arm_adi_v5: separate ROM table parsing from command output [2/3]

This change only targets the output of rtp_cs_component().
To easily propagate the coordinates of the CoreSight component,
add them in the struct that holds the register values.
While there, define a macro for the max depth of ROM tables.

Change-Id: I75e5ef4f9419da3192123aebcd61471c2af9374f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6820
Tested-by: jenkins

arm_adi_v5: separate ROM table parsing from command output [1/3]

In OpenOCD arm_adi_v5 we have already two implementations of code
for parsing the ADIv5 ROM table:
- in the commands "dap info" and "$dap_name info";
- in the function dap_lookup_cs_component().
Adding support for ADIv6 requires extending both implementations.

Moreover, current code does not handle few aspects of the ROM
parsing, e.g. the "Power Domain IDs".
To add such extensions both implementations should be touched.

I plan to add a command to parses (again) the ROM table and dump a
simple prototype of a configuration script for the target, useful
while analysing a new target.

Keeping aligned all these implementation would be too complex.

With focus to "dap info" command, decouple the part of code to
walk-through the ROM table from the code that creates the command
output.
The idea is to keep a single implementation for the walk-through
code, while parametrizing the output code to handle the generation
of a configuration script or the result of the function
dap_lookup_cs_component().

This change only targets the output of MEM-AP header
Further changes will target other parts of the code.

While there, add a message if MEM-AP is not accessible.

Change-Id: I112f637edfdb8688afb4e631297f6536da9604f1
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6819
Tested-by: jenkins

arm_adi_v5: handle faulting entry in ROM table

ARM IHI0031F "Arm Debug Interface Architecture Specification"
chapter C2.6.1 "BASE, Debug Base Address register" reports:
A debugger must handle the following situations as
non-fatal errors:
- ...
- An entry in the ROM Table points to a faulting location.
- ...
Typically, a debugger issues a warning if it encounters
one of these situations. However, Arm recommends that it
continues operating. An example of an implementation that
might cause errors of this type is a system with static
base address or ROM Table entries that enable entire
subsystems to be disabled, for example by a tie-off input,
packaging choice, fuse, or similar.

Don't halt ROM table parsing if one entry causes an error; log the
error condition and continue to next entry.
Not sure if we have to send an ABORT before continuing.

Change-Id: I94fdb5b175bfb07dde378149421582b7e7cd5b09
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6818
Tested-by: jenkins
Reviewed-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>

arm_adi_v5: report sysmem on class 0x9 ROM tables

As Class 0x1 ROM table, also Class 0x9 ROM tables encodes a flag
for system memory access.

Detect the flag in rtp_cs_component() and dump the same message
for both type of ROM tables.
Extend rtp_read_cs_regs() to read ARM_CS_C9_DEVID.

Change-Id: Ic85d1ea068ed706ceedfd65076ff4c96d04e9792
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6817
Tested-by: jenkins
Reviewed-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>

arm_adi_v5: split ROM table loop from generic coresight

During ROM table parsing, each ROM table entry points to a
CoreSight component that can, in turn, be another ROM table.

Split the specific code for ROM table handling from the generic
CoreSight code.
Log an error if a ROM table entry cannot be read.

Change-Id: I5ad106a99b9c21ddb48b5b162ae87101e4f49878
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6816
Tested-by: jenkins

arm_adi_v5: rework dap_read_part_id()

Rework dap_read_part_id() while preparing for reorganizing the
'ROM Table Parsing' (RTP):
- rename it with 'rtp' prefix;
- extends it to read other CoreSight registers, thus improving the
  overall speed by queuing more reads;
- reduce the list of arguments by using a struct;
- reorder the reads by increasing offset, potentially gaining
  speed using MEM_AP_REG_BDx and/or auto-increment;
- log a debug message in case of read error.

Change-Id: I6544ac7740b808a6c0fbacf97ac00b97f5bd3832
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6815
Tested-by: jenkins

arm_adi_v5: add support for display Class 0x9 ROM tables

ADI v5.1 and v6.0 permit the definition of CoreSight components (class 9
ROM entries).

dap_rom_display() is refactored a bit such that we always end up with
attempting to parse the ROM contents using the appropriate upper limit
for class 1 and 9 ROM types.

Change-Id: I4ba497b3807f1f11f06186eb6e61959ea3540c59
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6359
Tested-by: jenkins

arm_adi_v5: describe Class 0x9 Device Architecture register

Use the list of values from ARM IHI0029E to decode and print the
Device Architecture register.
Add attribute 'unused' to the function, not used yet.

Change-Id: I7b1dd204bd1db671578c588372b667e23611876c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6463
Tested-by: jenkins
Reviewed-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>

adi_v5_jtag: reduce verbosity on persistent WAIT

In case of AP not responding, e.g. not clocked, the first WAIT
reply is logged as:
DAP transaction stalled (WAIT) - slowing down
then OpenOCD retries the transaction few times, until it timeouts.
At each retry it prints the message:
DAP transaction stalled during replay (WAIT) - resending
Depending on JTAG speed and transport latency, the amount of log
messages can be quite annoying and not relevant.
The last printed line is at timeout:
Timeout during WAIT recovery

Reduce the verbosity.

Change-Id: I5a7a337527c98b2450de59066b13713511c2894f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6814
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>

arm_coresight: define ARM_CS_CIDR_CLASS()

Right now it has a single use but it will soon be used more.

Change-Id: I9a819c65df467fc859e4b5251035a17ed33daa35
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6813
Tested-by: jenkins
Reviewed-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>

cross-build.sh: fix build with capstone

since commit 12d1ad0c7529 : update capstone include path ...
the generated capstone.pc is not working

so fix the includedir in capstone.pc to get github action working

Change-Id: I7767e181a74c73a7514eeb6293cd556a794dbfe9
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6969
Tested-by: jenkins
Reviewed-by: Tim Newsome <tim@sifive.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

doc: fix typo s/Not/Note/

Append lacking e to word Note

Signed-off-by: Piotr Kasprzyk <ciri@ciri.pl>
Change-Id: Ibd40a2f93d11cf1945361f0c46329b88963d6826
Reviewed-on: https://review.openocd.org/c/openocd/+/6963
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target: document possibly unreachable target in deinit_target()

Change-Id: I95ff3d200bb2c8f5bc43a34c92726d9c47f8c172
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6953
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target: fix build with jimtcl 0.79

In jimtcl 0.80 the prototype of Jim_DictPairs() has changed.
The only code in OpenOCD that uses Jim_DictPairs() has been merged
recently and it only uses the current jimtcl syntax.

