Function riscv_program_exec() contains code for restoring
of register values after progbuf execution. This code is
not used anymore by current OpenOCD, and hence removed.

Related discussion can be found under:
https://github.com/riscv/riscv-openocd/issues/982



Change-Id: I4c79bec081522b6fc0d16367cef51ed19a131962
Signed-off-by: Jan Matyas <jan.matyas@codasip.com>

Revert "break from long loops on shutdown request"

Before this change, when we send an I2C Bulk data at the same
time while Jtag bitbanging functions execute, the microcontroller
puts JTAG bitbanging on wait and executes all I2C bitbanging
function, which causes problems like loss of Ack in DAP responses
and other errors.

With this commit, When I2C interruption occurs, it sets a variable
to true and continues JTAG bitbanging, when it finish it executes
the I2C bitbang.

Change-Id: Ia80bac21f8a259f4a1176b5346bf74ed0aa6e38b
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8074


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

Add GPIO extender initial configuration that is needed to configure
some important GPIOs and ensure that the dev board is ready to work.

Add i2c_write function that make a write transfer to any slave device.

Give a new Product ID to ANGIE to make it different than the non
programmed ANGIE.

Change-Id: I0a8dacb7fe218145b7d3ed1cb75f106ed6256714
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8072


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

Give ANGIE a new PID after renumeration to be able to distinguish
the two cases (programmed and not programmed)

Change-Id: I30a91d8ed2e8e261221488b98d40a027ca41da52
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7991


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

remove angie_clear_queue function from executing before the
angie_execute_queued_commands function and making it at the
end of the reset function.

Change-Id: Id8a0664fbd5b8f9730545ce0f8f272ae0b0e7e78
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7990


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

We want to keep the tri-state buffers located between the FPGA
and the board, in 'Z' state until we launch an i2c connection.

We launch an i2c start condition, make the SCL
direction 'OUT' to start the i2c protocol and at the end
of the i2c connection at the stop condition, we re-make
the tri-state buffers at 'Z' state.

Change-Id: Ic597a70d0427832547f6b539864c24ce20a18c64
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7989


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

Send bitstream size to firmware to initialize the GPIF count registers,
since we're going to send this size via GPIF, we need to give the exact
number of bytes to be sent, then the GPIF counter will decrement with
every clock cycle (every byte sent) until reaching zero and stops.

Change-Id: Ib4e8e0f95a6a4a95ef4888ba8a04a0ea45567f5a
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7988


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

Change the GPIF state machine, configuring only one of the 4 waveforms
to generate the clock signal (CCLK) used to program the FPGA, and send
one byte every cycle using an 8-bit bus.

Change-Id: I43cf5480b9d5c40cc2f6a62a52ecfe078b76458e
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7976


Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins

clang build triggers an error for an uninitialized value of the
variable 'instr'.
This is a false positive, as the macro
 #define MIPS32_CONFIG3_ISA_MASK (3 << MIPS32_CONFIG3_ISA_SHIFT)
guarantees the switch/case already covers all the possible values
with cases 0, 1, 2 and 3.

Silent clang by adding a useless default case to the switch.
While there, fix the indentation of the switch/case accordingly to
OpenOCD coding style.

Change-Id: I0ae316754ce7d091dd8366bf314b8e6ee780e313
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Fixes: 7de4b120 ("target/mips32: add cpu info detection")
Reviewed-on: https://review.openocd.org/c/openocd/+/8065


Tested-by: jenkins
Reviewed-by: Oleksij Rempel <linux@rempel-privat.de>

- Add extra error checking
- Cache PS; lower PS.INTLEVEL to allow  breakpoint trigger (LX)
- Xtensa algo support functional on LX per functional flash driver
- Test on NX via manual algo validation

Change-Id: Ie7cff4933979a0551308b382fa33c33c66376f25
Signed-off-by: ianst <ianst@cadence.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8075


Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Erhan Kurubas <erhan.kurubas@espressif.com>
Tested-by: jenkins

Some OpenOCD command gets fragment of TCL scripts as command-line
argument, fragments that will be kept and executed later on.
E.g. the command 'configure' gets the body of an OpenOCD event:
	$TARGET configure -event halted {TCL code}

These commands store the argument as a Jim_Obj and pass it to the
jimtcl interpreter when the TCL fragment has to be executed.
Using Jim_Obj as storage is relevant to let the jimtcl interpreter
to recover extra info of the TCL fragment, like the file-name and
the line-number that contain the fragment, that will be printed
out in case of run-time errors.

