Make the Thumb2 disassembler handle more 32-bit instructions:
A5.3.1 Data processing (modified immediate)
My small sample shows GCC likes to use many of these instructions.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2533 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Make the Thumb2 disassembler handle a bunch of 32-bit instructions:
A5.3.4 Branches and miscellaneous control
Note that this shifts some responsabililty out of helper functions,
making the code and layout simpler for 32-bit decoders: they only
need to know how to format the instruction and its parameters.
Also, technical note: with this patch, Thumb1 decoders could now
call the Thumb2 decoder if they wanted to get nicer treatment of
the exiting 32-bit B/BLX instructions.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2532 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Change layout of Thumb disassembly to work better with Thumb2:
- Move opcode to the left, allowing space for four hex bytes:
* after address, two spaces not one tab (taking 6 spaces)
* after 2-byte opcode, four spaces before tab
- Also, after opcode mnemonic use a tab not a space, to make
operands line up
Sample output (after some patches decoding a few 32-bit instructions):
0x00003e5a 0xf4423200 ORR r2, r2, #131072 ; 0x20000
0x00003e5e 0x601a STR r2, [r3, #0x0]
0x00003e60 0x2800 CMP r0, #0x00
0x00003e62 0xd1f3 BNE 0x00003e4c
0x00003e64 0xf008fa38 BL 0x0000c2d8
The affected lines of code now wrap at sane margins too.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2531 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Initial support for disassembling Thumb2 code. This works only for
Cortex-M3 cores so far. Eventually other cores will also need Thumb2
support ... but they don't yet support any kind of disassembly.
- Update the 16-bit Thumb decoder:
* Understand CPS, REV*, SETEND, {U,S}XT{B,H} opcodes added
by ARMv6. (It already seems to treat CPY as MOV.)
* Understand CB, CBNZ, WFI, IT, and other opcodes added by
in Thumb2.
- A new Thumb2 instruction decode routine is provided.
* This has a different signature: pass the target, not the
instruction, so it can fetch a second halfword when needed.
The instruction size is likewise returned to the caller.
* 32-bit instructions are recognized but not yet decoded.
- Start using the current "UAL" syntax in some cases. "SWI" is
renamed as "SVC"; "LDMIA" as "LDM"; "STMIA" as "STM".
- Define a new "cortex_m3 disassemble addr count" command to give
access to this disassembly.
Sanity checked against "objdump -d" output; a bunch of the new
instructions checked out fine.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2530 b42882b7-edfa-0310-969c-e2dbd0fdcd60
This patch correctly identifies a running target.
Patch made a tad more palatable by David Brownell <david-b@pacbell.net>
git-svn-id: svn://svn.berlios.de/openocd/trunk@2510 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Updates to private TAP state tables in amtjtagaccel interface driver.
The first change is the neccesary one to correct a long-standing bug that
caused the IDCODE to be shifted by one bit too many. This was caused by
an incorrect path from state RESET to state DRSHIFT.
The value of those 2 bytes were 0x8a and 0x04. This means that the
bitstream to do this transition is 0b 00100 01010 (send LSB first). This
will bring you from the reset state to the shift state; however, you
enter the shift-state twice, which explains why the ID-CODE that will be
read next will be shifted 1 bit. The fix changes these to 0x05 and 0x00.
This will send the bitstream 0b 00101 (send LSB first). This will bring
the TAP controller from the RESET state to the DRSHIFT state directly,
without entering the DRSHIFT state twice.
After checking the whole table, two other transitions were found that
could be optimized (5 bits in stead of 10 bits).
Summary off all changes:
From To Old values Old Bitstream New values New Bitstream Remarks
---- ------- ---------- ------------- ---------- ------------- -------
RESET DRSHIFT 0x8a 0x04 0b00100 01010 0x05 0x00 0b00101 1,2
IDLE DRSHIFT 0x85 0x08 0b01000 00101 0x04 0x00 0b00100 2
IDLE IRSHIFT 0x8b 0x08 0b01000 01011 0x06 0x00 0b00110 2
[1] Fixes the IDCODE bug
[2] Optimization
git-svn-id: svn://svn.berlios.de/openocd/trunk@2472 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Fix intermittent J-Link interface startup failures:
- Use usb_reset to ensure selected dongle is in known good state.
- Assert emulator reset durning status check to prevent supurious failures.
- Eliminate status check loop; not needed due to other fixes.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2471 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Restore some whitespace that got clobbered by over-aggressive
whitepace eradication patches a while back.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2446 b42882b7-edfa-0310-969c-e2dbd0fdcd60
This patch adds support for the Luminary Micro LM3S9B90 target and
LM3S9B92 Evaluation Kit. These kits include a new ft2232 adapter, the
Luminary In-Circuit Debug Interface (ICDI) Board, so this is added as a
new ft2232 layout called "luminary_icdi".
git-svn-id: svn://svn.berlios.de/openocd/trunk@2429 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Add "jtag names" command, mirroring "target names" but returning
TAP names instead of target names. This starts letting TAPs be
manipulated in scripts ... much like what works now for targets.
It's a bit limited just yet, since "jtag cget $TAPNAME" doesn't
expose all TAP attributes. "$TARGETNAME cget" is more functional.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2428 b42882b7-edfa-0310-969c-e2dbd0fdcd60
- Move repository URL output associate it with the version; they relate.
- 'openocd --version' output now appears much more terse, as expected.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2421 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Remove some bogus warnings during server startup for ARM926ejs
targets that were already halted for debug ... e.g. started up
a freshly built instance.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2417 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Warn when people (or scripts) use numeric identifiers for TAPs,
instead of dotted.name values. We want this usage to go away,
so that for example adding more TAPs doesn't cause config scripts
to break because some sequence number changed.
