target algo: do not write reg_param if direction is PARAM_IN

Without this change xxx_start_algorithm() writes all register
parameters no matter of their direction. It usually results
in writing of uninitialized reg_params[].value - possibly
reported by valgrind.

While on it fix the wrong parameter direction in
kinetis_disable_wdog_algo(). This bug did not have any
impact because of unconditional write of reg_params.

Change-Id: Ia9c6a7b37f77d5eb6e5f5463012dddd50471742b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4813
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

drivers/imx_gpio: fix polarity of srst and trst

The comment above the function is correct, but the code set
the GPIO with wrong (reversed) polarity.

Change-Id: Ifd09688150d3d2018af73521e0da3926bb1b7f84
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4847
Tested-by: jenkins
Reviewed-by: Grzegorz Kostka <kostka.grzegorz@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag: drivers: provide initial support for usb path filtering

With this patch drivers will be able to use usb path filtering.
The path format is identical to the format provided by linux kernel:
bus-port.port....
With this format it should be easier just to copy and paste
path found in dmesg.

Change-Id: I8bafa6fcb7a66ff68cc961a376f97f4f3dee35aa
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-on: http://openocd.zylin.com/4580
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

flash/nor: flash driver and cfg for SAM E54, E53, E51 and D51

The new Microchip (former Atmel) series powered by Cortex-M4 looks
very similar to older M0+ powered SAM D2x at the first sight.
Unfortunately the new series differs a lot in important details.
NVMCTRL has different register addresses, moved important bits
and even changed binary command set. An universal driver for all SAM D/E
would be very complicated. That's why a new driver was derived.

Tested on Microchip SAM E54 Xplained Pro kit (board cfg included).

Adjusted for the restructured dap support.
Checked by valgrind and clang static analyzer.

Change-Id: I26c67047a552076f4b207b9b89285a53d69b4ca4
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4272
Tested-by: jenkins
Reviewed-by: Andres Vahter <andres.vahter@gmail.com>

flash/nor/stm32f1x: use address instead of offset in stm32x_write_block()

stm32x_write_options() uses stm32x_write_block() at STM32_OB_RDP address.
The change eliminates subtracting bank->base (and later adding it back).

Change-Id: Icdd24afc7178b814be58e6033d4b93fdfb2d2a16
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4859
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Christopher Head <chead@zaber.com>

flash/nor/stm32f1x: fix minor error messages

Change-Id: I1e9e62979c4629c8ba1d5ae89ca7392259969eb6
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4858
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>

cortex_m: fix stepping on FPB rev 1

Stepping in the maskisr auto mode sets breakpoint to step over interrupt
service tasks. If the device has FPB rev 1, setting hard breakpoint
is impossible on address over 0x1fffffff.

Use soft type breakpoint for adresses over 0x1fffffff if FPB is rev 1.
This may eventually fail if the code memory is not writeable, but there
is nothing to do in such case.

Change-Id: Ibdeeb506903a35d550b64f82c24c37a668de62b3
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4857
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/nor/at91samd: add samr34j18.

samr34/r35 combine SAML21 and SX1276 (lora transceiver). This one was found on
xplaned pro evaluation kit. Ids for other r34/r35 chips are apparently not yet
documented.

Change-Id: I4054dd56ea53c9bae8d17abd5a3e4e65e1b9c8b1
Signed-off-by: Guillaume Revaillot <g.revaillot@gmail.com>
Reviewed-on: http://openocd.zylin.com/4872
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

arm_adi_v5: rewrite dap_to_jtag and dap_to_swd

The functions dap_to_jtag() and dap_to_swd() have been introduced by
3ef9beb52cd0 ("ADIv5 DAP ops switching to JTAG or SWD modes") in
arm_adi_v5.c by using the JTAG queue only.
Later, in 6f8b8593d63b ("ADIv5 transport support moves to separate
files") the functions has been moved in adi_v5_swd.c and adi_v5_jtag.c
but keeping the dependency from JTAG queue.
The functions does not work if the current transport is not JTAG.

Move back the functions in arm_adi_v5.c, replace the input parameter
"target" with "dap", use the transport to detect if the JTAG queue is
present, in case of SWD transport use the proper method, for other
transports report error.
Reuse the ADI v5 sequences already present in jtag/swd.h.
Also, OpenOCD does not support switching to another transport after
the initial selection, so do not change DAP's ops vector.

Change-Id: Ib681fbaa60cb342f732bc831eb92de25afa4e4db
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4852
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

aarch64: support for aarch32 ARM_MODE_SYS

Treat ARM_MODE_SYS like all other Aarch32 processor modes,
except for the special case of missing SPSR.

Change-Id: I60b21703659b264f552884cdc0f85fd45f7836de
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4494
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target: do not allow 'target create' after init

A target created after init lacks target_init_one() call
and is not added to gdb targets.

Steps to reproduce:
- start OpenOCD with a dap target
- connect by telnet
 target create ap0.mem mem_ap -dap $_CHIPNAME.dap -ap-num 0
 reset

Segmentation fault is rised because target->check_reset is NULL.

Change-Id: I2a62f3b450e4db3005c7041a22fb8f952e68c3b6
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4842
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

command: check command mode for native jim commands

The command mode was checked only for simple type of commands.
Native commands (handled by jim_handler) was treated as
they had mode COMMAND_ANY

Change-Id: Iab1d8cbb0b8c6f6b9f3cf942600432dec9a703ff
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4841
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/cortex_m: inform if an external reset occurs

Change-Id: I873e73012c44aac7af3b21b633bd096d8e299d07
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4840
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

command: Log the failed command by full name

Commit 44009186cfabe77fb260af221ebd6272d1e78f44 added logging
of failed cmd name but it used c->name only. It might be confusing:

Debug: 244 105 command.c:644 run_command(): Command 'init' failed with error
 code -4
User : 245 106 command.c:711 command_run_line():
Debug: 246 107 command.c:644 run_command(): Command 'init' failed with error
 code -4

The command on line 244 is 'dap init'
Use full name of cmd including parents.

