openocd.git
9 years agobuild: fix build from previous commit
Spencer Oliver [Wed, 8 Dec 2010 19:54:10 +0000 (19:54 +0000)]
build: fix build from previous commit

commit 740b9e25b410c164e661d0334a9ea4168406726b broke the drivers
for ftdi and parport due to retval already being defined.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
9 years agojtag: getting the JTAG speed can fail
Øyvind Harboe [Mon, 6 Dec 2010 08:21:23 +0000 (09:21 +0100)]
jtag: getting the JTAG speed can fail

If the JTAG speed has not been set, then it has no defined
value, add code to propagate the error.

No change to actual behavior as no new failure paths have
been introduced. This is a no-op patch to make subsequent patches
smaller.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
9 years agosvf: various improvements
Andrew Leech [Mon, 6 Dec 2010 21:05:56 +0000 (08:05 +1100)]
svf: various improvements

* added support for targeting particular tap
* improved file reading
* improved command line parsing
* added progress meter
* more readable time measurement output

9 years agolpc2478 target config: CCLK as (mandatory) parameter
Rolf Meeser [Fri, 3 Dec 2010 13:10:40 +0000 (14:10 +0100)]
lpc2478 target config: CCLK as (mandatory) parameter

9 years agoxscale: trace buffer remains enabled until explicitly disabled
Mike Dunn [Thu, 2 Dec 2010 19:08:28 +0000 (11:08 -0800)]
xscale: trace buffer remains enabled until explicitly disabled

Hi everyone,

Since a call went out for patches... been sitting on this for months.  For some
reason, the xscale trace buffer is automatically disabled as soon as a break
occurs and the trace data is collected.  This patch was a result of the
frustration of always re-enabling it, or else hitting a breakpoint and checking
the trace data, only to discover that I forgot to re-enable it before resuming.
Don't see why it should work this way.  There is no run-time penalty, AFAIK.

Along the way, I also cleaned up a little by removing the ugly practice of
recording wrap mode by setting the fill count variable to "-1", replacing it
with an enum that records the trace mode.

I've been using this for months.  Comments, criticisms gratefully received.

Mike

Signed-off-by: Mike Dunn <mikedunn@newsguy.com>
9 years agoAdd board config for Embedded Artists LPC2478-32
Rolf Meeser [Fri, 3 Dec 2010 13:03:28 +0000 (14:03 +0100)]
Add board config for Embedded Artists LPC2478-32

9 years agoFix flash name in Hitex LPC2929 board config
Rolf Meeser [Fri, 3 Dec 2010 12:54:47 +0000 (13:54 +0100)]
Fix flash name in Hitex LPC2929 board config

9 years agoFix sector layout for 504-KiB LPC2000 devices
Rolf Meeser [Fri, 3 Dec 2010 13:06:11 +0000 (14:06 +0100)]
Fix sector layout for 504-KiB LPC2000 devices

9 years agostm32: set default soft reset config
Spencer Oliver [Thu, 2 Dec 2010 13:12:48 +0000 (13:12 +0000)]
stm32: set default soft reset config

If no srst is configured then default to using sysresetreq to
reset the target.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
9 years agoluminary: remove unused config cmds.
Spencer Oliver [Thu, 2 Dec 2010 13:11:08 +0000 (13:11 +0000)]
luminary: remove unused config cmds.

Due to commit e40aee2954d2beabe1d8c530d9ff1e564fb01f48 we now honour the
targets 'reset_config' setting. Previously we ignored the srst setting
for luminary targets.

Luminary targets have never supported using srst to reset into debug mode
so remove the option from the target configs files.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
9 years agocortex_m3: change cortec_m3 reset_config behaviour
Spencer Oliver [Thu, 2 Dec 2010 12:58:50 +0000 (12:58 +0000)]
cortex_m3: change cortec_m3 reset_config behaviour

Currently the cmd 'cortex_m3 reset_config' will overide the default
target's 'reset_config'.
Chnage the behaviour to use the target 'reset_config' if configured and
fallback if not.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
9 years agoconfig: fix luminary jtag config
Spencer Oliver [Thu, 2 Dec 2010 22:01:44 +0000 (22:01 +0000)]
config: fix luminary jtag config

When this config was updated in commit e3773e3e3d1f1ee0dbb0b69e8babe8419784d1c1
the old jtag declaration was not removed.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
9 years agorename some target scripts to be consistent with the rest
Freddie Chopin [Wed, 1 Dec 2010 21:55:56 +0000 (22:55 +0100)]
rename some target scripts to be consistent with the rest

Rename Atmel target scripts which had wrong name ("at91" missing for ARM7 AT91SAM7..., "at" missing for AVR ATmega...)

