In 2016, ARM released the second edition of the semihosting specs
("Semihosting for AArch32 and AArch64"), adding support for 64-bits.
To ease the reuse of the semihosting logic for other platforms
(like RISC-V), the semihosting code was isolated from the ARM
target and updated to the latest specs.
The new code is already in use since January (in GNU MCU Eclipse
OpenOCD) and no problems were reported, neither for ARM nor for
RISC-V targets, after more than 7K downloads.
The 2 new files were formatted with uncrustify.
Change-Id: Ie84dbd86a547323bb8a5d24eab68fc7dad013d96
Signed-off-by: Liviu Ionescu <ilg@livius.net>
Reviewed-on: http://openocd.zylin.com/4518
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
gcc (GCC) 8.1.0 generates new warnings and thus fails the build.
The ARM disassembler warnings actually exposed a bug in SMALW, SMULW and
SMUL instructions decoding.
Reported by Eimers on IRC.
Change-Id: I200c70f75a9e07a1f13a592addc1c5fb37714440
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/4526
Tested-by: jenkins
Reviewed-by: Jiri Kastner <cz172638@gmail.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This improves startup time, which is important when connecting to
simulators. One problem is that triggers that are set when the debugger
connects are not cleared until enumeration happens. Execution may halt
due to a trigger set by a previous debug session, which could confuse
the user. If this happens, triggers will be instantly enumerated, so it
will only happen once per session.
Change-Id: I3396f713f16980a8b74745a1672fe8b8a2d4abae
I'm not sure what this check is adding, and it causes problems for implementations that take some time to report that they are halted out of reset (e.g. by executing Debug ROM).
Just remove our nop implementation. The default behavior when this is
left NULL does the same thing.
Change-Id: I865976c694d24661941584cb0efc92fc26612316
Neither the initial loop to clear dirty registers (which visits all
registers starting at R2 and counting upwards) nor the final explicit
flushes ensure a write-back in arm_dpm_write_dirty_registers.
This change makes sure that both our scratch registers (i.e. R0 and
R1) are written back to the target.
Change-Id: If65be4f371cd40af9a0cfa97f3730b070b92e981
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-on: http://openocd.zylin.com/4506
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Cortex-M target used 'auto_bp_type' mode. The requested type
of breakpoint was ignored and hard (FPB) breakpoints were set in
'code memory area' 0x00000000-0x1fffffff, soft breakpoints were set above
0x20000000.
The code memory area of Cortex-M does not mean the memory is flash and
vice versa. External flash (parallel or QSPI) is usually mapped above
code memory area. Cortex-M7 ITCM RAM is mapped at 0. Kinetis
has a RAM block under 0x20000000 boundary.
Remove 'auto_bp_type' mode, set breakpoints to requested type.
Change 'cortex_m maskisr auto' handling to use a hard temporary
breakpoint everywhere: it can also workaround not working soft breakpoints
on Cortex-M7 with ICache enabled.
Change-Id: I7a9f9464c5e10bfd7f17cba1037ed07a064fa2e8
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4429
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Various fixes for memory leaks, adds a target cleanup for aarch64
and ARM CTI objects.
Change-Id: I2267f0894df655fdf73d70c11ed03df0b8f8d07d
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4478
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested on PSoC6 (Cortex-M0+ core), onboard KitProg2 in CMSIS-DAP mode,
adapter_khz=1000.
Plain read:
flash read_bank 0 /dev/null
takes 48 seconds.
erase_check without this change:
flash erase_check 0
takes horrible 149 seconds!!
And the same command with the change applied takes 1.8 seconds.
Quite a difference.
Remove the erase-value=0 version of algorithm as the new one can check
for any value.
If the target is an insane slow clocked CPU (under 1MHz) algo
timeouts. Blocks checked so far are returned and the next call
uses increased timeout.
Change-Id: Ic0899011256d2114112e67c0b51fab4f6230d9cd
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4298
Tested-by: jenkins
Reviewed-by: Jonas Norling <jonas.norling@cyanconnode.com>
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>
'flash erase_check' command runs a check algorithm on a target
if possible. The algorithm is run repeatedly for each flash sector.
Unfortunately every start and stop of the algorithm impose not negligible
overhead.
In practice it means checking is faster than plain read only for
sectors of size approx 4 kByte or bigger. And checking sectors
as short as 512 bytes runs approx 4 times slower than plain read.
