openocd.git
7 years agostm32f1x: use async algorithm in flash programming routine
Andreas Fritiofson [Sun, 31 Jul 2011 08:31:56 +0000 (10:31 +0200)]
stm32f1x: use async algorithm in flash programming routine

Let the target algorithm be running in the background and buffer data
continuously through a FIFO. This reduces or removes the effect of latency
because only a very small number of queue executions needs to be done per
buffer fill. Previously, the many repeated target state changes, register
accesses (really inefficient) and algorithm uploads caused the flash
programming to be latency bound in many cases. Now it should scale better
with increased throughput.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
7 years agostm32f1x: use register base instead of register offset
Andreas Fritiofson [Sun, 17 Jul 2011 12:07:26 +0000 (14:07 +0200)]
stm32f1x: use register base instead of register offset

Access the different flash banks' registers using a bank specific register
base and a register specific offset. This is equivalent but feels more
natural.

Some accesses were discovered that maybe should not be hard coded to bank0
registers. Add a note about that.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
7 years agocortex_m3: use armv7m's async algorithm implementation
Andreas Fritiofson [Fri, 15 Jul 2011 20:21:34 +0000 (22:21 +0200)]
cortex_m3: use armv7m's async algorithm implementation

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
7 years agoarmv7m: implement async algorithm functions
Andreas Fritiofson [Fri, 15 Jul 2011 20:20:34 +0000 (22:20 +0200)]
armv7m: implement async algorithm functions

Split armv7m_run_algorithm into two pieces and use them to reimplement it.
The arch_info parameter is used to keep context between the two calls, so
both calls must refer to the same armv7m_algorithm struct. Ugly but works
for a proof-of-concept.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
7 years agotarget: add async algorithm entries to the target type
Andreas Fritiofson [Fri, 15 Jul 2011 20:18:39 +0000 (22:18 +0200)]
target: add async algorithm entries to the target type

On supported targets, this may be used to start a long running algorithm in
the background so the target may be interacted with during execution and
later wait for its completion.

The most obvious use case is a double buffered flash algorithm that can
upload the next block of data while the algorithm is flashing the current.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
7 years agoarm-jtag-ew: Send GDB keep_alive() messages when logging USB communication
Simon Barner [Fri, 16 Sep 2011 19:15:19 +0000 (21:15 +0200)]
arm-jtag-ew: Send GDB keep_alive() messages when logging USB communication

- Ticket: #35

7 years agoarm-jtag-ew: Formatting
Simon Barner [Fri, 16 Sep 2011 19:14:39 +0000 (21:14 +0200)]
arm-jtag-ew: Formatting

7 years agoarm-jtag-ew: In armjtagew_init(), set initial JTAG speed to 32 kHz (before TAP initia...
Simon Barner [Fri, 16 Sep 2011 19:13:39 +0000 (21:13 +0200)]
arm-jtag-ew: In armjtagew_init(), set initial JTAG speed to 32 kHz (before TAP initialization).

This prevents rare communication errors during startup.

7 years agoarm-jtag-ew: Emit a warning if interface firmware version != 1.6
Simon Barner [Fri, 16 Sep 2011 19:12:47 +0000 (21:12 +0200)]
arm-jtag-ew: Emit a warning if interface firmware version != 1.6

7 years agoarm-jtag-ew: Declare interface as `jtag_only'
Simon Barner [Fri, 16 Sep 2011 19:11:40 +0000 (21:11 +0200)]
arm-jtag-ew: Declare interface as `jtag_only'

7 years agoarm-jtag-ew: Provide armjtagew_speed_div() in order to fix interactive use of `adapte...
Simon Barner [Fri, 16 Sep 2011 19:10:23 +0000 (21:10 +0200)]
arm-jtag-ew: Provide armjtagew_speed_div() in order to fix interactive use of `adapter_khz'

7 years agoarm-jtag-ew: Fix setting interface speed (2/2)
Simon Barner [Fri, 16 Sep 2011 19:08:10 +0000 (21:08 +0200)]
arm-jtag-ew: Fix setting interface speed (2/2)

