addr >> 16;
b[4] = addr >> 8;
b[5] = addr;
if (stk600_xprog_command(pgm, b, 6, 2) < 0) {
avrdude_message(MSG_INFO, "%s: stk600_xprog_page_erase(): XPRG_CMD_ERASE(%d) failed\n",
progname, b[1]);
return -1;
}
return 0;
}
/*
* Modify pgm's methods for XPROG operation.
*/
static void stk600_setup_xprog(PROGRAMMER * pgm)
{
pgm->program_enable = stk600_xprog_program_enable;
pgm->disable = stk600_xprog_disable;
pgm->read_byte = stk600_xprog_read_byte;
pgm->write_byte = stk600_xprog_write_byte;
pgm->paged_load = stk600_xprog_paged_load;
pgm->paged_write = stk600_xprog_paged_write;
pgm->page_erase = stk600_xprog_page_erase;
pgm->chip_erase = stk600_xprog_chip_erase;
}
/*
* Modify pgm's methods for ISP operation.
*/
static void stk600_setup_isp(PROGRAMMER * pgm)
{
pgm->program_enable = stk500v2_program_enable;
pgm->disable = stk500v2_disable;
pgm->read_byte = stk500isp_read_byte;
pgm->write_byte = stk500isp_write_byte;
pgm->paged_load = stk500v2_paged_load;
pgm->paged_write = stk500v2_paged_write;
pgm->page_erase = stk500v2_page_erase;
pgm->chip_erase = stk500v2_chip_erase;
}
const char stk500v2_desc[] = "Atmel STK500 Version 2.x firmware";
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
void stk500v2_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "STK500V2");
/*
* mandatory functions
*/
pgm->initialize = stk500v2_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500v2_disable;
pgm->program_enable = stk500v2_program_enable;
pgm->chip_erase = stk500v2_chip_erase;
pgm->cmd = stk500v2_cmd;
pgm->open = stk500v2_open;
pgm->close = stk500v2_close;
pgm->read_byte = stk500isp_read_byte;
pgm->write_byte = stk500isp_write_byte;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
/*
* optional functions
*/
pgm->paged_write = stk500v2_paged_write;
pgm->paged_load = stk500v2_paged_load;
pgm->page_erase = stk500v2_page_erase;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
pgm->print_parms = stk500v2_print_parms;
pgm->set_vtarget = stk500v2_set_vtarget;
pgm->set_varef = stk500v2_set_varef;
pgm->set_fosc = stk500v2_set_fosc;
pgm->set_sck_period = stk500v2_set_sck_period;
pgm->perform_osccal = stk500v2_perform_osccal;
pgm->setup = stk500v2_setup;
pgm->teardown = stk500v2_teardown;
Mega-commit to bring in both, the STK500v2 support from Erik Walthinsen, as well as JTAG ICE mkII support (by me). Erik's submission has been cleaned up a little bit, mostly to add his name and the current year to the copyright of the new file, remove trailing white space before importing the files, and fix the minor syntax errors in his avrdude.conf.in additions (missing semicolons). The JTAG ICE mkII support should be considered alpha to beta quality at this point. Few things are still to be done, like defering the hfuse (OCDEN) tweaks until they are really required. Also, for reasons not yet known, the target MCU doesn't start to run after signing off from the ICE, it needs a power-cycle first (at least on my STK500). Note that for the JTAG ICE, I did change a few things in the internal API. Notably I made the serial receive timeout configurable by the backends via an exported variable (done in both the Posix and the Win32 implementation), and I made the serial_recv() function return a -1 instead of bailing out with exit(1) upon encountering a receive timeout (currently only done in the Posix implementation). Both measures together allow me to receive a datastreem from the ICE at 115 kbps on a somewhat lossy PCI multi-UART card that occasionally drops a character. The JTAG ICE mkII protocol has enough of safety layers to allow recovering from these events, but the previous code wasn't prepared for any kind of recovery. The Win32 change for this still has to be done, and the traditional drivers need to be converted to exit(1) upon encountering a timeout (as they're now getting a -1 returned they didn't see before in that case). git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@451 81a1dc3b-b13d-400b-aceb-764788c761c2
