openocd.git
11 years agospelling fix
oharboe [Mon, 28 Sep 2009 12:13:49 +0000 (12:13 +0000)]
spelling fix

git-svn-id: svn://svn.berlios.de/openocd/trunk@2766 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoShrink symbols exported from arm9tdmi.c and remove a forward ref.
dbrownell [Mon, 28 Sep 2009 04:55:23 +0000 (04:55 +0000)]
Shrink symbols exported from arm9tdmi.c and remove a forward ref.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2765 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoAdd list of JTAG adapter drivers with TAP_RESET statemove bug.
dbrownell [Sun, 27 Sep 2009 17:08:42 +0000 (17:08 +0000)]
Add list of JTAG adapter drivers with TAP_RESET statemove bug.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2764 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoUpdate FT2232 driver so that it reliably enters TAP_RESET.
dbrownell [Sun, 27 Sep 2009 16:55:52 +0000 (16:55 +0000)]
Update FT2232 driver so that it reliably enters TAP_RESET.

When the OpenOCD server starts up it records its state as TAP_RESET,
even though it could be anything.  Then when it starts to examine
the scan chain, it calls jtag_add_tlr() which sees it doesn't have
any work to do, and so it does nothing.  This can make the next
operations fail because they start from the wrong TAP state...

Instead of caring about the current recorded state, always enter
TAP_RESET by forcing five clocks with TMS high.

(NOTE:  it seems most other JTAG adapter drivers have this same bug.)

git-svn-id: svn://svn.berlios.de/openocd/trunk@2763 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDon't provide invalid OMAP5912 IR capture value/mask attributes
dbrownell [Sun, 27 Sep 2009 07:59:10 +0000 (07:59 +0000)]
Don't provide invalid OMAP5912 IR capture value/mask attributes

git-svn-id: svn://svn.berlios.de/openocd/trunk@2762 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoOn DM355 EVM board, associate NAND chips with $_TARGETNAME
dbrownell [Sat, 26 Sep 2009 22:08:50 +0000 (22:08 +0000)]
On DM355 EVM board, associate NAND chips with $_TARGETNAME
instead of a target number.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2761 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDiagnostics tweaks for jtag_examine_chain() failure paths.
dbrownell [Sat, 26 Sep 2009 22:01:24 +0000 (22:01 +0000)]
Diagnostics tweaks for jtag_examine_chain() failure paths.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2760 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoUpdates to the initial scanchain validation code:
dbrownell [Sat, 26 Sep 2009 19:18:42 +0000 (19:18 +0000)]
Updates to the initial scanchain validation code:
  - minor bug fixes
  - code cleanup
  - update comments
  - improve diagnostics
  - etc

git-svn-id: svn://svn.berlios.de/openocd/trunk@2759 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoStreamline Capture-IR handling and integrity test.
dbrownell [Sat, 26 Sep 2009 19:08:34 +0000 (19:08 +0000)]
Streamline Capture-IR handling and integrity test.

Change the handling of the "-ircapture" and "-irmask" parameters
to be slightly more sensible, given that the JTAG spec describes
what is required, and that we already require that conformance in
one place.  IR scan returns some bitstring with LSBs "01".

 - First, provide and use default values that satisfy the IEEE spec.
   Existing TAP configs will override the defaults, but those parms
   are no longer required.

 - Second, warn if any TAP gets set up to violate the JTAG spec.
   It's likely a bug, but maybe not; else this should be an error.
   Improve the related diagnostics to say which TAP is affected.

And associated minor fixes/cleanups to comments and diagnostics.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2758 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoUpdate DM355 target config to know about ICEpick.
dbrownell [Fri, 25 Sep 2009 17:02:59 +0000 (17:02 +0000)]
Update DM355 target config to know about ICEpick.
Still defaults to nonstandard EMU0/EMU1 settings.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2757 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMichael Hasselberg <mh@open-engineering.de> target configuration files for Toshiba...
oharboe [Fri, 25 Sep 2009 16:48:15 +0000 (16:48 +0000)]
Michael Hasselberg <mh@open-engineering.de> target configuration files for Toshiba TX09 familiy

git-svn-id: svn://svn.berlios.de/openocd/trunk@2756 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoTry/catch scheme. Typed up the functionality and regression tested.
oharboe [Fri, 25 Sep 2009 11:11:39 +0000 (11:11 +0000)]
Try/catch scheme. Typed up the functionality and regression tested.
Ready for discussion and tiny patches that tries out this scheme.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2755 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoWhen attaching GDB to OpenOCD, the target state is no longer affected.
oharboe [Thu, 24 Sep 2009 06:34:23 +0000 (06:34 +0000)]
When attaching GDB to OpenOCD, the target state is no longer affected.
Added gdb_sync feature that allows GDB to sync up to target state.
Issue "monitor gdb_sync" and the next stepi, will return immediately
with updated register values to GDB.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2754 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoStart handling the (second) SRST stage of reset better:
dbrownell [Wed, 23 Sep 2009 22:03:41 +0000 (22:03 +0000)]
Start handling the (second) SRST stage of reset better:
make sure that when there are two or more targets, their
various pre/post event reports are correctly ordered.

