flash/nor/stm32h7x: fix option bytes handling to work with both banks

To achieve that we need to avoid using FLASH_REG_BASE_B0, and use
bank registers instead:
   For dual bank devices, each option register is mapped in 2 addresses
   at the same offset from flash_bank_reg_base.
   This is true for OPTCR, OPTKEYR, OPTSR_CUR/PRG, OPTCCR according to
   RM0433 Rev6 (refer to section 3.9: FLASH registers)

In stm32x_write_options, according to RM0433 Rev6, after OBL launch we
should wait for OPTSR_CUR.BSY bit instead of FLASH_SR.QW

Change-Id: Ie24a91f069d03c9233797390fc2e925c737dad90
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/5291
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/stm32h7x: remove stm32x_options.protection2

Each bank had to store its options only, there is no need for bank1
to sneak into bank2 options.
Furthermore, some variants do not have a second bank.

Change-Id: I9229eb8ab4b5860ba2b0c5dbe626a54a84bca4d6
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/5290
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/nrf5: fix allocation of driver_priv and sector array

Drop static pointer to allocated struct nrf5_info, iterate over
the flash bank list to find previously allocated nrf5 instances.
nrf5 is swd only device, so static allocation makes no harm,
but we should avoid copying the wrong code to other flash drivers.

Free sector array before allocating it to avoid memory leak on
re-probing device.

Change-Id: I781d8f4418a91c043f2393e5ecc5278fc6df3566
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4910
Tested-by: jenkins

flash/nor/nrf5: show RAM size on old nRF51 devices (rev 1, 2)

Change-Id: I2452c084b9bd9e401bd49e15791428a53df1cd1d
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4868
Tested-by: jenkins

flash/nor/nrf5: do not check FICR PPFC on nRF52

Change-Id: I6beee9b85a542040f2495513b5ba51bd8e1389db
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4867
Tested-by: jenkins

flash/nor/nrf5: rename registers by nRF series

Change-Id: I70af671c52665b27a28508e06e7d3e5e40a621f7
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4866
Tested-by: jenkins

flash/nor/nrf5: refactor sector allocation to use alloc_block_array()

Change-Id: Ied8ea917cec492fc6bb8836a92d8c4ceaf3b499b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4865
Tested-by: jenkins

flash/nor/nrf5: implement BPROT protection check for nRF52810,832

Also refuse 'flash protect' on any nRF52.
Fail protection check on nRF52840 until ACL protection is implemented.

Change-Id: I84fcf117427e4894147c3ad92e2a3597566b4fcf
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4864
Tested-by: jenkins

flash/nor/nrf5: detect newer devices without HWID table

nrf5 flash driver detected devices by looking up the HWID in the table
of known devices. Unfortunately chips are produced with many different
HWIDs for each type.

All nRF52 devices have FICR INFO field suitable for device identification
without need of HWID lookup.
Some newer nRF51 devices have FICR INFO too although undocumented.
Use this information to identify the device.

nrf5_info() is reworked to show just concise info.
Decoding FICR and UICR registers was moved from nrf5_info()
to a new command 'nrf5 info' without functional changes.

The flash bank for UICR page has the same size as program flash sector.

Change-Id: I900095b9ae23ee995f8e2bef8539b75d00300da5
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4848
Tested-by: jenkins

flash/nor/nrf5: remove useless page padding and UICR autoerase

nRF5 flash controller can write a word at a time. Ask flash
infrastructure to handle alignment and padding.

Fix mixing of offset and address in nrf5_ll_flash_write()
- the original code worked just because NRF5_FLASH_BASE is 0

Change-Id: Ibe8bdf899a1764cf4117b2deda1a4618eeb16697
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4819
Tested-by: jenkins

helper: skip including sys/sysctl.h on Linux

Starting from glibc 2.30, the header file sys/sysctl.h gets
deprecated on Linux, after the commit 744e82963716 ("Linux:
Deprecate <sys/sysctl.h> and sysctl")
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=744e82963716

The associated NEWS reports
The Linux-specific <sys/sysctl.h> header and the sysctl
function have been deprecated and will be removed from a
future version of glibc.

Latest automake 1.16.1 still does not handle this case.
Current OpenOCD build fails with warning and requires configure
with "--disable-werror" to build.

Prevent including sys/sysctl.h on Linux build.

Change-Id: I5310976573352a96e5aef123352f73475f0c35fe
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5317
Tested-by: jenkins
Reviewed-by: Moritz Fischer <moritz.fischer.private@gmail.com>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>

CVE-2018-5704: Prevent some forms of Cross Protocol Scripting attacks

OpenOCD can be targeted by a Cross Protocol Scripting attack from
a web browser running malicious code, such as the following PoC:

var x = new XMLHttpRequest();
x.open("POST", "http://127.0.0.1:4444", true);
x.send("exec xcalc\r\n");

This mitigation should provide some protection from browser-based
attacks and is based on the corresponding fix in Redis:

https://github.com/antirez/redis/blob/8075572207b5aebb1385c4f233f5302544439325/src/networking.c#L1758

Change-Id: Ia96ebe19b74b5805dc228bf7364c7971a90a4581
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reported-by: Josef Gajdusek <atx@atx.name>
Reviewed-on: http://openocd.zylin.com/4335
Tested-by: jenkins
Reviewed-by: Jonathan McDowell <noodles-openocd@earth.li>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>

STM8 Target relicensing to GPLv2 and later

Change-Id: I21126945c0475399aaf12239b8972fde5fddd845
Signed-off-by: Ake Rehnman <ake.rehnman@gmail.com>
Reviewed-on: http://openocd.zylin.com/5331
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

jtag: usb_blaster: Add missing 'default' to switch statement

If a new JTAG command is added, then GCC will complain that
enumeration value not handled in switch. This is the only driver not
to have a default case, so add it.

