armv7a: read ttbcr and ttb0/1 at every entry in debug state

Commit bfc5c764df145f68835543119865eabe462e19c2 avoids reading
ttbcr and ttb0/1 at every virt2phys translation by caching them,
and it updates the cached values in armv7a_arch_state().
But the purpose of any (*arch_state)() method, thus including
armv7a_arch_state(), is to only print out and inform the user
about some architecture specific status.
Moreover, to reduce the verbosity during a GDB session, the
method (*arch_state)() is not executed anymore at debug state
entry (check use of target->verbose_halt_msg in src/openocd.c),
thus the state of translation table gets out-of-sync triggering
Error: Address translation failure
or even using a wrong address in the memory R/W operation.

In addition, the commit above breaks the case of armv7r by
calling armv7a_read_ttbcr() unconditionally.

Fixed by moving in cortex_a_post_debug_entry() the call to
armv7a_read_ttbcr() on armv7a case only.
Remove the call to armv7a_read_ttbcr() in armv7a_identify_cache()
since it is (conditionally) called only in the same procedure
cortex_a_post_debug_entry().

Fixes: bfc5c764df14 ("armv7a: cache ttbcr and ttb0/1 on debug
state entry")
Change-Id: Ifc20eca190111832e339a01b7f85d28c1547c8ba
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4601
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

server: Improve signal handling under Linux

Commit 5087a955 added custom signal handlers for the openocd
server process.

Before this commit, when openocd is run as a background process
having the same controlling terminal as gdb, Control-C would be
handled by gdb to stop target execution and return to the gdb prompt.

However, after commit 5087a955, the SIGINT caused by pressing
Control-C also terminates openocd, effectively crashing the
debugging session.  The only way to avoid this is run openocd in
a different controling terminal or to detach openocd from its
controlling terminal,
thus losing all job control for the openocd process.

This patch improves the server's handling of POSIX signals:
1) Keyboard generated signals (INT and QUIT) are ignored
   when server process has is no controlling terminal.
2) SIGHUP and SIGPIPE are handled to ensure that .quit functions
   for each interface are called if user's logs out of X
   session or there is a network failure.

SIG_INT & SIG_QUIT still stop openocd
when it is running in the foreground.

Change-Id: I03ad645e62408fdaf4edc49a3550b89b287eda10
Signed-off-by: Brent Roman <genosensor@gmail.com>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/3963
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

target/aarch64: Call aarch64_init_debug_access() earlier in aarch64_deassert_reset()

On Renesas R-Car, calling 'reset halt' and 'reset init' always made DAP inaccessible. Calling 'reset' and 'halt' seperatly worked fine.
The only differences seems to be the point in time when aarch64_init_debug_access() is called. This patch aligns the behaviour.

Change-Id: I2296c65e48414a7d9846f12a395e5eca315b49ca
Signed-off-by: Dennis Ostermann <dennis.ostermann@renesas.com>
Reviewed-on: http://openocd.zylin.com/4607
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/armv7m_trace: Use prefix for enums

Change-Id: I3f199e6053146a1094d96b98ea174b41bb021599
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/3905
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>

target/armv7m_trace: Fix typo in enum

Change-Id: I6364ee5011ef2d55c59674e3b97504a285de0cb2
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/3904
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>

target: Fix segfault for 'mem2array'

Call 'mem2array' without arguments to reproduce the segmentation
fault.

Change-Id: I02bf46cc8bd317abbb721a8c75d7cbfac99eb34e
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/4534
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Christopher Head <chead@zaber.com>

nrf5: Add HWID 0x139 (52832 rev E0)

Change-Id: I71b7471ccfcb8fcc6de30da57ce4165c7fb1f73f
Signed-off-by: James Jacobsson <slowcoder@gmail.com>
Reviewed-on: http://openocd.zylin.com/4604
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Avoid null target->semihosting references.

The new common semihosting code introduced a bug,
in certain conditions target->semihosting was
used without semihosting being initialised.

The solution was to explicitly test for
target->semihosting before dereferencing it.

Change-Id: I4c83e596140c68fe4ab32e586e51f7e981a40798
Signed-off-by: Liviu Ionescu <ilg@livius.net>
Reviewed-on: http://openocd.zylin.com/4603
Tested-by: jenkins
Reviewed-by: Jonathan Larmour <jifl@eCosCentric.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/stm32lx: Add revision 'V' for STM32L1xx Cat.3 devices

Change-Id: Ic92b0fb5b738af3bec79ae335876aa9e26f5f4cd
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/4600
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>

flash/nor/at91sam4: fix sam4sa16c flash banks and its gpnvms count

There was already a github fork that had this fixed, but as we try
to use the latest, non-modified version of all software we use,
I would like to have this fix in the next releases of OpenOCD so
that if people uses $packagemanager, they will not have issues flashing
the last part of the flash of sam4sa16c chips.

Additionally, I've added some more logging related to the flash
bank that was used, and the chip ID that was detected.

Change-Id: I7ea5970105906e4560b727e46222ae9a91e41559
Signed-off-by: Erwin Oegema <blablaechthema@hotmail.com>
Reviewed-on: http://openocd.zylin.com/4599
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

flash/nor: add support for TI MSP432 devices

Added msp432 flash driver to support the TI MSP432P4x and
MSP432E4x microcontrollers. Implemented the flash algo
helper as used in the TI debug and flash tools. This
implemention supports the MSP432E4, Falcon, and Falcon 2M
variants. The flash driver automatically detects the
connected variant and configures itself appropriately.
Added command to mass erase device for consistency with
TI tools and added command to unlock the protected BSL
region.