2005-05-10 19:17:12 +00:00
pgm->page_size = 256;
}
const char stk500pp_desc[] = "Atmel STK500 V2 in parallel programming mode";
void stk500pp_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "STK500PP");
/*
* mandatory functions
*/
pgm->initialize = stk500pp_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500pp_disable;
pgm->program_enable = stk500pp_program_enable;
pgm->chip_erase = stk500pp_chip_erase;
pgm->open = stk500v2_open;
pgm->close = stk500v2_close;
pgm->read_byte = stk500pp_read_byte;
pgm->write_byte = stk500pp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500pp_paged_write;
pgm->paged_load = stk500pp_paged_load;
pgm->print_parms = stk500v2_print_parms;
pgm->set_vtarget = stk500v2_set_vtarget;
pgm->set_varef = stk500v2_set_varef;
pgm->set_fosc = stk500v2_set_fosc;
pgm->set_sck_period = stk500v2_set_sck_period;
pgm->setup = stk500v2_setup;
pgm->teardown = stk500v2_teardown;
pgm->page_size = 256;
}
const char stk500hvsp_desc[] = "Atmel STK500 V2 in high-voltage serial programming mode";
void stk500hvsp_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "STK500HVSP");
/*
* mandatory functions
*/
pgm->initialize = stk500hvsp_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500hvsp_disable;
pgm->program_enable = stk500hvsp_program_enable;
pgm->chip_erase = stk500hvsp_chip_erase;
pgm->open = stk500v2_open;
pgm->close = stk500v2_close;
pgm->read_byte = stk500hvsp_read_byte;
pgm->write_byte = stk500hvsp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500hvsp_paged_write;
pgm->paged_load = stk500hvsp_paged_load;
pgm->print_parms = stk500v2_print_parms;
pgm->set_vtarget = stk500v2_set_vtarget;
pgm->set_varef = stk500v2_set_varef;
pgm->set_fosc = stk500v2_set_fosc;
pgm->set_sck_period = stk500v2_set_sck_period;
pgm->setup = stk500v2_setup;
pgm->teardown = stk500v2_teardown;
pgm->page_size = 256;
}
const char stk500v2_jtagmkII_desc[] = "Atmel JTAG ICE mkII in ISP mode";
void stk500v2_jtagmkII_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "JTAGMKII_ISP");
/*
* mandatory functions
*/
pgm->initialize = stk500v2_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500v2_disable;
pgm->program_enable = stk500v2_program_enable;
pgm->chip_erase = stk500v2_chip_erase;
pgm->cmd = stk500v2_cmd;
pgm->open = stk500v2_jtagmkII_open;
pgm->close = stk500v2_jtagmkII_close;
pgm->read_byte = stk500isp_read_byte;
pgm->write_byte = stk500isp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500v2_paged_write;
pgm->paged_load = stk500v2_paged_load;
pgm->page_erase = stk500v2_page_erase;
pgm->print_parms = stk500v2_print_parms;
pgm->set_sck_period = stk500v2_set_sck_period_mk2;
pgm->perform_osccal = stk500v2_perform_osccal;
pgm->setup = stk500v2_jtagmkII_setup;
pgm->teardown = stk500v2_jtagmkII_teardown;
pgm->page_size = 256;
}
const char stk500v2_dragon_isp_desc[] = "Atmel AVR Dragon in ISP mode";
void stk500v2_dragon_isp_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "DRAGON_ISP");
/*
* mandatory functions
*/
pgm->initialize = stk500v2_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500v2_disable;
pgm->program_enable = stk500v2_program_enable;
pgm->chip_erase = stk500v2_chip_erase;
pgm->cmd = stk500v2_cmd;
pgm->open = stk500v2_dragon_isp_open;
pgm->close = stk500v2_jtagmkII_close;
pgm->read_byte = stk500isp_read_byte;
pgm->write_byte = stk500isp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500v2_paged_write;
pgm->paged_load = stk500v2_paged_load;
pgm->page_erase = stk500v2_page_erase;
pgm->print_parms = stk500v2_print_parms;
pgm->set_sck_period = stk500v2_set_sck_period_mk2;