Change-Id: Icb838087bb7525d057a911bd256300e256da1668
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://openocd.zylin.com/5333
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag: jtag_vpi: Add missing 'default' to switch statement

If a new JTAG command is added, then GCC will complain that
enumeration value not handled in switch. Make this consistent with
other drivers, and add a 'default' case.

Change-Id: I66d6d0db3fcae93ea246f2d4882ffff5dec14693
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://openocd.zylin.com/5340
Tested-by: jenkins
Reviewed-by: Jan Matyas <matyas@codasip.com>
Reviewed-by: Franck Jullien <franck.jullien@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

libjaylink: Update to latest Git version

This version adds two new USB PIDs and fixes a build issue under MSYS2.

Change-Id: I753fab827783ea64e55e59d833742c9f70a28a2b
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5309
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

nrf5: update links to compatibility matrixes for nrf5x variants

Change-Id: If51aa992ccbb8c9a2e502b74827a36a62010546d
Signed-off-by: Mirko Vogt <mirko.vogt@sensorberg.com>
Reviewed-on: http://openocd.zylin.com/4843
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/cortex_a: Extract code to read/write from/to register to/from DCC

In preparation for supporting the ARM MCRR and MRRC commands which will
require using two 32-bit registers to read/write a 64-bit internal
register, extract the common logic to read/write from/to a register
to/from DCC and make that parameterized such that we can do this through
not just r0.

Change-Id: Iadb73f5cde8cf5961b5a18ddd198bf39d791e610
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5227
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

drivers/gw16012: remove useless cast on gw16012_port

The variable gw16012_port is of type uint16_t.
There is no need for a cast to print it.

Change the format modifier to PRIx16

Change-Id: I16fe688b9d235bae46525635d07849a00fba9548
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5195
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

gdb_server: Support vRun packet, allow setting cmdline from GDB

GDB uses the vRun packet if available to restart a running process in
extended remote mode. Support this like the R packet and set the
semihosting command-line to allow it to be specified from GDB.

Change-Id: I9cb812b22170630f782113c9927e46e0cd5b1f0f
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/5186
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

gdb-server: Create arch-specific structure type for every feature

As it is mentioned here [1] type's ID is unique name within containing feature.

That said if regs of the same type located in different features it's required
to insert type definition at least in each feature.

See more details in discussion here [2].

[1] https://sourceware.org/gdb/onlinedocs/gdb/Target-Description-Format.html#Types
[2] https://github.com/foss-for-synopsys-dwc-arc-processors/openocd/commit/2a5f5125ac8fa0e1359b6be03b209f9f5d1ade82#r33460077

Change-Id: Id92b061cfbf47d5c032a02c2c406b28affd0b02a
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Reviewed-on: http://openocd.zylin.com/5179
Tested-by: jenkins
Reviewed-by: Muhammad Omair Javaid <omair.javaid@linaro.org>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/cortex_a: use aligned accesses for read/write cpu memory slow

Armv7a is able to read and write memory at un-aligned address, but
only when bit SCTLR.A (Alignment check enable) is zero and the
address belongs to a memory space with attribute "Normal" (see [1]
chapter A3.2.1 "Unaligned data access"). In all the other cases
the memory access will trigger an alignment fault data abort
exception.
Memory attributes are explained in [1] chapter A3.5 "Memory types
and attributes and the memory order model".

Disabling the MMU cause a change in memory attribute, as explained
in [1] chapter B3.2 "The effects of disabling MMUs on VMSA
behavior".
This can cause several issues. e.g. a SW breakpoint on un-aligned
4-byte Thumb instruction, set when MMU is on, can be impossible to
remove when MMU turns off.

While is possible to check all the possible conditions before an
un-aligned memory access, it's clearly more maintainable to skip
such complexity and only perform aligned accesses.

Check the alignment and eventually modify the data size before
calling the functions cortex_a_{read,write}_cpu_memory_slow().
Change the comment in the two functions above to comply with the
new behaviour.

[1] ARM DDI 0406C.d - "ARM Architecture Reference Manual, ARMv7-A
    and ARMv7-R edition"

Change-Id: I57b4c11e7fa7e78aaaaee4406a5734b48db740ae
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5138
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

adi_v5_jtag: avoid RAM exhaustion by limiting jtag queue size

Issue has been found when I tried to read 64 MiB QSPI flash bank.
Bank is memory mapped, default_flash_read() is used for 'flash read_bank'
command. OpenOCD consumed as much as 6.8 GiB of RAM during this
process. Investigation showed that this happens because JTAG queue
is not limited in any way. OpenOCD queues 16 millions of AP reads
allocating all corresponding data structures.

Most of this memory is allocated in:
cmd_queue_alloc (commands.c) - 4.2 GiB
dap_cmd_new (adi_v5_jtag.c) - 2.25GiB

This patch implements a pool of "struct dap_cmd" objects using
linked list. Objects are taken from a pool in "dap_cmd_new()" and
returned to the pool when they are not needed. Size of the pool
is limited to 64K of objects, JTAG queue is forcibly executed
when this limit is reached.

Checked with Valgrind and Clang analyzer - no new warnings.