Previously, only the first target always saw its "pre"
method before SRST was asserted or deasserted.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2753 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoWhen setting up an ETM, cache its ETM_CONFIG register. Then
dbrownell [Wed, 23 Sep 2009 21:52:40 +0000 (21:52 +0000)]
When setting up an ETM, cache its ETM_CONFIG register.  Then
only expose the registers which are actually present.  They
could be missing for two basic reasons:

 - This version might not support them at all; e.g. ETMv1.1
   doesn't have some control/status registers.  (My sample of
   ARM9 boards shows all with ETMv1.3 support, FWIW.)

 - The configuration on this chip may not populate as many
   registers as possible; e.g. only two data value comparators
   instead of eight.

Includes a bugfix in the "etm info" command:  only one of the
two registers is missing on older silicon, so show the first
one before bailing.

Update ETM usage docs to explain that those registers need to be
written to configure what is traced, and that some ETM configs
are not yet handled.  Also, give some examples of the kinds of
constrained trace which could be arranged.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2752 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoStart cleaning up ETM register handling. On one ARM926 ETM+ETB
dbrownell [Wed, 23 Sep 2009 09:16:00 +0000 (09:16 +0000)]
Start cleaning up ETM register handling.  On one ARM926 ETM+ETB
system, removes 20 non-existent registers ... but still includes
over 45 (!) ETM registers which don't even exist there ...

 - Integrate the various tables to get one struct per register
 - Get rid of needless per-register dynamic allocation
 - Double check list of registers:
    * Remove sixteen (!) non-registers for data comparators
    * Remove four registers that imply newer ETM than we support
    * Change some names to match current architecture specs
 - Handle more register info
    * some are write-only
    * some are read-only
    * record which versions have them, just in case
 - Reorganize the registers to facilitate removing the extras
    * group e.g. comparator/counter #N registers together
    * add and use lookup-by-ID

git-svn-id: svn://svn.berlios.de/openocd/trunk@2751 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoInitial ETM cleanups. Most of these are cosmetic:
dbrownell [Wed, 23 Sep 2009 07:49:38 +0000 (07:49 +0000)]
Initial ETM cleanups.  Most of these are cosmetic:

 - Add a header comment
 - Line up the ETM context struct, pack it a bit
 - Remove unused context_id (this doesn't support ETMv2 yet)
 - Make most functions static
 - Remove unused string table and other needless lines of code
 - Correct "tracemode" helptext

Also provide and use an etm_reg_lookup() to find entries in the ETM
register cache.  This will help cope with corrected contents of that
cache, which doesn't include entires for non-existent registers.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2750 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoNico Coesel <ncoesel@dealogic.nl> fix warnings. . I'm wondering why these
oharboe [Wed, 23 Sep 2009 07:14:03 +0000 (07:14 +0000)]
Nico Coesel <ncoesel@dealogic.nl> fix warnings. . I'm wondering why these
didn't turn up earlier. Is everyone still using gcc 3.x? Or is the x86
version of gcc 4.x much more relaxed?

git-svn-id: svn://svn.berlios.de/openocd/trunk@2749 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years ago- fix build issue under win32 (cygwin/msys) from svn r2746
ntfreak [Tue, 22 Sep 2009 15:39:23 +0000 (15:39 +0000)]
- fix build issue under win32 (cygwin/msys) from svn r2746

git-svn-id: svn://svn.berlios.de/openocd/trunk@2748 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years ago- add missing svn props from previous commit
ntfreak [Tue, 22 Sep 2009 14:50:46 +0000 (14:50 +0000)]
- add missing svn props from previous commit

git-svn-id: svn://svn.berlios.de/openocd/trunk@2747 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMake it easier to erase or protect through to the end
dbrownell [Tue, 22 Sep 2009 05:39:06 +0000 (05:39 +0000)]
Make it easier to erase or protect through to the end
of a (NOR) flash chip: allow passing "last" as an alias
for the number of the last sector.

Improve several aspects of error checking while we're at it.

From: Johnny Halfmoon <jhalfmoon@milksnot.com>

git-svn-id: svn://svn.berlios.de/openocd/trunk@2746 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoUpdate presentation of TAP events and tap enable/disable.
dbrownell [Mon, 21 Sep 2009 21:35:56 +0000 (21:35 +0000)]
Update presentation of TAP events and tap enable/disable.