Change-Id: Iff131ce6454ef70b353ce1bc6d0a480b92820545
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4837
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Reviewed-by: Jean-Christian de Rivaz <jcamdr70@gmail.com>

hla_target: fix adapter_poll() to preserve TARGET_DEBUG_RUNNING state

Without this change TARGET_DEBUG_RUNNING changes to TARGET_RUNNING
after adapter_poll()

Change-Id: I1c965a43527b50fa723d78fb6eae56585a7ede03
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4820
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

command: initialize the command mode for every command

All the commands in OpenOCD have been inspected and have the
command mode initialize, apart for two of them.
This is not critical, because the uninitialized value (0) is
equivalent to the enum COMMAND_EXEC, that is also the correct
value for the two mentioned commands.

To keep the code consistent, initialize the command mode to
COMMAND_EXEC.

Change-Id: Iaf043364cbd1005418d787ed045a3ec653612382
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4861
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl: Support for Analog Devices ADSP-SC58x / ADSP-SC584-EZBRD

The original script was broken by changes to the Cortex-A code.  The
recent introduction of the mem_ap target provided a new mechanism to
allow the script to be fixed.  This also adds an example board script
for the ADSP-SC584-EZBRD.

Change-Id: I36bc1ac6b6c036539f4175f1e65223ba10a35355
Signed-off-by: Peter Lawrence <majbthrd@gmail.com>
Reviewed-on: http://openocd.zylin.com/4855
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag: cmsis-dap: use macro SWJ_PIN_SRST in place of magic value

While connecting under reset, use the already defined macro
SWJ_PIN_SRST to assert the srst pin.

Change-Id: Icebed462c0fe8f8c15f6522dc56625aa580b8858
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4846
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

rtos: Add RTOS task awareness for Chromium-EC

Add RTOS task awareness for Chromium-EC. Currently
only supports ARM Cortex-M0/M3/M4 based targets.

No new Clang Analyzer warnings.

Change-Id: Iea56fcb1be220e2437613922879b63d6e553703d
Signed-off-by: Moritz Fischer <moritz.fischer@ettus.com>
Reviewed-on: http://openocd.zylin.com/4685
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

Correct ZynqMP configuration to be appropriately named

The xilinx_ultrascale.cfg target is actually the configuration for a
ZynqMP, which is a combination of an UltraScale+ FPGA core and a quad
core A53. Update the filename/comments to reflect this, and include
the tap IDs for all known FPGA cores for this part.

Change-Id: I70dfcc99861a482b83b6a795e83021d9cf1fe047
Signed-off-by: Jonathan McDowell <noodles@earth.li>
Reviewed-on: http://openocd.zylin.com/4850
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

arm_opcode: fix encoding of ARMv5 breakpoint instruction

The encoding of BKPT instruction is 0xE12###7#, where the four '#'
characters should be replaced by the 16 bits immediate value.
The macro uses an incorrect shift amount, thus the immediate value
is not properly coded and bits 20~23 of the opcode could get
corrupted.

Fixed by using the proper shift amount.

Change-Id: I32db8224ab57aad6d3b002f92f9f259056593675
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4854
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

arm_adi_v5: fix and update sequences to spec IHI 0031E

Fix the SWD line reset sequence accordingly to Arm specification IHI
0031E that requires at least 2 idle clocks after the 50 clocks with
SWDIO high.
Fix the value of the activation code in the (currently unused)
sequence dormant-to-SWD.
Make each sequence's length multiple of 8, so it is compatible with
adapters that have such limitation (e.g. buspirate) and try to split
and comment each part of the sequence (when possible keep each part
byte aligned, inspired from commit 3ef9beb52cd0). This slightly
increases the sequence length but does not impact run-time
performance because these are rarely used sequences.
Add the missing sequence dormant-to-JTAG and JTAG-to-dormant, not
used yet.

On devices that implements the dormant state, IHI 0031E deprecates
the direct switching between SWD and JTAG, and recommends using a
transition through dormant. This is not implemented.

Change-Id: Iad18c0e736cfd9366be175d22658d664b0c61eab
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4851
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/arm_dap: fix segmentation fault in 'dap info' cmd

'dap info' command fails hard on a hla target.

Change-Id: Ia188b1afe527e0ed64512d1bddadd507f978e40b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4860
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

nrf5: add nrf5 device definition for HWID 0x00E3

This hardware id is e.g. used by the Insight SiP ISP1507-AX.

Change-Id: I82568d292f9882372ab061d8e3e36906b0cc5882
Signed-off-by: Mirko Vogt <mirko.vogt@sensorberg.com>
Reviewed-on: http://openocd.zylin.com/4845
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

cortex_m: fix bug in poll() machine state (external resume awareness)

This patch covers the fact that cortex_m could be resumed externally by
Cross Trigger Interface or by direct write to DHSCR ...

To reproduce:
 - halt the target
 - then run the core through DHCSR (mww 0xe000edf0 0xa05f0001)
 => this resumes the core, but target state in OpenOCD remains HALTED.

Change-Id: Ifa1ae18645bfeb863acc78a039bbf04873fd78fe
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/4817
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

cortex_a_poll: minor code factorization to enhance readability

cortex_a_debug_entry and update_halt_gdb are called in two consecutive
conditions which are complementary, so externalizing the common code
makes the conditions' body lighter

With the removal of LOG_DEBUG(" ") since it does not look too informative

Change-Id: I0c54e413619576bb3af164f2dcf256c5a862c5fd
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/4832
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

at91samd: added SAM-DA1 parts

Change-Id: I1b206461052ee1897432e379505fdf95372a9e24
Signed-off-by: Severin Junker <s.junker@cartelsol.com>
Reviewed-on: http://openocd.zylin.com/4835
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

helper/command: check for malloc failure in __command_name

If malloc fails in __command_name, the following strcpy will
segfault, thus preventing __command_name to return.
The actual calls to command_name() implement the correct check
for the NULL pointer, but propagate error -ENOMEM, that is not
an error value coherent within OpenOCD. Plus, in one case it
overwrites an already detected error.