Change-Id: I5aaaecce5ed71414f7965a2598f49742f6a6b2b5
Signed-off-by: Bohdan Tymkiv <bhdt@cypress.com>
Reviewed-on: http://openocd.zylin.com/4948
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

Fix wrong end-of-region calculation

Correct check for end-of-region is:
$BASE + $LEN > $ADDRESS

And it is currently (wrongly) calculated as:
$ADDRESS > $BASE - $LEN

Change-Id: If10bfee19b0c7dbc085731ac1eda943f5d8a36a3
Signed-off-by: Piotr Kasprzyk <ciri@ciri.pl>
Reviewed-on: http://openocd.zylin.com/4798
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

nrf5: Fix misuse of flash bank number

Make driver_priv point directly into the corresponding chip bank structure
and add a pointer to it to get back to its chip when it's needed. This
removes the need to keep track of any bank number, either global or chip-
local.

In addition, it simplifies the cases where the chip structure was just used
to access the chip bank fields; now they are directly accessible.

Change-Id: Iaa353cd4fa7d8ff94c2ef69028c7cb32fade0420
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/4775
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/target: Add Infineon TLE987x

Prepare a config for Infineon TLE9879.

Change-Id: Ic667ae822fd514cac365993bc3f39b4185f1a118
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-on: http://openocd.zylin.com/4339
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

stm32l0|l1: don't corrupt RCC registers

instead of overwriting Reset settings, let's
read-modify-write RCC registers.

Change-Id: I21b7e26e6007d3c3d73803c681c980c6947f5910
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Reviewed-on: http://openocd.zylin.com/3601
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

target: Switch to target_read_buffer() in verify_image fallback

The current code checks the count to determine whether to read bytes or
words. However it fails to consider whether the base address is suitably
aligned.

Instead use the target_read_buffer() function which is for exactly this
purpose and generates optimal accesses with natural alignment.

Change-Id: I32ab5417890ee2219902df1529bc220fe353b4c7
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3217
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Update FTDI C232HM cfg, and add two new cfgs from cable modem research

Change-Id: Idbeffcd5ff4380b1e7c9fd5ef6ba3ca77cc22d99
Signed-off-by: Al Dyrius <aldyrius42@gmail.com>
Reviewed-on: http://openocd.zylin.com/5307
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

src/flash/startup.tcl: Add preverify to program command

The preverify option allows to check whether flashing is necessary.
If the target is flashed often/automatically this can save time and
preserve the flash. This is expecially helpful in CI environments.

Change-Id: Iead0a269e1a772b751d4dd9e8b53b2fecc874624
Signed-off-by: Moritz 'Morty' Strübe <moritz.struebe@redheads.de>
Reviewed-on: http://openocd.zylin.com/5292
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>

doc: emphasize the role of 'reset init' before flash commands

Change-Id: I4a4061ad0fa6e5dfb1e33f01d62145ca9bf12148
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/5304
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

Add complete JTAG debug logging.

Sample output, with default_interface_jtag_execute_queue replaced by
dijeq to satisfy commit message line length check:
Debug: 646 18 core.c:847 dijeq(): JTAG IR SCAN to RUN/IDLE
Debug: 647 18 core.c:852 dijeq():   5b out: 11
Debug: 648 18 core.c:847 dijeq(): JTAG DR SCAN to RUN/IDLE
Debug: 649 18 core.c:852 dijeq():   40b out: 4400000001
Debug: 650 18 core.c:857 dijeq():   40b in: 4400000000

Signed-off-by: Tim Newsome <tim@sifive.com>
Change-Id: I014e9e3a77755413b36edfcede2ab8f6aa08061b
Reviewed-on: http://openocd.zylin.com/4451
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Add wall clock timeout warning to mpsse_flush()

I think that libusb_handle_events_timeout_completed is supposed to make
progress or time out, but sometimes we hit a case where it makes no
progress, and mpsse_flush() loops forever. This wall clock timeout
notifies the user that this is going on.

When I wrote this code, this bug would reproduce every hour or two, but
right now it's not happening for me.

Change-Id: I7eb66f43462298e263a48048aa0c8769095661eb
Signed-off-by: Tim Newsome <tim@sifive.com>
Reviewed-on: http://openocd.zylin.com/4767
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

efm32: use device-specific MSC base for EFM32TG11B

According to the reference manual it should be 0x40000000. Flashing (and
booting) a firmware with this MSC base was successful.

Change-Id: I739e67d36555b8170a3b8e26f54cf1c09ce8424b
Signed-off-by: Christian Meusel <christian.meusel@posteo.de>
Reviewed-on: http://openocd.zylin.com/5263
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

src/jtag/aice: Fix obviously incorrect bit op.

Fix expression "(pin_status | 0x4)" which was always true rather than
testing a bit. Untested - was clearly not expressing the intent of the
author by inspection. Found by automated tooling and rtrieu@google.com.

Signed-off-by: Seth LaForge <sethml@google.com>
Change-Id: I4bb91e60e8ce9757bf21976cc48de6f85a39c68d
Reviewed-on: http://openocd.zylin.com/5301
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

helper/command: clear errno before calling parser

The C standard says that errno is set to ERANGE if an out-of-range value
is returned by strtol, strtoul, et. al., but it does not say that errno
is cleared if the function is successful (and, indeed, it is not on
glibc). This means that, if errno is ERANGE before strtol is called, and
if the value to be converted is exactly the maximum (or, for a signed
conversion, the minimum) legal value, COMMAND_PARSE_NUMBER will
erroneously indicate that the value is out of range.