Highlight that the "post-reset" event kicks in before the
scan chain is validated, which limits what can be done
in a post-reset handler.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2745 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoRemove annoying end-of-line whitespace from doc/* files.
dbrownell [Mon, 21 Sep 2009 18:52:45 +0000 (18:52 +0000)]
Remove annoying end-of-line whitespace from doc/* files.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2744 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoRemove annoying end-of-line whitespace from tcl/* files
dbrownell [Mon, 21 Sep 2009 18:48:22 +0000 (18:48 +0000)]
Remove annoying end-of-line whitespace from tcl/* files

git-svn-id: svn://svn.berlios.de/openocd/trunk@2743 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoRemove annoying end-of-line whitespace from most src/*
dbrownell [Mon, 21 Sep 2009 18:40:55 +0000 (18:40 +0000)]
Remove annoying end-of-line whitespace from most src/*
files; omitted src/httpd

git-svn-id: svn://svn.berlios.de/openocd/trunk@2742 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoEthan Eade <ethan@evolution.com> board config script for Cogent CSB732 i.MX35 (arm1136)
oharboe [Mon, 21 Sep 2009 09:25:52 +0000 (09:25 +0000)]
Ethan Eade <ethan@evolution.com> board config script for Cogent CSB732 i.MX35 (arm1136)

git-svn-id: svn://svn.berlios.de/openocd/trunk@2741 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoEnsure that DaVinci chips can't start with a too-fast JTAG clock.
dbrownell [Mon, 21 Sep 2009 00:37:58 +0000 (00:37 +0000)]
Ensure that DaVinci chips can't start with a too-fast JTAG clock.
It can be sped up later, once it's known the PLLs are active.

Note that modern tools from TI all use adaptive clocking; and
that if that's done with OpenOCD, "too fast" is also a non-issue.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2740 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoUpdate the jtag-examine_chain() logic to verify that there's no
dbrownell [Mon, 21 Sep 2009 00:04:35 +0000 (00:04 +0000)]
Update the jtag-examine_chain() logic to verify that there's no
garbage after the expected data (from the TAPs' BYPASS or IDCODE
registers).

NOTE that there was previously some code that looked like it was
trying to do this ... which didn't work, because it was looping
over the list of expected TAPs, and never checked *after* that
list completed!  That could hide some *nasty* reset issues...

Also replace a now-obsolete scanchain length test with one that
behaves correctly; and update reporting of unexpected IDCODEs.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2739 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoEnhancement: stm32 flash protection error message
dbrownell [Sun, 20 Sep 2009 23:54:13 +0000 (23:54 +0000)]
Enhancement: stm32 flash protection error message

This patch modifies an error message which, in its original state,
I find somewhat unhelpful. So a small hint was added.

Signed-off-by: Johnny Halfmoon <jhalfmoon at milksnot.com>
git-svn-id: svn://svn.berlios.de/openocd/trunk@2738 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoUpdate the User's Guide to cover the scan chain verification step
dbrownell [Sun, 20 Sep 2009 21:17:08 +0000 (21:17 +0000)]
Update the User's Guide to cover the scan chain verification step
done on exit from the config stage, how JTAG clocking issues can
trigger errors there, and how to avoid such problems.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2737 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDebug message updates:
dbrownell [Sun, 20 Sep 2009 07:55:11 +0000 (07:55 +0000)]
Debug message updates:

 - Shrink messaging during resets, primarily by getting rid of
   "nothing happened" noise that hides *useful* information.

 - Improve: the "no IDCODE" message by identifying which tap only
   supports BYPASS; and the TAP event strings.

Related minor code updates:

 - Remove two needless tests when examining the chain:  we know
   we have a TAP, and that all TAPs have names.

 - Clean up two loops, turning "while"s into "for"s which better
   show what's actually being done.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2736 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMinor regression bugfix for the jtag_tap_handle_event() case
dbrownell [Sun, 20 Sep 2009 07:46:22 +0000 (07:46 +0000)]
Minor regression bugfix for the jtag_tap_handle_event() case
for disabling TAPs.  We don't actually know how to make any
JRCs which do that yet; but when we do, this will matter.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2735 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoAdded CPUDBG_WCR_BASE define
mlu [Sat, 19 Sep 2009 20:38:38 +0000 (20:38 +0000)]
Added  CPUDBG_WCR_BASE define

git-svn-id: svn://svn.berlios.de/openocd/trunk@2734 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoAvoid cache invalidation when writing to hardware debug registers
mlu [Sat, 19 Sep 2009 14:50:18 +0000 (14:50 +0000)]
Avoid cache invalidation when writing to hardware debug registers