The patch changes API call target_blank_check_memory() and related
to take an array of sectors (or arbitrary memory blocks).
Changes in target-specific checking routines are kept minimal.
They use only the first block from the array and process it by
the unchanged algorithm.
default_flash_blank_check() routine repeats target_blank_check_memory()
until all blocks are checked, so it works with both multi-block
and single-block based checkers.
Change-Id: I0e6c60f2d71364c9c07c09416b04de9268807f5e
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4297
Tested-by: jenkins
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>
On SMP targets, the "target smp" command creates a list of targets
that belong to the SMP cluster. Free this list when a target gets
destroyed on shutdown. For simplicity, the complete list is free'd as
soon as the first target of the SMP cluster is destroyed instead of
individually removing targets from the list.
Change-Id: Ie217ae1efb2e819c288ff3b1155aeaf0a19b06be
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4481
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
apcsw command was limited to SPROT bit only.
Now user can manipulate any bit except size and addrinc fields.
Can be used e.g. to set bus signal 'cacheable' on Cortex-M7
Change-Id: Ia1c22b208e46d1653136f6faa5a7aaab036de7aa
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4431
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The CTRL/STAT register in the ARM DAP DP has a debug power up
ack bit and a system power up ack bit. Some devices do not set
the system power up ack bit until sometime later. To avoid having
the initial target examination fail due to this or to have a
sticky bit error report claim power failure due to this a user
can now specify that this bit should be ignored.
Change-Id: I2451234bbe904984e29562ef6f616cc6d6f60732
Signed-off-by: Eric Katzfey <eric.katzfey@mentalbee.com>
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3710
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
It's not implemented for 0.11 because we don't need it. Returning error
caused 0.11 targets to not be debuggable since change
848062d0d1.
Change-Id: I8b04a1fcf3c3e8bf8340cbf39aaf475d2a213519
Neither the initial loop to clear dirty registers (which visits all
registers starting at R2 and counting upwards) nor the final explicit
flushes ensure a write-back in arm_dpm_write_dirty_registers.
This change makes sure that both our scratch registers (i.e. R0 and
R1) are written back to the target.
Change-Id: If65be4f371cd40af9a0cfa97f3730b070b92e981
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-on: http://openocd.zylin.com/4506
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
* Enforce OpenOCD style guide.
Change-Id: I579a9f54ed22a774bf52f6aa5bc13bcbd2e82cd8
* Fail if `git diff` fails
Change-Id: I57256b0a24247f6123cb0e25a89c1b59867cb3f9
* Maybe every line gets its own shell?
Change-Id: I1a6f83e9f3d7cfd39f8933f0dba13c3cf76f71f6
* Maybe this will error properly.
Change-Id: I50803cfc229e61158569fb6b609195f7191ecac9
* Take different approach than merge-base
Change-Id: I345cbc4eecc4755c7127e8e36e403f7b727010b1
* Fix style issues.
Change-Id: I90e71f710858524812d0ab160b25c486b7b099e7
Cortex-M target used 'auto_bp_type' mode. The requested type
of breakpoint was ignored and hard (FPB) breakpoints were set in
'code memory area' 0x00000000-0x1fffffff, soft breakpoints were set above
0x20000000.
The code memory area of Cortex-M does not mean the memory is flash and
vice versa. External flash (parallel or QSPI) is usually mapped above
code memory area. Cortex-M7 ITCM RAM is mapped at 0. Kinetis
has a RAM block under 0x20000000 boundary.
Remove 'auto_bp_type' mode, set breakpoints to requested type.
Change 'cortex_m maskisr auto' handling to use a hard temporary
breakpoint everywhere: it can also workaround not working soft breakpoints
on Cortex-M7 with ICache enabled.
Change-Id: I7a9f9464c5e10bfd7f17cba1037ed07a064fa2e8
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4429
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Various fixes for memory leaks, adds a target cleanup for aarch64
and ARM CTI objects.
Change-Id: I2267f0894df655fdf73d70c11ed03df0b8f8d07d
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4478
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested on PSoC6 (Cortex-M0+ core), onboard KitProg2 in CMSIS-DAP mode,
adapter_khz=1000.
Plain read:
flash read_bank 0 /dev/null
takes 48 seconds.
erase_check without this change:
flash erase_check 0
takes horrible 149 seconds!!