Change-Id: I8a8b50a815b408a38235968f1c1d70297ea1a6aa
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/5298
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/tcl: Fix usage of 'flash erase_sector' command

Change-Id: I2141e377a0531cab8d1140049a2ee7721d30cfdc
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5299
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/board: Add Rigado BMD-300 Evaluation Kit

Change-Id: Iba8e12818e2041e51214dab413eb57f0e5bf3f75
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5218
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

esirisc_flash: Rename PAGE_SIZE to FLASH_PAGE_SIZE

PAGE_SIZE is defined in system includes on some systems, this would
avoid the unintended conflict

Fixes
| src/flash/nor/esirisc_flash.c:95:9: error: 'PAGE_SIZE' macro redefined [-Werror,-Wmacro-redefined]
| #define PAGE_SIZE                       4096
|         ^
| /mnt/a/yoe/build/tmp/work/core2-64-yoe-linux-musl/openocd/0.10+gitrAUTOINC+7ee618692f-r0/recipe-sysroot/usr/inclu
de/limits.h:89:9: note: previous definition is here
| #define PAGE_SIZE PAGESIZE

Change-Id: I195b303fc88a7c848ca4e55fd6ba893796df55cc
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Reviewed-on: http://openocd.zylin.com/5180
Tested-by: jenkins
Reviewed-by: Steven Stallion <sstallion@gmail.com>
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

flash/nor/stm32h7x: remove unused 'pages_per_sector' from stm32h7x_part_info

Change-Id: I1b79c25cada574e3a9849f506443c836bd707604
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/5289
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Christopher Head <chead@zaber.com>

flash/nor/stm32h7x: remove flash size information from device name

There is no sense in displaying the max size (2M) as there is variants
of this device with reduced flash size

Change-Id: I40574064d75fdf2a038044c81038a6d7abc6c4dd
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/5288
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor: flash driver for Synwit SWM050 MCUs

SWM050 is a series of MCU product by Foshan Synwit Tech, which is
available in TSSOP-8 or SSOP-16 packages.

Adds flash driver for the internal 8KiB flash of the MCU. The registers
are based on reverse engineering the J-Flash blob provided by the
vendor.

Also adds a pre-made cfg file.

Change-Id: I0b29f0c0d062883542ee743e0750a4c6b6609ebd
Signed-off-by: Icenowy Zheng <icenowy@aosc.io>
Signed-off-by: Caleb Szalacinski <contact@skiboy.net>
Reviewed-on: http://openocd.zylin.com/4927
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

target/dsp563xx: dsp563xx restore reg support

Added "exist=true" field to the reg_list struct to make access to the
dsp563xx registers again possible. Without it defaults to exist=false
and all the reg related functions will return nothing.

Fixes regression from b5964191f0d2fc3ace607af001df3d57cbfbaf2b

Change-Id: I9c256346735b8d66919c4ba83f528a8afca46ff9
Signed-off-by: Han Hartgers <han.hartgers@gmail.com>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/5239
Tested-by: jenkins

src/flash/nand: Fix some operator precedence bugs.

Fix two expressions where precedence of operator | vs ?: was clearly confused.
Untested - was clearly not expressing the intent of the author by inspection.
Found by automated tooling and rtrieu@google.com.

Change-Id: I46f190154797f8affc761caf3a15a1a9db53d702
Signed-off-by: Seth LaForge <sethml@google.com>
Reviewed-on: http://openocd.zylin.com/5281
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

gdb_server, rtos: Fine-grained RTOS register access

1. Add get_thread_reg() to rtos. It's used in rtos_get_gdb_reg() to read
the value of a single register, instead of reading all register values
by calling get_thread_reg_list().
2. Add set_reg() to rtos. gdb_server uses this to change a single
register value for a specific thread.
3. Add target_get_gdb_reg_list_noread() so it's possible for gdb to get
a list of registers without attempting to read their contents.

The clang static checker doesn't find any new problems with this change.

Change-Id: I77f792d1238cb015b91527ca8cb99593ccc8870e
Signed-off-by: Tim Newsome <tim@sifive.com>
Reviewed-on: http://openocd.zylin.com/5114
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

flash/nor/core: fix some minor typo

Change-Id: I03832b3e4a6eaadfd87729a3a898e0a2cd30931a
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/5264
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

contrib/rpc_examples: Adapt to new command line handling

Change-Id: I844ef7fbf57a22097a936f4614b4a4c7e980bef6
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5248
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

mflash: Remove this broken flash driver

This is causing repeated build failures. Its design is so fundamentally
broken that if someone actually wants to use it, a full rewrite is the
only option. So it's not even worth deprecating in the hope that someone
will notice and fix it, just get rid of it.

Change-Id: I513069919a3873bd69253110f7fb6f622ee7d061
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/5243
Tested-by: jenkins
Reviewed-by: Jeffrey Booher-Kaeding <Jeff.Booher-Kaeding@arm.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

configure.ac: Fix ST-Link adapter description

The ST-Link driver supports not only JTAG but also SWD and SWIM.

Change-Id: I9f0e7b018cae54ed8e73a724151647e050e7bb49
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5247
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins

cortex_m: set C_DEBUGEN in soft_reset_halt

The command "soft_reset_halt" is deprecated since mid 2013 with
the commit 146dfe32956d ("cortex_m: deprecate soft_reset_halt").
Nevertheless it is still extremely useful with multicore chips
where it allows to reset only one of the cores, option not
available through asserting the chip-wide srst.
Without a better replacement of the command, it's worth fixing it.