To allow compiling OpenOCD master branch with older versions of
jimtcl, detect the version of jimtcl and use the appropriate
syntax.

Change-Id: I6fc78303b6a4db064a97f326c46119f4568e88f3
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reported-by: dullfire@yahoo.com
Reviewed-on: https://review.openocd.org/c/openocd/+/6948
Tested-by: jenkins

bluenrg: add support for bluenrg-lps device and board

Added bluenrg-lps support
Added file for the board steval-idb012v1
Fixed size_info information using a mask
Changed the if condition in bluenrg-x.cfg to be valid only for bluenrg-1 and bluenrg-2

Signed-off-by: Salvatore Giorgio PECORINO <salvatore-giorgio.pecorino@st.com>
Change-Id: Ic0777ec0811ee6fac7d5e1d065c4629e47d84a1f
Reviewed-on: https://review.openocd.org/c/openocd/+/6928
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

server/gdb: fix gdb remote monitor cmd on multi-target

Commit 5ebb1bdea1df ("server/gdb: fix return of gdb remote monitor
command") replaces the call to command_run_line() with call to
Jim_EvalObj() but does not properly set the "context".
In multi-target environment, his can cause the erroneously
execution of the command on the wrong target.

Copy from the code in command_run_line() the proper setup before
executing Jim_EvalObj().

Change-Id: I56738c80779082ca146a06c01bc30e28bc835fd3
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reported-by: Bohdan Tymkiv <bohdan200@gmail.com>
Fixes: 5ebb1bdea1df ("server/gdb: fix return of gdb remote monitor command")
Reviewed-on: https://review.openocd.org/c/openocd/+/6966
Tested-by: jenkins
Reviewed-by: Bohdan Tymkiv <bohdan200@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Tim Newsome <tim@sifive.com>

smp: deprecate legacy SMP core switching support

The deprecation was already in the documentation since v0.11.0
through commit 85ba2dc4c6ab ("rtos/hwthread: add hardware-thread
pseudo rtos") but OpenOCD was not informing the user printing a
runtime message.

Remove the deprecated method from the documentation and print a
deprecated message at runtime.
There is no reliable way to print the same message in GDB console,
so we have to rely on user noticing it in the OpenOCD log.
Target is to remove the functionality after v0.12.0.

Change-Id: Idd2d9e3b6eccc92dcf0432c3c7de2f8a0fcabe9f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6862
Tested-by: jenkins

drivers/cmsis-dap: Correct the DAP protocol parameter parsing error

Fixes: 01030fb89354 (drivers/cmsis-dap: tidy up buffer access)
Change-Id: Id192d3930a89980d641058b6444d12caec19ce6f
Signed-off-by: Hongbin Ji <longma@orbbec.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6954
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: jihongbin <jhb_ee@163.com>

target/disassembler:  update capstone include path to <capstone.h>

on macos (homebrew base) `pkg-config --cflags capstone` output with
`-I/opt/homebrew/Cellar/capstone/4.0.2/include/capstone`
gcc not find headers on parent "include" path,
causes build error `fatal error: 'capstone/capstone.h' file not found`
it's ok to change to <capstone.h> for all platforms.

Signed-off-by: fatalc <cnfatal@gmail.com>
Change-Id: Ia2e2058024d4fc1a57a8b4ea847c664d74f67efb
Reviewed-on: https://review.openocd.org/c/openocd/+/6946
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/arm_cti: Fix error handling in 'cti create'

Handle JIM_CONTINUE return value of adiv5_jim_mem_ap_spot_configure(),
otherwise OpenOCD silently quits when an unknown option is provided.

Change-Id: I9b1351c0911e74999d8dd1260ede9760088510d7
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6944
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

bcm2835gpio: Make buffer an output before the GPIO connected to it

The correct ordering is required to prevent two outputs connected
together.

Change-Id: I634a9ca7e0ccf337d1723011b8aee1f2d81efbcf
Signed-off-by: Steve Marple <stevemarple@googlemail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6937
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

linuxgpiod: add SWDIO buffer

The SWDIO buffer requires a direction pin to select input or output
direction. Output is selected by a high logic level (matches
bcm2835gpio driver).

Change-Id: I240cb99a5dfea08121bb33d4b5e2108ce7597468
Signed-off-by: Steve Marple <stevemarple@googlemail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6936
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/nor/stm32f1x: add can_load_options flag for GD32F1x0, F3x0 and E23x

According to GigaDevice user manuals the devices have OBRLD bit in FMC_CTL
register which is functionally compatible with OBL_LAUNCH @ FLASH_CR
of STM32 counterparts.

Change-Id: I84d231b38815fcb6452fd73b9153b269cce3b737
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6759
Tested-by: jenkins
Reviewed-by: Andrzej Sierżęga <asier70@gmail.com>

tcl/target/gd32vf103: add flash bank

The flash is compatible with stm32f1x, reuse the driver.

Extend the size of work area to RAM size of the smallest device.

Stop watchdogs before flash programming.

Change-Id: I67a7654a6e196f9d4b2409edaa7990c53334437e
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6711
Tested-by: jenkins
Reviewed-by: Tim Newsome <tim@sifive.com>

flash/stm32f1x: add support for RISC-V GigaDevice GD32VF103

The device has compatible flash macro with STM32F1 family, reuse
stm32f1x driver code.

Detect non-ARM target - for simplicy test target type name 'riscv'
and the address has 32 bits.

In case of RISC-V CPU use simple chunked write algo - async algo
cannot be used as the core implemented in this device doesn't
allow memory access while running.