Signed-off-by: Freddie Chopin <freddie_chopin@op.pl>
9 years agoprofile: use 100Hz as a default sampling frequency
Øyvind Harboe [Wed, 1 Dec 2010 14:28:56 +0000 (15:28 +0100)]
profile: use 100Hz as a default sampling frequency

it's a lie that is somewhere in the vicinity of the
truth. Certainly 64MHz confuses gprof and produces
zero output and no error messages.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
9 years agolpc2900.c: Improve debug output for device probing
Rolf Meeser [Wed, 1 Dec 2010 07:28:43 +0000 (08:28 +0100)]
lpc2900.c: Improve debug output for device probing

9 years agolpc2900.c: Add support for new device LPC2926
Rolf Meeser [Wed, 1 Dec 2010 07:22:29 +0000 (08:22 +0100)]
lpc2900.c: Add support for new device LPC2926

9 years agoLPC2900: Avoid false positive reprobing if first probing fails
Rolf Meeser [Sat, 27 Nov 2010 22:09:59 +0000 (23:09 +0100)]
LPC2900: Avoid false positive reprobing if first probing fails

9 years agoSome cosmetic fixes to the Lisa/L layout support functions.
Piotr Esden-Tempski [Mon, 29 Nov 2010 20:58:30 +0000 (12:58 -0800)]
Some cosmetic fixes to the Lisa/L layout support functions.

9 years agoAdded support for the blinking leds on Floss-JTAG v0.3 and newer.
Piotr Esden-Tempski [Mon, 29 Nov 2010 20:56:21 +0000 (12:56 -0800)]
Added support for the blinking leds on Floss-JTAG v0.3 and newer.

9 years agoUpdated Floss-JTAG config file to support v0.3 and newer. Also added noeeprom version...
Piotr Esden-Tempski [Mon, 29 Nov 2010 20:35:39 +0000 (12:35 -0800)]
Updated Floss-JTAG config file to support v0.3 and newer. Also added noeeprom version of the config file for older versions of Floss-JTAG.

10 years agobuild: fix make install with jimtcl
Spencer Oliver [Fri, 26 Nov 2010 14:43:44 +0000 (14:43 +0000)]
build: fix make install with jimtcl

Update subproject jimtcl to fix issue with make install.
see jimtcl commit 373b721510fd2d0754a41cc70a3b7cfd02e929bd

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agobuild: correct configure help message
Spencer Oliver [Fri, 26 Nov 2010 13:35:41 +0000 (13:35 +0000)]
build: correct configure help message

As we default to building jimtcl the help text should show that.
No change in functionality or configure args required.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agoFLASH/NOR: Rename spearsmi.c to stmsmi.c
Antonio Borneo [Mon, 22 Nov 2010 04:25:09 +0000 (12:25 +0800)]
FLASH/NOR: Rename spearsmi.c to stmsmi.c

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: rename from spearsmi to stmsmi
Antonio Borneo [Mon, 22 Nov 2010 04:21:31 +0000 (12:21 +0800)]
FLASH/NOR: rename from spearsmi to stmsmi

STMicroelectronics controller SMI is not SPEAr specific.
Rename it and change name to every symbol in the code.

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoSTR750: Add SMI interface support
Antonio Borneo [Wed, 17 Nov 2010 03:28:46 +0000 (11:28 +0800)]
STR750: Add SMI interface support

Modified spearsmi driver to include support for STR75x
Added missing initialization in tcl file for STR750

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoNOR/SPEARSMI: fix segfault
Antonio Borneo [Thu, 18 Nov 2010 07:01:03 +0000 (15:01 +0800)]
NOR/SPEARSMI: fix segfault

If flash chip is not listed in the table, or if no flash is
connected, pointer must be properly initialized.