Interface expects speed in Hz, not kHz

- Ticket #34

7 years agoarm-jtag-ew: Fix setting interface speed (1/2)
Simon Barner [Fri, 16 Sep 2011 19:06:02 +0000 (21:06 +0200)]
arm-jtag-ew: Fix setting interface speed (1/2)

CMD_SET_TCK_FREQUENCY message length is 5, not 4

- Ticket: #34

7 years agoAdd udev rules for openmoko neo1973 debug board
Eugeniy Meshcheryakov [Tue, 4 Oct 2011 13:14:02 +0000 (15:14 +0200)]
Add udev rules for openmoko neo1973 debug board

7 years agoSTM32L: Added flash driver and target
Clément Burin des Roziers [Fri, 16 Sep 2011 13:55:54 +0000 (15:55 +0200)]
STM32L: Added flash driver and target

Added the flash driver for the STM32L family, which highly differ from the STM32F family.
Added the TCL target file for JTAG access.

7 years agoVerdex: Add support for Gumstix Verdex boards.
Ash Charles [Fri, 30 Sep 2011 23:17:21 +0000 (16:17 -0700)]
Verdex: Add support for Gumstix Verdex boards.

Gumstix Verdex is a PXA270-based series of computer-on-modules. This
configuration file is based off the voipac.cfg configuration with
a different flash memory configuration. This has been tested flyswatter
adapter to reflash a Gumstix Verdex XL6P board.

7 years agou8500 : config for L2 cache
Michel Jaouen [Mon, 19 Sep 2011 14:49:23 +0000 (16:49 +0200)]
u8500 : config for L2 cache

7 years agoarmv7a ,cortex a : add L1, L2 cache support, va to pa support
Michel Jaouen [Thu, 29 Sep 2011 15:17:27 +0000 (17:17 +0200)]
armv7a ,cortex a : add L1, L2 cache support, va to pa support

7 years agojim-nvp is moving from jimtcl to openocd
Steve Bennett [Thu, 29 Sep 2011 06:12:41 +0000 (16:12 +1000)]
jim-nvp is moving from jimtcl to openocd

The jim-nvp code is specific to openocd, so it belongs in openocd,
not in the core jimtcl.

Signed-off-by: Steve Bennett <steveb@workware.net.au>
7 years agoAM/DM37x: add ES1.2 silicon type into account
Vladimir Zapolskiy [Wed, 28 Sep 2011 16:41:23 +0000 (19:41 +0300)]
AM/DM37x: add ES1.2 silicon type into account

The missing value for ES1.2 silicon revision is mentioned in
sprugn4m.pdf, and the recent TI Beagleboard XM is powered by it,
so let support the revision.

7 years agoadd target events, run algorithm and default r/w buffer api
Mathias K [Wed, 21 Sep 2011 17:18:26 +0000 (19:18 +0200)]
add target events, run algorithm and default r/w buffer api

Target events are added to get better gdb support. The run
algorithm functionality are implemented to support feature
fast flash write functionality. The new r/w buffer api is now
used to support the special memory address handling. The output
of the md command was fixed.

7 years agokinetis : fix deadlock on device having hasidcode false.
Michel Jaouen [Tue, 20 Sep 2011 07:37:52 +0000 (09:37 +0200)]
kinetis : fix deadlock on device having hasidcode false.

8 years agokinetis cpu flash driver
Mathias K [Sat, 17 Sep 2011 08:09:50 +0000 (10:09 +0200)]
kinetis cpu flash driver

Initial release of the freescale kinetis cpu flash driver.

8 years agocontrib: fix udev rules for tty based adaptors
Luca Bruno [Tue, 13 Sep 2011 13:11:01 +0000 (15:11 +0200)]
contrib: fix udev rules for tty based adaptors

Most serial adaptors are identified by udev with SUBSYSTEM=tty and
without DEVTYPE. This patch fix udev rules to work with any listed
tty-based adaptor. It has been tested with a FTDI-based Bus Pirate.