Accordingly to ARM DDI 0403E.d, chapter C1.4.1 "Entering Debug
state on leaving reset state", to halt the core at reset both bits
DHCSR.C_DEBUGEN and DEMCR.VC_CORERESET must be set.
Current code only sets the latter bit, relying on having C_DEBUGEN
already set through other commands, e.g. "halt". This prevents the
command "soft_reset_halt" to work if issued as very first command.

Set the bit C_DEBUGEN in command "soft_reset_halt".

Change-Id: I66bfd6a0da1fca5049dea037b4d258cf6f842966
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4987
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

helper/options: simplify the code using command_run_linef()

Thanks to command_run_linef() there is no need to pre-build the
command using alloc_printf().

Change-Id: Iccfebd6063d1ac162f090fe2309b1f51bebf0214
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5226
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

helper/command: make command_run_line reentrant

The `command_run_line` function contains a comment saying it should be
reentrant. However, it isn’t: it NULLs out `current_target_override` and
doesn’t restore it before returning, and it changes the `context`
associated data of the `interp` object and then deletes that associated
data before returning rather than restoring it to its previous value.

Change-Id: I84fd46ef7173f08cf7c57b9a5b76e4986a60816f
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/5223
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

doc: makeinfo extra whitespace

Extra whitespace in file creating build errors with makeinfo.

Change-Id: Ib764850c1c8ff596d3c753eadd8e27f8c5982d20
Signed-off-by: Kevin Gillespie <kgills@gmail.com>
Reviewed-on: http://openocd.zylin.com/5229
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/board: Add SAML11 Xplained Pro Evaluation Kit

Change-Id: I118929cdecd9ba1f39d6e2791c114ac7e347b3f5
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5206
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

tcl/target: Add initial Microchip SAML1x support

There is not flash bank support at the moment.

Change-Id: I833c009d9d21cdeb70b57d67eb557d50ed0fb4de
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5205
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/cortex_m: Add support for AHB5-AP

The AHB5-AP is implemented in Cortex-M23/33 based devices.

Change-Id: I505954a2e2c6462ce0aa96eba1d55b016c5028b9
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5232
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Tested-by: jenkins

target/arm_adiv5: Add type for AHB5-AP

This access port type comes with the AMBA 5 protocol specification,
see 'C1.3 AP' in ARM IHI 0031D.

Change-Id: I3c4f0a69230daaf4f5f979de6213fe3c025a089a
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5231
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Muhammad Omair Javaid <omair.javaid@linaro.org>

Configs for ARM corelink SSE-200 target and Musca A board

This patch adds configuration files for ARM CoreLink SSE-200 SoCs. Also
adds configuration file for SSE-200 based Musca A board. Flash programming
support for Musca A QSPI flash is still not functional. This configuration
will be updated once that support lands into OpenOCD.

Please refer to ARM documentation for more information about SSE-200 and
Musca A.

Change-Id: Id3783c34d6e2609d659ef91c0bf7252c39439874
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/5006
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

Support for debugging on ARMv8-M CPUs

This patch adds ARMv8-M CPUs detection logic in ARMv7m target specific code.

Also adds a slightly different watchpoint manipulation logic for ARMv8-M.

This is based on ARMv8-M architecture reference manual.

Tested on ARM Musca A board.

Change-Id: I0652560954ef02c378a7067fab586edf39d3e9cc
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/4997
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/cortex_a: remove dependency from jtag queue

Replace jtag specific API jtag_add_reset() with transport
independent API adapter_{de}assert_reset().

Change-Id: I1b917a4c1205115c4e0315373d81a9305e931258
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4944
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/aarch64: remove dependency from jtag queue

Replace jtag specific API jtag_add_reset() with transport
independent API adapter_{de}assert_reset().

Change-Id: I32c43e2e47366363521fa3f387de9e2fb1c20852
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4943
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

jtag: fix error on TCL command "return" in jtag event handler

The TCL command "return" always returns error code JIM_RETURN, to
indicate that the effective error code and message are elsewhere.

In the current implementation, the caller of jtag's event only
checks for return code JIM_OK and considers any other value,
including JIM_RETURN, as an error condition.

It can be tested running openocd on a jtag target and adding a
jtag event "setup" with a single line "return", e.g.
openocd -f board/ti_cc3200_launchxl.cfg \
-c 'jtag configure cc32xx.cpu -event setup return'
to get the message:
../src/jtag/core.c:1599: Error:
in procedure 'jtag_init' called at file "../src/jtag/core.c",
line 1599

Modify jtag_tap_handle_event() to detect the specific return value
of the "return" command and to test the real error code that is,
eventually, specified to the TCL "return" command.

Change-Id: I6d6febc15ef169638afffbffc1810e0b84fcf5c8
Reported-by: Tomas Vanek <vanekt@fbl.cz>
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5199
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target: fix error on TCL command "return" in target event handler

The TCL command "return" always returns error code JIM_RETURN, to
indicate that the effective error code and message are elsewhere.

In the current implementation, the caller of target's event only
checks for return code JIM_OK and considers any other value,
including JIM_RETURN, as an error condition, thus dumping the
call-trace. The execution is not stopped because the error is not
further propagated, but the error message is annoying and
misleading.

