patch #9110: Let reserved fuse bits to be read as *don't care*
* avr.c (compare_memory_masked): New function
* libavrdude.h: declare compare_memory_masked(); also, insist on C99
so <stdint.h> is required now
* main.c: Use compare_memory_masked() in safemode comparisons
C99 / stdint.h has basically already been required before, as types
like uint8_t are in use in a number of other locations throughout the
source.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1470 81a1dc3b-b13d-400b-aceb-764788c761c2
* misleading indentation after spin loop
* EM_AVR32 might be defined in system <libelf.h> already
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1431 81a1dc3b-b13d-400b-aceb-764788c761c2
patch #9507: Fix UPDI chip erase
* libavrdude.h (PROGRAMMER): add unlock method
* avr.c (avr_unlock): Generic unlock function
* jtag3.c: Implement unlock feature; avoid calling
jtag3_edbg_prepare() and jtag3_edbg_signoff() on XplainedMini
boards to work around a bug in early firmware versions;
implement "userrow" memory region
* main.c: Call avr_unlock() rather than avr_chip_erase() when
encountering a locked UPDI chip
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1417 81a1dc3b-b13d-400b-aceb-764788c761c2
* avr.c: avoid division by zero in report_progress(), eg. when
writing an empty eeprom file were total becomes 0
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1340 81a1dc3b-b13d-400b-aceb-764788c761c2
serial.h fileio.h safemode.h update.h pgm_type.h config.h confwin.h
lists.h) into a single header that can be included by anyone wanting
to link against the library.
Adapt everything to cope with this situation.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1311 81a1dc3b-b13d-400b-aceb-764788c761c2
This change was made for the shared library, since library functions
should not write to std-streams directly. Instead avrdude_message()
has to be implemented by the library user. For the avrdude application
this function is implemented in main.c.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1305 81a1dc3b-b13d-400b-aceb-764788c761c2
yielded 0, resulting in a division by 0 in report_progress().
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1221 81a1dc3b-b13d-400b-aceb-764788c761c2
* avr.c (avr_read, avr_write): Change the progress reporting for
paged read/write from per-address to per-considered-page. This
ought to give a realistic estimation about the time still to be
spent.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1220 81a1dc3b-b13d-400b-aceb-764788c761c2
Generic function enabling external programming on TPI devices. This
function sets the required guard time (which is passed in as parameter),
checks the TPI identification register, writes SKEY command + SKEY, and
finally polls the NVMEN bit in TPISR.
avr.h: Add prototype definitions of avr_tpi_program_enable() and
avr_tpi_chip_erase().
avrftdi_tpi.c: Removes tpi_skey_cmd array, containing the SKEY command
and the SKEY bytes.
tpi.h: Adds tpi_skey_cmd array, containing the SKEY command and the SKEY
bytes, but in the reverse order of tpi_skey.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1167 81a1dc3b-b13d-400b-aceb-764788c761c2
supported error
* stk500v2.c: Handle all Xmega memory sections (except
"prodsig" which is not documented in AVR079)
* fileio.c: Treat the "boot", "application", and "apptable"
regions (which are actually subregions of "flash") all as
being flash, i.e. suppress trailing 0xFF bytes when reading
them
* avr.c: (Dito.)
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1075 81a1dc3b-b13d-400b-aceb-764788c761c2
functions. So a negative return code led never to a fallback to byte
functions.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1042 81a1dc3b-b13d-400b-aceb-764788c761c2
their programmer doesn't provide a (low-level) cmd_tpi method;
instead, fall back to the normal programmer methods which are
supposed to handle the situation.
This fixes a regression where the recent bitbang-TPI implementation
broke TPI handling of STK600/AVRISPmkII.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1008 81a1dc3b-b13d-400b-aceb-764788c761c2
Each memory image byte is now tagged as it's being read from a file.
Only bytes read from a file will be written or verified (modulo page
granularity requirements).
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1007 81a1dc3b-b13d-400b-aceb-764788c761c2
programmers)
* config_gram.y: Introduce new keyword "is_at90s1200".
* lexer.l: (Ditto.)
* avrdude.conf.in: Applew new keyword to the AT90S1200 device.
* avrpart.h: Introduce new flag AVRPART_IS_AT90S1200, reflecting
the is_at90s1200 configuration keyword.
* bitbang.c (bitbang_initialize): Replace existing test for
AT90S1200 by AVRPART_IS_AT90S1200
* avr.c (avr_write_byte_default): Avoid the pre-write reading for
the AT90S1200, as this appears to sometimes corrupt the high byte
by pre-programming the low byte just written into it.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@1002 81a1dc3b-b13d-400b-aceb-764788c761c2
* avrdude.conf.in: Correctly declare EEPROM page sizes for
all Xmega devices (0x20 instead of 0x100).
* avr.c: If a memory region has a page size declared, try
using the paged IO routines regardless of the target memory
name. Xmega EEPROM requires to be written in paged mode.
Correctly use a long (rather than unsigned long) variable to
evaluate the success status of the paged mode write attempt.
* stk500v2.c: Don't apply TIF space offsets twice (bug #27995:
AVRDUDE 5.8svn fails to program and read XMEGA); use
stk500v2_loadaddr() prior to paged mode (EEPROM and flash) writes,
otherwise programming of flash areas will fail; while being there,
check the return value of stk500v2_loadaddr() everywhere; use the
correct write/erase mode bits (same as AVR Studio does).