Signed-off-by: Luca Bruno <lucab@debian.org>
8 years agokinetis auto mass erase on secured devices
Mathias K [Mon, 12 Sep 2011 19:24:17 +0000 (21:24 +0200)]
kinetis auto mass erase on secured devices

This is a proof of concept to get access to the debug port of a
secured kinetis cpu. On full flash erase the cpu is automatically
secured and the debug port is not accessible.
To get this to work the srst line is needed and the necessary
configuration should be added to the configuration file.

8 years agoFixes and spellchecks for various Buspirate output messages
Luca Bruno [Sat, 10 Sep 2011 15:01:46 +0000 (17:01 +0200)]
Fixes and spellchecks for various Buspirate output messages

Signed-off-by: Luca Bruno <lucab@debian.org>
8 years agoAdd definition for the STEVAL-PCC010 board with the STM32F207
Uwe Bonnes [Thu, 1 Sep 2011 15:14:35 +0000 (17:14 +0200)]
Add definition for the STEVAL-PCC010 board with the STM32F207

8 years agoImplementation of a new jtag remote_bitbang driver.
Richard Uhler [Thu, 25 Aug 2011 20:29:33 +0000 (13:29 -0700)]
Implementation of a new jtag remote_bitbang driver.

The driver sends ascii encoded bitbang commands over unix sockets or TCP to
another process. This driver is useful for debugging software running on
processors which are being simulated.

8 years agoULINK driver: Remove typedefs in ulink.c
Martin Schmölzer [Wed, 31 Aug 2011 15:22:21 +0000 (17:22 +0200)]
ULINK driver: Remove typedefs in ulink.c

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Remove typedefs in OpenULINK firmware: Use typedefs from stdint.h ...
Martin Schmölzer [Wed, 31 Aug 2011 14:51:41 +0000 (16:51 +0200)]
ULINK driver: Remove typedefs in OpenULINK firmware: Use typedefs from stdint.h (uint8_t, uint16_t) instead of custom typedefs in shorttypes.h (u8, u16, ...)

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Remove typedefs in OpenULINK firmware USB descriptor structures
Martin Schmölzer [Tue, 30 Aug 2011 22:20:35 +0000 (00:20 +0200)]
ULINK driver: Remove typedefs in OpenULINK firmware USB descriptor structures

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agousleep to jtag_sleep
Rodrigo L. Rosa [Wed, 31 Aug 2011 22:43:11 +0000 (15:43 -0700)]
usleep to jtag_sleep

this will help avoid platform specific timing issues

8 years agostatic for some functions
Rodrigo L. Rosa [Wed, 31 Aug 2011 19:37:20 +0000 (12:37 -0700)]
static for some functions

made two functions into static, since they are not required by anything external

8 years agospeed up, relocate function
Rodrigo L. Rosa [Wed, 31 Aug 2011 19:32:14 +0000 (12:32 -0700)]
speed up, relocate function

added an attempt to use the non-reseting halting sequence. if it fails, then the full sequence will be attempted. this makes things a bit faster most of the time.
changed the location of a function, avoiding a forward def

8 years agoMerge branch 'dsp5680xx_cherry' of git://repo.or.cz/openocd/dsp568013 into fix
Øyvind Harboe [Wed, 31 Aug 2011 14:27:09 +0000 (16:27 +0200)]
Merge branch 'dsp5680xx_cherry' of git://repo.or.cz/openocd/dsp568013 into fix

8 years agoULINK driver: Update firmware image to reflect the latest changes in source code.
Martin Schmoelzer [Tue, 23 Aug 2011 15:22:53 +0000 (17:22 +0200)]
ULINK driver: Update firmware image to reflect the latest changes in source code.

