Scan-build is unable to correctly follow the deferred loading of
queued read, finalized by the atomic write, thus it incorrectly
claims that the arrays d_u_ccsidr[] and i_ccsidr[] could carry
not initialized values:
armv7m_cache.c:154:31: warning: 1st function call argument
is an uninitialized value [core.CallAndMessage]
cache->arch[cl].d_u_size = decode_ccsidr(d_u_ccsidr[cl]);
armv7m_cache.c:172:29: warning: 1st function call argument
is an uninitialized value [core.CallAndMessage]
cache->arch[cl].i_size = decode_ccsidr(i_ccsidr[cl]);
Initialize the arrays to zero to hide these false positive.
Change-Id: I6d1e88093cb8807848643139647a571c1b566aa8
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Fixes: 04da6e2c62 ("target: cortex-m: add support for armv8m caches")
Reviewed-on: https://review.openocd.org/c/openocd/+/9167
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Use the macro 'LOG_LEVEL_IS()' to test 'debug_level'.
While there, use the macro 'LOG_LVL_*' in place of the numeric
value.
Skip all riscv code, as it is going to be updated soon from the
external fork.
Change-Id: Icad7e879e040d3b9cf1cc004c433f28725017493
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9070
Tested-by: jenkins
Add a comment in the breakpoint code to clarify the check for the
odd breakpoint length of 3 bytes, introduced by [1].
[1]: commit 0a5e03c12a ("cortex_m.c: Use two byte breakpoint for
32bit Thumb-2 request").
Change-Id: I024863d10078b5d9062c876aa59ccf70a81bf641
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9139
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins
Cores like Cortex-M7, Cortex-M55 and Cortex-M85 can have either
D-Cache and/or I-Cache.
Using SW breakpoints in RAM requires handling these caches.
Detect the presence of cache at examine.
Detect cache state (enable/disable) at debug entry.
Take care of caches synchronization through the PoC (usually the
SRAM) while setting and removing SW breakpoints.
Add command 'cache_info' to check cache presence and size.
Change-Id: Ice637c215fe3042c8fff57edefbab1b86515ef4b
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9077
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins
Do not echo the selected reset config. This is one of many changes to
make the behavior of Tcl commands more consistent.
This also avoids stray and confusing messages in the output of OpenOCD.
For example, the "reset_config" line here:
Open On-Chip Debugger 0.12.0+dev-00802-gb7f0145fc-dirty
Licensed under GNU GPL v2
For bug reports, read
http://openocd.org/doc/doxygen/bugs.html
cortex_m reset_config sysresetreq
Info : Listening on port 6666 for tcl connections
Info : Listening on port 4444 for telnet connections
While at it, fix some coding style and command handling issues.
Change-Id: I3b3d8687af1d23a2dc1764f29b52dc607b80cb59
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8638
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
We need to cleanup watchpoints on all targets in SMP group when GDB
connects. Otherwise, the targets will not be consistent.
Once thats fixed, both *_clear_target functions clearly duplicate
the corresponding *_remove_all functions.
Change-Id: I8e85dbc66fd3e596990d631ed2aed22959a8ca60
Signed-off-by: Samuel Obuch <samuel.obuch@espressif.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9086
Reviewed-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Some target, like aarch64, has more than one bank of registers.
Let the commands 'get_reg' and 'set_reg' to search the register
name in all banks, as is already done in command 'reg'.
Change-Id: Iae350a52f993790c5546925a2f7f81fbdb3f49b8
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Fixes: e8e62c5aca ("target/tcl: Add get_reg function")
Fixes: da73280101 ("target/tcl: Add set_reg function")
Reviewed-on: https://review.openocd.org/c/openocd/+/9122
Tested-by: jenkins
Reviewed-by: zapb <dev@zapb.de>
The coding style requires the 'case' to be at the same indentation
level of its 'switch' statement.
Align the code accordingly.
While there:
- add space around the operators;
- drop useless empty line.
Skip all riscv code, as it is going to be updated soon from the
external fork.
No changes are reported by
git log -p -w --ignore-blank-lines --patience
Change-Id: I2691dfdd2b6734143e14160b46183623e9773539
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9051
Tested-by: jenkins
To prepare for aligning switch and case statements, fix in advance
some checkpatch error due to existing code:
- remove useless parenthesis;
- remove useless 'break';
- join spit lines;
- add space around operators;
- remove 'else' after exit() and return.