git-svn-id: svn://svn.berlios.de/openocd/trunk@2733 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoReduced sleep time after reset
mlu [Sat, 19 Sep 2009 14:47:53 +0000 (14:47 +0000)]
Reduced sleep time after reset

git-svn-id: svn://svn.berlios.de/openocd/trunk@2732 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoStyle cleanup for the updated ft2232_add_pathmove().
dbrownell [Sat, 19 Sep 2009 06:54:01 +0000 (06:54 +0000)]
Style cleanup for the updated ft2232_add_pathmove().
Indentation, whitespace, line lengths.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2731 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMinor behavior fixes for the two JTAG reset events (C/internal,
dbrownell [Sat, 19 Sep 2009 06:49:40 +0000 (06:49 +0000)]
Minor behavior fixes for the two JTAG reset events (C/internal,
and Tcl/external):

 - Reorder so *both* paths (TCK/TMS or TRST) can enable TAPs with
   ICEpick ... first C code flags TAPs that got disabled, then call
   any Tcl code that might want to re-enable them.

 - Always call the C/internal handlers when JTAG operations can be
   issued; previously that wasn't done when TRST was used.

Plus some small cleanups (whitespace, strings, better messaging
during debug and on some errors) to reset-related code.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2730 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMinor tweaks to description of JTAG adapter/dongle issues.
dbrownell [Sat, 19 Sep 2009 06:42:50 +0000 (06:42 +0000)]
Minor tweaks to description of JTAG adapter/dongle issues.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2729 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMove Cortex A8 debug access initialisation from omap3530.cfg to cortex_a8.c
mlu [Fri, 18 Sep 2009 15:11:46 +0000 (15:11 +0000)]
Move Cortex A8 debug access initialisation from omap3530.cfg to cortex_a8.c

git-svn-id: svn://svn.berlios.de/openocd/trunk@2728 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoremove unused externs
oharboe [Fri, 18 Sep 2009 07:09:44 +0000 (07:09 +0000)]
remove unused externs

git-svn-id: svn://svn.berlios.de/openocd/trunk@2727 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoTweak TCL reset script ... mostly improving descriptions of
dbrownell [Fri, 18 Sep 2009 00:11:51 +0000 (00:11 +0000)]
Tweak TCL reset script ... mostly improving descriptions of
the various steps, but also calling [target names] only once.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2726 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoAddress codereview comment from Steve Grubb <sgrubb@redhat.com>:
dbrownell [Thu, 17 Sep 2009 19:20:18 +0000 (19:20 +0000)]
Address codereview comment from Steve Grubb <sgrubb@redhat.com>:
avoid a duplicate test.

Plus other cleanup in the same code: be "static", sane line lengths
for source and diagnostics, and fix misleading variable names.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2725 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMinor fixes to NAND code and docs
dbrownell [Thu, 17 Sep 2009 18:56:17 +0000 (18:56 +0000)]
Minor fixes to NAND code and docs

Erase logic:
 - command invocation
    + treat "nand erase N" (no offset/length) as "erase whole chip N"
    + catch a few more bogus parameter cases, like length == 0 (sigh)
 - nand_erase() should be static
 - on error
    + say which block failed, and if it was a bad block
    + don't give up after the first error; try to erase the rest
 - on success, say which nand device was erased (name isn't unique)

Device list ("nand list"):
 - say how many blocks there are
 - split summary into two lines
 - give example in the docs

Doc tweaks:
 - Use @option{...} for DaVinci's supported hardware ECC options

For the record, I've observed that _sometimes_ erasing bad blocks causes
failure reports, and that manufacturer bad block markers aren't always
erasable (even when erasing their blocks doesn't trigger an error report).

git-svn-id: svn://svn.berlios.de/openocd/trunk@2724 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoFix coredump seen in some code paths.
dbrownell [Thu, 17 Sep 2009 18:52:32 +0000 (18:52 +0000)]
Fix coredump seen in some code paths.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2723 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agomichal smulski <michal.smulski@ooma.com> fix regression in jtag_add_pathmove() which...
oharboe [Thu, 17 Sep 2009 12:25:53 +0000 (12:25 +0000)]
michal smulski <michal.smulski@ooma.com> fix regression in jtag_add_pathmove() which broke arm11 in r1825. Other uses of jtag_add_pathmove are svn + xsvf + xscale...