Check the pointer returned by malloc and, in case of failure,
issue an error message and return the NULL pointer.
Let the caller of command_name() to keep the already detected
error or to return ERROR_FAIL in case of end of memory.

Change-Id: I151a24569409777dd5bc09a3daf5dba2b8e2829b
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4838
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/stm8: add missing destroy_reg_param()

Change-Id: Ibd8a423a4400226790cfbb9a6f113b7ea762c436
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4814
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Ake Rehnman <ake.rehnman@gmail.com>

drivers/sysfsgpio: fix usage messages

The notation "(tck tms tdi tdo)* " is incorrect, because it means the
quadruple of gpio can be repeated on the command-line.
The correct syntax of the command requires instead to provide either
all the four gpio numbers (in order to set the values) or to pass an
empty command-line (to dump the values previously set).

Change the .usage field to "[tck tms tdi tdo]".
Change similarly the corresponding .usage field for SWD command.
Add the .usage field for the commands that individually set each gpio.

Change-Id: If5b3c618097b71dfe7fcf988fb3c1499ae03a6d5
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4833
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Add flash support for SiFive's Freedom E platforms

Valgrind and Clang Static Analyzer have no complaints about this change.

Change-Id: I7757615ec52448372bdc57729cdf97c7016d97e8
Signed-off-by: Tim Newsome <tim@sifive.com>
Reviewed-on: http://openocd.zylin.com/4656
Tested-by: jenkins
Reviewed-by: Philipp Guehring <pg@futureware.at>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

stlink_usb: Submit multiple USB URBs at once to improve performance

Commands to stlink devices are typically comprised of multiple
transactions with each transaction completing before moving to the next.
This change allows for multiple USB transactions to be issued at once
followed by a check that all transactions completed successfully.  This
improves performance on some machines where there is a large turn-around
time between USB transfers such as is seen on some virtual machines.

This change is only supported when compiled with libusb1 as libusb1
supports and asynchronous interface.

Multi-transaction queueing introduced in this change paves the way for
improving speed of other transactions in the future such as memory and
register reads where multiple USB transactions in succession are
required to complete a command.  Multiple USB transactions can be
submitted at once using jtag_libusb_bulk_transfer_n function.

Change-Id: I924e049217a789ef445b14e00aa1983576970fbf
Signed-off-by: Austin Phillips <austin_phillips@hotmail.com>
Reviewed-on: http://openocd.zylin.com/4484
Tested-by: jenkins
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

libusb: add debug message on adapter not found due to wrong serial

When few adapters of the same type are in use, the serial string is
the way to select the right one.
Currently a serial string that does not match any of the connected
adapters will just fail the open, without specific information to
track the issue.

Add a specific message to highlight that the open failure is caused
by a serial mismatch.

Change-Id: I5cb77f1045cc746e532d395b2e5ced40a23ab638
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4701
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

jtag/drivers/usb_common: return oocd error values

Where possible, keep the same style for returning error.

Change-Id: I3a04220c0b9f129a36e9fe83038b7c19dd57fe61
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4699
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

libusb0: return oocd error values

Commit d1b74376336814266236054f925a9964b87dd8a5 fixes libusb1 to
return OpenOCD error values instead of negative errors in Linux
kernel's style.
The same fix should be applied to libusb0 too.

Fix return value of libusb0 to uniform it to OpenOCD style.

Change-Id: I68478c29c91c6be720074f58c432fe51477e03ed
Fixes: d1b743763368 ("libusb: return oocd error values")
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4698
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target/aarch64: add support for multi-architecture gdb

GDB can be built for multi-architecture through the command
./configure --enable-targets=all && make
Such multi-architecture GDB requires the target's architecture to
be selected either manually by the user through the GDB command
"set architecture" or automatically by the target description sent
by the remote target (i.e. OpenOCD).

Commit e65acd889c61a424c7bd72fdee5d6a3aee1d8504 ("gdb_server: add
support for architecture element") already provides the required
infrastructure to support multi-architecture gdb.

aarch64-linux-gnu-gdb 8.2 uses "aarch64" as default architecture,
but also supports the value "aarch64:ilp32" and all the values
supported by arm-none-eabi-gdb.
These values can be displayed on arm gdb prompt by typing
"set architecture " followed by a TAB for autocompletion.

Set the gdb architecture value for aarch64 target to "aarch64".

Change-Id: I63e9769f47d8e73f048eb84fa73e082dd1c8e52c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4755
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target/arm: add support for multi-architecture gdb

GDB can be built for multi-architecture through the command
./configure --enable-targets=all && make
Such multi-architecture GDB requires the target's architecture to
be selected either manually by the user through the GDB command
"set architecture" or automatically by the target description sent
by the remote target (i.e. OpenOCD).

Commit e65acd889c61a424c7bd72fdee5d6a3aee1d8504 ("gdb_server: add
support for architecture element") already provides the required
infrastructure to support multi-architecture gdb.

arm-none-eabi-gdb 8.2 uses "arm" as default architecture, but also
supports the following values: "arm_any", "armv2", "armv2a",
"armv3", "armv3m", "armv4", "armv4t", "armv5", "armv5t", "armv5te",
"armv5tej", "armv6", "armv6k", "armv6kz", "armv6-m", "armv6s-m",
"armv6t2", "armv7", "armv7e-m", "armv8-a", "armv8-m.base",
"armv8-m.main", "armv8-r", "ep9312", "iwmmxt", "iwmmxt2", "xscale".
These values can be displayed on arm gdb prompt by typing
"set architecture " followed by a TAB for autocompletion.

Set the gdb architecture value for all arm targets to "arm".

Change-Id: I176cb89878606e1febd546ce26543b3e7849500a
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4754
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target/stm8: add support for multi-architecture gdb

GDB can be built for multi-architecture through the command
./configure --enable-targets=all && make
Such multi-architecture GDB requires the target's architecture to
be selected either manually by the user through the GDB command
"set architecture" or automatically by the target description sent
by the remote target (i.e. OpenOCD).