pgm->setup = stk500v2_jtagmkII_setup;
pgm->teardown = stk500v2_jtagmkII_teardown;
pgm->page_size = 256;
}
const char stk500v2_dragon_pp_desc[] = "Atmel AVR Dragon in PP mode";
void stk500v2_dragon_pp_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "DRAGON_PP");
/*
* mandatory functions
*/
pgm->initialize = stk500pp_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500pp_disable;
pgm->program_enable = stk500pp_program_enable;
pgm->chip_erase = stk500pp_chip_erase;
pgm->open = stk500v2_dragon_hv_open;
pgm->close = stk500v2_jtagmkII_close;
pgm->read_byte = stk500pp_read_byte;
pgm->write_byte = stk500pp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500pp_paged_write;
pgm->paged_load = stk500pp_paged_load;
pgm->print_parms = stk500v2_print_parms;
pgm->set_vtarget = stk500v2_set_vtarget;
pgm->set_varef = stk500v2_set_varef;
pgm->set_fosc = stk500v2_set_fosc;
pgm->set_sck_period = stk500v2_set_sck_period_mk2;
pgm->setup = stk500v2_jtagmkII_setup;
pgm->teardown = stk500v2_jtagmkII_teardown;
pgm->page_size = 256;
}
const char stk500v2_dragon_hvsp_desc[] = "Atmel AVR Dragon in HVSP mode";
void stk500v2_dragon_hvsp_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "DRAGON_HVSP");
/*
* mandatory functions
*/
pgm->initialize = stk500hvsp_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500hvsp_disable;
pgm->program_enable = stk500hvsp_program_enable;
pgm->chip_erase = stk500hvsp_chip_erase;
pgm->open = stk500v2_dragon_hv_open;
pgm->close = stk500v2_jtagmkII_close;
pgm->read_byte = stk500hvsp_read_byte;
pgm->write_byte = stk500hvsp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500hvsp_paged_write;
pgm->paged_load = stk500hvsp_paged_load;
pgm->print_parms = stk500v2_print_parms;
pgm->set_vtarget = stk500v2_set_vtarget;
pgm->set_varef = stk500v2_set_varef;
pgm->set_fosc = stk500v2_set_fosc;
pgm->set_sck_period = stk500v2_set_sck_period_mk2;
pgm->setup = stk500v2_jtagmkII_setup;
pgm->teardown = stk500v2_jtagmkII_teardown;
pgm->page_size = 256;
}
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
const char stk600_desc[] = "Atmel STK600";
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
void stk600_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "STK600");
/*
* mandatory functions
*/
pgm->initialize = stk500v2_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500v2_disable;
pgm->program_enable = stk500v2_program_enable;
pgm->chip_erase = stk500v2_chip_erase;
pgm->cmd = stk500v2_cmd;
pgm->open = stk600_open;
pgm->close = stk500v2_close;
pgm->read_byte = stk500isp_read_byte;
pgm->write_byte = stk500isp_write_byte;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
/*
* optional functions
*/
pgm->paged_write = stk500v2_paged_write;
pgm->paged_load = stk500v2_paged_load;
pgm->page_erase = stk500v2_page_erase;
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
pgm->print_parms = stk500v2_print_parms;
pgm->set_vtarget = stk600_set_vtarget;
pgm->set_varef = stk600_set_varef;
pgm->set_fosc = stk600_set_fosc;
pgm->set_sck_period = stk600_set_sck_period;
pgm->perform_osccal = stk500v2_perform_osccal;
pgm->setup = stk500v2_setup;
pgm->teardown = stk500v2_teardown;
pgm->page_size = 256;
}
const char stk600pp_desc[] = "Atmel STK600 in parallel programming mode";
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
void stk600pp_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "STK600PP");
/*
* mandatory functions
*/
pgm->initialize = stk500pp_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500pp_disable;
pgm->program_enable = stk500pp_program_enable;
pgm->chip_erase = stk500pp_chip_erase;
pgm->open = stk600_open;
pgm->close = stk500v2_close;
pgm->read_byte = stk500pp_read_byte;
pgm->write_byte = stk500pp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500pp_paged_write;