Change-Id: Ie3886fbd8573652691f91a02335812a7300689f7
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6704
Tested-by: jenkins
Reviewed-by: Tim Newsome <tim@sifive.com>

rtos: zephyr: do not use deprecated symbols name

Zephyr plan to remove openocd specific symbols in favour
of more generic one.

These generic symbols has been introduced in Zephyr 2.6.

Signed-off-by: Julien Massot <julien.massot@iot.bzh>
Change-Id: I89418c9c378fb8b8baa29763fc6f1b6e652dc7ef
Reviewed-on: https://review.openocd.org/c/openocd/+/6844
Tested-by: jenkins
Reviewed-by: Stephanos Ioannidis <root@stephanos.io>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/image: fix - p_flags field in ELF64 segment headers is 64 bits wide

Fixed the reading of p_flags in ELF64 segment headers - that field
is 64 bits wide.

Change-Id: I053ca57d36efb54b7c638484acd6c7a2fbcbd05a
Signed-off-by: Jan Matyas <matyas@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6927
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

server/gdb: fix return of gdb remote monitor command

Current implementation for gdb remote monitor command uses the
command_run_line() to execute the command.
While command_run_line() has several advantages, it unfortunately
hides the error codes and outputs the result of the command
through LOG_USER(), which is not what gdb requires. See 'qRcmd' in
https://sourceware.org/gdb/onlinedocs/gdb/General-Query-Packets.html

Replace command_run_line() with Jim_EvalObj() and parse the output
to provide the proper result to gdb.

Can be tested by defining in OpenOCD:
proc a {} {return hello}
proc b {} {return -code 4}
proc c {} {return -code 4 "This is an error!"}
then by executing in gdb console:
monitor a
monitor b
monitor c
monitor foo

Change-Id: I1b85554d59221560e97861a499e16764e70c1172
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reported-by: Torbjorn Svensson <torbjorn.svensson@st.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6886
Tested-by: jenkins

tcl/target/stm32l4x: align format/order/comments with stm32f4x

Change-Id: Ie97bb2f56b582bc735c238af5f160fcb28a61eb0
Signed-off-by: Markus Reiter <me@reitermark.us>
Reviewed-on: https://review.openocd.org/c/openocd/+/6933
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/stm32l4x: switch to new TPIU/SWO support

Change-Id: I3362fa7292eae7a3ba119cf6183f8bc4cbd5cbd4
Signed-off-by: Markus Reiter <me@reitermark.us>
Reviewed-on: https://review.openocd.org/c/openocd/+/6932
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/stm32l4x: set default WORKAREASIZE to smallest device

Change-Id: Ia8bfb664ff28bd0579492032ce513b010e71c593
Signed-off-by: Markus Reiter <me@reitermark.us>
Reviewed-on: https://review.openocd.org/c/openocd/+/6931
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/target/stm32f4x: fix name

Change-Id: I9baa79d8cf402991e6638c255a91728b8a77020c
Signed-off-by: Markus Reiter <me@reitermark.us>
Reviewed-on: https://review.openocd.org/c/openocd/+/6930
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

nds32: deprecate it, together with aice adapter driver

The target nds32 and its companion adapter aice have not received
any real improvement since 2013.
It has been hard to keep them aligned during the evolution of
OpenOCD code, with no way for maintainers to really check if they
are still working.
No real documentation is present for them in OpenOCD.

The arch nds32 has been dropped from Linux kernel v5.18-rc1.

Deprecate both nds32 target and aice adapter with the target of
dropping them for v0.13.0.
Remove automatic build of aice, forcing user to select it.

Change-Id: Ib465d676246fa3b4e95c3d399ba9a5cf1f8b3baf
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6887
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

flash/nor/stm32f1x: lock flash in case of error

The current code locks the flash controller in case of error during
flash write only. An error in other flash operations may cause the
flash is left unlocked.

Implement locking also after error in erase, mass erase, options
write and erase.

Change-Id: I26c2ed7914e7847122306f29b777b9eefd1dc580
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6710
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

flash/nor/stm32f1x: unify flash error reporting

stm32x_wait_status_busy() has two side effects in case of flash programming error:
- reports error
- clears error bit in status register

Use stm32x_wait_status_busy() to report also flash error during target
algo flash write.

While on it use more descriptive error codes in stm32x_wait_status_busy().

Change-Id: I6e1cffc2aa5411b918a23ed62d5194910888a9d1
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6709
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

flash/nor/stm32f1x: tidy up async algo supporting code

Use target_get_working_area_avail() instead of try-fail iteration.

Call destroy_reg_param() in a for cycle.

Change-Id: I1891d1ffdea99010c6ab66b9578400b9d7922e20
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6708
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

flash/nor/stm32f1x: remove write alignment code

Use flash infrastructure to ensure writes are halfword aligned.

Change-Id: Iddca3a256ace3486a23e1a9cb6a31c7a91ee58bf
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6707
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

flash/nor/stm32f1x: allow write fallback for flash options

Mostly refactoring.

Rename original stm32x_write_block() to stm32x_write_block_async()
as it uses target async algo.

Introduce new stm32x_write_block() and move slow, host controlled
fallback flash write there.

The change allows stm32x_write_options() to use slow flash write fallback.

While on it rename variables where halfword count is stored.

Change-Id: I386ae15cf052b1490461ed8f7eea5b4403d466f7
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6706
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

tcl/board: Add NXP FRDM-K64F

Change-Id: I4b8fbfb2948c4295c2a34d641dd59a73c512d9fa
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6884
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

drivers/cmsis-dap: Remove stray whitespace

Change-Id: I7b60f9e87af2f582864ce94198d0343acf7d45f2
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6883
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

libusb_helper.h: Increase USB timeout

When we debug a target that works as a USB device, halting
the target causes the USB communication with the USB host to
become unresponsive. The host will try to reconnect/reset/setup
the unresponsive device during which communication with other
devices on the same USB bus can get stalled for several seconds.
If the JTAG adapter is on the same bus, we need to make sure
openOCD will wait for packets at least as long as the host USB
stack. Otherwise the USB stack might deliver a valid packet, but
openOCD would ignore it due to the timeout. The xHCI spec uses 5
sec timeouts, so let's use that in openOCD with some margin.