It can be tested running
openocd -f ./test.cfg
using the following script "test.cfg". You can replace the board
file in line 1, to use a board available in your lab:
  1 source [find board/st_nucleo_f4.cfg]
  2 [target current] configure -event reset-start {}
  3 [target current] configure -event reset-end {return}
  4 init
  5 proc a {} {[target current] invoke-event reset-start}
  6 proc b {} {[target current] invoke-event reset-end}
  7 proc c {} {a;b;echo "arrived at the end"}
  8 c
  9 shutdown
The execution produces:
./test.cfg:7: Error:
in procedure 'c' called at file "./test.cfg", line 8
in procedure 'b' called at file "./test.cfg", line 7

arrived at the end
that shows the call-trace but does not halt the execution.

The developer can avoid using the "return" command in the event
body by defining a TCL procedure that implements the handler and
that contains the "return" command, reducing the handler body to
a simple call to the procedure above. But this approach is either
not documented nor always intuitive while writing the handler,
causing waste of time to look for the false error.

Modify target_handle_event() to detect the specific return value
of the "return" command and to test the real error code that is,
eventually, specified to the TCL "return" command.

Change-Id: I2b860bab7233c6ed13ee4098e348d7533e1c4626
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4974
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/stm32h7x: fix register names to comply with RM0399 Rev2 and RM0433 Rev6

Change-Id: I085d86a2a47f4aeef93a99238e3b80ee294d46df
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/5192
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag: drivers: buspirate: chunk SWD switch sequence transfer.

Commit c2e18bfaeafd changed the size of the JTAG-to-SWD sequence
from 15 bytes to 17 bytes. This broke SWD switch sequence transfer
for buspirate, since buspirate packets can only hold a payload of up
to 16 bytes and we tried to fit the whole sequence in a single packet.

Splitting up the sequence transfer in appropriately sized packets
makes buspirate SWD work again (successfully tested with buspirate
firmwares v6.1 and v7.0).

Change-Id: Ib5b412b9e77287d705d2762e31c16d30318b50e3
Signed-off-by: Tilman Sauerbeck <tilman@code-monkey.de>
Reviewed-on: http://openocd.zylin.com/5200
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag/drivers/jtag_usb_common: fix typo

Change-Id: If1f56fd5d610b993a4ecbc900fac9f90638037c9
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/5202
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target: make target re-configuration possible again

Before commit 877cec20dca6e78f9f029f0f173879cda101a6c2
("command: check command mode for native jim commands") all the jim commands
were erroneously treated as they had mode COMMAND_ANY.

The command '$_TARGET configure -xxx' was therefore applicable on running
OpenOCD to change the target configuration. It is handy e.g. for changing
an event handler or changes of the work area.

Change 'configure' command .mode to COMMAND_ANY to make it possible again.

The only parameter which cannot be re-configured after init is -gdb-port.
Test the command mode and refuse setting of gdb port after init.

Change-Id: I88493ac10a46647dc52a88fbc9f8ce6b5ba3bcd0
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/5214
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

fix XScale register access

since b5964191f0, all XScale-specific registers were missing, breaking
config scripts.

Change-Id: Ia56f3ca17500ba54bd08f417e9a5aaaa8a1be8c4
Signed-off-by: Michael Schwingen <spam-openocd@discworld.dascon.de>
Reviewed-on: http://openocd.zylin.com/5136
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

mem_ap: fix format of logged addresses

The macro TARGET_ADDR_FMT, defined in helper/types.h, already
includes the prefix "0x" in front of the hexadecimal number,
being defined as:
#define TARGET_ADDR_FMT "0x%8.8" TARGET_PRIxADDR
An additional "0x" is present in mem_ap; it prints debug messages
with a double "0x" before the address:
Debug: 2921 34180 mem_ap.c:153 mem_ap_write_memory():
Writing memory at physical address 0x0x5000000c; size 4;
count 1

Remove the incorrect hexadecimal prefix.

Change-Id: I38f19ed2a2f542bd5df53e947a2604f1cbe80e08
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5222
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

ARMv8: Update rtos_reg storage from 8 to 16 bytes

To support 128 bit registers, the rtos_reg structure value
array needs to be updated from 8 to 16 bytes.

Tested by reading ARMv8 NEON FP regs on an Ampere eMAG 8180 with GDB.

Change-Id: I7f3fe1a5b2def599d021787fbe9cdd51f92859a4
Signed-off-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>
Reviewed-on: http://openocd.zylin.com/5209
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

armv7a_mmu: Add support for decoding Super Sections

The ARMv7-A architecture supports super sections which allows mapping
physical addresses up to 40-bit into a 32-bit virtual address using the
short descriptor format (see ARM DDI 0406C.c section B4.1.112 for
details).

Change-Id: I8e64d0e93e36ae7a7da7b7bf2a8342856bb044f1
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5212
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

armv7a_mmu: Do not restrict virtual addresses to uint32_t

In preparation for adding super section decoding, do not restrict
armv7a_mmu_translate_va_pa() to 32-bit virtual addresses since ARMv7-A
processors with VMSA extensions (including LPAE) can issue wider
physical addresses. Update casting to uint32_t where necessary.

Change-Id: Id1c3d0d5ac324cbdc334259d9ea75fe4981671a1
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5211
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

armv7a_mmu: Remove armv7a_mmu_translate_va()

This function is not used anywhere in the tree, remove it, such that we
only have a single function moving forward that might need to deal with
short vs. long format specifics.

Change-Id: I80e81cd7eba1e028d1afaeaedb675b46c0ca6fa1
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5213
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

armv7a_mmu: Check earlier for PAR read

Check earlier that the read of the PAR register was successful instead
of starting the decoding and then checking for an error reading that
register.