It's been deprecated since late 2008, but putting a warning on
this should help us remove it (say, in June 2010) by helping to
phase out old (ab)usage in config scripts.
Other than in various config files, the only code expecting such
a number was the almost unused str9xpec driver. This code was
changed to use the TAP it was passed, instead of making its own
dubious lookup and ignoring that TAP.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2415 b42882b7-edfa-0310-969c-e2dbd0fdcd60
- Adds new source files to encapsulate static/dynamic module handling.
- Further work should implement the jtag_interface_modules_load routine,
to populate the jtag_interfaces list from shared libraries in a path.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2413 b42882b7-edfa-0310-969c-e2dbd0fdcd60
- issue is gdb stdin buffer gets full before we redirect openocd output
git-svn-id: svn://svn.berlios.de/openocd/trunk@2350 b42882b7-edfa-0310-969c-e2dbd0fdcd60
This is minimal patch to support FA526 ARMv4 compatible core.
Since it is very similar to ARM920T I tried to reuse as much
code as possible.
CPU and board configs will follow soon.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2292 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Clean up the PLD files:
- Get rid of some extraneous whitespace
- Make various functions static
- Wrap overlong lines
git-svn-id: svn://svn.berlios.de/openocd/trunk@2272 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Let jtag_call_event_callbacks() behave when the callback removes itself.
Oddly, this crashed on x86_32 but not x86_64.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2264 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Fix parsing bugs for "$target_name mww addr data [count]" ... it was
always requiring the count, instead of just defaulting it to one.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2263 b42882b7-edfa-0310-969c-e2dbd0fdcd60
- Use macro to eliminate duplicate body definitions.
- Rename okay as is_okay; add parenthesis to help "clarify" logic.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2253 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Fix a bug preventing ICEpick "enable that TAP" code from working:
the "runtest" command wrongly finished with a JTAG reset, discarding
the work the TAP enable handler just finished! Instead, JTAG should
stay in RUN/IDLE state.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2252 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Extend the internal JTAG event handlers to cover enable/disable,
and use those events to make sure that targets get "examined" if
they were disabled when the scan chain was first set up:
- Remove "enum jtag_tap_event", merge with "enum jtag_event",
so C code can now listen for TAP enable/disable events.
- Report those events so they can trigger callbacks.
- During startup, make target_examine() register a handler to
catch ENABLE events for any then-disabled targets.
This fixes bugs like "can't halt target after enabling its TAP".
One class of unresolved bugs: if the target has an ETM hooked
up to an ETB, nothing activates the ETB. But starting up the
ETM without access to the ETB registers fails...
git-svn-id: svn://svn.berlios.de/openocd/trunk@2251 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Tighten error handling on TAP enable/disable paths a bit:
- Don't enable/disable unless it's necessary. Those event
handlers could have nasty side effects...
- Don't *succeed* enables/disables if there was no code which
could have implemented that action. This prevents bugs like
wrongly acting as if the scan chain changed.
- Minor whitespace cleanup in enable/disable command code.
The big problem is still the lack of code to verify scan chains
were actually updated as requested; this adds a comment on that.
I suspect the best we can do near term will be to verify IDCODE.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2250 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Fix bug in a warning. It warned about "huge IRlength" for an
older JRC with a two bit instruction register ... wrong!
git-svn-id: svn://svn.berlios.de/openocd/trunk@2249 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Fix some polling issues:
- Don't background-poll disabled TAPs ... this was just a bug
waiting to happen. (And then it happened!)
- Don't fail command line polls of disabled taps; that's not any
kind of error, it's just that you can't do much. But do show
that tap-disabled status.
- Spell "continuous" correctly in the variable name. ;)
Not resolved by this patch: the need for an interlock whereby
other code (like the JTAG layer) can block all other access to
the JTAG layer, e.g. while enabling or disabling TAPs. And
that interlock needs to be timer-safe...
git-svn-id: svn://svn.berlios.de/openocd/trunk@2247 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Doc update: say "jtag newtap ... -disable" records the
state after exiting the RESET state, matching the only
implementation we're working with so far (TI ICEpick-C).
Matching code updates. Now we can be sure that the
"enabled" flag value is correct after JTAG resets.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2246 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Fix a memory leak in jtag_tap_free(): unregister the event
callback too.
Also fix the associated conceptual bug in unregistering JTAG
event callbacks: since the same callback procedure is used
many times with different callback data (a TAP handle), that
data must be considered when unregistering any callback.
This could fix some crashes after TAP registration errors,
by making sure the reset event handler doesn't scribble over
memory that's now used by something else.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2245 b42882b7-edfa-0310-969c-e2dbd0fdcd60
Minor jtag cleanup:
- remove hidden assumption about JTAG event numbering
- move function declarations to a header
- some end'o'line whitespace
- use "calloc" not "malloc + memset"
git-svn-id: svn://svn.berlios.de/openocd/trunk@2244 b42882b7-edfa-0310-969c-e2dbd0fdcd60
OpenOCD doesn't actually *need* to be keeping all TCP ports
active ... creating security issues in some network configs.
Instead, let config file specify e.g. "tcl_port 0" (or gdb_port,
telnet_port) to disable that particular remote access method.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2240 b42882b7-edfa-0310-969c-e2dbd0fdcd60
- Make fast_load_image use the helper coverage the standard load_image.
- Improve whitespace in the moved lines.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2239 b42882b7-edfa-0310-969c-e2dbd0fdcd60
- Simplify argument parsing logic using switch statement.
- Use parse_u32 to ensure all values parse properly.
- Return syntax error when mode argument fails to parse.
git-svn-id: svn://svn.berlios.de/openocd/trunk@2236 b42882b7-edfa-0310-969c-e2dbd0fdcd60
zwelch
oharboe
ntfreak
duane