git-svn-id: svn://svn.berlios.de/openocd/trunk@2722 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoprefix zy1000_reboot command to avoid name conflicts
oharboe [Thu, 17 Sep 2009 11:44:14 +0000 (11:44 +0000)]
prefix zy1000_reboot command to avoid name conflicts

git-svn-id: svn://svn.berlios.de/openocd/trunk@2721 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agosrst_gates_jtag option. at91sam9260 needs retesting, and possibly srst_gates_jtag...
oharboe [Thu, 17 Sep 2009 11:23:41 +0000 (11:23 +0000)]
srst_gates_jtag option. at91sam9260 needs retesting, and possibly srst_gates_jtag added to reset_config. Could i.MX27 be a case where srst does not pull trst, but really srst gates jtag clock?

git-svn-id: svn://svn.berlios.de/openocd/trunk@2720 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoThe "arm9tdmi.c" file is more of a generic ARM9 support file:
dbrownell [Thu, 17 Sep 2009 08:02:43 +0000 (08:02 +0000)]
The "arm9tdmi.c" file is more of a generic ARM9 support file:

 - update comments to say so.
 - update docs to clarify that the "arm9tdmi" command prefix
   is a misnomer.
 - bugfix some messages that wrongly assume only ARM9TDMI
   based processors use this code.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2719 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDoc update: mention how ARM's WFI instruction affects
dbrownell [Thu, 17 Sep 2009 07:56:24 +0000 (07:56 +0000)]
Doc update:  mention how ARM's WFI instruction affects
JTAG clocking by gating the core clock, and workarounds.
Most details are with the "halt" command, which is one
of the first places this issue will be noticed.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2718 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoRemove unused varables (moved to armv7a)
mlu [Wed, 16 Sep 2009 18:18:38 +0000 (18:18 +0000)]
Remove unused varables (moved to armv7a)

git-svn-id: svn://svn.berlios.de/openocd/trunk@2717 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoUse a variable armv7a->debug_base instead of hardedcoded OMAP3530_DEBUG_BASE
mlu [Wed, 16 Sep 2009 18:17:26 +0000 (18:17 +0000)]
Use a variable armv7a->debug_base instead of hardedcoded  OMAP3530_DEBUG_BASE

git-svn-id: svn://svn.berlios.de/openocd/trunk@2716 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoRolf Meeser <rolfm_9dq@yahoo.de> adds flash support for NXP's LPC2900 family (ARM968E).
oharboe [Wed, 16 Sep 2009 12:38:26 +0000 (12:38 +0000)]
Rolf Meeser <rolfm_9dq@yahoo.de> adds flash support for NXP's LPC2900 family (ARM968E).

git-svn-id: svn://svn.berlios.de/openocd/trunk@2715 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDefine debug_base, debug_ap, memory_ap in armv7a_common_t
mlu [Tue, 15 Sep 2009 16:20:39 +0000 (16:20 +0000)]
Define debug_base, debug_ap, memory_ap in armv7a_common_t

git-svn-id: svn://svn.berlios.de/openocd/trunk@2714 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoUpdated mode string list.
mlu [Tue, 15 Sep 2009 15:50:09 +0000 (15:50 +0000)]
Updated mode string list.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2713 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDefiny symbolic values for VA to PA address translation operations
mlu [Tue, 15 Sep 2009 15:41:14 +0000 (15:41 +0000)]
Definy symbolic values for VA to PA address translation operations

git-svn-id: svn://svn.berlios.de/openocd/trunk@2712 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agofix bug when using 32 instead of 64 bit value in callback, caught by -Wall
oharboe [Tue, 15 Sep 2009 14:20:37 +0000 (14:20 +0000)]
fix bug when using 32 instead of 64 bit value in callback, caught by -Wall

git-svn-id: svn://svn.berlios.de/openocd/trunk@2711 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoadded embedded ice programming while srst is asserted todo item
oharboe [Tue, 15 Sep 2009 09:41:09 +0000 (09:41 +0000)]
added embedded ice programming while srst is asserted todo item

git-svn-id: svn://svn.berlios.de/openocd/trunk@2710 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoCheck return values to avoid infinite wait in loop on error.
mlu [Mon, 14 Sep 2009 22:41:47 +0000 (22:41 +0000)]
Check return values to avoid infinite wait in loop on error.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2709 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoCache invalidation when writing to memory
mlu [Mon, 14 Sep 2009 22:36:27 +0000 (22:36 +0000)]
Cache invalidation when writing to memory

git-svn-id: svn://svn.berlios.de/openocd/trunk@2708 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoerror message upon recursive invocation of reset from reset event handlers
oharboe [Mon, 14 Sep 2009 13:54:49 +0000 (13:54 +0000)]
error message upon recursive invocation of reset from reset event handlers

git-svn-id: svn://svn.berlios.de/openocd/trunk@2707 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agofix email address
oharboe [Mon, 14 Sep 2009 08:03:57 +0000 (08:03 +0000)]
fix email address