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoDocumentation: fix typo
Antonio Borneo [Mon, 22 Nov 2010 11:32:57 +0000 (19:32 +0800)]
Documentation: fix typo

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoflash: iterating over an address range now handles multiple banks
Øyvind Harboe [Mon, 22 Nov 2010 10:16:40 +0000 (11:16 +0100)]
flash: iterating over an address range now handles multiple banks

e.g. flash erase_address now works across an address
range that spans multiple flash chips.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agofastload: fix error handling upon running out of memory
Øyvind Harboe [Mon, 22 Nov 2010 08:15:47 +0000 (09:15 +0100)]
fastload: fix error handling upon running out of memory

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agobuild: remove AC_CONFIG_AUX_DIR macro
Spencer Oliver [Fri, 19 Nov 2010 11:53:17 +0000 (11:53 +0000)]
build: remove AC_CONFIG_AUX_DIR macro

This was used during testing and should have been removed.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agobuild: disable jimtcl lineedit
Spencer Oliver [Fri, 19 Nov 2010 11:24:11 +0000 (11:24 +0000)]
build: disable jimtcl lineedit

This is a recent jimtcl feature but it currently breaks mingw
builds as this system does not have termios.h etc.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agobuild: fix subconfigure parameter issue
Spencer Oliver [Fri, 19 Nov 2010 09:57:37 +0000 (09:57 +0000)]
build: fix subconfigure parameter issue

When passing CFLAGS for example through to the jimtcl subconfigure the
quotes were not being preserved.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agobuild: update bootstrap comments
Spencer Oliver [Fri, 19 Nov 2010 09:08:09 +0000 (09:08 +0000)]
build: update bootstrap comments

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agobuild: prepend --with-jim-ext=nvp to jimtcl configure
Spencer Oliver [Fri, 19 Nov 2010 09:07:43 +0000 (09:07 +0000)]
build: prepend --with-jim-ext=nvp to jimtcl configure

This allows us to add options to jimtcl configure.
The default autoconf AC_CONFIG_SUBDIRS does not currently support this.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agobuild: add autobuild jimtcl to configure scripts
Spencer Oliver [Wed, 17 Nov 2010 10:26:21 +0000 (10:26 +0000)]
build: add autobuild jimtcl to configure scripts

Rather than having to configure/build jimtcl openocd
will do this as part of its own build.

To use an external jimtcl lib specify disable-internal-jimtcl
to the configure step.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agobuild: add common.mk
Spencer Oliver [Fri, 12 Nov 2010 11:41:55 +0000 (11:41 +0000)]
build: add common.mk

Rather than specifying common makefile variables move
them all to a common.mk.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agoFLASH/NOR: Remove useless file tms470.h
Antonio Borneo [Wed, 17 Nov 2010 13:28:34 +0000 (21:28 +0800)]
FLASH/NOR: Remove useless file tms470.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file str9xpec.h
Antonio Borneo [Wed, 17 Nov 2010 13:25:57 +0000 (21:25 +0800)]
FLASH/NOR: Remove useless file str9xpec.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file str9x.h
Antonio Borneo [Wed, 17 Nov 2010 13:23:04 +0000 (21:23 +0800)]
FLASH/NOR: Remove useless file str9x.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file str7x.h
Antonio Borneo [Wed, 17 Nov 2010 13:20:02 +0000 (21:20 +0800)]
FLASH/NOR: Remove useless file str7x.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file stm32x.h
Antonio Borneo [Wed, 17 Nov 2010 13:16:23 +0000 (21:16 +0800)]
FLASH/NOR: Remove useless file stm32x.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file stellaris.h
Antonio Borneo [Wed, 17 Nov 2010 13:12:37 +0000 (21:12 +0800)]
FLASH/NOR: Remove useless file stellaris.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file pic32mx.h
Antonio Borneo [Wed, 17 Nov 2010 13:04:37 +0000 (21:04 +0800)]
FLASH/NOR: Remove useless file pic32mx.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file lpc288x.h
Antonio Borneo [Wed, 17 Nov 2010 13:00:25 +0000 (21:00 +0800)]
FLASH/NOR: Remove useless file lpc288x.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file lpc2000.h
Antonio Borneo [Wed, 17 Nov 2010 12:55:40 +0000 (20:55 +0800)]
FLASH/NOR: Remove useless file lpc2000.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file avrf.h
Antonio Borneo [Wed, 17 Nov 2010 10:01:07 +0000 (18:01 +0800)]
FLASH/NOR: Remove useless file avrf.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoflash: fix bug with multiple back-to-back flash chips
Øyvind Harboe [Mon, 15 Nov 2010 13:43:16 +0000 (14:43 +0100)]
flash: fix bug with multiple back-to-back flash chips

flash programming via flash write_image or gdb load would
produce a bogus error message that the flash chip was to
small.