Use this value in all libusb calls. HID API might have a libusb
backend and would probably be victim to the same bug, so it
should use this timeout, too.

Ticket: https://sourceforge.net/p/openocd/tickets/343/
Signed-off-by: Gabor Csapo <gaborcsapo@google.com>
Change-Id: Ia3dc1356e676fe550f57a4c72f7a24ba296b6af2
Reviewed-on: https://review.openocd.org/c/openocd/+/6882
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

gdb_server: Improve logging of GDB-remote packets

- Print also the target name, not just the packet contents.
  This is important when there are more GDB servers (more
  debug-able targets) active in one OpenOCD session.

- Log also the received Ctrl-C requests coming from GDB
  (one byte 0x3), ACKs ("+") and NACKs ("-").

- Do not print zero-length incoming packets (this occurred
  when Ctrl-C packets were received).

- Removed a stray apostrophe "'" that got printed
  in gdb_log_outgoing_packet()

Signed-off-by: Jan Matyas <matyas@codasip.com>
Change-Id: If68fe0a8aa635165d0bbe6fa0e48a4645a02da67
Reviewed-on: https://review.openocd.org/c/openocd/+/6879
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

helper/bits: add BIT_ULL and GENMASK macros

To support 64 bits bit and masks
Replace local definition of BIT in rtos/chromium-ec

Change-Id: I1f268d6e8790f1b07bf798680b797878ce81064b
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6857
Tested-by: jenkins

openocd: include config.h in every file .c

Including config.h as first is required for every C file.
Add it to the C files that still miss it.

Change-Id: I1a210e7d3a854958a85a290b086ad8a9f5176425
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6856
Tested-by: jenkins

arm_tpiu_swo: fix autodetection of SWO pin frequency

While the documentation reports that SWO pin frequency can be
omitted to let the adapter autodetect the value, the code wrongly
drops an error when pin frequency is not specified.

Don't require the pin frequency to be set at "enable", but verify
that the adapter has properly changes it to a valid value.

Change-Id: I3dfbe3256e8887ef4f03512769b06381cdc9db0d
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reported-by: Karl Palsson <karlp@tweak.net.au>
Fixes: 184724d14e12 ("arm_tpiu_swo: add support for independent TPIU and SWO")
Reviewed-on: https://review.openocd.org/c/openocd/+/6310
Tested-by: jenkins
Reviewed-by: zapb <dev@zapb.de>

tcl/tools: Add function to measure the speed of ARM Cortex-M devices

Tested on an EFM32PG12 Starter Kit.

Change-Id: I2cbc36fe0d2ad2089bf8c1e7d2260daaae4ddbb4
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/5353
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

semihosting: fix mode flags for local host open()

Commit dbbac5f11d66 ("semihosting: use open mode flags from GDB,
not from sys/stat.h") fixes the conversion of the mode flags from
ARM semihosting encoding for SEMIHOSTING_SYS_OPEN to GDB mapping
for open().
Doing this, it breaks the conversion to local host's OS mapping
for open().

Split the conversion array to one for GDB and one for local host.
The local host conversion array is taken directly from the old
code.

Change-Id: I385321ddd32c3ac5cf6da3f1ce9eff76b05dd527
Fixes: dbbac5f11d66 ("semihosting: use open mode flags from GDB, not from sys/stat.h")
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reported-by: Erhan Kurubas <erhan.kurubas@espressif.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6870
Tested-by: jenkins

breakpoints: fix build on -fno-inline

Some configuration of GCC could default to -fno-inline, causing
the build to fail after commit fb43f1ff4e2f ("target: Rework 'set'
variable of break-/watchpoints").

Switch the new inline functions to 'static inline', as it's widely
used in the rest of the code.

Change-Id: I8bf31045a137bd34ed825f4b2a9338eb3c70046d
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Fixes: fb43f1ff4e2f ("target: Rework 'set' variable of break-/watchpoints")
Reviewed-on: https://review.openocd.org/c/openocd/+/6881
Tested-by: jenkins

target: Rework 'set' variable of break-/watchpoints

The 'set' variable name suggests a boolean data type which determines
whether a breakpoint (or watchpoint) is active. However, it is also
used to store the number of the breakpoint.

This encoding leads to inconsistent value assignments: boolean and
integer values are mixed. Also, associated hardware comparator
numbers, which are usually numbered from 0, cannot be used directly.
An additional offset is required to store the comparator numbers.

In order to make the code more readable and the value assignment more
consistent, change the variable name to 'is_set', its data type to 'bool'
and introduce a dedicated variable for the break-/watchpoint
number.

In order to make the review easier, the data types of various related
variables (e.g. number of breakpoints) are not changed.

While at it, fix a few coding style issues.

Change-Id: I2193f5639247cce6b80580d4c1c6afee916aeb82
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6319
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/stm32l4x: fix auto-probe when RDP is promoted from 0 to 0.5

Considering this use case: (using STM32 L5 or U5)
 1- first probe : TZEN enabled, RDP level 0
    flash_regs_base |= STM32L5_REGS_SEC_OFFSET => 0x50022000
 2- the user promotes the RDP to level 0.5
 3- the second probe, fails to read OPTR using secure flags_regs_base:
    used OPTR address is 0x50022040

Step 3 fails because when RDP is level 0.5, we should use Non-Secure
flash registers.
To fix this, always use NS flash regs to read OPTR in probe functions.

Fixes: 80d323c6e82b (flash/stm32l4x: introduce auto-probe when OPTR is changed)
Change-Id: I296aa633972b0c410b927488c999584a07b912d3
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6864
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

tcl/stm32u5x: fix clock config used at 'reset init'

Change-Id: If004a04b93be47439809ea3fa336b14de7a12277
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6597
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

stlink: enable queuing with stlink-server API v3

ST-Link Server 2.1.0-1 fixes concurrency issue with RW_MISC command
Starting from this version the ST-Link Server API is now v3.

In this change we save the ST-Link Server version, and check if the
API is greater or equal to 3 to enable the queuing.