Commit e65acd889c61a424c7bd72fdee5d6a3aee1d8504 ("gdb_server: add
support for architecture element") already provides the required
infrastructure to support multi-architecture gdb.

The gdb patches for stm8 are still not merged in the official
repository and are temporarily hosted in
https://stm8-binutils-gdb.sourceforge.io/
The latest patch set
stm8-binutils-gdb-sources-2018-03-04.tar.gz
define only one possible value ("stm8") for this architecture; it
can be displayed typing "set architecture " followed by a TAB for
autocompletion in gdb for stm8.

Set the gdb architecture value for stm8 to "stm8".

Change-Id: I643ceba662de46cecf061d1dc672b9178a077f1b
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4753
Tested-by: jenkins
Reviewed-by: Ake Rehnman <ake.rehnman@gmail.com>
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

armv7m: always set xPSR.T=1 when starting an algorithm

xPSR.T sets the processor to Thumb mode when set to 1. ARMv7-M only
supports execution of Thumb instructions, so it must always be set to 1.

If xPSR.T is set to 0 on armv7m, a usage fault is generated when a
instruction execution is attempted.

On armv7m, issuing a reset causes the vector table to be examined. PC
and xPSR.T  are loaded from the vector table at byte offset 4. xPSR.T is
taken from the least significant bit this value, PC from the remaining
bits.  This occurs even with `reset halt`, as the reset itself causes
this load to occur without the execution of any instructions.

As a result of this, following a reset with a "bad" value programmed in
the vector table, openocd would be unable to run algorithms on the
target, as running them would immediately result in a usage fault due to
xPSR.T being unset (0).

Allow algorithms to run regardless of the content of the vector table by
explicitly setting xPSR so that xPSR.T=1 prior to executing an
algorithm. One can think of this as openocd more closely emulating a
reset or branch instruction in executing it's algorithms.

Ticket: https://sourceforge.net/p/openocd/tickets/203/
Signed-off-by: Cody P Schafer <openocd@codyps.com>
Change-Id: I4dc3427ab195d06c3fd780ea768027fefccc4c28
Reviewed-on: http://openocd.zylin.com/4658
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

icepick.cfg: add cancel reset bit to TAP register writes

The Agama family of devices (CC26x2/CC13x2) required an
additional bit to be set when adding the core's TAP into
the scan chain. The cancel reset bit 0x10000 tells the
ICEPick to take the bus out of reset so that the other
bits will take effect. This bit is a NOP on other devices
and ICEPicks, so the change shouldn't adversely affect
other devices.

Change-Id: I9245eef0936ea7eea28ae84ab5e8ce05fa63af40
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4789
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

drivers: xds110: Add support for XDS110 stand-alone probe

The XDS110 stand-alone version has the ability to supply
voltage to the target board via it's AUX FUNCTIONS port.
Added command to enable setting the voltage on the XDS110
stand-alone.

Change-Id: I2f21c4a3d15ed99e649f3a83973c5e724c4bfeb6
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4793
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

HACKING: add note about refs/for/master

The fact that one needs to always push contributions to a single
remote reference (refs/for/master) might seem odd to people unfamiliar
with Gerrit. GitHub, for instance, hosts personal repositories where
developers typically create topic branches for each contribution and
use a proprietary mecanism to request a review (the "pull request").
More generally, one normally does not expect to be able to push
non-fast-forwarding stuff to a remote branch.

This commit adds a clarifying note to the patch guidelines.

Change-Id: Ia750b815b82b18e92b6109c07f451000dcbecf9b
Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-on: http://openocd.zylin.com/4806
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

HACKING: replace refs/publish/master with refs/for/master

refs/publish/master is deprecated and gives a warning in newer Gerrit.
Replace with refs/for/master.

Change-Id: I56871cc6e80c014ba81f4458230cd67dc318ecb3
Suggested-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-on: http://openocd.zylin.com/4810
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

NOR: lpc2000 Add support for LPC84x devices

These devices differ from LPC8xx devices in that they have a different
IAP entry point, but everything else is the same.  Using Tcl to pass
different IAP entry point.
no new Clang analyser warnings and no new build sanitizers issues.

Change-Id: I2d654dd250f416e74262c0228cad8713a283402f
Signed-off-by: Rod Boyce <developer@teamboyce.co.uk>
Reviewed-on: http://openocd.zylin.com/4684
Reviewed-by: Jean-Christian de Rivaz <jcamdr70@gmail.com>
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash: nor: ath79: remove base calculation

Currently it is impossible to flash ELF with correct offsets. The reason
is a bogus offset calculation extracted from base.
Since any other spi drivers do not care about base, do the same for
ath79 as well.

Change-Id: I9e46e01c9e7a709c2d07da9203c634f302603afd
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Reviewed-on: http://openocd.zylin.com/4821
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/nrf5: set correct timeout for nvmc operations

Longest erase all FLASH for nRF5 series is 295.3 ms for nRF52832.
Timeout period now is set to 340 ms (295.3 + 15%)

Change-Id: Iae00ed7b634f111b9798db11e35e4e066d4aaa95
Signed-off-by: Jānis Skujenieks <janis.skujenieks@gmail.com>
Reviewed-on: http://openocd.zylin.com/4822
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/arm_cti : export CTI APPPULSE and INACK register

this permits the full control of CTI from config files

Change-Id: Ia27ac8e12e08ec72da05f26dcbd81d24fa1a0f6f
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/4815
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/arm_cti: add debug message when an incorrect CTI register name is used

the patch also contains some typo fixes

Change-Id: Ia4267036068455144cdcbfdffed15518d48f445e
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/4816
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

command: Log the failed command name

Change-Id: I03938a845110002755636a9514b17a213bf1cc72
Signed-off-by: Jean-Christian de Rivaz <jcamdr70@gmail.com>
Reviewed-on: http://openocd.zylin.com/4808
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

Add LPC8Nxx and NHS3xx support.