Change-Id: I8a87a0ea104205d087dcb8cbf4c67ff13a47742f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9050
Tested-by: jenkins
Reviewed-by: zapb <dev@zapb.de>
The code falls-through in the default case, making it not easy to
read.
Add the explicit break to improve the readability.
Change-Id: I4784b883e0e82258de17018dfdfb59b4042ac743
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9049
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The code in an event handler can use the command '$target_name
configure' to add a new event or to remove or modify an existing
event.
Such operation impacts the list of event of the target and also
modify the event itself, causing OpenOCD to access memory already
deallocated or not anymore valid.
Use the safe version of list_for_each_entry() to iterate on the
list of events.
Make a local copy of the current event, to avoid issues if it gets
deallocated.
Use Jim_IncrRefCount() to guarantee that the body of the event
handler don't gets deallocated when the event is removed.
Change-Id: I936e35adddc030ba7cec6e2fc0c7d3b1b5c4a863
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9063
Reviewed-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Tested-by: jenkins
Format strings are often split to allow using the conversion
specifiers macros from <inttypes.h>.
When the format string ends with one of such macros, there is no
need to add an empty string "" after the macro.
In current code we have 203 cases of empty string present, against
1159 cases of string ending with the macro.
Uniform the style across OpenOCD by removing the empty string.
Don't modify the files 'angie.c' and 'max32xxx.c' as they are
already changed by other independent commits.
Change-Id: I23f1120101ce1da67c6578635fc6507a58c803e9
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9065
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The function keep_alive() is optimized and return immediately if
has nothing to do.
There is no need to overly-complicate the code with extra counters
or time computation plus the relative checks to reduce the number
of calls to keep_alive().
Drop such extra code.
Change-Id: I4574a3f154b5779f44105936c74af8fca1d2c49c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/9064
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins
Reviewed-by: Lucien Buchmann <lucien.buchmann@dufour.aero>
When we read unaligned memory there is an offset in the albuff buffer,
that we account for when copying back to original buffer. But in case
the first access failed, the retry call already removed the offset,
so doing it a second time shifts the returned memory.
Change-Id: Ie255c367ca6a001bfe7038a76cf8a6443e398c51
Signed-off-by: Samuel Obuch <samuel.obuch@espressif.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8987
Tested-by: jenkins
Reviewed-by: Ian Thompson <ianst@cadence.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Consider target endianness when reading control block and channel
information. Current implementation fails on big-endian devices.
Tested on TMS570 (big-endian) and on nRF52 (little-endian).
Note that in its current implementation RTT does not work properly on
TMS570 due to its missing support for background memory access.
Change-Id: Iab58804c42c85a932a750201a69ded35cebedd5d
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8993
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
The variable is already used in some parts of the code as boolean value
but have the wrong data type.
Change-Id: I50ccbf84c6f33a3034de989789c6b17312458ea8
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8989
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
The variables are already used in some parts of the code as boolean
value but have the wrong data type.
Change-Id: I2c4955a6ed463fabf63a1dbd79145cb63bc7a99c
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8988
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
While on it warn about screwed SWD diagrams in ADI spec
and add reference to a SWD timing diagram.
Change-Id: I628d707ebf8ce7c22ba19bdcfd06028d4eaa60f8
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/8690
Tested-by: jenkins
Running target algorithms on ARMv8M may require core in secure
mode with SAU and MPU off (as set after reset).
cortex_m_set_secure() forces this mode with optional save of
the previous state.
cortex_m_security_restore() restores previously saved state.
Change-Id: Ia71826db47ee7b0557eaffd55244ce13eacbcb4b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/8959
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Use LOG_TARGET_xxx() for log messages as it is used for other targets.
While at it, rework the log messages. For example by removing spaces or
punctuation marks at the end of the message.
Change-Id: I295001876d40527ec8f35c2aec8d562a29e57b26
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8967
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Print a debug message about missing multiprocessing support rather than
an error message.
Change-Id: Ia1581f7284747d8a92096d6f5515f891c8069f71
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8965
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Use LOG_TARGET_xxx() to indicate which target the message belongs to.
While at it, rework the log messages. For example, using correct format
specifiers.
Change-Id: I05031e0ae25fe9e7bc38dfb781b6623a967fd533
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8964
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Instead of using NULL terminated arrays to determine the last
element of the array, use the size of the array.