pgm->paged_load = stk500pp_paged_load;
pgm->print_parms = stk500v2_print_parms;
pgm->set_vtarget = stk600_set_vtarget;
pgm->set_varef = stk600_set_varef;
pgm->set_fosc = stk600_set_fosc;
pgm->set_sck_period = stk600_set_sck_period;
pgm->setup = stk500v2_setup;
pgm->teardown = stk500v2_teardown;
pgm->page_size = 256;
}
const char stk600hvsp_desc[] = "Atmel STK600 in high-voltage serial programming mode";
Add initial support for the Atmel STK600, for "classic" AVRs (AT90, ATtiny, ATmega) in both, ISP and high-voltage programming modes. * Makefile.am: Add -lm. * avrdude.conf.in: Add stk600, stk600pp, and stk600hvsp. * config_gram.y: Add support for the stk600* keywords. * lexer.l: (Ditto.) * pgm.h: Add the "chan" parameter to set_varef(). * stk500.c: (Ditto.) * serial.h: Add USB endpoint support to struct filedescriptor. * stk500v2.c: Implement the meat of the STK600 support. * stk500v2.h: Add new prototypes for stk600*() programmers. * stk500v2_private.h: Add new constants used in the STK600. * term.c: Add AREF channel support. * usb_libusb.c: Automatically determine the correct write endpoint ID, the STK600 uses 0x83 while all other tools use 0x82. Propagate the EP to use through struct filedescriptor. * usbdevs.h: Add the STK600 USB product ID. * tools/get-stk600-cards.xsl: XSL transformation for targetboards.xml to obtain the list of socket and routing card IDs, to be used in stk500v2.c (for displaying the names). * tools/get-stk600-devices.xsl: XSL transformation for targetboards.xml to obtain the table of socket/routing cards and their respective AVR device support for doc/avrdude.texi. * avrdude.1: Document all the STK600 stuff. * doc/avrdude.texi: Ditto. Added a new chapter for Programmer Specific Information. Thanks to Eirik Rasmussen from Atmel Norway for his support in getting this code running within that short amount of time! git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@768 81a1dc3b-b13d-400b-aceb-764788c761c2
2008-03-14 13:00:08 +00:00
void stk600hvsp_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "STK600HVSP");
/*
* mandatory functions
*/
pgm->initialize = stk500hvsp_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500hvsp_disable;
pgm->program_enable = stk500hvsp_program_enable;
pgm->chip_erase = stk500hvsp_chip_erase;
pgm->open = stk600_open;
pgm->close = stk500v2_close;
pgm->read_byte = stk500hvsp_read_byte;
pgm->write_byte = stk500hvsp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500hvsp_paged_write;
pgm->paged_load = stk500hvsp_paged_load;
pgm->print_parms = stk500v2_print_parms;
pgm->set_vtarget = stk600_set_vtarget;
pgm->set_varef = stk600_set_varef;
pgm->set_fosc = stk600_set_fosc;
pgm->set_sck_period = stk600_set_sck_period;
pgm->setup = stk500v2_setup;
pgm->teardown = stk500v2_teardown;
pgm->page_size = 256;
}
const char stk500v2_jtag3_desc[] = "Atmel JTAGICE3 in ISP mode";
void stk500v2_jtag3_initpgm(PROGRAMMER * pgm)
{
strcpy(pgm->type, "JTAG3_ISP");
/*
* mandatory functions
*/
pgm->initialize = stk500v2_jtag3_initialize;
pgm->display = stk500v2_display;
pgm->enable = stk500v2_enable;
pgm->disable = stk500v2_jtag3_disable;
pgm->program_enable = stk500v2_program_enable;
pgm->chip_erase = stk500v2_chip_erase;
pgm->cmd = stk500v2_jtag3_cmd;
pgm->open = stk500v2_jtag3_open;
pgm->close = stk500v2_jtag3_close;
pgm->read_byte = stk500isp_read_byte;
pgm->write_byte = stk500isp_write_byte;
/*
* optional functions
*/
pgm->paged_write = stk500v2_paged_write;
pgm->paged_load = stk500v2_paged_load;
pgm->page_erase = stk500v2_page_erase;
pgm->print_parms = stk500v2_print_parms;
pgm->set_sck_period = stk500v2_jtag3_set_sck_period;
pgm->perform_osccal = stk500v2_perform_osccal;
pgm->setup = stk500v2_jtag3_setup;
pgm->teardown = stk500v2_jtag3_teardown;
pgm->page_size = 256;
}