Change-Id: I239eb81024700514c607a269b66651f457206faa
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6876
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

semihosting: permit redirection of semihosting I/O to TCP

This command permits the usage of a TCP port to perform debug and stdio
operations:
 - debug : READC, WRITEC and WRITE0
 - stdio : READ, WRITE

This will permit the separation of semihosting message from OpenOCD log,
and separate semihosting messages per core.

syntax: arm semihosting_redirect (disable | tcp <port> [debug|stdio|all])

this allows to select which operations to be performed via TCP (debug,
stdio or all (default)).

Note: for stdio operations, only I/O from/to ':tt' file descriptors are
redirected.

tested using netcat on ubuntu

Change-Id: I37053463667ba109d52429d4f98bc98d0ede298d
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/5562
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

keep-alive: drop link with log framework

OpenOCD implements the GDB keep-alive by sending empty strings as
output for GDB client. This has been implemented as part of the
log framework, creating an odd dependency.

Move the keep-alive notifications out of log framework.
For the moment, keep keep_alive() inside log.c, but it should be
moved in server.c

This should also fix an old issue with KDE Konsole when tab alert
for activity is enabled. The empty strings is sent to all the
connections, including telnet, and causes the tab running OpenOCD
telnet to continuously show activity even when no new text is
printed. Anyway, I cannot replicate this issue anymore.

Change-Id: Iebb00b00fb74b3c9665d9e1ddd3c055275bfbd43
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6840
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

gdb_server: simplify logic to enable/disable gdb_log_callback()

GDB client cannot always display generic messages from OpenOCD.
The callback gdb_log_callback() is continuously added and removed
to follow the GDB status and thus enabling/disabling sending the
OpenOCD output to GDB.
While this is a nice stress test for log_{add,remove}_callback(),
it is also a waste of computational resources that could impact
the speed of OpenOCD during GDB user interactions.

Add a connection-level flag to enable/disable the log callback and
simply change the flag instead of adding/removing the callback.

Use an enum for the flag instead of a bool. This improves code
readability and allows setting other states, e.g. keep-alive
through asynchronous notification https://review.openocd.org/4828/

Change-Id: I072d3c6928dedfd0cef0abe7acf9bdd4b89dbf5b
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6839
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

server: change prototype of add_service()

To easily add new methods to a service, pass all the methods
through a struct.
While there, drop the typedef for the methods and add currently
unused new methods to support keep-alive and connections during
keep-alive.

No change in functionality.

Change-Id: I2b5e7140db95021f6e7201e9d631ee340c60b453
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6838
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

log: drop global current_time

The value of this variable is not shared across functions, so the
variable can be local.

Change-Id: I00b0444209e81c07bb57fb732f47052ad0596728
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6837
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

server: fix: remove kept_alive() from server loop

The kept_alive() action is specific of a server that enjoyed an
unscheduled keep_alive and want to communicate it to the keep
alive logic to reschedule next keep_alive().
In server loop we are not expected to call kept_alive().
Remove it!

This call was erroneously added in commit 94e75e0c06c4.
Later, commit 7442b26d45dc properly added the same call in
gdb_put_packet(), but incorrectly left the older in place.

Change-Id: If476410f870eebfbdaccdb1366ba2e9254e2fdf6
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6836
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

Remove all occurrences of 'mem2array' and 'array2mem'

Replace deprecated commands 'mem2array' and 'array2mem' with
new Tcl commands 'read_memory' and 'write_memory'.

Change-Id: I116d995995396133ca782b14cce02bd1ab917a4e
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6859
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target: Deprecate 'array2mem' and 'mem2array''

Replace 'mem2array' and 'array2mem' with a Tcl wrapper that
internally uses 'read_memory' and 'write_memory'.

The target-specific 'mem2array' and 'array2mem' functions
remain for now.

Change-Id: If24c22a76ac72d4c26916a95f7f17902b41b6d9e
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6308
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/tcl: Add 'read_memory' and 'write_memory'

These functions are meant as replacement for 'mem2array' and
'array2mem'.

The main benefits of these new functions are:

 * They do not use Tcl arrays but lists which makes it easier
   to parse (generate) the data. See the Python Tcl RPC code
   in contrib as a negative example.

 * They do not operate on Tcl variables but instead return (accept)
   the Tcl list directly. This makes the C and Tcl code base
   smaller and cleaner.

 * The code is slightly more performant when reading / writing
   large amount of data. Tested with a simple Python Tcl RPC
   benchmark.

Change-Id: Ibd6ece3360c0d002abaadc37f078b10a8bb606f8
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6307
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

jimtcl: add configure flag for build maintainer mode

When jimtcl is built in maintainer mode, it runs extra tests at
exit to look for memory leak due to jim objects not properly freed
either through Jim_IncrRefCount()/Jim_DecrRefCount() or by passing
it to a jim API.

Add optional OpenOCD configure flag '--enable-jimtcl-maintainer'
to enable jimtcl maintainer mode.

Modify the implementation of macro AX_CONFIG_SUBDIR_OPTION to
allow expanding a variable passed as second argument.

Change-Id: Id1a39b25cee3773b172faf70803fa150182f0cd6
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6871
Tested-by: jenkins
Reviewed-by: zapb <dev@zapb.de>

board: Add LS1046ARDB

This adds support for the LS1046A Reference Design Board. There are
several JTAG headers accessable once the case is opened, but this config
is for the externally-accessable CMSIS DAP.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Change-Id: I0f83470da3758f0c4512ce47348c4db7de17b27e
Reviewed-on: https://review.openocd.org/c/openocd/+/6855
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target: Add LS1046A

The LS1046A is a quad-core processor from NXP in the layerscape family.
This SoC is a bit tricky to program: while the AArch64 CPUs are
little-endian, most of the peripherals are big-endian. Care must be
taken when interpreting memory reads/writes. This processor is in the
same family as the ls1012a, so the setup is similar.

If you use OpenOCD to attach early in the boot process, only the cpu0
may be available. Trying to halt other CPUs will fail. To avoid this,
defer examination of cpus 1-3, and provide a core_up helper (like e.g.
zynqmp).