Tested using MSP432E401Y, MSP432P401R, and MSP432P4111
LaunchPads.
Tested with embedded XDS110 debug probe in CMSIS-DAP
mode and with external SEGGER J-Link probe.

Removed ti_msp432p4xx.cfg file made obsolete by this
patch.
Change-Id: I3b29d39ccc492524ef2c4a1733f7f9942c2684c0
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4153
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

drivers: cmsis-dap: Print version info when available

No need to wait until after connecting, might help diagnose part information by
printing earlier.

Change-Id: I51eb0d584be306baa811fbeb1ad6a604773e602c
Signed-off-by: Karl Palsson <karlp@tweak.net.au>
Reviewed-on: http://openocd.zylin.com/3404
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

drivers: cmsis-dap: pull up common connect code

Just a minor deduplication

Change-Id: Idd256883e5f6d4bd4dcc18462dd5468991f507b3
Signed-off-by: Karl Palsson <karlp@tweak.net.au>
Reviewed-on: http://openocd.zylin.com/3403
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/tcl: Distinguish between sectors and blocks in status messages

Use the right word in flash protect command status messages based on
whether the target bank defines num_prot_blocks. Minor message style
tidy-up.

Change-Id: I5f40fb5627422536ce737f242fbf80feafe7a1fc
Signed-off-by: Dominik Peklo <dom.peklo@gmail.com>
Reviewed-on: http://openocd.zylin.com/4573
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Christopher Head <chead@zaber.com>

nds32: Avoid detected JTAG clock

AICE2 doesn't support scan for the maximum clock frequency of
JTAG chain. It will cause USB command timeout.

Change-Id: I41d1e3be387b6ed5a4dd0be663385a5f053fbcf9
Signed-off-by: Hellosun Wu <wujiheng.tw@gmail.com>
Reviewed-on: http://openocd.zylin.com/4292
Tested-by: jenkins
Reviewed-by: Hsiangkai Wang <hsiangkai@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Adds SAMD11D14AU flash support.

Corrects names of SAMD11D14AM and SAMD11D14ASS per datasheet.

Change-Id: I8beb15d5376966a4f8d7de76bfb2cbda2db440dc
Signed-off-by: Christopher Hoover <ch@murgatroid.com>
Reviewed-on: http://openocd.zylin.com/4597
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/board: add SAMD11 Xplained Pro evaluation board

Change-Id: Id996c4de6dc9f25f71424017bf07689fea7bd3af
Signed-off-by: Peter Lawrence <majbthrd@gmail.com>
Reviewed-on: http://openocd.zylin.com/4507
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target: atmel samd10 xplained mini

cortex m0+ on a tiny board, with an mEDBG (CMSIS-DAP) debug interface.

Change-Id: Iaedfab578b4eb4aa2d923bd80f220f59b34e6ef9
Signed-off-by: Karl Palsson <karlp@tweak.net.au>
Reviewed-on: http://openocd.zylin.com/3402
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

psoc5lp: fix erase check, add free_driver_priv

psoc5lp_erase_check() was not properly adapted to the new
armv7m_blank_check_memory() in the hot fix 53376dbbede4f0bf42e724ff
This change fixes handling of num_sectors in dependecy of ecc_enabled.
Also add comments how ecc_enabled influences num_sectors.

Add pointer to default_flash_free_driver_priv() to all psoc5lp flash
drivers to keep valgrind happy.

Change-Id: Ie1806538becd364fe0efb7a414f0fe6a84b2055b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4569
Tested-by: jenkins

target/stm32f7x: Clear stuck HSE clock with CSS

Change-Id: Ica0025ea465910dd664ab546b66f4f25b271f1f5
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/4570
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

target/image: make i/j unsigned to avoid ubsan runtime error

src/target/image.c:1055:15: runtime error: left shift of 128 by 24 places cannot be represented in type 'int'

Change-Id: I322fd391cf3f242beffc8a274824763c8c5e69a4
Signed-off-by: Cody Schafer <openocd@codyps.com>
Reviewed-on: http://openocd.zylin.com/4584
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Christopher Head <chead@zaber.com>

gdb_server: set current_target from connection's one

In a multi-target environment we are supposed to have a single
gdb server for each target (or for each group of targets within
a SMP node).
By default, the gdb attached to a server sends its command to
the target (or to the SMP node targets) linked to that server.

This is working fine for the normal gdb commands, but it is
broken for the native OpenOCD commands executed through gdb
"monitor" command. In the latter case, gdb "monitor" commands
will be executed on the current target of OpenOCD configuration
script (that is either the last target created or the target
specified in a "targets" command).

Fixed in gdb_new_connection() by replacing the current target
in the connection's copy of command context.

Change-Id: If7c8f2dce4a3138f0907d3000dd0b15e670cfa80
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4586
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Christopher Head <chead@zaber.com>

gdb_server: only trigger once the event gdb-detach at gdb quit

When GDB quits (e.g. with "quit" command) we first execute
gdb_detach() to reply "OK" then, at GDB disconnect (either TCP
or pipe connection type), we execute gdb_connection_closed().
In case GDB is killed or it crashes, OpenOCD only executes the
latter when detects the disconnection.
Both gdb_detach() and gdb_connection_closed() trigger the event
TARGET_EVENT_GDB_DETACH thus getting it triggered twice on clean
GDB quit.

Do not trigger the event TARGET_EVENT_GDB_DETACH in gdb_detach()
and let only gdb_connection_closed() to handle it.