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Add '-lm' linker flag when building this driver (required for correct...
Martin Schmoelzer [Thu, 4 Aug 2011 20:30:25 +0000 (22:30 +0200)]
ULINK driver: Add '-lm' linker flag when building this driver (required for correct calculation of JTAG TCK speed setting)

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Implement variable TCK frequency in OpenOCD driver
Martin Schmoelzer [Thu, 4 Aug 2011 17:03:59 +0000 (19:03 +0200)]
ULINK driver: Implement variable TCK frequency in OpenOCD driver

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Implement variable TCK frequency in OpenULINK firmware
Martin Schmölzer [Mon, 4 Jul 2011 17:03:29 +0000 (19:03 +0200)]
ULINK driver: Implement variable TCK frequency in OpenULINK firmware

Also, speed up jtag_clock_tck() significantly (150 kHz -> 375 kHz)

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Fix whitespace in OpenULINK firmware usb.c module (trivial)
Martin Schmoelzer [Tue, 23 Aug 2011 12:21:56 +0000 (14:21 +0200)]
ULINK driver: Fix whitespace in OpenULINK firmware usb.c module (trivial)

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Implement JTAG_PATHMOVE command
Martin Schmoelzer [Tue, 23 Aug 2011 14:53:03 +0000 (16:53 +0200)]
ULINK driver: Implement JTAG_PATHMOVE command

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Implement JTAG_STABLECLOCKS command
Martin Schmölzer [Mon, 4 Jul 2011 10:29:02 +0000 (12:29 +0200)]
ULINK driver: Implement JTAG_STABLECLOCKS command

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Implement command to manually force downloading firmware image from...
Martin Schmölzer [Mon, 4 Jul 2011 17:50:46 +0000 (19:50 +0200)]
ULINK driver: Implement command to manually force downloading firmware image from arbitrary location

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Re-order queue functions to reflect the order in commands.h
Martin Schmölzer [Mon, 4 Jul 2011 10:38:47 +0000 (12:38 +0200)]
ULINK driver: Re-order queue functions to reflect the order in commands.h

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Properly propagate return values in ulink_execute_queue()
Martin Schmölzer [Mon, 4 Jul 2011 10:30:13 +0000 (12:30 +0200)]
ULINK driver: Properly propagate return values in ulink_execute_queue()

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agoULINK driver: Update some comments, fix some coding mistakes
Martin Schmölzer [Sun, 26 Jun 2011 23:00:15 +0000 (01:00 +0200)]
ULINK driver: Update some comments, fix some coding mistakes

Signed-off-by: Martin Schmölzer <martin.schmoelzer@student.tuwien.ac.at>
8 years agovsllink driver compile fails with'vsllink_debug_buffer' defined but not used
simonqian.openocd [Tue, 30 Aug 2011 17:52:20 +0000 (01:52 +0800)]
vsllink driver compile fails with'vsllink_debug_buffer' defined but not used

USB communication is handled by code under versaloon directory.
So _DEBUG_USB_COMMS_  should not be used in vsllink.c.
Attachment is the patch.

8 years agoremoved trailing whitespaces
Rodrigo L. Rosa [Wed, 31 Aug 2011 04:14:33 +0000 (21:14 -0700)]
removed trailing whitespaces

emacs is awesome.
replace-regexp RET [ ]+$ RET RET
and it's done

8 years agofix enter debug mode for locking
Rodrigo L. Rosa [Wed, 31 Aug 2011 04:08:59 +0000 (21:08 -0700)]
fix enter debug mode for locking

added an alternative way to enter debug mode, which does not require restarting the chip.
this will not always work, but in general it will (failure 0.3%), and failure is not a dramatic issue, simply have to use the full sequence.
the user can only access "halt", which uses the full sequence, so the user should not have any problems.
restarting the chip requires reconfiguring the flash module. the doc is very poor, so i'd rather have the two methods, and live with the 0.3%.

8 years agofix irlen handling
Rodrigo L. Rosa [Tue, 30 Aug 2011 21:46:20 +0000 (14:46 -0700)]
fix irlen handling

sometimes the master tap will be enabled, since tap switching is required during halt/lock/unlocking procedures.
now irscan handles this, avoiding unnecessary warnings and preventing errors.