Change-Id: I0bdbca8dd9b234aca355230af7269463c9f70bd1
Signed-off-by: Jean-Christian de Rivaz <jcamdr70@gmail.com>
Reviewed-on: http://openocd.zylin.com/4515
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

add w600 support

w600 is a wifi soc from winner micro(www.winnermicro.com).

Change-Id: Ib8ccd6e52baefca6547fb97d29db75db0ee73948
Signed-off-by: Simon Qian <versaloon@simonqian.com>
Reviewed-on: http://openocd.zylin.com/4801
Tested-by: jenkins
Reviewed-by: yichen <wdyichen@wdyichen.cn>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/spi: add adesto AT25DF081A

8Mbit SPI flash on SAM D21 Xplained board

Change-Id: Iec087f5d889c1cbdd4fed90863e73511f6101cec
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4802
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target: allow moving and resizing working area on running target

target configure -work-area-xxx calls target_free_all_working_areas()
and sets the desired new parameter. Without this change the working area
does not get reallocated if it has been allocated before.
target_free_all_working_areas() results in work area containing one block
marked as free.

Completely free working area in target_free_all_working_areas()

Change-Id: I79c681082f32f2a96a2b40eb3b8751e427549693
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4797
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target: move all working_area functions to one block

The block of code moved without any changes

Change-Id: I70b82dc3315dcc3f34de0537b362bee230007d02
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4796
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target/cortex_m: fix cortex_m reset_config help and check for syntax error

Remove option 'srst' which is not recognized from on-line help and texi.
Check parameter and return syntax error if wrong option is entered.

Change-Id: I87daa423a9f53193a0b015080594820b933628f5
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4795
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target/cortex_m: do not use VECTRESET on Cortex-M0, M0+ and M1

Cortex-M0, M0+ and M1 do not support VECTRESET bit in AIRCR.
Without this change the 'reset' command silently fails if VECTRESET
is requested.

Detect these cores, show warning if VECTRESET is about to use
and use SYSRESETREQ instead.

Change-Id: Ief174373e3ef0e6b287c57911c0aca4dfa8209f2
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4794
Tested-by: jenkins
Reviewed-by: Jean-Christian de Rivaz <jcamdr70@gmail.com>
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target: armv8: Add TARGET_HALTED check for gdb connect

This commit adds TARGET_HALTED check in armv8_get_core_reg32()
and armv8_set_core_reg32() to void a crash issue when gdb connects
but fails to halt the ARM core. Similar logic can be found in
armv8_get_core_reg() and armv8_set_core_reg().

Below is the call stack information of this case when gdb connects.

(gdb) bt
    regnum=regnum@entry=0, dpm=0x990110) at src/target/armv8_dpm.c:657
    r=0x9c7240, regnum=0, mode=<optimized out>) at src/target/armv8_dpm.c:974
    at src/target/armv8.c:1487
    packet=0x8ec8e0 <gdb_packet_buffer.9962> "g",
    packet_size=<optimized out>, connection=<optimized out>)
    at src/server/gdb_server.c:1200
    at src/server/gdb_server.c:3180
    command_context=command_context@entry=0x935010)
    at src/server/server.c:566
...

Change-Id: I159837b533f110998184f910a0abe48409bd58f1
Signed-off-by: Liming Sun <lsun@mellanox.com>
Reviewed-on: http://openocd.zylin.com/4758
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/cortex_a: enable DSCR_HALT_DBG_MODE during examine

Arm architecture reference manual DDI0406C reports at page 2024 in
table C3-1 the processor behaviour on debug events depending on
the debug-mode (none, monitor or halt), mode selected through the
bits MDBGen and HDBGen in DSCR register.

The halt request is served independently from the debug-mode. Thus
it's useless to enable the halt debug-mode in cortex_a_halt() by
setting the bit HDBGen (macro DSCR_HALT_DBG_MODE).

On the other side, halting for a breakpoint, a watchpoint or a
vector catch requires being in halt debug-mode.
Today HDBGen is set only in cortex_a_halt(), so we are forced to
halt the core at least once before it can be halted for hitting a
breakpoint/watchpoint/vector-catch. This is annoying since there
is no need to halt the target to set a HW breakpoint.

Move in cortex_a_init_debug_access() the selection of the halt
debug-mode, so the mode is set during examine.
To prevent a misconfigured hardware breakpoint/watchpoint/vector
catch to halt the target when OpenOCD has already quit, return to
debug-mode none at OpenOCD exit.

Change-Id: I68a1c51de3572ca1b89e90caf7eb20374268e926
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4783
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

adi_v5_jtag: fix build break when open DEBUG_WAIT macro

DEBUG_WAIT is useful to debug adi_jtag issue,
and the WCR register is replaced by DLCR for DP registers update
in commit 150b7d26f213398d717bf46744811b48834a3744.

Change-Id: I3faa9ea8a6adacd3d5275e40382801da731db32f
Signed-off-by: YanLin Zhu <zhuyanlin@pinecone.net>
Reviewed-on: http://openocd.zylin.com/4804
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Christopher Head <chead@zaber.com>

helper/startup.tcl: fix execution stack frame of wrapped commands

The OpenOCD commands that have been wrapped with 'ocd_bouncer' are
executed within two levels of nested proc's:
# see register_command_handler() in src/helper/command.c
proc my_command {args} {eval ocd_bouncer my_command $args}

# see ocd_bouncer in src/helper/startup.tcl
proc ocd_bouncer {name args} {
... [eval ocd_my_command $args] ...
}
This causes the stack frame of 'ocd_my_command' to be the same one
of proc 'ocd_bouncer', thus two levels below the stack frame of the
caller of 'my_command'. This is an issue with commands that receive
a variable by name and have to resolve them to access the value.

E.g. the command
<target> mem2array arrayname bitwidth address count
is wrapped; it receives the name of the array but fails to resolve
it in the current stack frame. Instead, the commands
mem2array arrayname bitwidth address count
ocd_<target> mem2array arrayname bitwidth address count
are not wrapped and can directly access the array because they share
the same stack frame of the caller.
Same situation with the symmetric commands 'array2mem'.