Change-Id: Iacf035c855b8b3e2239c1c0e259c279688b418ee
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4585
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl: board: Add NXP Freedom FRDM-LS1012A config

An update for the K20 CMSIS-DAP firmware can be found here:
https://community.nxp.com/thread/387080?commentID=840141#comment-840141

Change-Id: I149d7f8610aa56daf1aeb95f14ee1bf88f7cb647
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-on: http://openocd.zylin.com/4595
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

tcl: target: Add NXP LS1012A config

As seen on the FRDM-LS1012A board.

Change-Id: Ifc9074b3f7535167b9ded5f544501ec2879f5db7
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-on: http://openocd.zylin.com/4594
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target: armv8: Avoid semihosting segfault on halt

Avoid a NULL pointer dereference when halting an aarch64 core.

Change-Id: I333d40475ab26e2f0dca5c27302a5fa4d817a12f
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-on: http://openocd.zylin.com/4593
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

GDB fileIO stdout support

This patch fixes gdb fileio support to allow gdb console to be used as stdout.

Now we can do something like
gdb <inferior file>

(gdb) tar ext :3333
(gdb) load
(gdb) monitor arm semihosting enable
(gdb) monitor arm semihosting_fileio enable
(gdb) continue

Here: Output from inferior using puts, printf etc will be routed to gdb console.

Change-Id: I9cb0dddda1de58038c84f5b035c38229828cd744
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/4538
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

Add ARM v8 AArch64 semihosting support

This patch implements semihosting support for AArch64. This picks
code from previously submitted AArch64 semihosting support patch
and rebases on top of reworked semihosting code. Tested in AArch64
mode on a Lemaker Hikey Board with NewLib and GDB.

Change-Id: I228a38f1de24f79e49ba99d8514d822a28c2950b
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/4537
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target: fix 'bp' command help message

"asid" and "length" are separate arguments of the command.
Put space between them.

Change-Id: I36cfc1e3a01caafef4fc3b26972a0cc192b0b963
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4511
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/stm32h7: Fix incorrect comment

The name of the bit according to the reference manual is inconsistency
error, not increment error.

Change-Id: Ie3b73c0312db586e35519e03fd1a5cb225673d97
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/4521
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>

flash/nor/stm32: Eliminate working area leak

On a specific early-return path, an allocated working area was not
freed. Free it.

Change-Id: I7c8fe51ff475f191624086996be1c77251780b77
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/4520
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/stm32: Report errors in wait_status_busy

Flash operation errors that occur during algorithm programming are
reported via the algorithm return value. However, Flash operation
errors that occur during non-algorithm work (erasing, programming
without a work area, programming the last non-multiple-of-32-bytes on
an H7, etc.) generally end with a call to stm32x_wait_status_busy,
which reads the status register and clears the error flags but fails
to actually report that something went wrong should an error flag
(other than WRPERR) be set. Return an error status from
stm32x_wait_status_busy in those cases. Correct a log message
accordingly.

Change-Id: I09369ea5f924fe58833aec1f45e52320ab4aaf43
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/4519
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/cortex_m: return error if breakpoint address is out of range

If the "Flash Patch and Breakpoint" unit is rev.1 then it can only
accept breakpoint addresses below 0x1FFFFFFF.
Detailed info in "ARM v7-M Architecture Reference Manual", DDI0403E
at chapter "C1.11 Flash Patch and Breakpoint unit".

Print a message and return error if the address of hardware
breakpoint cannot be handled by the breakpoint unit.

Change-Id: I95c92b1f058f0dfc568bf03015f99e439b27c59b
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4535
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Christopher Head <chead@zaber.com>

jtag/drivers/cmsis-dap: fix connect in cmsis_dap_swd_switch_seq()

The proc cmsis_dap_swd_switch_seq() is part of the SWD API for
this interface driver. It is valid only when the interface is
used in SWD mode.
In this proc there is the need to call, in sequence, first
cmsis_dap_cmd_DAP_Disconnect() then cmsis_dap_cmd_DAP_Connect().
The latter call requires the connection mode as parameter, that
inside cmsis_dap_swd_switch_seq() can only be CONNECT_SWD.

The current implementation is not correct and in some cases can
pass mode CONNECT_JTAG. Moreover, JTAG is optional in CMSIS-DAP
and passing mode CONNECT_JTAG triggers an error with SWD-only
interfaces.

Use mode CONNECT_SWD in SWD specific cmsis_dap_swd_switch_seq().

Change-Id: Ib455bf5b69cb2a2d146a6c8875387b00c27a5690
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4571
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/board: Factor out common R-Car Gen2 code

Factor out the code shared by all R-Car Gen2 boards into a single
file to get rid of the duplication.

Change-Id: I70b302c2e71f4e6fdccb2817dd65a5493bb393d8
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Reviewed-on: http://openocd.zylin.com/4533
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

tcl/board: Add Renesas R-Car R8A7794 E2 Silk board

Add configuration for the Renesas R-Car R8A7794 E2
based Silk evaluation board.

Change-Id: I504b5630b1a2791ed6967c6c2af8851ceef9723f
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
---
NOTE: This requires SW7[1] in position 1 (default is 0)
Reviewed-on: http://openocd.zylin.com/4532
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

tcl/board: Add Renesas R-Car R8A7791 M2W Porter board

Add configuration for the Renesas R-Car R8A7791 M2W
based Porter evaluation board.

Change-Id: Iaadb18f29748f890ebb68519ea9ddbd18e7649af
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Reviewed-on: http://openocd.zylin.com/4498
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

tcl/board: Add Renesas R-Car R8A7790 H2 Stout board

Add configuration for the Renesas R-Car R8A7790 H2
based Stout ADAS board.