git-svn-id: svn://svn.berlios.de/openocd/trunk@2706 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoRolf Meeser <rolfm_9dq@yahoo.de> warning fix in previous commit was wrong. target_cod...
oharboe [Mon, 14 Sep 2009 07:48:28 +0000 (07:48 +0000)]
Rolf Meeser <rolfm_9dq@yahoo.de> warning fix in previous commit was wrong. target_code_size needs the real value later.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2705 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agofix warning
oharboe [Mon, 14 Sep 2009 06:06:35 +0000 (06:06 +0000)]
fix warning

git-svn-id: svn://svn.berlios.de/openocd/trunk@2704 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMagnus Lundin <lundin@mlu.mine.nu> Disable poll while core register initialization
oharboe [Sun, 13 Sep 2009 20:07:13 +0000 (20:07 +0000)]
Magnus Lundin <lundin@mlu.mine.nu> Disable poll while core register initialization

git-svn-id: svn://svn.berlios.de/openocd/trunk@2703 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMore CortexA8 debug register definitions.
mlu [Sun, 13 Sep 2009 17:26:07 +0000 (17:26 +0000)]
More CortexA8 debug register definitions.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2702 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoFix argument passing in cortex_a8_write_cp.
mlu [Sun, 13 Sep 2009 13:57:50 +0000 (13:57 +0000)]
Fix argument passing in cortex_a8_write_cp.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2701 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDirk Behme <dirk.behme@googlemail.com> document post TAP reset event
oharboe [Sat, 12 Sep 2009 13:05:20 +0000 (13:05 +0000)]
Dirk Behme <dirk.behme@googlemail.com> document post TAP reset event

git-svn-id: svn://svn.berlios.de/openocd/trunk@2700 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDavid Brownell <david-b@pacbell.net>
oharboe [Sat, 12 Sep 2009 08:11:45 +0000 (08:11 +0000)]
David Brownell <david-b@pacbell.net>
Update the board config for the DaVinci DM355 EVM so the reset-init
event handler does the rest of the work it should do:

 - minor PLL setup bugfixes
 - initialize the DDR2 controller
 - probe both NAND banks
 - initialize UART0
 - enable the icache

git-svn-id: svn://svn.berlios.de/openocd/trunk@2699 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDavid Brownell <david-b@pacbell.net>
oharboe [Sat, 12 Sep 2009 08:10:19 +0000 (08:10 +0000)]
David Brownell <david-b@pacbell.net>
Cleanup some the downloaded ARM target algorithm code:

 - Provide more complete disassembly of the DCC bulk write code

 - Make code blocks "static const", in case GCC doesn't

 - Fix some tabbing/layout issues

 - Make some arm7_9_common.h flags be "bool" not "int"; and compact
   the layout a bit (group most bools together)

git-svn-id: svn://svn.berlios.de/openocd/trunk@2698 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDavid Brownell <david-b@pacbell.net> some early todo items on run_algorithm
oharboe [Fri, 11 Sep 2009 21:14:31 +0000 (21:14 +0000)]
David Brownell <david-b@pacbell.net> some early todo items on run_algorithm

git-svn-id: svn://svn.berlios.de/openocd/trunk@2697 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agotap post reset event added. Allows omap3530 to send 100 runtest idle tickle's after...
oharboe [Fri, 11 Sep 2009 18:34:15 +0000 (18:34 +0000)]
tap post reset event added. Allows omap3530 to send 100 runtest idle tickle's after a TAP_RESET.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2696 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years ago- revert change made to sheevaplug.cfg in rev2573
ntfreak [Fri, 11 Sep 2009 14:08:28 +0000 (14:08 +0000)]
- revert change made to sheevaplug.cfg in rev2573

git-svn-id: svn://svn.berlios.de/openocd/trunk@2695 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoNicolas Pitre <nico@cam.org> put feroceon target definition at the end so to avoid a
oharboe [Fri, 11 Sep 2009 08:04:50 +0000 (08:04 +0000)]
Nicolas Pitre <nico@cam.org> put feroceon target definition at the end so to avoid a
bunch of useless forward declarations.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2694 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoNicolas Pitre <nico@cam.org> Dragonite support
oharboe [Fri, 11 Sep 2009 08:03:46 +0000 (08:03 +0000)]
Nicolas Pitre <nico@cam.org> Dragonite support

git-svn-id: svn://svn.berlios.de/openocd/trunk@2693 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agospelling mistake
oharboe [Fri, 11 Sep 2009 07:46:50 +0000 (07:46 +0000)]
spelling mistake