How to test:
within a telnet connection, run the following set of commands,
eventually replacing the address 0x08000000 with a valid readable
address of your <target>,
unset -nocomplain v1 v2 v3
info vars v?
mem2array v1 32 0x08000000 1
<target> mem2array v2 32 0x08000000 1
ocd_<target> mem2array v3 32 0x08000000 1
info vars v?
and notice that only v1 and v3 are now allocated. The array v2 has
been allocated in the temporarily stack frame of proc ocd_bouncer,
together with its local variables, and then lost when proc ended.

Fixed by executing the wrapped commands with the command 'uplevel'
instead of 'eval'. The amount of levels to skip is checked to avoid
errors in the unusual case 'ocd_bouncer' is called directly without
the first level of wrapper.

Change-Id: Iff90fb8921faf9b5ab04f61062a530578cc20d78
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4731
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/core: Fix chunk size calculation in default_flash_mem_blank_check

Slow version of blank check procedure reads target memory sector-by-sector
using 1 KB chunks. Due to bug in chunk size calculation algorithm the actual
size of the chunk is always 1 KB even if sector size is smaller.
This causes out-of-boundary read of the last sector.

Steps to reproduce:
1) Use target with small sectors (e.g. psoc6 with 512-byte sectors)
2) set WORKAREASIZE_CM0 0
3) flash erase_check 1

Running slow fallback erase check - add working memory
Info : SWD DPIDR 0x6ba02477
Error: Failed to read memory at 0x14008000
unknown error when checking erase state of flash bank #1 at 0x14000000
Bank is erased

Change-Id: I03d0d5fb3a1950ae6aac425f5e24c7fd94b38325
Signed-off-by: Bohdan Tymkiv <bhdt@cypress.com>
Reviewed-on: http://openocd.zylin.com/4785
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/esirisc: refactor register write protection

This patch removes use of register write protection in protect() and
protect_check() now that Change 4765 has merged.

Change-Id: I42c429dc283c5b53989a6b98ebfc58214274ff16
Signed-off-by: Steven Stallion <stallion@squareup.com>
Reviewed-on: http://openocd.zylin.com/4791
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

esirisc: support eSi-Trace

This patch adds support for instruction tracing to eSi-RISC targets. The
command interface is borrowed heavily from ETM; eSi-Trace uses a less
sophisticated model for tracing, however the setup and usage is similar.
This patch also cleans up the command interfaces of the other esirisc
command groups and adds additional debugging information to log messages
when dealing with CSRs.

This patch "finalizes" support for 32-bit eSi-RISC targets.

Change-Id: Ia2a9de79a3c7c066240b5212721fb1b7584a9a45
Signed-off-by: Steven Stallion <stallion@squareup.com>
Reviewed-on: http://openocd.zylin.com/4780
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

fix cc32xx related changes

original CC3200 launchpad works only with ti-icdi driver
which stopped to work after merging to master commit
d02de3a8a92091b9761ebaf44dff1a71f5b2edcb

Change-Id: I247b5d99831fa744de1fdc5b8a7cffdf49fe953c
Signed-off-by: Jiri Kastner <cz172638@gmail.com>
Reviewed-on: http://openocd.zylin.com/4792
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

armv7a_mmu: s/LOG_ERROR/LOG_WARNING/ on address translation failure

When GDB analyses the status of the target it try to guess the
current stack frame and issues few memory read.
E.g. on ARM targets GDB uses R11 value as a potential frame-pointer
and reads at the address pointed by R11.
The address of such memory read is not always valid and can trigger
an address translation failure.

Replace LOG_ERROR with LOG_WARNING in case the virtual address does
not have a hit in TTB; print the virtual address in the warning
message and discriminate the two identical messages with [1]/[2].

Change-Id: I288b8cd26bec2543c4f1c16b7c06dc47d5d843d1
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4602
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

libusb0: add compatibility define for transfer type bulk

For compatibility with libusb1, define LIBUSB_TRANSFER_TYPE_BULK
in libusb0.
Remove the #ifdef HAVE_LIBUSB1 in jtag/driver/aice

This also fixes a compile error in jtag/drivers/openjtag with
libusb0.

Change-Id: I827b77eac10216759eb31aab461b2b63cabaf195
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4700
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

flash/nor: Add support for sector erase in stm32l4x.c for L4+ family

Updates support for L4+ device id: 0x470 added by #4310
Extends #4641 to account for L4+ use of multiple DBANK option bits
Enables L4+ 1M and 2M devices to be programmed using sector erase

Change-Id: I42bb379d7d97986f4506423e3da503d07c787c6b
Signed-off-by: bob <rea952@gmail.com>
Reviewed-on: http://openocd.zylin.com/4777
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor: update cc3220sf for issue found in code review

Remove inappropriate use of bank_number.

Change-Id: I11be1f2540cb09a3ccede35312f90bc8276af338
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4788
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor: update CC26xx/CC13xx support

Added fixes found in additional code reviews.

Remove inappropriate use of bank_number field and updated
documentation to reflect the change.

Restored functionality to cc2538.cfg file because previous
change removed the cc26xx.cfg file because the flash support
changes made it obsolete. Rolled the previous cc26xx.cfg
file into cc2538.cfg and updated it to work with other
recent changes.  Tested using a SmartRF06 Evaluation
board with embedded XDS100v3 and external XDs110.

Change-Id: Ia19d00cf8055c5c0f1acc53aa23fd06a80fd2ebc
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4787
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag: sysfsgpio: clean up swd gpios

All SWD GPIOs should be un-exported when no longer needed, not just
srst.

Change-Id: I998377afe43b72446cab3da2d4406fc2912ff8c3
Signed-off-by: Reto Schneider <code@reto-schneider.ch>
Reviewed-on: http://openocd.zylin.com/4784
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

helper/log: remove dead code in log_forward()

Commit f1be0e6af2e204805a8bc2e8cadf828b9fa46c98 ("HELPER/LOG:
review unused symbols") removed a set of unused functions and,
as consequence, there is no code anymore that sets the variable
"log_forward_count". But, the commit above did not removed the
code (now dead) that depends on "log_forward_count" set.