Change-Id: Id96c2b2f76d2d1c745fcfa55ad4c1e6db92106f9
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5215
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

contrib/rpc_examples: remove 'ocd_' command prefix from haskell example

The prefixed commands has been removed in commit 0840414f0e57
("helper/command: do not replace new commands with ocd_ prefix").

Change-Id: I9f101beb85533973041386896bbb215bb141962f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5191
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>

contrib/rpc_examples: Remove 'ocd_' command prefix

The prefix is not necessary anymore.

Change-Id: Ie0df06a70ff51e6719d7564396739d28618b0196
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5190
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/armv7m: fix register number in armv7m_get_core_reg()

armv7m_get_core_reg() calls arm->read_core_reg()
arm->read_core_reg() expects the register number as an index
to core reglist, not an ARMv7M specific register code.
Use reg->number instead of armv7m_reg->num.

The change solves assert
  src/target/armv7m.c:222: armv7m_read_core_reg: Assertion
  `num < (int)armv7m->arm.core_cache->num_regs' failed.
when gdb 'info reg' is issued on a Cortex-M target and
no cortex_m_debug_entry() has been called since OpenOCD start
(target was halted before OpenOCD start).

Change-Id: I32a2294693ef979b613be93aeceb3b0eb06ee6df
Ticket: https://sourceforge.net/p/openocd/tickets/216/
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/5203
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

jtag: set default "jtag_only" to uninitialized transports

For legacy support, drivers that do not define a list of
transports get identified as jtag_only.

Cleanup this old crust and initialize properly the transports
field in the jtag_interface for all the drivers.

Change-Id: I9c86064e5d05bd0212bc18f4424414e615e617fe
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4893
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

gdb_server: remove call to jtag_execute_queue()

In the initial commit 6c9b804d6187edda4f46f8458deec0b17ec76bb9
in 2007's svn://svn.berlios.de/openocd/trunk@246
a target script gdb_program_config was called before gdb flash
programming. To guarantee the script does not left any pending
command in the jtag queue, a call to jtag_execute_queue() was
inserted after the execution of the script.

In following commit ef1cfb23947bd32798077c6abb5c25a049460ae9
in 2008's svn://svn.berlios.de/openocd/trunk@975
the script was replaced by the event "old-gdb_program_config" and
the call to jtag_execute_queue() get executed in every case, even
if the event handler was not present.

At last, commit bb3793c9a4ccd232c4ee3ce0a36bf200589ca0bb
("target: remove legacy target events") stripped away the
obsolete event but left the call to jtag_execute_queue(), now
completely useless.

Remove the call to jtag_execute_queue() and clean-up the code
around it.

Change-Id: I284f54d656d431ad6cdc25ca18218c09db31bd25
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4911
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/cortex_m: remove dependency from jtag queue

Since the first commit 09883194f867 that introduced cortex_m, the
code has a delay of 50ms after srst has been asserted.
The specific delay is implemented through the JTAG_SLEEP command
sent in the jtag queue.

To remove the dependency from the jtag queue, replace the delay
with a transport independent function.

In case of jtag transport, this change keeps the same behaviour
only if the jtag queue has been flushed before the delay. This
does not happen if the call to dap_dp_init(), few lines above,
fails while calling a dap_queue_dp_{read,write}(); in this case
the jtag queue will be flushed later, after the delay, while in
the original code the delay would follow the flushing of the
commands already queued. Anyway, this different behavior would
only happen in case of DAP already not responsive so anticipating
the delay in such error condition is not supposed to add further
problems.

Change-Id: If15978246764e4266b10e707d86c03e5ed907de7
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4912
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

helper: add bitmap helper primitives

Mainly copied/inspired from Linux kernel code in
- include/linux/types.h
- include/linux/bitmap.h
- include/linux/bitops.h

Change-Id: I317b542993ab81530c86553f339b79505d0fef0f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4894
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag: simplify management of non-implemented handlers

There is just a single entry point for the jtag API .khz(),
.speed_div(), .power_dropout(), .srst_asserted().

Simplify the code by in-lining the default handler.
The overall code behaviour is not changed.
This change prevents modifying at run-time the content of
struct jtag_interface.

Inspired from change http://openocd.zylin.com/943
by Evan Hunter <ehunter@broadcom.com>

Change-Id: I09aeb76d614db57b1884ac7ee9f00c152cd77849
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4892
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

swd: remove unused API frequency()

The specific SWD API to change/query the adapter speed is never
called because the equivalent JTAG API is used in place of it.
In the restructure of struct jtag_interface, the JTAG API is
promoted as global adapter API, thus a specific SWD one is not
anymore required.

Change-Id: I1e810d255b4dfcd5791b4fac8ae1260c31a057fd
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4891
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Convert DEBUG_JTAG_IO to LOG_DEBUG_IO

Change-Id: Ifee9723a57fea93a7022be3299f69680860f236b
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3910
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

gdb_server: fix GDB_BUFFER_SIZE usage, fix unaligned access during bulk transfers

Currently size of the GDB buffer is 16384 bytes but it is treated as
nul-terminated string in most of the code, so effective size of the
buffer is actually 16383 bytes. OpenOCD responds with `PacketSize=3fff`
to qSupported request. Result of GDB's `m` command is encoded in hex so
each data byte uses two bytes in the buffer. As a result GDB will split
bulk read requests into chunks 0x1fff bytes each. This causes troubles
on targets (or memory regions) which support only aligned, word-sized
access (such as MMIO buffers).