Change-Id: Ib880b5d2e1fab5c8c0bc0dbcedcdce8055463fe2
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Reviewed-on: http://openocd.zylin.com/4497
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

tcl/target: Add Renesas R-Car R8A7794 E2 target

Add configuration for the Renesas R-Car R8A7794 E2 target.
This is an SoC with two Cortex A7 ARMv7a cores, both A7
cores are supported.

Change-Id: Ic1c81840e3bfcef8ee1de5acedffae5c83612a5e
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Reviewed-on: http://openocd.zylin.com/4531
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/arm_adi_v5: keep CSW and TAR cache updated

The call to dap_queue_ap_write() can fail and the value in CSW and
TAR becomes unknown.

Invalidate the OpenOCD cache if dap_queue_ap_write() fails.

Change-Id: Id6ec370b4c5ad07e454464780c1a1c8ae34ac870
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4564
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/arm_adi_v5: sync CSW and TAR cache on apreg write

When using apreg to change AP registers CSW or TAR we get internal
cached value not valid anymore.

Reuse the setup functions for CSW and TAR to write them.
Invalidate the cached value before the call to force the write, thus
keeping original apreg behaviour.

Change-Id: Ib14fafd5e584345de94f2e983de55406c588ac1c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4565
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

mips_m4k.c: Fix build with --disable-target64

Replace PRIx64 with TARGET_PRIxADDR to avoid build problems
when --disable-target64 is used during configure.

Change-Id: I054a27a491e86c42c9386a0488194320b808ba96
Signed-off-by: Liviu Ionescu <ilg@livius.net>
Reviewed-on: http://openocd.zylin.com/4566
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Tim Newsome <tim@sifive.com>

target/arm: Add PLD command to ARM disassembler.

Updates the ARM disassembler to handle PLD (PreLoad Data) commands.
Previously handled by printing a TODO message. There are three forms of
the command: literal, register, and immediate. Simply decode based off
of the A1 encoding for the instructions in the ARM ARM. Also fixes mask
to handle PLDW commands.

Change-Id: I63bf97f16af254e838462c7cfac80f6c4681c556
Signed-off-by: James Marshall <jcmarsh@gwmail.gwu.edu>
Reviewed-on: http://openocd.zylin.com/4348
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>

flash/nor/psoc5lp: fix compile issue on GCC 8.1.0

Issue already identified by Alex https://sourceforge.net/u/alexbour/
in ticket #191 https://sourceforge.net/p/openocd/tickets/191/

src/flash/nor/psoc5lp.c:237:2: error: ‘strncpy’ output
truncated before terminating nul copying 2 bytes from a
string of the same length [-Werror=stringop-truncation]

Fix it by assigning the value to the array elements.

Change-Id: I22468e5700efa64ea48ae8cdec930c48b4a7d8fb
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4563
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/target/stm32f7x: configure faster system clock in reset-init

STM32F7xx devices need faster clock for flash programming
over JTAG transport. Using reset default 16 MHz clock
resulted in lot of DAP WAITs and substantial decrease
of flashing performance.

Adapted to the restructured dap support
(see 2231da8ec4e7d7ae9b652f3dd1a7104f5a110f3f).

Change-Id: Ida6915331dd924c9c0d08822fd94c04ad408cdc5
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4464
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>

src/flash/tms470: remove testing of sectors[].is_erased state

The erase check routine checked sectors only if is_erased != 1

Check sector unconditionally.

While on it fix clang static analyzer warnings.

Change-Id: I9988615fd8530c55a9b0c54b1900f89b550345e9
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4401
Tested-by: jenkins

flash/nor/nrf5: remove is_erased setting and autoerase before write

Cached flash erase state in sectors[].is_erased is not reliable as running
target can change the flash.

Autoerase was issued before flash write on condition is_erased != 1
Remove autoerase completely as it is a quite non-standard feature.

Change-Id: I19bef459e6afdc4c5fcaa2ccd194cf05be8a42b6
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4400
Tested-by: jenkins

flash/nor: Add support for TI CC26xx/CC13xx flash

Added cc26xx flash driver to support the TI CC26xx and CC13xx
microcontrollers. Driver is capable of determining which MCU
is connected and configures itself accordingly. Added config
files for four specific variants: CC26x0, CC13x0, CC26x2, and
CC13x2.

Note that the flash loader code is based on the sources used
to support flash in Code Composer Studio and Uniflash from TI.

Removed cc26xx.cfg file made obsolete by this patch.

Change-Id: Ie2b0f74f8af7517a9184704b839677d1c9787862
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4358
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Fredrik Hederstierna <fredrik@hederstierna.com>

psoc5lp: Add NV Latch flash driver

Erasing is not supported by the hardware, it can be written directly.

Tested on CY8CKIT-059, except modifying-write.

Change-Id: I6e920ed930dcd5c7f0b10c5b1b4791a828d9080a
Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/3434
Tested-by: jenkins

flash/nor/psoc5lp: fix bad commit 2d5f2ede55150235352773a976166c3ab68297bc

Change #3432 was merged into git master without adapting it
to #4297 "prepare infrastructure for multi-block blank check".
This is a fast fix of PSoC5LP specific blank check.
Not tested on real PSoC5LP device.

Change-Id: I7dc13ee7bd1f07b2bfe5a93a5030c0c482d30f00
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4557
Tested-by: jenkins

drivers: xds110: Remove unnecessary and deprecated libusb function.

libusb-1.0.22 has deprecated the libusb_set_debug function. This
function was present in the XDS110 driver as an artifact left over
from early debugging. It isn't required because logging is disabled
by default. Removing it to simplify the code and no longer need to
have libusb version specific #if's in the code.