The solution is to limit the current flash programming
run to the current chip.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoFLASH/NOR: Remove useless file at91sam7.h
Antonio Borneo [Wed, 17 Nov 2010 02:53:30 +0000 (10:53 +0800)]
FLASH/NOR: Remove useless file at91sam7.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file at91sam3.h
Antonio Borneo [Wed, 17 Nov 2010 02:25:55 +0000 (10:25 +0800)]
FLASH/NOR: Remove useless file at91sam3.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoFLASH/NOR: Remove useless file spearsmi.h
Antonio Borneo [Wed, 17 Nov 2010 00:21:22 +0000 (08:21 +0800)]
FLASH/NOR: Remove useless file spearsmi.h

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoNOR/SPEARSMI: Add comments about SPI
Antonio Borneo [Wed, 17 Nov 2010 00:12:45 +0000 (08:12 +0800)]
NOR/SPEARSMI: Add comments about SPI

SMI interface hides the real SPI bus between SPEAr and
external flash.
Added comments to highlight the SPI operation, to help a
future rework in SPI generic and SPEAr specific drivers.

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoTCL/SPEAr: Added Serial flash in board file
Antonio Borneo [Thu, 11 Nov 2010 06:16:52 +0000 (14:16 +0800)]
TCL/SPEAr: Added Serial flash in board file

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoNOR/SPEAr: Add support for Serial NOR
Antonio Borneo [Thu, 11 Nov 2010 06:12:31 +0000 (14:12 +0800)]
NOR/SPEAr: Add support for Serial NOR

Add support and documentation for STMicroelectronics
SPEAr Serial Memory Interface (SMI).
Code tested on SPEAr3xx only.

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agohttpd: retire this server
Øyvind Harboe [Thu, 11 Nov 2010 07:15:49 +0000 (08:15 +0100)]
httpd: retire this server

this never panned out and there are enough mistakes in
the code that probably nobody used this.

Use the tcl server and implement a standalone http
app instead works fine.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoAdd comments and tiny improvements to STM32 flash loader algorithm
Freddie Chopin [Sat, 13 Nov 2010 14:42:00 +0000 (15:42 +0100)]
Add comments and tiny improvements to STM32 flash loader algorithm

Add comments to assembly flash loader for STM32. Add tiny improvement in
size of the algorithm (40 vs 48 bytes) and tiny speed improvement (~1.5%,
as time is wasted on waiting for end of operation anyway).

Signed-off-by: Freddie Chopin <freddie_chopin@op.pl>
10 years agogdb: fix occasional crash when flash probe failed
Øyvind Harboe [Sat, 13 Nov 2010 12:03:29 +0000 (13:03 +0100)]
gdb: fix occasional crash when flash probe failed

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agogdb: improve error message when gdb connect fails
Øyvind Harboe [Thu, 11 Nov 2010 07:50:22 +0000 (08:50 +0100)]
gdb: improve error message when gdb connect fails

gdb connect can fail when the flash has not been probed.

During gdb connect, the flash layout is reported, but this
can not be automatically detected for a target that is
powered up and OpenOCD supports connecting to gdb server
even if the target is powered down.

The solution is to turn of the gdb_memory_map feature.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agocortex_m3: report detected error condition in poll
Øyvind Harboe [Tue, 9 Nov 2010 08:14:21 +0000 (09:14 +0100)]
cortex_m3: report detected error condition in poll

If the CPU crashed at some point, poll will discover this.

Previously the poll fn would clear the error and print a warning,
rather than propagating the error.

The new behavior is to report the error back up, but still
clear the error.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agotarget: document that target_poll() will report and clear sticky errors
Øyvind Harboe [Tue, 9 Nov 2010 08:17:49 +0000 (09:17 +0100)]
target: document that target_poll() will report and clear sticky errors

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoTCL scripts: replace "puts" with "echo"
Antonio Borneo [Mon, 8 Nov 2010 09:23:49 +0000 (17:23 +0800)]
TCL scripts: replace "puts" with "echo"

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoJIM: document "echo" command
Antonio Borneo [Mon, 8 Nov 2010 07:13:28 +0000 (15:13 +0800)]
JIM: document "echo" command

Document "-n" option in manual;
Modify "echo" command definition as COMMAND_HANDLER to
easily add help message
Add help message aligned with manual.