And the same command with the change applied takes 1.8 seconds.
Quite a difference.
Remove the erase-value=0 version of algorithm as the new one can check
for any value.
If the target is an insane slow clocked CPU (under 1MHz) algo
timeouts. Blocks checked so far are returned and the next call
uses increased timeout.
Change-Id: Ic0899011256d2114112e67c0b51fab4f6230d9cd
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4298
Tested-by: jenkins
Reviewed-by: Jonas Norling <jonas.norling@cyanconnode.com>
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>
'flash erase_check' command runs a check algorithm on a target
if possible. The algorithm is run repeatedly for each flash sector.
Unfortunately every start and stop of the algorithm impose not negligible
overhead.
In practice it means checking is faster than plain read only for
sectors of size approx 4 kByte or bigger. And checking sectors
as short as 512 bytes runs approx 4 times slower than plain read.
The patch changes API call target_blank_check_memory() and related
to take an array of sectors (or arbitrary memory blocks).
Changes in target-specific checking routines are kept minimal.
They use only the first block from the array and process it by
the unchanged algorithm.
default_flash_blank_check() routine repeats target_blank_check_memory()
until all blocks are checked, so it works with both multi-block
and single-block based checkers.
Change-Id: I0e6c60f2d71364c9c07c09416b04de9268807f5e
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4297
Tested-by: jenkins
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>
On SMP targets, the "target smp" command creates a list of targets
that belong to the SMP cluster. Free this list when a target gets
destroyed on shutdown. For simplicity, the complete list is free'd as
soon as the first target of the SMP cluster is destroyed instead of
individually removing targets from the list.
Change-Id: Ie217ae1efb2e819c288ff3b1155aeaf0a19b06be
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4481
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
apcsw command was limited to SPROT bit only.
Now user can manipulate any bit except size and addrinc fields.
Can be used e.g. to set bus signal 'cacheable' on Cortex-M7
Change-Id: Ia1c22b208e46d1653136f6faa5a7aaab036de7aa
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4431
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The CTRL/STAT register in the ARM DAP DP has a debug power up
ack bit and a system power up ack bit. Some devices do not set
the system power up ack bit until sometime later. To avoid having
the initial target examination fail due to this or to have a
sticky bit error report claim power failure due to this a user
can now specify that this bit should be ignored.
Change-Id: I2451234bbe904984e29562ef6f616cc6d6f60732
Signed-off-by: Eric Katzfey <eric.katzfey@mentalbee.com>
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3710
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This saves us from re-reading s0 before doing just about anything
program buffer related.
Improves DebugBreakpoint from 3.01s to 2.89s. Feels like the improvement
should be larger than that. Maybe my metric isn't very good.
Change-Id: I85e1a1ddbf09006d76c451a32048be7b773dcfe9
Avoid ever overflowing the DWT_COMPARATOR array by allocating space for
16 comparators (the field is masked by 0xf).
On a stm32f767zi chip (on a nucleo-767zi board) I've been seeing crashes
with address sanitizer enabled due to its (apparent) 10 present
comparators. This appears to be due to
https://sourceforge.net/p/openocd/tickets/178/.
In non-address sanitizer builds, this would likely cause some random
memory to be written to in some cases. (see above bug for observations).
Change-Id: I2b7d599eb326236dbc93f74b350c442c9a502c4b
Signed-off-by: Cody P Schafer <openocd@codyps.com>
Reviewed-on: http://openocd.zylin.com/4458
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
This works around some side effects of the -rtos hack, namely that we
were unable to set hardware breakpoints on harts whose misa differed
from the first one. There may be other bugs like this one lurking
elsewhere. The only proper solution is for gdb to have a better user
interface when talking to a server that exposes multiple targets, but
that's a very big project.
This fixes#194.
Change-Id: I81aedddeaa922d220e936730e9c731545953ae21
This replaces the earlier mechanism which would propagate errors only
for targets that decided they wanted to. It was suggested by Matthias
Welwarsky from the OpenOCD team.
Change-Id: Ibe8e97644abb47aff26d74b8280377d42615a4d3
Kamal Dasu
Liviu Ionescu
Tim Newsome
Paul Fertser
Megan Wachs
Dan Robertson
Philipp Tomsich
Tomas Vanek
Matthias Welwarsky
Darius Rad
Ryan Macdonald
Cody P Schafer