Change-Id: I9c3f16c039da3f3fad9c4a1169978b0d85a1b45c
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4553
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

psoc5lp: Add EEPROM flash driver

Tested on CY8CKIT-059.

Change-Id: Ib02262e8eebf0df3d29492b8a7daa65b262da580
Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/3433
Tested-by: jenkins

flash/nor: Add PSoC 5LP flash driver

Always probe for ECC mode and display ECC sectors if disabled.
Non-ECC write is implemented as zeroing the ECC/config bytes.
Erasing ECC sectors is ignored, erase-checking takes them into account.

Tested with CY8CKIT-059 (CY8C5888), except ECC mode.

Change-Id: If63b9ffca7ad8de038be3c086c49712b629ec554
Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Signed-off-by: Forest Crossman <cyrozap@gmail.com>
Reviewed-on: http://openocd.zylin.com/3432
Tested-by: jenkins

flash/nor: Add support for TI CC3220SF internal flash

Added cc3220sf flash driver to support the TI CC3220SF
microcontrollers. Implemented flash driver to support the
internal flash of the CC3220SF. The implementation does not
support the serial flash of the CC32xx family that requires
connection over UART, and not via JTAG/SWD debug. Added config
files for both CC32xx devices (no flash) and CC3220SF (with
flash).

Updated to implement comments from code review.
Additional updates to handle remaining comments from review.
Additional updates per review.

Added code to only request aligned writes and full 32-bit
words down to flash helper algorithm. Updated for recent
changes in OpenOCD flash code.

Removed cc32xx.cfg file made obsolete by this patch.
Change-Id: I58fc1478d07238d39c7ef02339f1097a91668c47
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4319
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Fix Semihosting FileIO for targets using vcont packet

This patch fixes a bug where target fails to resume after completing GDB FileIO.
We need to update target last run control information to decide resumption. This
was not being done for vcont packets.

Change-Id: I44bea31720f8b877dba97d77a202303d546ea5bd
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/4539
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/target.c: free semihosting member

When destroying the target, if the semihosting
pointer is set, free it.

Change-Id: I07d34918bb3fddab1eee11219dd66f4842708ec1
Signed-off-by: Liviu Ionescu <ilg@livius.net>
Reviewed-on: http://openocd.zylin.com/4552
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

Fix libusb-1.0.22 deprecated libusb_set_debug with libusb_set_option

libusb_set_debug is deprecated in libusb >= 1.0.22 therefore replace
with equivalent libusb_set_option()  API

Change-Id: Ic0287a1bc0ccc90afe8e4c688085f0b25a7fb004
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Reviewed-on: http://openocd.zylin.com/4540
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/cortex_m: constify some variables

Change-Id: I1f3fddd89597333fc1bb9535c0cd76ca3e008324
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/4503
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor/virtual: copy missing fields from master flash_bank structure

Change-Id: I4ac71ad4b46ed613192310d85eb385b727649a11
Signed-off-by: Bohdan Tymkiv <bhdt@cypress.com>
Reviewed-on: http://openocd.zylin.com/4505
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins

flash/nor/core: fix double-free crash with 'virtual' flash banks

flash_bank structure of 'virtual' flash driver is a full copy of
the master flash_bank structure including bank->sectors and
bank->prot_blocks pointers. These pointers point to memory
locations allocated by the master driver and thus master driver
is responsible for deallocating them.
Do not free bank->sectors and bank->prot_blocks of 'virtual'
driver since they were already released by master flash driver.

Change-Id: I01f373d4adb3fc79e2724964926b9276442c5c52
Signed-off-by: Bohdan Tymkiv <bhdt@cypress.com>
Reviewed-on: http://openocd.zylin.com/4504
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

breakpoints: simplify the test to find a breakpoint

The test is overly complicated and unreadable.
Simplify it while keeping the exact same behaviour.

Change-Id: I6b22291ca454e1eddeeab4024d3983dc4c603d3a
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4512
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target: fix syntax in help message

The name of the command is already printed before the string
in ->usage, thus it appears twice.
E.g.
> help examine
  stm32f4x.cpu arp_examine arp_examine ['allow-defer']
        used internally for reset processing
  stm32f4x.cpu examine_deferred examine_deferred
        used internally for reset processing
  stm32f4x.cpu was_examined was_examined
        used internally for reset processing

Remove the command name from the string in ->usage.

Change-Id: If3b1368ffff8a94eb629ae3679e2e5f2f11ae92e
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4536
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target: aarch64: Adding mcr, mrc 32-bit coprocesor read/write support

Adding mrc and mcr support for 32-bit boot mode on aarch64 target.

Change-Id: I3879908253ccdf79509dcad752871f422526ec64
Signed-off-by: Kamal Dasu <kdasu.kdev@gmail.com>
Reviewed-on: http://openocd.zylin.com/4483
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

Rework/update ARM semihosting

In 2016, ARM released the second edition of the semihosting specs
("Semihosting for AArch32 and AArch64"), adding support for 64-bits.

To ease the reuse of the semihosting logic for other platforms
(like RISC-V), the semihosting code was isolated from the ARM
target and updated to the latest specs.

The new code is already in use since January (in GNU MCU Eclipse
OpenOCD) and no problems were reported, neither for ARM nor for
RISC-V targets, after more than 7K downloads.

The 2 new files were formatted with uncrustify.

Change-Id: Ie84dbd86a547323bb8a5d24eab68fc7dad013d96
Signed-off-by: Liviu Ionescu <ilg@livius.net>
Reviewed-on: http://openocd.zylin.com/4518
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

jtag: drivers: sysfsgpio: lift upper gpio number limit

Recent Linux maps GPIOs from 1023 downwards so do not limit the number
to 1000.