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Change-Id: If5a1a9441fb35fea3e05dc708b42e0cb3bbf2a54
Reviewed-on: https://review.openocd.org/c/openocd/+/6854
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/nor/efm32: Use Cortex-M 'core_info' field

Change-Id: I5e477036e5cb7518c35df88878d53261311deb40
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6868
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/nor/sim3x: Fix typo

Change-Id: I2143c81d44b49bed9585c4aaee2bb6e2165345f2
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6869
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/stm32h7x: fix FLASH_WPSN_PRG mask used for protection

STM32H7Ax/7Bx devices have a different WPSN mask (0xFFFFFFFF),
(0xFF for STM32H74x/75x and STM32H72x/73x devices).

And when supporting STM32H7Ax/7Bx devices, stm32x_protect() was
not updated accordingly.

Change-Id: I081217af3e5ed815b67bfdfec7f4ebaa3152a865
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Fixes: 0b7eca17691a (flash/stm32h7x: add support of STM32H7Ax/H7Bx devices)
Reviewed-on: https://review.openocd.org/c/openocd/+/6858
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

jtag_vpi: Minor cleanup in jtag_vpi driver

Multiple smaller items addressed in jtag_vpi:

- Several log prints adjusted to make them more clear to the user.

- Ensured that command handlers return ERROR_COMMAND_SYNTAX_ERROR
  on incorrect number of arguments.

- Fix in "jtag_vpi set_address": Leave the previously set address
  intact on error. Do not revert it to default.

- Minor update of help messages for the TCL commands.

- Updated macro names: SERVER_ADDRESS --> DEFAULT_SERVER_ADDRESS,
  the same for SERVER_PORT

Change-Id: Ibe386403a179adab5edb69c77fa408aef55701bd
Signed-off-by: Jan Matyas <matyas@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6845
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl: don't use 'set' to retrieve the value of a variable

Original TLC syntax uses 'set varname' to retrieve the value of
variable 'varname'. Such archaic syntax is still valid, but the
shorter '$varname' makes the code easier to read.

Replace 'set varname' with '$varname'.
While there, remove some useless curly brackets.

Change-Id: I27310e8c05afe56ea8bd0e41d4ae2c34447b725c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6863
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/board: Add AM625 EVM basic support

Add basic connection details with am625 EVM/SK

For further details, see https://www.ti.com/lit/zip/sprr448

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: Ibd23203ea98e34d03d2f55dac3565aa15aad744b
Reviewed-on: https://review.openocd.org/c/openocd/+/6799
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Add AM625 SoC

Add support for the latest in TI k3 family AM625 SoC.

For further details, see https://www.ti.com/lit/pdf/spruiv7

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: Ia54d0eab1c30a973afb1c2c61f4c5a72d29d9b78
Reviewed-on: https://review.openocd.org/c/openocd/+/6798
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/board: Add J721s2 EVM basic support

Add basic connection details with J721s2 EVM.

For further details, see https://www.ti.com/lit/zip/sprr439

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I68f8818c492ea6e07c14f2da305671c26da801cb
Reviewed-on: https://review.openocd.org/c/openocd/+/6797
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Add J721S2 SoC

Add support for the latest in TI k3 family J721S2 SoC.

For further details, see http://www.ti.com/lit/pdf/spruj28

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I608ab4513ffb6b5c4166ba423e7d0dddbbb3bbfd
Reviewed-on: https://review.openocd.org/c/openocd/+/6796
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Add a gdb-attach event hook for armv8 and simplify startup function

Since we can detect the type of target as well, reuse the _cpu_no_smp_up
function name and use the target name to simplify the _up function and
maintain consistency with what we introduced for r5.

Lets introduce gdb-attach event in a much cleaner fashion.

NOTE: we add a halt 1000 to retain the default gdb-attach hook behavior

While at it, fix a minor type of s/are/as in "Set Default target are
core 0" and simplify the foreach usage.

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I3259b7c3ae4c71b06d921edfaefe17c03bb673dc
Reviewed-on: https://review.openocd.org/c/openocd/+/6616
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Add a gdb-attach event hook for r5 and simplify startup function

Since we can detect the type of target as well, make the attach
function name generic for the follow on cleanup patch on armv8 to use
as well.

Lets introduce gdb-attach event in a much cleaner fashion. We can
introduce a simpler r5_up function since we now have more descriptive
core names making the individual descriptive procs redundant.

NOTE: we add a halt 1000 to retain the default gdb-attach hook behavior

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I31506bb2b86e63638082640eb72aa7c4c9575e93
Reviewed-on: https://review.openocd.org/c/openocd/+/6617
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Rename R5 targets to be more descriptive

R5 targets are currently named r5.0..n and the only way for user to
determine the actual type is external documentation. Lets just rename
the target names to make them descriptive to not require external
documentation for finding which R5 to connect to.

NOTE: we leave the _mcu_r5_cores _main0_r5_cores _main1_r5_cores alone
for now to allow existing startup proc functions to work, but we will
drop it in the follow on patch.

Previously:
Info : starting gdb server for j721e.cpu.r5.0 on 3336
Info : Listening on port 3336 for gdb connections
Info : starting gdb server for j721e.cpu.r5.1 on 3337
Info : Listening on port 3337 for gdb connections
Info : starting gdb server for j721e.cpu.r5.2 on 3338
Info : Listening on port 3338 for gdb connections
Info : starting gdb server for j721e.cpu.r5.3 on 3339
Info : Listening on port 3339 for gdb connections
Info : starting gdb server for j721e.cpu.r5.4 on 3340
Info : Listening on port 3340 for gdb connections
Info : starting gdb server for j721e.cpu.r5.5 on 3341
Info : Listening on port 3341 for gdb connections

With this patch:
Info : starting gdb server for j721e.cpu.mcu_r5.0 on 3336
Info : Listening on port 3336 for gdb connections
Info : starting gdb server for j721e.cpu.mcu_r5.1 on 3337
Info : Listening on port 3337 for gdb connections
Info : starting gdb server for j721e.cpu.main0_r5.0 on 3338
Info : Listening on port 3338 for gdb connections
Info : starting gdb server for j721e.cpu.main0_r5.1 on 3339
Info : Listening on port 3339 for gdb connections
Info : starting gdb server for j721e.cpu.main1_r5.0 on 3340
Info : Listening on port 3340 for gdb connections
Info : starting gdb server for j721e.cpu.main1_r5.1 on 3341
Info : Listening on port 3341 for gdb connections

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I2989efe3ae3e16754f98fa1dc9363ec4c898f7c3
Reviewed-on: https://review.openocd.org/c/openocd/+/6627
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Rename m4 target as general purpose mcu

The MCU is present on few of the SoCs and is meant as General Purpose
(GP) MCU of the system. Lets rename it to make clear what we are
debugging - esp when multiple MCUs are present in the system.