Remove the code dependant on "log_forward_count" set and the
variable itself.

Change-Id: I6efe93d1dccbe13c409c5bc55ba47a2684c0e3ac
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4779
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

jtag: cmsis-dap: developer tool - enter a command by hex nuber

Handy to test vendor commands in a CMSIS-DAP adapter.

Change-Id: Ieeaa276edf770b1a3076a186e9056b4e5180362a
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/3103
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

doc: add ARMv6-M info regarding TAP Declaration

Change-Id: I89095fd0c830fdc2ce40e5d23f8af98502b0ff50
Signed-off-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-on: http://openocd.zylin.com/4782
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

doc: fix some typo about STMicroelectronics name

s/ST/STMicroelectronics/
s/ST Micro/STMicroelectronics/
s/ST Microelectronics/STMicroelectronics/

Change-Id: I62aef331d172db22a70841a089c1889e37997cde
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4718
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add support for STLINK-V3

Extend the driver to include the minimal functionality to support
the HLA model.
Due to the small change in the name (ST-LINK/V2 => STLINK-V3), fix
the existing names in the comments in udev rules.

Change-Id: Ied33e38063a6da81d9bf249ed195444d7cdf4f03
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4717
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Tested-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: fix printed version for new STM8 nucleo board

The ST-LINK/V2.1 embedded in the new nucleo boards for STM8 does
not follow the normal versioning rules, and puts mass-storage and
swim version in the field normally used respectively for jtag and
mass-storage version.

Apply the check suggested by STMicroelectronics to discriminate
the two cases and print the correct version.

Change-Id: I0dd1da11013be3f1e56084489e28cfba98bb07af
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4716
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Tested-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: dump version in the same format of ST firmware upgrade tool

This should help comparing the versions before deciding for a
firmware update.

Change-Id: I7012a60587587539b2fd1a5f9c3b1ce72877793f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4715
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Tested-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add STLINK_F_HAS_GETLASTRWSTATUS2

Starting from stlink V2J15 the API
STLINK_DEBUG_APIV2_GETLASTRWSTATUS
is obsoleted and replaced by the new API
STLINK_DEBUG_APIV2_GETLASTRWSTATUS2.

Manage the two cases to be prepared for an eventual future
removal of the obsolete API.

Change-Id: Ic4498a652865b2eb6148825138c2f6855a06ba47
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4714
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: remove reset pulse when entering in JTAG

Until version J14 the behaviour of ST-Link was to send a reset
pulse to the target when the debug connection is started in JTAG
mode. No reset pulse is sent, instead, in SWD mode.
Version J15 introduces a new parameter to avoid the reset pulse in
JTAG mode, aligning the behaviour with SWD.

This reset from the ST-Link, if propagated to the target, prevents
attaching a running target.

Actually this reset pulse is very short (few microsecond) and can
be easily filtered out by an on-board capacitor, usually present
on the reset wire (mainly to filter the bounces of the reset
button). Moreover, most of the use cases for ST-Link are with SWD
(not with JTAG) and this has probably further masked this JTAG
specific behaviour.

OpenOCD can tolerate it but requires the flag "connect_assert_srst"
to the command "reset_config", but the flag is not present in any
configurations in folder tcl. This enforces the guess it was not
noticed due to on-board capacitors or missing connection of reset
pin or ST-Link only used in SWD; so it's safe applying this patch.

Change the default behaviour to avoid reset in JTAG at connection.
There is no need to manage the ST-Link version here, since every
parameter that is not recognized by older ST-Link is treated as
"connect in JTAG with reset pulse", keeping backward compatibility.

Change-Id: Idc97a1457279e3970fd0839cadbff22d9b0302d4
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4713
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: simplify api version handling

The API version is hardcoded in the STLINK device and depends on
the specific firmware version. There is no way to switch between
different API. Thus, it make no sense to keep in two separate
variables the current and the max API supported because they are
always equal to "the only" API supported by the specific STLINK.

Collapse h->version.jtag_api_max and h->jtag_api in a single
variable h->version.jtag_api
Remove an extra print of api version; this is already advertised
while printing the STLINK version.

Change-Id: I0519694a34748e5a0aa330932c1256080a6d5f20
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4712
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add STLINK_F_HAS_MEM_16BIT

Change-Id: I51a160a105342ddbab99c4182d89650a436f5945
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4711
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Tested-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add STLINK_F_HAS_JTAG_SET_FREQ

Change-Id: I591543f218c6bb30aaeb6321ba8a9425a2b5b6e2
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4710
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Tested-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add STLINK_F_HAS_SWD_SET_FREQ

Change-Id: Ibba786c7e18bf15a04b85a2071c79f631a252ccf
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4709
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Tested-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add STLINK_F_HAS_TARGET_VOLT

Use an alias to STLINK_F_HAS_TRACE, both are added in the same version

Change-Id: Iabf2fdac407f5c4737c3da942323d60ee50c3470
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4708
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Tested-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add STLINK_F_HAS_TRACE

Fix the condition for ST-Link/V1; it does not support trace.
Trace is added on ST-Link/V2 and ST-Link/V2.1 only.

Change-Id: I17416edd05a2c8ece605560b7533edac4f982cfc
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4707
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: simplify maintenance of version and features

The number of stlink firmware version is growing, each carrying new
features. Today's code has several check distributed here and there
and it's already hard to track them and verify the correctness.
The introduction of STLINK-V3 will make the situation much worst,
and the code much less readable.

Add a "flags" bitmask in the struct stlink_usb_version to allow
setting individual bits for each feature available or for specific
quirks and workarounds.
This patch does not implement setting nor testing "flags"; it would
be introduced in following patches, one bit at a time.

Change-Id: I09d78202646a6c8330731f8aa96dc9d295fa5655
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4706
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: check for SWD support

The old ST-Link/V1 can either run an obsolete firmware with API_V1 or
a more recent firmware (from V1J11 to last V1J13) with API_V2.
SWD is only provided by the latter API.