Steps to reproduce (psoc6 target):
gdb> dump binary memory dump.bin 0x040320000 (0x040320000 + 65536)

OpenOCD:
Error: Failed to read memory at 0x40321ffe
Error: Failed to read memory at 0x40321000
Error: Failed to read memory at 0x40323000
Error: Failed to read memory at 0x40325ffe
Error: Failed to read memory at 0x40329ffa
Error: Failed to read memory at 0x40329ffc
Error: Failed to read memory at 0x4032bffc
Error: Failed to read memory at 0x4032dffa

Consolidate GDB_BUFFER_SIZE usage: ensure size of each buffer is
(GDB_BUFFER_SIZE + 1), add explicit comment that additional byte is used
for nul-termination. Report correct size of the buffer to GDB (0x4000)
as recommended in GDB's docummentation: `if the stub stores packets in a
NUL-terminated format, it should allow an extra byte in its buffer for
the NUL`

Checked with clang-asan, clang-analyzer, valgrind - no new errors.

Change-Id: I909e8a2c6b010c5d4a304641808d4a807a4ec18d
Signed-off-by: Bohdan Tymkiv <bhdt@cypress.com>
Reviewed-on: http://openocd.zylin.com/5109
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jlink: add usb location support

this patch needs latest extended libjaylink version

Change-Id: Ib688ba18740717ccce7863e08c09425975b5f153
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/4596
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag/drivers/jtag_usb_common: Fix variable name

Change-Id: I3773afa75724dc71801af39cc9135b7b9585bc47
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5154
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

libjaylink: Update to latest Git version

This version introduces jaylink_device_get_usb_bus_ports()

Change-Id: Ib3fea95c3545c1340a6798e100f5ca644ec89510
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-on: http://openocd.zylin.com/4629
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

jtag/drivers/jtag_usb_common: Remove warning

The warning appears for every possible device during enumeration
and provides no useful information for an end-user.

Change-Id: I56e7889a5d9d4656de13ad9e21be3a6e1906e05f
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5155
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

arm_adi_v5: Split CSW bits into AHB/APB/AXI

The implementation-defined bits have different semantics for each bus
and different recommended defaults.

Change-Id: I562fe24643bb1f3abc696339e382a75ccf2f2873
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-on: http://openocd.zylin.com/5124
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target: make handle_md_output() global

Remove a copy of handle_md_output() from src/target/dsp563xx.c

Change-Id: Iadd003fd1dcdbc7990d46a58ee2e7c30826ac6af
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/5175
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

configure.ac: fix minor typo

s/overide/override/

Change-Id: Iecacf0d6839a4d23cb1b029f867d7e1a21d482a4
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5164
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

guess-rev.sh: fix minor typo

s/reecognized/recognized/

Change-Id: I7129090d464d780632c0b7504e48fb90e847b30d
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5163
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

HACKING: fix minor typos

s/additonal/additional/
s/seperately/separately/
Use uppercase after a '.'

Change-Id: I933894ae44924363d634fa07d3de4ceeb4b329e2
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5162
Tested-by: jenkins
Reviewed-by: Jonathan McDowell <noodles-openocd@earth.li>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

cortex_m: set the debug reason to DBGRQ when NVIC_DFSR indicates EXTERNAL

By definition the EXTERNAL bit in Debug Fault Status Register indicates
that an external debug request (EDBGRQ) signal was asserted.

Usage example: this could be done by CTI in multicore devices in order to
halt all the cores together.

Change-Id: I7830455ce5da6702b7d08c8fa7bfe80e4d8a5055
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/5157
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

helper/command: remove dead code in command_unknown()

Commit 89fa493a3bc34d22eeca06fa4e78523ac3b766a8 removes the top
level command "unknown" and adds in command_unknown() some unclear
code that should detect a user-implemented "unknown" command.
But, all the commands that trigger the execution of the function
command_unknown() are registered in register_command_handler() and
have the command name prefixed with "ocd_".
Due to such "ocd_" prefix in cmd_name, the condition
if (strcmp(cmd_name, "unknown") == 0) { ... }
is always false.
Nobody complained for almost 10 years so we can proceed to remove
this dead code, before merging any change that obsoletes the
prefix "ocd_" and potentially makes the condition true.

Remove the dead code guarded by the condition above.

Change-Id: Icf11a956bb2d68fc84e6eb5779edf8e35db8fa53
Fixes: 89fa493a3bc3 ("remove unknown handler")
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5096
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target: remove unused function target_buffer_get_u8()

Left unused after commit "target: unify memory read/write
commands", can be removed.

Change-Id: Iea6ef9204c8071283a66a679b3d6edbb0c929c5c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5153
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

stm32f7x: Use CHIPNAME-specific name for ITCM bank

Change-Id: Icf67eaecd56ac3fb88bcfa2b7084b846109454e6
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/5102
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>

doc/openocd.texi: fix bad aarch64 merge

The documentation added for commit b3d29cb5441ee5d38e8f7b561a58f03eb269dbe4
was merged after the end of the eSi-RISC section rather than AARCH64.
This patch relocates this hunk to the correct location.

Change-Id: I46a2d24442556e9e8000b46a5e1af03b83de6d98
Signed-off-by: Steven Stallion <stallion@squareup.com>
Reviewed-on: http://openocd.zylin.com/5181
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

tcl/target: Fix V3M/V3H SoC chipname

The V3M SoC is R8A77970 while the V3H SoC is R8A77980 . Update the
CHIPNAME and swap the SoCs to keep the list sorted.

Change-Id: I7e1777c0c7181e5e0beac98333f2047cb443d0df
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Reviewed-on: http://openocd.zylin.com/5140
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)