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@907 81a1dc3b-b13d-400b-aceb-764788c761c2
PDI mode of the STK600 is supported. Single-byte EEPROM
(and flash) updates do not work yet.
* avr.c: "boot" memory is a candidate memory region for paged
operations, besides "flash" and "eeprom".
* avrdude.conf.in: add ATxmega128A1 and ATxmega128A1revD
* avrpart.h: add the AVRPART_HAS_PDI flag (used to distinguish
ATxmega parts from classic AVRs), the nvm_base part field, and
the offset field for a memory region.
* config_gram.y: add "has_pdi", "nvm_base", and "offset"
* lexer.l: (Ditto.)
* main.c: disable auto_erase for ATxmega parts
* stk500v2.c: implement the XPROG functionality, and divert to
this for ATxmega parts
* avrdude.1: Document the changes.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@777 81a1dc3b-b13d-400b-aceb-764788c761c2
paged_write backend functions iff the memory area in question has
a page_size != 0.
This is supposed to fix bug #19234: avrdude-5.3.1 segfaults when
stk500v1 tries to program an ATtiny15
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@739 81a1dc3b-b13d-400b-aceb-764788c761c2
- Make all internal functions "static".
- Make sure each module's header and implementation file match.
- Remove all library-like functionality from main.c, so only
the actual frontend remains in main.c.
- Add C++ brackets to all header files.
That effectively leaves the various module C files as something like
an "avrdude library", with main.c being the currently only frontend
program for that library. In theory, it should be possible to write
different frontends using the same library backend functions though.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@722 81a1dc3b-b13d-400b-aceb-764788c761c2
declarations in each file by a central header file "avrdude.h".
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@721 81a1dc3b-b13d-400b-aceb-764788c761c2
avr_write(), respectively, in case the paged_load()/paged_write()
methods succeeded, rather than only 0.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@696 81a1dc3b-b13d-400b-aceb-764788c761c2
fallback from each programmer's paged_load() or paged_write()
method, respectively. The return value needs to be checked for
being greater or equal than 0 rather equal to 0 in order to
assume the operation has been successful.
Fixes bug #18489: avrdude is too slow (20 byte/s)
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@695 81a1dc3b-b13d-400b-aceb-764788c761c2
avr_read_byte_default() and avr_write_byte_default (resp.) by directly
calling the latter functions from within all programmers that don't
implement their own read_byte()/write_byte() methods. In turn, make the
read_byte() and write_byte() methods mandatory, and the cmd() method
(direct ISP command) optional instead (it's effectively mandatory for
any programmer using avr_read_byte_default()/avr_write_byte_default()
though). Remove all the pointless cmd() method stubs from those programmers
that don't need it.
Eliminate avr_read_byte() as it was now completely identical to
pgm->read_byte().
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@684 81a1dc3b-b13d-400b-aceb-764788c761c2
definitions that are organized in pages.
* avr.c (avr_write_byte_default): Consider using the loadpage
instructions only if the respective memory is marked "paged".
Closes bug #17199: EEPROM fails verification on ATmega645 with
pony-stk200 hardware
Closes bug #16849: EEPROM write fails for AT90USB1287 with mode 0x41
Closes bug #15146: stk500v2_paged_write: loadpage instruction not
defined for part
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@622 81a1dc3b-b13d-400b-aceb-764788c761c2
jtag2 and bitbang programmers are working, stk500v2
still needs to be done.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@581 81a1dc3b-b13d-400b-aceb-764788c761c2
* add support for parts with just 'fuse' memory
* if any fuse memories are altered, reflect those changes in the
post-programming safemode check so that safemode no longer
complains about fuses which were intentionally altered; this
eliminates the need to completely disable safemode using -u in
order to program fuses.
* provide -s option which will not ask to restore fuses, it will
just do it
Submitted by: Colin O'Flynn <coflynn@newae.com>
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@519 81a1dc3b-b13d-400b-aceb-764788c761c2
*avr.h: Removed prototypes for moved functions
*avrpart.h: Added prototypes for functions in avrpart.c
*Makefile.am: Added new file avrpart.c
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@388 81a1dc3b-b13d-400b-aceb-764788c761c2
resides above the last non-0xff data value in the address space. Only
do this for flash memory since writing a 0xff to flash is a no-op.
This has the affect of creating smaller output files when dumping
memory contents from flash if the program in flash does not consume
the whole memory space. It also results in shorter programming times
when avrdude is asked to load a file into flash that has lots of 0xff
filled data past the last non-0xff data value.
I think this is basically where Alexey was going with his s-record
routine, but this should have a similar affect for all the I/O
routines. The main difference is that Alexey's also optimized 0xff
from the beginning of the address space and was not limited to flash.
I think that these optimizations should be limited to the flash since
it is currently the only memory that treats 0xff as special.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@331 81a1dc3b-b13d-400b-aceb-764788c761c2
Have avr_read_byte() call pgm->read_byte() or avr_read_byte_default().
Add avr_write_byte_default().
Have avr_write_byte() call pgm->write_byte or avr_write_byte_default().
* pgm.c: Initialize pgm->write_byte and pgm->read_byte.
* pgm.h: Add write_byte and read_byte fields to struct programmer_t.
git-svn-id: svn://svn.savannah.nongnu.org/avrdude/trunk/avrdude@296 81a1dc3b-b13d-400b-aceb-764788c761c2
Joerg Wunsch
Rene Liebscher
Axel Wachtler
Hannes Weisbach
Brian S. Dean
Jan-Hinnerk Reichert
Theodore A. Roth