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I16132d321daf6e9b1d893fe6f92026d5aa9eb152
Reviewed-on: https://review.openocd.org/c/openocd/+/6619
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Rename m3 target as sysctrl

The M3 is the system controller of the system. Lets rename it to make
clear what we are debugging - esp when multiple MCUs are present in the
system.

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I4cd03b6068b8ce140fd254f9dd88151c4c7006d7
Reviewed-on: https://review.openocd.org/c/openocd/+/6618
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Add a gdb-attach event hook for m3 and m4

Add gdb-attach event to call the "up" function of m3 and m4 allowing for
more seamless integration with gdb for end users. We still retain _up
functions for non-gdb functionality.

NOTE: we add a halt 1000 to retain the default gdb-attach hook behavior

Suggested-by: Antonio Borneo <borneo.antonio@gmail.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I2e51fdbd8756f156551e589c748c3a338afa655c
Reviewed-on: https://review.openocd.org/c/openocd/+/6615
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/target/ti_k3: Remove args from m3 and m4_up

args serve no purpose, so drop them.

Signed-off-by: Nishanth Menon <nm@ti.com>
Change-Id: I136394307016453d576cf524b0f02227ba26ef8a
Reviewed-on: https://review.openocd.org/c/openocd/+/6626
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

cortex_m: use LOG_TARGET_XXX

Change-Id: I8be0f67442644031e6e8df3090d81af195caf82b
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6672
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

tcl/stm32l5x|u5x: refactor common tcl code

both stm32l5x and stm32u5x configs are almost identical except
clock config.

while at there rename target procs to avoid issues with JTAG daisy
chaining.

Change-Id: Ibbb1dfeb91a7f8d5d45744cf57dca2877f60e0c5
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6596
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

flash/stm32f1x,f2x: fix endianess in slow fallback flash write

Use target_write_memory() instead of target_write_u16()

Change-Id: I2389fe7a5fa18c9bc9c1aad8b8ddd64608bf2566
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6705
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/arm_tpiu: Fix usage of 'tpiu create'

Change-Id: I1ffad65a9e6d76f4d7fbbe249d8af3beb7e7692b
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6860
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/arm_tpiu: Fix 'tpiu create' parameter check

The current implementation crashes when executing 'tpiu create'
without an object name due to an invalid memory access. Pass 'argv'
instead 'goi.argv' to fix the problem.

While at it, match the style of the error message to the style used for
other Tcl commands. Especially, make the 'name' parameter mandatory.

Change-Id: Ib2b233f8556934af61608ae93d6405585c2c40b7
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6329
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins

target/arm_tpiu: Make error message easier to understand

Change-Id: Idddc31e34a67641c32d041c89d01fe2126ec5ddb
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/6328
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/tcl: Add get_reg function

Change-Id: Id1be9554d1df2c07cec3161a0fd3a586fdf18246
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/5312
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/tcl: Add set_reg function

Change-Id: I97a01b93046cb7af289792489f77f5580312585a
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/5313
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/stm32l4x: fix maybe-uninitialized compiler error

using gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0 we get:
error: ‘retval’ may be used uninitialized in this function

fixes: 13cd75b6ecfd (flash/nor/stm32xx: fix segfault accessing Cortex-M part number)
Change-Id: I897c40c5d2233f50a5385d251ebfa536023e5cf7
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6861
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

gdb_server: check target examined while combining reg list

Commit 6541233aa78d ("Combine register lists of smp targets.")
assumes that all the targets in the SMP cluster are already
examined and unconditionally call target_get_gdb_reg_list_noread()
that will in turn return error if the target is not examined yet.

Skip targets not examined yet.
Add an additional check in case the register list cannot be built,
e.g. because no target in the SMP cluster is examined. This should
never happen, but it's better to play safe.

Change-Id: I8609815c3d5144790fb05a870cb0c931540aef8a
Fixes: 6541233aa78d ("Combine register lists of smp targets.")
Reported-by: Michele Bisogno <michele.bisogno.ct@renesas.com>
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6853
Tested-by: jenkins
Reviewed-by: Michele Bisogno <michele.bisogno.ct@renesas.com>
Reviewed-by: Tim Newsome <tim@sifive.com>

gdb_server: fix double free

Commit 6541233aa78d ("Combine register lists of smp targets.")
unconditionally assigns the output pointers of the function
smp_reg_list_noread(), even if the function fails and returns
error.
This causes a double free from the caller, that has assigned NULL
to the pointers to simplify the error handling.

Use local variables in smp_reg_list_noread() and assign the output
pointers only on success.