8 years agofix debug mode,lock,unlock
Rodrigo L. Rosa [Tue, 30 Aug 2011 22:13:49 +0000 (15:13 -0700)]
fix debug mode,lock,unlock

got new info regarding setting the chip to debug mode, and locking/unlocking flash memory.
the newer implementation is a bit slower, but always works.
the previous implementation would randomly (as once every 25k-70k times) get the chip into a state where the freescale tool would be necessary. this is fixed now.

added functions to play around with the jtag state machine. they are not the happiest, but are necessary to be able to execute the halting/locking/unlocking sequences.

Conflicts:

src/target/dsp5680xx.c

8 years agofix error handling during halt
Rodrigo L. Rosa [Tue, 30 Aug 2011 21:26:05 +0000 (14:26 -0700)]
fix error handling during halt

the user can execute halt, but no enter_debug_mode. modified the error handling to suite this.
the new implementation of unlocking will use enter_debug_mode, and should not get the same errors as the user would, because not being able to enter debug mode is actually success when checking for locked flash.

8 years agooptional crc for flash writing
Rodrigo L. Rosa [Tue, 30 Aug 2011 21:19:09 +0000 (14:19 -0700)]
optional crc for flash writing

crc check was always performed on newly flashed data, now it is optional
flash mem can be locked by writing a specific word to a specific address in flash.
to verify flash, target must be halted, and this will (when the new halt sequence is implemented) require reseting the chip. if the target is reset after writing the lock words, then it will lock, hence the CRC will fail because it is not possible to read stuff from the target.

also added a function that resets the jtag state machine.
this is not used yet, but will be soon.
it is implemented to allow strict control over JTAG state machine, necessary to implement to halt and unlocking sequences.

8 years agodef syntax to match tap irlen
Rodrigo L. Rosa [Tue, 30 Aug 2011 20:49:44 +0000 (13:49 -0700)]
def syntax to match tap irlen

the master tap has a 4 bit irlen
changed the instructions to be 4 bit, ie, removed the zeros.
it makes it clearer to interpret.

8 years agodsp568013 disable polling by default
Rodrigo L. Rosa [Tue, 16 Aug 2011 02:50:59 +0000 (19:50 -0700)]
dsp568013 disable polling by default

8 years agocortex_a hybrid & context breakpoints
Heythem Bouhaja [Mon, 9 May 2011 08:40:35 +0000 (10:40 +0200)]
cortex_a hybrid & context breakpoints

8 years agoFix off by one bug in FreeRTOS
Evan Hunter [Fri, 26 Aug 2011 06:30:45 +0000 (16:30 +1000)]
Fix off by one bug in FreeRTOS

8 years agoAdd Valuline HD to config file
Jonathan Dumaresq [Wed, 24 Aug 2011 20:13:35 +0000 (16:13 -0400)]
Add Valuline HD to config file

This will add the BSTAP for the medium and high density devices

8 years agoThis will add the Value Line HD of stm32 devices.
Jonathan Dumaresq [Wed, 24 Aug 2011 20:08:18 +0000 (16:08 -0400)]
This will add the Value Line HD of stm32 devices.

This has been tested on STM32F100VE

8 years agoremove target argument from gdb packet handling functions
Jie Zhang [Wed, 24 Aug 2011 15:23:04 +0000 (11:23 -0400)]
remove target argument from gdb packet handling functions

8 years agoAdd suspended task list to FreeRTOS support
Evan [Mon, 18 Jul 2011 03:47:24 +0000 (13:47 +1000)]
Add suspended task list to FreeRTOS support

8 years agoFix FreeRTOS thread list parsing
Evan Hunter [Tue, 23 Aug 2011 04:02:24 +0000 (14:02 +1000)]
Fix FreeRTOS thread list parsing

8 years agoFix redbee config files
Jim Paris [Mon, 22 Aug 2011 20:48:14 +0000 (16:48 -0400)]
Fix redbee config files