This should fix #183.

Change-Id: I6d4f493b670be9ed9b82759f0fb686a9faddbbf5
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/4502
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

jtag/drivers: Add support for TI XDS110 debug probe

Add support for the XDS110 debug probe using the APIs in the
probe's firmware. Includes support for older versions of the
firmware (with reduced performance) and support for a newer
version that includes OpenOCD specific APIs. Tested on various
TI LauchPads including MSP432P4, MSP432E4, CC2650, CC2652, and
CC3220SF.

Updated to add better support for swd switch. Removed issues found with
clang static analysis.

Updated to add rules entry for the XDS110 probe and Tiva DFU mode (which
affects both XDS110 and ICDI probes).

Change-Id: Ib274143111a68e67e80003797c6a68e3e80976b2
Signed-off-by: Edward Fewell <efewell@ti.com>
Reviewed-on: http://openocd.zylin.com/4322
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

Fix warnings exposed by GCC8

gcc (GCC) 8.1.0 generates new warnings and thus fails the build.

The ARM disassembler warnings actually exposed a bug in SMALW, SMULW and
SMUL instructions decoding.

Reported by Eimers on IRC.

Change-Id: I200c70f75a9e07a1f13a592addc1c5fb37714440
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/4526
Tested-by: jenkins
Reviewed-by: Jiri Kastner <cz172638@gmail.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

configure.ac: Fix required libjaylink version

Change-Id: I9ede86eeef8991db52ffa188ff293c14bd50a1f4
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/4450
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>

doc: fix several typos within manual documents

Mostly trivial fixes spotted by spell checker

Change-Id: I3af693faf5506c866e2abe253a0e37aea3c4d284
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4510
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

arm_dpm: flush both scratch registers (R0 and R1)

Neither the initial loop to clear dirty registers (which visits all
registers starting at R2 and counting upwards) nor the final explicit
flushes ensure a write-back in arm_dpm_write_dirty_registers.

This change makes sure that both our scratch registers (i.e. R0 and
R1) are written back to the target.

Change-Id: If65be4f371cd40af9a0cfa97f3730b070b92e981
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-on: http://openocd.zylin.com/4506
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

doc: fix several typos in openocd.texi

Mostly trivial fixes spotted by spell checker
One fix s/are/is/
No changes in the content of the document

Change-Id: Ic2d8696860c540e901e8c5190f8f1e7dce80545f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4402
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

ChibiOS thread states: Update thread state to label mapping

Fixed style issue.
Removed #define with list of strings, and just put the
strings in the array initialization directly.
Removed empty space at the start of line.

Change-Id: I76580be203d7d69b8c5b5440f820156543e0d5cc
Signed-off-by: Faisal Shah <faisal.shah@gmail.com>
Reviewed-on: http://openocd.zylin.com/4488
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

psoc6: Run flash algorithm asynchronously to improve performance

Existing psoc6 driver starts flash algorithm for each Flash row. This is
suboptimal from performance point of view, starting/stopping flash
algorithm for each row adds significant overhead. This change starts
flash algorithm and leaves it running asynchronously while driver
performs flash operations.

Performance gain is 170...250% depending on probe:

flash write_image img_256k.bin    | w/o this change | with this change |
----------------------------------|-----------------|------------------|
KitProg2/CMSIS-DAP, SWD @ 1 MHz   |     4 KiB/s     |     10 KiB/s     |
J-Link Ultra, SWD @ 1 MHz         |    17 KiB/s     |     31 KiB/s     |
J-Link Ultra, SWD @ 4 MHz         |    33 KiB/s     |     57 KiB/s     |

Change-Id: I5bd582584b35af67600c4d197829eb7aeeec7e3f
Signed-off-by: Bohdan Tymkiv <bhdt@cypress.com>
Reviewed-on: http://openocd.zylin.com/4472
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

flash/nor, contrib/loaders: add stm32 loaders Makefile and generated .inc

Flash loaders refactored to the new style - use generated .inc
instead of hexadecimal machine code in the flash driver source.

Change-Id: If65a2099589e210f9450819b467d67819fd841fc
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4439
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

board: add configuration for stm32f103c8 "Blue Pill"

The "Blue Pill" is a popular development board with an STM32F103C8
micro controller. According to sources, it has a 128kB Flash on board
even though the option bytes only report 64kB. This patch therefore also
modifies target/stm32f1x.cfg to take an optional FLASH_SIZE variable into
account which the board file sets to 0x20000.

Change-Id: I8a78ccd2b5faf637c539ee3cf8136789ee15c95d
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4495
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

tcl/target: Add Renesas R-Car R8A7790 H2 target

Add configuration for the Renesas R-Car R8A7790 H2 target.
This is an SoC with four Cortex A15 and four Cortex A7
ARMv7a cores, only the four A15 cores are supported.

Change-Id: I6099b257cc0f04e6858ed5f5f8c8d8ad82ef7650
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Reviewed-on: http://openocd.zylin.com/4490
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

jtag: adapter: fix indentation in handle_interface_command

Minor fix, no code change, just align it to the block it belongs to.

Change-Id: I4c3b0d0bd00a55d5109d3723e5c4bfb2fc72e366
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4492
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target: add Cortex-M4 target to VF6xx target

The Vybrid VF6xx SoCs contain an additional Cortex-M4
core connected to AP number 3 of the main DAP.