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoJIM: Add "-n" option to "echo"
Antonio Borneo [Mon, 8 Nov 2010 04:08:56 +0000 (12:08 +0800)]
JIM: Add "-n" option to "echo"

With the new JIMTCL, "puts" only writes to stdout.
To write on telnet port too, "echo" must be used.
This patch gives to "echo" similar commandline option of "puts".

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agostm32: return early upon block write failure
Øyvind Harboe [Mon, 8 Nov 2010 15:53:24 +0000 (16:53 +0100)]
stm32: return early upon block write failure

only if we do not have enough ram do we continue.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agostm32: return error when failing to read
Øyvind Harboe [Mon, 8 Nov 2010 15:26:58 +0000 (16:26 +0100)]
stm32: return error when failing to read

add missing error handling.

Output warning when assuming maximum flash size in the
family when failing to read.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agostm32: add error propagation on writes
Øyvind Harboe [Mon, 8 Nov 2010 15:22:22 +0000 (16:22 +0100)]
stm32: add error propagation on writes

catch problems earlier.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agostm32: sharpen error handling for timeouts
Øyvind Harboe [Mon, 8 Nov 2010 15:02:07 +0000 (16:02 +0100)]
stm32: sharpen error handling for timeouts

delete lots of crud by handling this all in one spot.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agolpc3131: target definition
Andrew Leech [Tue, 9 Nov 2010 07:05:02 +0000 (08:05 +0100)]
lpc3131: target definition

10 years agoTCL scripts: add support for ST SPEAr310
Antonio Borneo [Thu, 4 Nov 2010 08:53:28 +0000 (16:53 +0800)]
TCL scripts: add support for ST SPEAr310

Initial support for ST SPEAr310 and for the evaluation
board EVALSPEAr310 Rev. 2.0.
Scripts are split in generic for SPEAr3xx family and
specific for SPEAr310. This should easily allow adding
new members of the family.

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoAdd EfikaMX smarttop board support
Marek Vasut [Fri, 29 Oct 2010 01:06:16 +0000 (03:06 +0200)]
Add EfikaMX smarttop board support

This patch finally adds support for i.MX51 based Genesi USA EfikaMX smarttop
board.

Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
10 years agoCortexA8: Introduce Freescale i.MX51 variant
Marek Vasut [Fri, 29 Oct 2010 00:57:32 +0000 (02:57 +0200)]
CortexA8: Introduce Freescale i.MX51 variant

This patch introduces support for Cortex A8 based Freescale i.MX51 CPU. This CPU
has the Debug Access Port located at a different address (0x60008000) than TI
OMAP3 series of CPUs.

i.MX51 configuration file based on OMAP3 configuration file and an email from
Alan Carvalho de Assis <acassis@gmail.com>.

Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
10 years agoCortexA8: Implement debug base autodetection
Marek Vasut [Fri, 29 Oct 2010 00:57:32 +0000 (02:57 +0200)]
CortexA8: Implement debug base autodetection

Implement autodetection of debug base. Also, implement a function solving
various hardware quirks (like iMX51 ROM Table location bug).

Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
10 years agoADIv5: Implement function to lookup CoreSight component
Marek Vasut [Sun, 31 Oct 2010 06:11:47 +0000 (07:11 +0100)]
ADIv5: Implement function to lookup CoreSight component

This patch implements "dap_lookup_cs_component()", which allows to lookup CS
component by it's identification.

Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
10 years agoADIv5: Introduce function to detect ROM Table location
Marek Vasut [Sun, 31 Oct 2010 04:24:36 +0000 (05:24 +0100)]
ADIv5: Introduce function to detect ROM Table location

This patch adds function called "dap_detect_debug_base()", which should be
called to get location of the ROM Table. By walking ROM Table, it's possible to
discover the location of DAP.

Sadly, some CPUs misreport this value, therefore I had to introduce an fixup
table, which will be used in case such CPU is detected.

Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
10 years agoarm964e: Add support for ARM946E target.
ddraskovic [Thu, 4 Nov 2010 13:33:10 +0000 (14:33 +0100)]
arm964e: Add support for ARM946E target.

So far most of the people have been using existing ARM966E in the
place of ARM946E, because they have practically the same scan chains.

However, ARM946E has caches, which further complicates JATG handling
via scan-chain. this was preventing single-stepping for ARM946E when
SW breakpoints are used.