Currently the board/redbee-*.cfg files incorrectly include the
interface definition.  Move the interfaces to interface/,
and create a single board/redbee.cfg that is common to both boards.
Intended usage is now:
  openocd -f interface/redbee-econotag.cfg -f board/redbee.cfg

8 years agonon_cfi: add SST39WF1601 support
Andreas Bießmann [Fri, 19 Aug 2011 14:53:40 +0000 (16:53 +0200)]
non_cfi: add SST39WF1601 support

Invented by jknick in sparkfun forum:
http://forum.sparkfun.com/viewtopic.php?t=19788

Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
8 years agoFix Sam3u flash bank 1 issue
olivier Schonken [Sat, 20 Aug 2011 14:31:35 +0000 (16:31 +0200)]
Fix Sam3u flash bank 1 issue

8 years agocfi: add EN29LV800BB support
Gunnar Henne [Thu, 18 Aug 2011 08:52:16 +0000 (09:52 +0100)]
cfi: add EN29LV800BB support

Posted by telekatz@gmx.de in the bettyhacks forum for openocd
0.4.0 and integrated into 0.5.0 by Gunnar Henne.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
8 years agoremove white space before TAB
Jie Zhang [Wed, 17 Aug 2011 15:21:22 +0000 (11:21 -0400)]
remove white space before TAB

8 years agobuild: check buspirate build host
Spencer Oliver [Tue, 16 Aug 2011 16:33:19 +0000 (17:33 +0100)]
build: check buspirate build host

buspirate has never supported building on native windows (mingw).
configure will now check this is not the case.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
8 years agoversaloon driver update
SimonQian [Mon, 11 Jul 2011 17:48:05 +0000 (01:48 +0800)]
versaloon driver update

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
8 years agobuild: check guess-rev.sh can be found/executed
Spencer Oliver [Tue, 16 Aug 2011 11:06:57 +0000 (12:06 +0100)]
build: check guess-rev.sh can be found/executed

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
8 years agobuild: rename configure.in to configure.ac
Spencer Oliver [Tue, 16 Aug 2011 11:00:27 +0000 (12:00 +0100)]
build: rename configure.in to configure.ac

configure.ac is the correct name to use with modern autotools.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
8 years agojim: update to fix mingw/msys build issues
Spencer Oliver [Tue, 16 Aug 2011 10:58:06 +0000 (11:58 +0100)]
jim: update to fix mingw/msys build issues

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
8 years agoshow git commit number even when cross-compiling
Jie Zhang [Mon, 15 Aug 2011 14:52:11 +0000 (10:52 -0400)]
show git commit number even when cross-compiling

AC_CHECK_FILE will die when cross-compiling. So don't use it to test the existence of guess-rev.sh.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
8 years agojimtcl: delete OpenOCD's broken 'stacktrace' command
Øyvind Harboe [Sun, 14 Aug 2011 16:28:12 +0000 (18:28 +0200)]
jimtcl: delete OpenOCD's broken 'stacktrace' command

Use "info stacktrace" instead. This fixes build problems with
latest Jim Tcl.

8 years agomips: fix potential alignment error
Stefan Mahr [Wed, 10 Aug 2011 21:42:28 +0000 (23:42 +0200)]
mips: fix potential alignment error

8 years agomips: fix reading uint32 and uint16 when running on big endian host
Stefan Mahr [Wed, 10 Aug 2011 21:29:20 +0000 (23:29 +0200)]
mips: fix reading uint32 and uint16 when running on big endian host

8 years agotarget: add helper functions to get/set u16 or u32 array from/to buffer
Stefan Mahr [Wed, 10 Aug 2011 20:34:41 +0000 (22:34 +0200)]
target: add helper functions to get/set u16 or u32 array from/to buffer

8 years agoftd2xx: fix build warnings
Spencer Oliver [Wed, 10 Aug 2011 21:23:29 +0000 (22:23 +0100)]
ftd2xx: fix build warnings