Return error is SWD is selected on adapters that does not support it.

Change-Id: Iac4ba54d191ba80fb445a7cd3d8c2621dc7a4846
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4705
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add support for 16 bit memory read/write

Recent stlink version V2J26 introduces support for 16 bit memory
read/write.
Add the new API and modify the wrappers stlink_usb_{read|write}_mem
For older version of stlink keep the same behavior as in current
code.
While there, fix some minor typo in comments inside the wrappers.
This fixes ticket #204 ("st-link can't flash stm32f2x based flash
without target algorithms").

Change-Id: Id7d404e588f10a4b0f8a93d7ca44cef8e4c49b4f
Author: Laurent LEMELE <laurent.lemele@st.com>
Rebased-by: Antonio Borneo <borneo.antonio@gmail.com>
Ticket: https://sourceforge.net/p/openocd/tickets/204/
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4704
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

stlink: add JTAG speed selection

ST-Link uses two separate API for setting the interface speed in SWD
and JTAG.
Add the missing API for JTAG and run the proper API depending on the
selected mode.

Change-Id: I71e955e23c1d82c2ea1c8bfade7612c5a0377a74
Author: Laurent LEMELE <laurent.lemele@st.com>
Rebased-by: Antonio Borneo <borneo.antonio@gmail.com>
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4703
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

stlink: add usb pid for v2.1 without mass storage device

New version of ST-Link/V2.1 without mass storage device.
From debug point of view, it is compatible with existing ST-Link/V2.1
It uses a new USB PID because the USB endpoints and interfaces are
different from usual ST-Link/V2.1

Add the new PID in the driver, in the tcl interface script and in the
udev configuration script.

Change-Id: Id2e1b5a5d0347c5d951a86a9cdb76be52cfd4ea3
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4702
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

flash/stmsmi: fix byte order for big-endian host

The original code was written for and tested on little-endian
host only.

Rewrite it to be independent by host endianess.
Not tested on real HW; I don't own anymore a SPEAr device.

Change-Id: I2f427a804693f56cb9dea4936c525eb814c48c28
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reported-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4778
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

mem_helper: add mrh command

This patch adds support for reading halfword values from memory. This
command compliments existing support for writing halfwords (mwh).

Change-Id: I8ec628e65c05a7f00aa57e3af0f228eb8bd4d14e
Signed-off-by: Steven Stallion <stallion@squareup.com>
Reviewed-on: http://openocd.zylin.com/4781
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

jtag: drivers: ft232r: unhardcoded

FT232R: introducing configurable parameters for pinout,
initial buffer size, state of RS232 signals at exit with
option to reattach kernel driver. All this was hardcoded before.
New parameters are documented in "openocd.texi" file.
Allows hopfully self-explainable and user-friendly adaptation to
various pinouts and similar chips like FT230X and FT231X.

Change-Id: Ib807f2ea3d4c1a164d351d65aeacd1978318d217
Signed-off-by: EMARD <vordah@gmail.com>
Reviewed-on: http://openocd.zylin.com/4681
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>

flash/nor: use default_flash_blank_check() instead of dummy

Some flash drivers had a dummy method for erase_check.
Use default_flash_blank_check() instead if possible.

Change-Id: Iddfeff45ce477007328d061fcb5c553d93c3be98
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4766
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>

SPI table updates (some new devices and new info)

read_cmd, qread_cmd, pprog_cmd added as some recent high densities devices
use variants for 4-byte addressing.

Some new flash and FRAM device ids added. FRAMs don't have write pages nor
erase commands or sector sizes. The corresponding entries are marked as
"not used" (i. e. zero). Checks in existing SPI flash drivers added to
handle these cases gracefully.

Change-Id: I5104bce7c815ac22f98bc32c1bb6db66b984404a
Signed-off-by: Andreas Bolsch <hyphen0break@gmail.com>
Reviewed-on: http://openocd.zylin.com/4773
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Set TCP_NODELAY for local connections to jtag_vpi.

This increases performance drematically for local connections, which is the
most likely arrangement for a VPI connection.

Change-Id: Id15b29ae663f5d8100b2175357649bd03d05b7c8
Signed-off-by: Darius Rad <darius@bluespec.com>
Reviewed-on: http://openocd.zylin.com/4549
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor: consolidate flash protect/protect_check

Make flash_driver methods protect() and protect_check()
optional.

Remove dummy definitions of these methods from the drivers
which do not implement protection handling.

Some drivers did not define protect method. It raised segfault
before this change and now it is handled properly.

Lot of drivers returned ERROR_OK from dummy protect()
- now flash_driver_protect() returns an error if not handled by the driver.

Change-Id: I2d4a0da316bf03c6379791b1b1c6198fbf22e66c
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4765
Tested-by: jenkins
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>
Reviewed-by: Christopher Head <chead@zaber.com>

tcl/target, board: remove useless gdb-attach event definitions

Since commit bae76053dc515252dc5c8235b9a848e461080c66
gdb-attach event is defined as halt by default.
Remove useless and in case of bcm281xx wrong definitions of the event.

Change-Id: I8e69780a93722eb9392673303f54d502e71eceb6
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4763
Tested-by: jenkins
Reviewed-by: Steven Stallion <sstallion@gmail.com>

stm32l4x: Fix stm32l4x dual bank support

The dual bank option was being incorrectly read and the
bank b base incorrectly set. This is tested with 512kB
dual bank configuration but needs checking with other
configurations (e.g. 256kb).

This fix should remove the need to use a mass_erase command
prior to programming with OpenOCD

Change-Id: I6e920f11b794c4c1fd34c0e44fb8fa01e7fe8f85
Signed-off-by: Alex J Lennon <alex.lennon@s19.tech>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4641
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-by: Thomas Søhus <soehus@gmail.com>

tcl/target: ti_tms570.cfg restructure dap support

ti_tms570 was probably omitted in commit
2231da8ec4e7d7ae9b652f3dd1a7104f5a110f3f

Change-Id: Idd4828fd5ea3641bda6c73c7f07a598c1e512ef6
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4762
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

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)