This patch thus introduces :
1) Correct cache handling on memory write
2) Possibility to flush whole cache and turn it off during debug, or
just to flush affected lines (faster and better)
3) Correct SW breakpoint handling and correct single-stepping
4) Corrects the bug on CP15 read and write, so CP15 values
are now correctly R/W

10 years agobootstrap: add quick start build instructions
Øyvind Harboe [Fri, 29 Oct 2010 13:05:15 +0000 (15:05 +0200)]
bootstrap: add quick start build instructions

leave some bread-crumbs for users to figure out how
to build OpenOCD

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agojimtcl: 0.63 release
Øyvind Harboe [Fri, 15 Oct 2010 07:51:08 +0000 (09:51 +0200)]
jimtcl: 0.63 release

as a git module.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agojim tests: use installed
Øyvind Harboe [Sun, 10 Oct 2010 17:17:03 +0000 (19:17 +0200)]
jim tests: use installed

Delete obsolete jim that comes with OpenOCD.

10 years agosrc: add loader src description
Spencer Oliver [Thu, 28 Oct 2010 09:08:16 +0000 (10:08 +0100)]
src: add loader src description

 - add comment where to find the various loaders src files.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agocontrib: add ram loader src code
Spencer Oliver [Thu, 28 Oct 2010 08:19:37 +0000 (09:19 +0100)]
contrib: add ram loader src code

Add src code for ram loaders to contrib directory.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agoimx31pdk: use rclk w/1MHz fallback
Øyvind Harboe [Thu, 28 Oct 2010 06:34:41 +0000 (08:34 +0200)]
imx31pdk: use rclk w/1MHz fallback

measure_clk indicates ca. 3-4MHz, so 1MHz should be safe.

Added self_test proc used to test that rclk worked.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agohelp: help now works on configuration commands
Øyvind Harboe [Wed, 27 Oct 2010 06:37:54 +0000 (08:37 +0200)]
help: help now works on configuration commands

help would not show help for commands when the command
interpreter was in the wrong mode, which means that
e.g. "help newtap" didn't work, it wouldn't show the
"jtag newtap" help as it was a configuration command.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoMake systesetreq typos read sysresetreq instead
Peter Stuge [Sun, 24 Oct 2010 14:28:48 +0000 (16:28 +0200)]
Make systesetreq typos read sysresetreq instead

Signed-off-by: Peter Stuge <peter@stuge.se>
10 years agoRemove srst_pulls_trst from LPC1768 target
Peter Stuge [Sun, 24 Oct 2010 14:28:47 +0000 (16:28 +0200)]
Remove srst_pulls_trst from LPC1768 target

srst_pulls_trst may be true on some (broken) LPC1768 boards but is
not true in general for the LPC1768.

Signed-off-by: Peter Stuge <peter@stuge.se>
10 years agobuspirate: change handling of communication speed setting + create serial port open...
Michal Demin [Fri, 22 Oct 2010 21:26:43 +0000 (23:26 +0200)]
buspirate: change handling of communication speed setting + create serial port open function

Signed-off-by: Michal Demin <michaldemin@gmail.com>
10 years agobuspirate: Serial port was not correctly closed.
Michal Demin [Fri, 22 Oct 2010 19:06:31 +0000 (21:06 +0200)]
buspirate: Serial port was not correctly closed.

Signed-off-by: Michal Demin <michaldemin@gmail.com>
10 years agoTARGET: review handle_load_image_command()
Antonio Borneo [Tue, 12 Oct 2010 08:47:24 +0000 (16:47 +0800)]
TARGET: review handle_load_image_command()

Collect variable definitions.
Report syntax error to command dispatcher.
Propagate error when unable to open file.

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoTARGET: fix segfault in handle_dump_image_command()
Antonio Borneo [Tue, 12 Oct 2010 08:08:35 +0000 (16:08 +0800)]
TARGET: fix segfault in handle_dump_image_command()

The struct fileio is used after fileio_close().
Move fileio_close() after last usage.

Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
10 years agoswj-dp.tcl (SWD infrastructure #1)
David Brownell [Sun, 10 Oct 2010 21:41:11 +0000 (14:41 -0700)]
swj-dp.tcl (SWD infrastructure #1)

Provide new helper proc that can set up either an SWD or JTAG DAP
based on the transport which is in use -- mostly for SWJ-DP.