While converting the commands to COMMAND_HANDLER, we should avoid
storing the argument as C strings otherwise we will loose precious
info in case of run-time errors making challenging the debugging
of such TCL fragments.

Extend the struct command_invocation to contain the array that
points to the Jim_Obj of the command arguments.
This will be used while converting commands to COMMAND_HANDLER.

Change-Id: If37c5f20e9a71349f77ba1571baf1e6778e28aa5
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8057
Tested-by: jenkins

Simplify the command execution by inlining run_command() inside
exec_command().

Change-Id: Id932b006846720cfd867d22d142cd35831dbd1a2
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8056
Tested-by: jenkins

The jimtcl interpreter guarantees that the Jim objects in argv[]
are not deallocated during the command execution. Thus, there is
no need to copy the string content of argv[].

Simplify exec_command() by inlining its two sub-functions and
dropping the strdup().

While there, add a LOG_ERROR() for out of memory.

Change-Id: I3e21ed7da50ca0bd072edbd49fca9740c81f95b0
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8055
Tested-by: jenkins

The output parameter nwords is always equal to the input parameter
argc, when the function succeeds.

Drop the parameter nwords and let the caller use directly the
value in argc.

While there, convert some 'unsigned' to 'unsigned int'.

Change-Id: Ie3d8ce1351792f3c07fe39cdcbcd180fd24dc928
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8054
Tested-by: jenkins

In both functions script_debug() and script_command_args_alloc()
the variable len is never used, and Jim_GetString() does not
mandate it.

Drop the variable and pass NULL to Jim_GetString().

Change-Id: I754b27a59c6087cde729496be42609d2a7145b0c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8053
Tested-by: jenkins

Commit e370e06b

 ("target: Deprecate 'array2mem' and
'mem2array''") has already replaced the deprecated root versions
of commands mem2array and array2mem with TCL proc's that use
'read_memory' and 'write_memory'. It has left the deprecated code
of the target's version of the commands because the effort to code
the TCL replacement was not considered valuable.

To drop the last jim_handler commands, I consider much easier and
less error-prone to code them in TCL instead of converting the
deprecated code to COMMAND_HANDLER.

Drop the code in target.c and extend the TCL proc's.
While there, add the TCL procs to _telnet_autocomplete_skip.

Change-Id: I97d2370d8af479434ddf5af68541f90913982bc0
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8052
Tested-by: jenkins

The SPDX tag 'GPL-2.0' has been deprecated in
https://spdx.org/licenses/GPL-2.0.html
and the preferred tag is now 'GPL-2.0-only'
https://spdx.org/licenses/GPL-2.0-only.html



Update the LICENSES documents and the SPDX of the only file that
reports the deprecated tag.

Change-Id: I3c3215438bc4378ff470bb9fa8fa962505a9ae50
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8064
Tested-by: jenkins

When requested by the core code (handle_breakpoints = true),
arc_resume() should be able to advance over a potential breakpoint set
at the resume address instead of getting stuck in one place. This is
achieved by removing the breakpoint, executing one instruction,
resetting the breakpoint, then proceeding forward as normal.

With this patch applied, openocd is now able to resume from a
breakpoint halt when debugging ARCv2 targets via telnet.

This has previously been committed to the Zephyr project's openocd repo
(see https://github.com/zephyrproject-rtos/openocd/pull/31

).

Change-Id: I17dba0dcea311d394b303c587bc2dfaa99d67859
Signed-off-by: Evgeniy Didin <didin@synopsys.com>
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Signed-off-by: Artemiy Volkov <artemiy@synopsys.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7817


Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins

Presently, we rely on gdb to restore break/watchpoints upon resuming
execution in arc_resume(). To match this behavior in absence of gdb
(more specifically, when handle_breakpoints is true), this patch
explicitly re-enables all breakpoints and watchpoints in arc_resume().

This has previously been committed to the Zephyr project's openocd repo
(see https://github.com/zephyrproject-rtos/openocd/pull/31

).

Change-Id: I59e9c91270ef0b5fd19cfc570663dc67a6022dbd
Signed-off-by: Evgeniy Didin <didin@synopsys.com>
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Signed-off-by: Artemiy Volkov <artemiy@synopsys.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7816


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Oleksij Rempel <linux@rempel-privat.de>

- Read and display EXCCAUSE on exe error
- Clean up error messages
- Clarify "xtensa exe" documentation

Signed-off-by: ianst <ianst@cadence.com>
Change-Id: I90ed39f6afb6543c0c873301501435384b4dccbe
Reviewed-on: https://review.openocd.org/c/openocd/+/7982


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

For Armv8.1-M based cores, detect if the core implements the optional
M-profile vector extension (MVE), using MVFR1 register.