Change-Id: Ic0fd2f26520566cf322f0190780e15637c01cfae
Fixes: 6541233aa78d ("Combine register lists of smp targets.")
Reported-by: Michele Bisogno <michele.bisogno.ct@renesas.com>
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6852
Tested-by: jenkins
Reviewed-by: Michele Bisogno <michele.bisogno.ct@renesas.com>
Reviewed-by: Tim Newsome <tim@sifive.com>

board: Add NXP LS1088ARDB

This adds a board file for the NXP LS1088ARDB. This only covers the
"primary" JTAG header J55, and not the PCIe header (J91). The only
oddity is that the LS1088A and CPLD are muxed by adding/removing a
jumper from J48. Unfortunately, it doesn't look like OpenOCD supports
this CPLD beyond determining the irlen, so it's not very useful. Those
who are interested in experimenting can define CWTAP to access the CPLD,
but the default is to access the CPU.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Change-Id: Ia07436a534f86bd907aa5fe2a78a326a27855a24
Reviewed-on: https://review.openocd.org/c/openocd/+/6849
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target: ls1088a: Add service processor

Normally the service processor is not necessary for debugging. However,
if you are using the hard-coded RCW or your boot source is otherwise
corrupt, then the general purpose processors will never be released from
hold-off. This will cause GDB to become confused if it tries to attach,
since they will appear to be running arm32 processors. To deal with
this, we can release the CPUs manually with the BRRL register. This
register cannot be written to from the axi target, so we need to do it
from the service processor target. This involves halting the service
processor, modifying the register, and then resuming it again. We try
and determine what state the service processor was in to avoid resuming
it if it was already halted.

The reset vector for the general purpose processors is determined by the
boot logation pointer registers in the device configuration unit.
Normally these are set using pre-boot initialization commands, but if
they are not set then they default to 0. This will cause the CPU to
almost immediately hit an illegal instruction. This is fine because we
will almost certainly want to attach to the processor and load a program
anyway.

I considered adding this as an event handler for either gdb-attach or
reset-init. However, this command shouldn't be necessary most of the
time, and so I don't think we should run it automatically.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Change-Id: I1b725292d8a11274d03af5313dc83678e10e944c
Reviewed-on: https://review.openocd.org/c/openocd/+/6850
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target: Add support for ls1088a

The LS1088A is an octo-core aarch64 processor from NXP in the layerscape
family. The JTAG is undocumented, but I was able to figure things out
from the output of `dap info`. This is the first in-tree example of
using the hwthread rtos (as far as I know), so hopefully it can serve as
an example to other developers. There are some ETMs, but I was unable to
try them out because I got 'invalid command name "etm"' when trying to
test things out.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Change-Id: I9b0791d27d8c41170a413a8d86431107a85feba2
Reviewed-on: https://review.openocd.org/c/openocd/+/6848
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

cpld: altera-epm240: Increase adapter speed

According to the datasheet, the minimum clock period with Vccio1 = 1.5V
(the lowest voltage supported) is 143ns, or around 6MHz. Set the default
adapter speed to 5 MHz.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Change-Id: I21cad33fa7f1e25e81f43b5d2214d1fa4ec924de
Reviewed-on: https://review.openocd.org/c/openocd/+/6847
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

cpld: altera-epm240: Add additional IDCODEs

This adds some additional IDCODEs from the datasheet. It also adds
support for customizing the tap name.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Change-Id: I7cda10b92c229b61836c12cd9ca410de358ede2e
Reviewed-on: https://review.openocd.org/c/openocd/+/6846
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/nor/stm32xx: fix segfault accessing Cortex-M part number

Some of STM32 flash drivers read Cortex-M part number from
cortex_m->core_info.
In corner cases the core_info pointer was observed uninitialised
even if target_was_examined() returned true. See also [1]

Use the new and safe helper to get Cortex-M part number.

While on it switch also target_to_cm()/target_to_armv7m() to the safe
versions. This prevents a crash when the flash bank is misconfigured
with non-Cortex-M target.

Add missing checks for target_was_examined() to flash probes.

[1] 6545: fix crash in case cortex_m->core_info is not set
    https://review.openocd.org/c/openocd/+/6545

Change-Id: If2471af74ebfe22f14442f48ae109b2e1bb5fa3b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Fixes: f5898bd93ff8 (flash/stm32fxx.c: do not read CPUID as this info is stored in cortex_m_common)
Reviewed-on: https://review.openocd.org/c/openocd/+/6752
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

target/cortex_m: add Cortex-M part number getter

The getter checks the magic numbers in arch_info to detect eventual
type mismatch.

Change-Id: I61134b05310a97ae9831517d0516c7b4240d35a5
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6751
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

target/armv7m,cortex_m: introduce checked arch_info cast routines

target_to_armv7m() and target_to_cm() do not match the magic number
so they are not suitable for use outside of target driver code.

Add checked versions of pointer getters. Match the magic number
to ensure the returned value points to struct of the correct type.

Change-Id: If90ef7e969ef04f0f2103e0da29dcbe8e1ac1c0d
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6750
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/cortex_m: fix target_to_cm() helper

The third parameter of container_of() should point to the same member
as target->arch_info points to, struct arm.

It worked just because struct arm is the first member in
struct armv7m_common.
If you move arm member from the first place, OpenOCD fails heavily.

Change-Id: I0c0a5221490945563e17a0a34d99a603f1d6c2ff
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: https://review.openocd.org/c/openocd/+/6749
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

1024 SHA256:YKx8b7u5ZWdcbp7/4AeXNaqElP49m6QrwfXaqQGJAOk gerrit-code-review@openocd.zylin.com (DSA)
384 SHA256:jHIbSQa4REvwCFG4cq5LBlBLxmxSqelQPem/EXIrxjk gerrit-code-review@openocd.org (ECDSA)
521 SHA256:UAOPYkU9Fjtcao0Ul/Rrlnj/OsQvt+pgdYSZ4jOYdgs gerrit-code-review@openocd.org (ECDSA)
256 SHA256:A13M5QlnozFOvTllybRZH6vm7iSt0XLxbA48yfc2yfY gerrit-code-review@openocd.org (ECDSA)
256 SHA256:spYMBqEYoAOtK7yZBrcwE8ZpYt6b68Cfh9yEVetvbXg gerrit-code-review@openocd.org (ED25519)
+--[ED25519 256]--+
|=..              |
|+o..   .         |
|*.o   . .        |
|+B . . .         |
|Bo. = o S        |
|Oo.+ + =         |
|oB=.* = . o      |
| =+=.+   + E     |
|. .=o   . o      |
+----[SHA256]-----+
2048 SHA256:0Onrb7/PHjpo6iVZ7xQX2riKN83FJ3KGU0TvI0TaFG4 gerrit-code-review@openocd.zylin.com (RSA)