git-svn-id: svn://svn.berlios.de/openocd/trunk@2692 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agodo not use dynamically sized stack arrays, not compatible with embedded OS's
oharboe [Fri, 11 Sep 2009 07:43:36 +0000 (07:43 +0000)]
do not use dynamically sized stack arrays, not compatible with embedded OS's

git-svn-id: svn://svn.berlios.de/openocd/trunk@2691 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoregistering a target event twice caused infinite loop. Same bug as in jtag/core.c...
oharboe [Fri, 11 Sep 2009 06:58:49 +0000 (06:58 +0000)]
registering a target event twice caused infinite loop. Same bug as in jtag/core.c copy & pasted.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2690 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agosyntax error fix
oharboe [Fri, 11 Sep 2009 06:08:51 +0000 (06:08 +0000)]
syntax error fix

git-svn-id: svn://svn.berlios.de/openocd/trunk@2689 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoAlexei Babich <a.babich@rez.ru> cleanup
oharboe [Fri, 11 Sep 2009 06:01:28 +0000 (06:01 +0000)]
Alexei Babich <a.babich@rez.ru>  cleanup

git-svn-id: svn://svn.berlios.de/openocd/trunk@2688 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoNicolas Pitre <nico@cam.org> tighten error checking in bulk_write
oharboe [Fri, 11 Sep 2009 05:57:51 +0000 (05:57 +0000)]
Nicolas Pitre <nico@cam.org> tighten error checking in bulk_write

git-svn-id: svn://svn.berlios.de/openocd/trunk@2687 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoeol-style:native
oharboe [Thu, 10 Sep 2009 13:35:08 +0000 (13:35 +0000)]
eol-style:native

git-svn-id: svn://svn.berlios.de/openocd/trunk@2686 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoAlexei Babich <a.babich@rez.ru> imx31 nand flash controller support
oharboe [Thu, 10 Sep 2009 13:17:25 +0000 (13:17 +0000)]
Alexei Babich <a.babich@rez.ru>  imx31 nand flash controller support

git-svn-id: svn://svn.berlios.de/openocd/trunk@2685 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoAlexei Babich <a.babich@rez.ru> fix problems with unecessary tailend byte accesses...
oharboe [Thu, 10 Sep 2009 13:17:05 +0000 (13:17 +0000)]
Alexei Babich <a.babich@rez.ru> fix problems with unecessary tailend byte accesses. Use 16 bit access on tailend of a memory read if possible.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2684 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agomichal smulski <michal.smulski@ooma.com> telo target/board scripts
oharboe [Thu, 10 Sep 2009 08:06:22 +0000 (08:06 +0000)]
michal smulski <michal.smulski@ooma.com> telo target/board scripts

git-svn-id: svn://svn.berlios.de/openocd/trunk@2683 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoRolf Meeser <rolfm_9dq@yahoo.de>
oharboe [Wed, 9 Sep 2009 16:11:33 +0000 (16:11 +0000)]
Rolf Meeser <rolfm_9dq@yahoo.de>
This patch adds target algorithm support for those flash devices that do not support DQ5 polling. So far they could only be programmed with host algorithm, but this was way too slow.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2682 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years ago- Fix bug-in-waiting when adding more than one TAP event type
oharboe [Wed, 9 Sep 2009 07:09:14 +0000 (07:09 +0000)]
- Fix bug-in-waiting when adding more than one TAP event type
- Infinite loop bugfix when running tap configure a second time

git-svn-id: svn://svn.berlios.de/openocd/trunk@2681 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDavid Brownell <david-b@pacbell.net>
oharboe [Wed, 9 Sep 2009 06:28:49 +0000 (06:28 +0000)]
David Brownell <david-b@pacbell.net>
Optionally shave time off the armv4_5 run_algorithm() code:  let
them terminate using software breakpoints, avoiding roundtrips
to manage hardware ones.

Enable this by using BKPT to terminate execution instead of "branch
to here" loops.  Then pass zero as the exit address, except when
running on an ARMv4 core.  ARM7TDMI, ARM9TDMI, and derived cores
now set a flag saying they're ARMv4.

Use that mechanism in arm_nandwrite(), for about 3% speedup on a
DaVinci ARM926 core; not huge, but it helps.  Some other algorithms
could use this too (mostly flavors of flash operation).

git-svn-id: svn://svn.berlios.de/openocd/trunk@2680 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDavid Brownell <david-b@pacbell.net>
oharboe [Wed, 9 Sep 2009 06:27:47 +0000 (06:27 +0000)]
David Brownell <david-b@pacbell.net>
Fix docs on ARM11 MCR and MRC coprocessor commands:
correct read-vs-write; and describe the params.