Change-Id: I3cdc0f6aef8a5110073aeef333c439e61fc54032
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8952
Tested-by: jenkins
Reviewed-by: Brandon Martin
[1] removed target_halt() from cortex_m_assert_reset()
It broke debug_reason tracking and the previous reason
was shown after reset halt.
Set debug_reason to DBG_REASON_DBGRQ during reset halt
preparation.
Fixes: [1] commit 226085065b ("target/cortex_m: drop useless target_halt() call")
Reported-by: Marc Schink <dev@zapb.de>
Change-Id: I685618ed158abde11f6e00eeeee1dfa8ed90952d
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/8945
Tested-by: jenkins
Reviewed-by: zapb <dev@zapb.de>
Update MPIDR decode to support the multithreading (MT) bit.
If detected, socket, cluster, core and multithread affinity levels are
decoded and displayed.
Change-Id: I43569141fa0eef8ee8fc16c187a4af3c23e97db8
Signed-off-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7190
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
For aarch32 secure EL3
- Change target_el to 3 for SVC/ABT/IRQ/FIQ/UND/SYS for aarch32 secure
EL3
- Do not update SPSR for SYS, behavior is UNPREDICTABLE (ARMv8-A F5.1.121)
- Do not execute DRPS for SYS, behavior is UNPREDICTABLE (ARMv8-A
F5.1.51)
Change-Id: Ic1484665cd53afcccb5c20b152993a3f0407f8a2
Signed-off-by: Kevin Yang <kangyang@google.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/5854
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Plamena Marinova <pmarinova@hilscher.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Fix endianness for cortex_r4 and cortex_r5 when inserting software
breakpoints. Because the cortex_a target is used by the cortex_r
architecture and some chips start in BE by default (e.g. TMS570)
Change-Id: I68b7fe7c4604de67fee2e64fff0fad2691659a58
Signed-off-by: Lucien Dufour <lucien.buchmann@dufour.aero>
Reviewed-on: https://review.openocd.org/c/openocd/+/8909
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Since the initial commit f444c57bf2 ("arm_cti: add cti command
group") the helper handle_cti_dump() return JIM error codes.
Fix it by returning standard OpenOCD error codes.
Change-Id: Ia36b82083d213aff90fe22fcfe7fbe26172806a3
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8904
Tested-by: jenkins
Rewrite only the command, but still use the old jimtcl specific
code in arm_tpiu_swo_configure(), shared with command 'create'.
Change-Id: If2258f048403f54faf229e602d9b395b71894f97
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8900
Tested-by: jenkins
Rewrite only the command, but still use the old jimtcl specific
code in arm_tpiu_swo_configure(), shared with commands 'configure'
and 'cget'.
Change-Id: I39c69b1cdc23f7b5f875df3e15be987c715b0bcf
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8899
Tested-by: jenkins
Rewrite only the command, but still use the old jimtcl specific
code in cti_configure().
Change-Id: I29fb952a7c8148416b301cbf78b6e342979af7d3
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8898
Tested-by: jenkins
Rewrite only the command, but still use the old jimtcl specific
code in dap_configure().
Change-Id: I3360884616367aae52f5b32247d9864000c53fdc
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8897
Tested-by: jenkins
On ARMv7a/r the single-step is implemented through a HW breakpoint
that hits instructions at any address except the address of the
current instruction.
The method above fails in case of an infinite loop coded by a
single instruction that jumps on itself; in such case, the same
instruction (at the same address) is executed over and over and
the breakpoint never hits. In current code this case is wrongly
considered as an error.
Reduce the timeout while waiting for the HW breakpoint being hit,
then halt.
The jump on itself would be executed several times before the
timeout and the halt, but this is not an issue. There are few
"pathological" instructions in ARMv7a/r that jumps on itself and
that can have side effects if executed more than once. They are
listed in the code. We do not consider these as real use cases
generated by a compiler.
Document the method in the code.
Report that the single-step function is not properly managing the
HW breakpoints if it exits on error. To be fixed in the future.
Change-Id: I9641a4a3e2f68b83897ccf3a12d3c34e98a7805c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8871
Tested-by: jenkins
`target create` calls can get quite long and an indication what is the
option that caused the error can be helpful.
Also, there can be multiple `-event` options for different events,
therefore indicating which one is it is also helpful.
Change-Id: I5ea61437ca9705e790ed8343183883a3fdfebc80
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8861
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