Change-Id: I59c020fdfc53e909b1f0dac1a8627a62cdaa74f2
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3640
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

target/cortex_m: allow setting the type of a breakpoint

Cortex-M target used 'auto_bp_type' mode. The requested type
of breakpoint was ignored and hard (FPB) breakpoints were set in
'code memory area' 0x00000000-0x1fffffff, soft breakpoints were set above
0x20000000.

The code memory area of Cortex-M does not mean the memory is flash and
vice versa. External flash (parallel or QSPI) is usually mapped above
code memory area. Cortex-M7 ITCM RAM is mapped at 0. Kinetis
has a RAM block under 0x20000000 boundary.

Remove 'auto_bp_type' mode, set breakpoints to requested type.

Change 'cortex_m maskisr auto' handling to use a hard temporary
breakpoint everywhere: it can also workaround not working soft breakpoints
on Cortex-M7 with ICache enabled.

Change-Id: I7a9f9464c5e10bfd7f17cba1037ed07a064fa2e8
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4429
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

armv8: valgrind memleak fixes

Various fixes for memory leaks, adds a target cleanup for aarch64
and ARM CTI objects.

Change-Id: I2267f0894df655fdf73d70c11ed03df0b8f8d07d
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4478
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target: add configuration for NXP MC-IMX8M-EVK

includes target configuration for i.MX8M SoC family,
board file needs to set up CHIPNAME and CHIPCORES
to match the actual hardware configuration

Change-Id: Ieb6d89cab2477a58f85d0ef9cd242710950191c0
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4434
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

tcl/board: add board configuration for NXP IMX7SABRE

configuration also contains a reset-init procedure that disables the
watchdog and initilizes the boards DDR memory so that you can upload
baremetal (e.g. boot loader) code into DDR and start it from there.

Change-Id: I4d2311b3708a5fcb5174a3447f34ae3904de7243
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4227
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

flash/nor/at91sam: implement flash bank deallocation for SAM series

Microchip (former Atmel) SAM drivers allocate a struct per chip.

at91sam3, at91sam34:
Deallocate all chip structs from the list at once, on the first bank
deallocation.

at91samd and at91sam4l drivers do not handle more than one bank.
Convert them to simple driver_priv allocation and use
default_flash_free_driver_priv().

Change-Id: I49d7200f38a4568c7e12f306c27d1b1b72646736
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4416
Tested-by: jenkins

flash/nor/kinetis: implement flash bank deallocation

Change-Id: I8ef80eae646d3b3eb7f6dd42067f8516adc5abef
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4415
Tested-by: jenkins

flash/nor: handle flash write alignment/padding in the infrastructure

Most of flash drivers have to ensure proper flash write block alignment
and padding. As there was no support for it in the flash infrastructure,
each driver does it its own way. Sometimes this part of code is not properly
tested and contains bugs.

flash_write(_unlock) joins all image sections targeted to one flash bank
using padded areas as a glue. This solves alignment problems on section
boundaries but imposes other problems.

Introduce new flash bank parameters write_start_alignment,
write_end_alignment and minimal_write_gap.
New flash drivers can just properly set these values instead of handling
alignment by its own.

Adapt infrastructure (namely flash_write_unlock(), handle_flash_fill_command()
and handle_flash_write_bank_command()) to prepare write data padded
to an alignment required by the flash bank.

Rework flash_write_unlock() to discontinue write block when the gap
between sections is bigger than minimum specified in minimal_write_gap.
minimal_write_gap is set to one sector by default.

Change-Id: I4368dd402dfaf51c193bcbf1332cffff092b239b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4399
Tested-by: jenkins
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>

target armv7m: multi-block erase check

Tested on PSoC6 (Cortex-M0+ core), onboard KitProg2 in CMSIS-DAP mode,
adapter_khz=1000.
Plain read:
flash read_bank 0 /dev/null
takes 48 seconds.

erase_check without this change:
flash erase_check 0
takes horrible 149 seconds!!

And the same command with the change applied takes 1.8 seconds.
Quite a difference.

Remove the erase-value=0 version of algorithm as the new one can check
for any value.

If the target is an insane slow clocked CPU (under 1MHz) algo
timeouts. Blocks checked so far are returned and the next call
uses increased timeout.

Change-Id: Ic0899011256d2114112e67c0b51fab4f6230d9cd
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4298
Tested-by: jenkins
Reviewed-by: Jonas Norling <jonas.norling@cyanconnode.com>
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>

target, flash: prepare infrastructure for multi-block blank check

'flash erase_check' command runs a check algorithm on a target
if possible. The algorithm is run repeatedly for each flash sector.
Unfortunately every start and stop of the algorithm impose not negligible
overhead.
In practice it means checking is faster than plain read only for
sectors of size approx 4 kByte or bigger. And checking sectors
as short as 512 bytes runs approx 4 times slower than plain read.

The patch changes API call target_blank_check_memory() and related
to take an array of sectors (or arbitrary memory blocks).

Changes in target-specific checking routines are kept minimal.
They use only the first block from the array and process it by
the unchanged algorithm.

default_flash_blank_check() routine repeats target_blank_check_memory()
until all blocks are checked, so it works with both multi-block
and single-block based checkers.

Change-Id: I0e6c60f2d71364c9c07c09416b04de9268807f5e
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4297
Tested-by: jenkins
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>

tcl/target: Add Renesas R-Car R8A7791 M2W target

Add configuration for the Renesas R-Car R8A7791 M2W target.
This is an SoC with two Cortex A15 ARMv7a cores, both cores
are supported.

This patch is based on initial submission by Adam Bass and
improvements by Niklas Söderlund.