Due to build warnings introduced in newer versions of ftd2xx we
use strings to report errors rather than result codes. This also
gives us the same behaviour as libftdi.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
8 years agoremove useless pxref to SMP subsection
Jie Zhang [Wed, 10 Aug 2011 19:32:09 +0000 (15:32 -0400)]
remove useless pxref to SMP subsection

8 years agoMerge branch 'dsp5680xx_cherry' of git://repo.or.cz/openocd/dsp568013 into HEAD
Øyvind Harboe [Thu, 11 Aug 2011 14:23:15 +0000 (16:23 +0200)]
Merge branch 'dsp5680xx_cherry' of git://repo.or.cz/openocd/dsp568013 into HEAD

8 years agoEvaluate 'script' in the global scope
Steve Bennett [Thu, 11 Aug 2011 02:10:54 +0000 (12:10 +1000)]
Evaluate 'script' in the global scope

Scripts sourced via 'script' should evaluate in the global
scope to make it easy to set and reference global variables.

Signed-off-by: Steve Bennett <steveb@workware.net.au>
8 years agofix return code from dsp5680xx_read
Rodrigo L. Rosa [Fri, 5 Aug 2011 18:40:05 +0000 (11:40 -0700)]
fix return code from dsp5680xx_read

it returned ERROR_OK even though it actually failed.
this made the Tcl interface report success, though it had not succeeded.

8 years agorenamed for clarity
Rodrigo L. Rosa [Fri, 5 Aug 2011 18:38:12 +0000 (11:38 -0700)]
renamed for clarity

i had started my code from dsp5683xx, i renamed a bunch of stuff to names i consider to be better.
i believe no one is using this code, so nobody should be affected. (it's not too late to do this change)

8 years agorlink: read only the expected number of bytes
Andreas Fritiofson [Sat, 16 Jul 2011 21:37:40 +0000 (23:37 +0200)]
rlink: read only the expected number of bytes

After correcting the reply size counter, it should be safe to rely on it
for the number of bytes expected in the USB read, instead of reading the
endpoint maximum. This doesn't make things go any faster but it's nicer and
removes the local buffer.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
8 years agorlink: simplify and optimize queue fill level checks
Andreas Fritiofson [Sat, 16 Jul 2011 23:02:59 +0000 (01:02 +0200)]
rlink: simplify and optimize queue fill level checks

Add a helper function for running the queue if it would overflow otherwise.
Use it to simplify the queue fill level checks and optimize in a few cases
that would previously run the queue prematurely.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
8 years agorlink: remove redundant text from log messages
Andreas Fritiofson [Sat, 16 Jul 2011 21:41:44 +0000 (23:41 +0200)]
rlink: remove redundant text from log messages

__FILE__ and __LINE__ are already printed using the log macros.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
8 years agorlink: remove duplicated code
Andreas Fritiofson [Sat, 16 Jul 2011 14:32:50 +0000 (16:32 +0200)]
rlink: remove duplicated code

After the reply_index handling is fixed, there's no need to special case
the out scan.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
8 years agorlink: fix reply counter to enable sending full buffers
Andreas Fritiofson [Sat, 16 Jul 2011 12:30:20 +0000 (14:30 +0200)]
rlink: fix reply counter to enable sending full buffers

dtc_queue.reply_index was wrongly being increased during out scans, causing
the queue to be sent before the out buffer was full. This patch increases
raw upload speed by 50% or so.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
8 years agorlink: more indentation fixes
Andreas Fritiofson [Sat, 16 Jul 2011 12:09:08 +0000 (14:09 +0200)]
rlink: more indentation fixes

Remove unnecessary block scopes to reduce indentation level.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
8 years agorlink: fix indentation errors
Andreas Fritiofson [Sat, 16 Jul 2011 12:03:39 +0000 (14:03 +0200)]
rlink: fix indentation errors

Indentation was inconsistent and some lines not indented at all. Quickfix
using Eclipse's auto-indentation.

Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
8 years agomips32 : Fixed memory byte access
Drasko DRASKOVIC [Mon, 25 Jul 2011 12:23:35 +0000 (14:23 +0200)]
mips32 : Fixed memory byte access

Function mips_m4k_write_memory() does endianess byte swap,
but this procedure break one byte access (temporary array
overwrites content in buffer).
As a fix, this endianess swap and buffer affecting
is preformed only on hword and word accesses (not on byte access).

8 years agomips32: Sync Caches to Make Instr Writes Effective
Drasko DRASKOVIC [Thu, 7 Jul 2011 15:59:13 +0000 (17:59 +0200)]
mips32: Sync Caches to Make Instr Writes Effective

Pprogram that loads another program into memory is actually writing the
D- side cache.
The instructions it has loaded can't be executed until they reach the
I-cache.

After the instructions have been written, the loader should arrange to
write back any containing D-cache line and invalidate any locations
already in the I-cache.

For the MIPS Architecture Release2 cores, we can use synci command
that does this job.
For Release1 we must use "cache" instruction.

8 years agomips32: Added CP0 coprocessor R/W routines
Drasko DRASKOVIC [Thu, 7 Jul 2011 15:41:20 +0000 (17:41 +0200)]
mips32: Added CP0 coprocessor R/W routines

This patch adds MIPS32 CP0 coprocessor R/W routines,
as well as adequate commands to use these routines via
telnet interface.

Now is becomes possible to affect CP0 internal registers
and configure CPU directly from OpenOCD.

8 years agomips_m4k: common_magic should be unsigned
Drasko DRASKOVIC [Thu, 7 Jul 2011 14:38:38 +0000 (16:38 +0200)]
mips_m4k: common_magic should be unsigned

For all architectures we use distinct common magic number,
and this should be a uint32_t type.
Otherwise, comparison with macros will yield compilation
warning.

8 years agomips32: Removed Unnecessary JTAG Queue Flush
Drasko DRASKOVIC [Tue, 5 Jul 2011 15:37:05 +0000 (17:37 +0200)]
mips32: Removed Unnecessary JTAG Queue Flush

jtag_execute_queue() is executed as a part of mips_ejtag_drscan_32().
No need for this to be done before - removed for optimisation.

8 years agofix tapenabler return code
Rodrigo L. Rosa [Tue, 9 Aug 2011 18:25:23 +0000 (11:25 -0700)]
fix tapenabler return code

if tap enable/disable failed then a warning was written to the log, but JIM_OK was returned. if using openocd via a TCP interface to the TCL port, there is no way to catch that the command failed (it didn't enable the tap, so it failed)
now it return an error if it fails.

8 years agofix return error msj
Rodrigo L. Rosa [Tue, 9 Aug 2011 18:24:42 +0000 (11:24 -0700)]
fix return error msj

retval was not correctly propagated

8 years agoflash/nor/cfi: fix TopBottom for atmel chips
Andreas Bießmann [Tue, 12 Jul 2011 10:38:22 +0000 (12:38 +0200)]
flash/nor/cfi: fix TopBottom for atmel chips

There are some older atmel nor chips which have negated logic for
TopBottom detection. This patch adds a special handling for the old
chips. This is the same mechanism as implemented in linux kernel.

Signed-off-by: Andreas Bießmann <andreas.devel@gmail.com>
8 years agoRevert "dsp5680xx: disable for now, it generates warnings"
Øyvind Harboe [Tue, 9 Aug 2011 18:14:50 +0000 (20:14 +0200)]
Revert "dsp5680xx: disable for now, it generates warnings"

This reverts commit d567df02b9f3e7d2e7e78b3c2907ecad9aa4bbd4.

8 years agodsp5680xx fix constante ref
Rodrigo L. Rosa [Wed, 20 Jul 2011 03:04:46 +0000 (20:04 -0700)]
dsp5680xx fix constante ref

a counter was incorrectly set
when i added the macros i incorrectly called them.
fixed that.