While at there rework armv7m->fp_feature detection based on MVFR0
and MVFR1 registers.

Change-Id: I92d5b1759aea9f7561d285f46acdec51d6efb7b4
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6950


Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins

By queuing multiple jtag transfers the connection speed between
JTAG-Host and JTAG-Hub is improved. This is due to much less
calls to OS functions. An improvement of about x30 has been
measured with ftdi-based jtag adapters

For this to work the JTAG-Host server needs to know if flow control
is enabled on the JTAG-Hub ports. This is possible with newer
JTAG-Hub/JtagCDC. For old JTAG-Hubs the queuing is not enabled so
this change is backwards compatible.

Change-Id: I8a5108adbe2a2c1e3d3620b5c9ff77a546bfc14e
Signed-off-by: Daniel Anselmi <danselmi@gmx.ch>
Reviewed-on: https://review.openocd.org/c/openocd/+/7978


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

On some systems (at least Windows/CYGWIN and macOS) libusb_dev_mem_alloc()
simply returns NULL. The helper can fall-back to malloc() to allocate
CMSIS-DAP pending command/response buffers.

Fixes: fd75e9e5

 (jtag/drivers/cmsis_dap_bulk: use asynchronous libusb transfer)
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Change-Id: I89660f6747ad9d494b8192711cbbee5764e058fa
Reviewed-on: https://review.openocd.org/c/openocd/+/8044


Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins

On some systems (at least Windows/CYGWIN and macOS) libusb_dev_mem_alloc()
simply returns NULL.

Use the result of the very first libusb_dev_mem_alloc() call to decide
if the underlining system supports dev mem allocation or we should
fall-back to plain heap malloc().

From the decision time on, keep using the selected type of memory allocator
and deallocator.

Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Change-Id: Ia1f0965cea44b4bb6d936b02ec43f5a16a46f080
Reviewed-on: https://review.openocd.org/c/openocd/+/8059


Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

This reverts commits 2e920a21 and
8dbb1250

.

The reason is, after `openocd_is_shutdown_pending()` becomes true,
arbitrary command may be executed:
* In `target_destroy()` and the corresponding
  `target->type->deinit_target()`.
* In user-specifyed `pre_shutdown_commands` list.

Change-Id: Icd00d1d954cd45e255880a6f76c3a74c098d6a17
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>

target/riscv: move read redirection for `priv` to `riscv-013.c`

target/riscv: report info about target during `poll`

target/riscv: cleanup `get_riscv_debug_reg_ctx()`

Addresses issue #196.

Change-Id: I71146c7bc769cb9727e57da33e9f514eedef9ce4
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>

The reason for the change is a conflict: `dcsr[5]` is `dcsr.v` in
current spec, but it is `dcsr.debugint` in 0.11. This causes `priv`
register to be read incorrectly.

Change-Id: If2d8fdcd8536afa4c7149c453101b00ce0df1ce0
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>

target/riscv: remove `riscv_hart_count()`

target/riscv: use defined constants in `dmi_*_t` enums (non-functional change)

target/riscv: fix DM register address checks in `dm_read`/`dm_write`

This commit makes the function safe to use throughout the lifetime of a
target.

Change-Id: I7a573e5d3b70daec2cf8f47a2aa1e30e39321549

There was an error in `dm_read`/`dm_write`: DMI address was checked
against DM registers disregarding DM base address.

To solve the issue `dmi_address()` function was introduced.

Change-Id: Ia3be619417b5f5b53db5dfe302db05170d6787c9
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>

The motivalion for the change:
* `riscv_hart_count()` is used only once to print the value into the log
  during exmination.
* The returned value is a bit confusing: it's not the total number of
targets on the TAP. It is the number of targets accessable through the
same DM. So the name of the function is misleading.
* This value is already reported on `-d3` level.

So the function seems redundant and can be safely removed.

Change-Id: Iac9021af59ba8dba2cfb6b9dd15eebc98fe42a08
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>

Merge up to adcc8ef8 from upstream.

Break from long loops on shutdown request

Change-Id: Ia45da0e7f3e24dbeafc41c0213cf28d469641fe8
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>