Change-Id: I297da62b9ce71ad222a401d98e6bcb8502427673
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Adam Bass <adam.bass@renesas.com>
Cc: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-on: http://openocd.zylin.com/4485
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

HACKING: document practices to improve code quality

Change-Id: I58a7d978b7d5bca3037c4535f06746b9f4411950
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/4343
Tested-by: jenkins

configure: disable all drivers when zy1000 is enabled

This also fixes the transport_is_hla FIXME.

Change-Id: I33960f373f11e3e203f9aed9c6d02bf7ca48ac97
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/4473
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

gdb_server: gdb_memory_map() rework

Use sector sizes instead of bank size.

Detect a gap between sectors and emit xml blocks accordingly.

Detect sector overflow over the bank size.

Change-Id: If0e0e44b0c3b93067b4d717c9c7b07c08582e57b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4436
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

rtos: Fix XPSR_OFFSET for cortex_m4f stacking

Structures rtos_standard_Cortex_M4F_stacking and
rtos_standard_Cortex_M4F_FPU_stacking in rtos_standard_stackings.c
where using rtos_standard_Cortex_M3_stack_align for the stack-align
function. This function calls rtos_Cortex_M_stack_align with
XPSR_OFFSET = 0x3c. This offset is correct for cortex-M3 but not for
cortex-M4F and cortex-M4F with fpu. This patch adds stack_align
functions for M4F an M4F_FPU

Change-Id: If6a90b1898fccbb85619a10f3aef5277dd88ce47
Signed-off-by: Armin van der Togt <armin@otheruse.nl>
Reviewed-on: http://openocd.zylin.com/4037
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target: free target SMP list on shutdown

On SMP targets, the "target smp" command creates a list of targets
that belong to the SMP cluster. Free this list when a target gets
destroyed on shutdown. For simplicity, the complete list is free'd as
soon as the first target of the SMP cluster is destroyed instead of
individually removing targets from the list.

Change-Id: Ie217ae1efb2e819c288ff3b1155aeaf0a19b06be
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4481
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

target/arm_adi_v5: extend apcsw command to accept arbitrary bits

apcsw command was limited to SPROT bit only.

Now user can manipulate any bit except size and addrinc fields.
Can be used e.g. to set bus signal 'cacheable' on Cortex-M7

Change-Id: Ia1c22b208e46d1653136f6faa5a7aaab036de7aa
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4431
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

Add gdb_report_register_access_error command

With this option enabled (it's disabled by default) errors accessing
registers are returned to gdb. Otherwise they are ignored and success is
reported to gdb. (This is the current behavior.)

We want this for RISC-V, but there's still some cleanup that needs to be
done before that can be upstreamed.

Signed-off-by: Tim Newsome <tim@sifive.com>
Change-Id: I7e56109ea52d18b780c14a07fb35f9e6e8979da4
Reviewed-on: http://openocd.zylin.com/4452
Reviewed-by: Steven Stallion <sstallion@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Tested-by: jenkins

arm_adi_v5: Add ability to ignore the CSYSPWRUPACK bit

The CTRL/STAT register in the ARM DAP DP has a debug power up
ack bit and a system power up ack bit. Some devices do not set
the system power up ack bit until sometime later. To avoid having
the initial target examination fail due to this or to have a
sticky bit error report claim power failure due to this a user
can now specify that this bit should be ignored.

Change-Id: I2451234bbe904984e29562ef6f616cc6d6f60732
Signed-off-by: Eric Katzfey <eric.katzfey@mentalbee.com>
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3710
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>

flash/nor/at91samd: Add "nvmuserrow" command.

Add option "nvmuserrow" to "at91samd" for changing and reading the register at 0x804000 which represents various fuses.

Change-Id: I6382cc4ac15e6b9681e2f30b0ae60397a6289c3b
Signed-off-by: Stefan Arnold <sarnold@sh-sw.de>
Reviewed-on: http://openocd.zylin.com/4260
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

nrf51: Add HWID 0x008F again

HWID originally added in commit 7829f31a6dd61297e97d8e94fe98a1658eac833e
was accidentally omited during refactoring in commit
52885d2b538dcd4184aae14cf2706fb97acccbd9

While on it move old ingeneering sample of 51822 to block of 51822 rev 1

Change-Id: Ie9f15563792a27a72e71df6edbcc6b04490370ed
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4437
Tested-by: jenkins

drivers/kitprog: workaround KitProg firmware bug of missing ZLP

KitProg firmware does not send a zero length packet at the end of the bulk-in
transmission of a length divisible by a bulk packet size. This is inconsistent
with the USB specification and results in jtag_libusb_bulk_read()
waits forever when a transmission of specific size is received.

Limit bulk read size to expected number of bytes for problematic tranfer sizes.
Use 1 second timeout as the last resort.

Change-Id: Ice80306424afd76e9fbc6851911ffd5109c84501
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4426
Tested-by: jenkins
Reviewed-by: Bohdan Tymkiv <bhdt@cypress.com>

target/cortex_m: avoid dwt comparator overflow

Avoid ever overflowing the DWT_COMPARATOR array by allocating space for
16 comparators (the field is masked by 0xf).

On a stm32f767zi chip (on a nucleo-767zi board) I've been seeing crashes
with address sanitizer enabled due to its (apparent) 10 present
comparators. This appears to be due to
https://sourceforge.net/p/openocd/tickets/178/.

In non-address sanitizer builds, this would likely cause some random
memory to be written to in some cases. (see above bug for observations).

Change-Id: I2b7d599eb326236dbc93f74b350c442c9a502c4b
Signed-off-by: Cody P Schafer <openocd@codyps.com>
Reviewed-on: http://openocd.zylin.com/4458
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>

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)