 Also update some SWJ-DP based chips/targets to use it.  The goal
is making SWD-vs-JTAG transparent in most places.  SWJ-DP based chips
really need this flexible configuration to cope with debug adapters
that support different transports, without needing new target configs
for each transport or adapter.

For JTAG-DP, callers will use "jtag newtap" directly, as today; only
one chip-level transport option exists.

For SW-DP (e.g. LPC1[13]xx or EFM32, they'll use "swd newdap" directly
(part of an upcoming SWD transport patch).  Again, only one transport
option exists, so hard-wiring is appropriate there.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
10 years agoadd JLINK protocol doc ref
David Brownell [Sun, 10 Oct 2010 21:27:35 +0000 (14:27 -0700)]
add JLINK protocol doc ref

Provide URL for a recent
version of JLINK protocol.
Signed-Off-By: David Brownell <dbrownell@users.sourceforge.net>
10 years agobuild: remove warn_unused_result errors
Spencer Oliver [Tue, 5 Oct 2010 15:29:37 +0000 (16:29 +0100)]
build: remove warn_unused_result errors

Remove any build errors for strtol when building release
version of openocd.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agogdbderver: fix gdb pipe startup overflow
Spencer Oliver [Tue, 5 Oct 2010 14:00:40 +0000 (15:00 +0100)]
gdbderver: fix gdb pipe startup overflow

When usng gdb pipes we need to keep openocd output at a minimum,
otherwise the gdb stdin will overflow and fail.
Make the calls to gdb_port and log_output synchronous to stop this.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agogdbserver: fix gdb_port memory leak
Spencer Oliver [Mon, 4 Oct 2010 19:35:33 +0000 (20:35 +0100)]
gdbserver: fix gdb_port memory leak

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agoserver: fix server pipe windows support
Spencer Oliver [Mon, 4 Oct 2010 19:34:47 +0000 (20:34 +0100)]
server: fix server pipe windows support

commit 50d5441e2a615fb2c44b41a777e4373901f7a2e6 caused
native windows build to fail.

Firstly this patch fixes the build issue, but it also disables support
for named pipes under Windows. Windows does not support posix named
pipes.
A cross-platfom access layer will need creating before support can be
enabled again.

Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
10 years agozy1000: add : port number syntax for tftp filing system
Øyvind Harboe [Sun, 5 Sep 2010 19:20:15 +0000 (21:20 +0200)]
zy1000: add : port number syntax for tftp filing system

Allows using non-standard port number. Default to port 69.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agopipes: add documentation for pipes
Øyvind Harboe [Mon, 27 Sep 2010 20:55:30 +0000 (22:55 +0200)]
pipes: add documentation for pipes

Stick with the name "gdb_port" even if this command
can be used for other things(disable, named pipes,
anonymous stdin/out pipe). "port" is correct for
probably more than 90% of use cases, if not more.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoserver: add support for pipes
Øyvind Harboe [Mon, 27 Sep 2010 06:50:49 +0000 (08:50 +0200)]
server: add support for pipes

-p/--pipe is now deprecated. Use '-c "gdb_port pipe;log_output openocd.log"'
instead. Warning logged.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoserver: specify port as a string
Øyvind Harboe [Mon, 27 Sep 2010 06:48:31 +0000 (08:48 +0200)]
server: specify port as a string

This will allow switching to using named pipes.

Split this out as a seperate commit to make changes
easier to follow.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoserver: read/write now goes through connection fn's
Øyvind Harboe [Mon, 27 Sep 2010 07:24:51 +0000 (09:24 +0200)]
server: read/write now goes through connection fn's

depending on whether the connection is over a socket
or pipe, the read is done differently.

pipes can return -1 when writing 0 bytes, make 0 byte
writes a successful no-op. 0 byte writes falls out
naturally of tcl server code.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoserver: split file descriptors in in/out fd's
Øyvind Harboe [Mon, 27 Sep 2010 06:26:31 +0000 (08:26 +0200)]
server: split file descriptors in in/out fd's

pipes have different fd's for in/out. This makes the
code more orthogonal and prepares for adding pipes.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
10 years agoserver: rely on ctrl-c to stop openocd
Øyvind Harboe [Sun, 26 Sep 2010 17:30:31 +0000 (19:30 +0200)]
server: rely on ctrl-c to stop openocd

there was special support to support pressing 'x' to quit
openocd. ctrl-c is sufficient. The main server loop is already
complicated enough.

Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>