(ARM920 and ARM926 have cp15-specific commands; this
approach is more generic.  MCR2, MRC2, MCRR, MCRR2,
MRRC, and MRRC2 instructions could also get exposed.)

git-svn-id: svn://svn.berlios.de/openocd/trunk@2679 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoReport correct core instruction state for ARMv/A targets
mlu [Tue, 8 Sep 2009 15:32:18 +0000 (15:32 +0000)]
Report correct core instruction state for ARMv/A targets

git-svn-id: svn://svn.berlios.de/openocd/trunk@2678 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoLoad PC with bit 0 set to 1 when resuming to say in Thumb instruction state.
mlu [Tue, 8 Sep 2009 15:31:24 +0000 (15:31 +0000)]
Load PC with bit 0 set to 1 when resuming  to say in Thumb instruction state.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2677 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDavid Brownell <david-b@pacbell.net>
oharboe [Tue, 8 Sep 2009 06:18:45 +0000 (06:18 +0000)]
David Brownell <david-b@pacbell.net>
Provide an "armv7a disassemble" command.  Current omissions include
VFP (except as coprocessor instructions), Neon, and various Thumb2
opcodes that are not available in ARMv7-M processors.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2676 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDavid Brownell <david-b@pacbell.net>
oharboe [Tue, 8 Sep 2009 06:17:33 +0000 (06:17 +0000)]
David Brownell <david-b@pacbell.net>
lean up some loose ends with the ARM disassembler

 - Add a header comment describing its current state and uses
   and referencing the now-generally-available V7 arch spec

 - Support some mode switch instructions:
    * Thumb to Jazelle (BXJ)
    * Thumb to ThumbEE (ENTERX)
    * ThumbEE to Thumb (LEAVEX)

 - Improve that recent warning fix (and associated whitespace goof)

 - Declare the rest of the internal code and data "static".  A
   compiler may use this, and it helps clarify the scope of these
   routines (e.g. what changes to them could affect).

git-svn-id: svn://svn.berlios.de/openocd/trunk@2675 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoImproved handling of instruction set state, helps for debugging Thumb state.
mlu [Mon, 7 Sep 2009 20:19:17 +0000 (20:19 +0000)]
Improved handling of instruction set state, helps for debugging Thumb state.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2674 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMahr, Stefan <Stefan.Mahr@sphairon.com> removes the endianness swapping in mips_m4k...
oharboe [Fri, 4 Sep 2009 19:35:10 +0000 (19:35 +0000)]
Mahr, Stefan <Stefan.Mahr@sphairon.com> removes the endianness swapping in mips_m4k.c Swapping is already done in target.c

git-svn-id: svn://svn.berlios.de/openocd/trunk@2673 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agouse "armv4_5 core_state arm" instead of soft_reset_halt, fewer side effects
oharboe [Fri, 4 Sep 2009 11:03:26 +0000 (11:03 +0000)]
use "armv4_5 core_state arm" instead of soft_reset_halt, fewer side effects

git-svn-id: svn://svn.berlios.de/openocd/trunk@2672 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDirk Behme <dirk.behme@googlemail.com> retire jtag_speed usage
oharboe [Fri, 4 Sep 2009 08:27:27 +0000 (08:27 +0000)]
Dirk Behme <dirk.behme@googlemail.com> retire jtag_speed usage

git-svn-id: svn://svn.berlios.de/openocd/trunk@2671 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoDirk Behme <dirk.behme@googlemail.com> Add default fall back freqency.
oharboe [Fri, 4 Sep 2009 08:27:08 +0000 (08:27 +0000)]
Dirk Behme <dirk.behme@googlemail.com> Add default fall back freqency.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2670 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoset ARM mode using explicit command rather than soft_reset_halt which has lots of...
oharboe [Fri, 4 Sep 2009 08:23:24 +0000 (08:23 +0000)]
set ARM mode using explicit command rather than soft_reset_halt which has lots of side effects.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2669 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agoMatt Hsu <matt@0xlab.org> This patch simply enables the halting debug mode.
oharboe [Fri, 4 Sep 2009 08:22:02 +0000 (08:22 +0000)]
Matt Hsu <matt@0xlab.org> This patch simply enables the halting debug mode.
By enabling this bit, the processor halts when a debug event
such as breakpoint occurs.

git-svn-id: svn://svn.berlios.de/openocd/trunk@2668 b42882b7-edfa-0310-969c-e2dbd0fdcd60

11 years agomore debug output for breakpoints
oharboe [Fri, 4 Sep 2009 08:21:18 +0000 (08:21 +0000)]
more debug output for breakpoints

git-svn-id: svn://svn.berlios.de/openocd/trunk@2667 b42882b7-edfa-0310-969c-e2dbd0fdcd60