Now internal watch/breakpoint will not be removed in case
of error during removing triggers from hardware.
Also change signature of some functions (for deletion
bp/wp) to print message in case of some error.
Change-Id: I71cd1f556a33975005d0ee372fc384fddfddc3bf
Signed-off-by: Kirill Radkin <kirill.radkin@syntacore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7738
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Previously, an unknown semihosting operation number
was logged as debug. This patch changes it and few
other places to be logged as error instead.
Change-Id: I83cae5ca1e3daed440f92b08bd372bfffbbad63c
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
This patch disables software breakpoints of size 2 for targets
which don't support compressed instructions.
Change-Id: I8200b22a51c97ba2aa89e6328beadde8dd35cdd5
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
Previously, if a pin was configured as ADAPTER_GPIO_INIT_STATE_INPUT and
its drive value was ADAPTER_GPIO_DRIVE_MODE_PUSH_PULL, initialize_gpio
would configure the pin as an output.
The set_gpio_value function is optimized to not set the direction for
pins configured as ADAPTER_GPIO_DRIVE_MODE_PUSH_PULL as it only needs to
be set once. When initialize_gpio performs this setup, it checked only
that the drive value was ADAPTER_GPIO_DRIVE_MODE_PUSH_PULL to set the
output direction but did not exclude input pins which have already had
their direction set.
Now, input pins are ignored when initialize_gpio checks for
ADAPTER_GPIO_DRIVE_MODE_PUSH_PULL to set the mode to output.
Change-Id: I4fc7a8132a6b00c7f213ec9fd05c7bbb37ee5f20
Fixes: 0dd969d83b ("drivers/bcm2835gpio: Migrate to adapter gpio commands")
Signed-off-by: Brandon Pupp <bpupp@xes-inc.com>
[vfazio: update commit message]
Signed-off-by: Vincent Fazio <vfazio@xes-inc.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7862
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
add new i2c bit-banging feature, we can now connect in JTAG with the SoC
target and in i2c with the main board components at the same time.
Change-Id: I8e4516fe1ad5238e0373444f1c3c9bc0814d0f52
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7796
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
add a line that checks the returned value of set signals function
add two VIDs of other original boards (have onboard angie architecture)
so angie driver can connect to them and change their VID after
renumeration.
Change-Id: Ide4f1f6f38168a410191bf3ff75bcd59dcf7ef50
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7795
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Introduce the ability to detect CPUs based on CP0 PRId register and
apply cpu specific quirks, which alter the default ejtag behavior.
First of those is EJTAG_QUIRK_PAD_DRET, which makes sure extra NOPs are
placed after the DRET instruction on exit from debug mode. This fixes
resume behavior on Ingenic JZ4780 SoC.
The proper detection of some (currently unsupported) CPUs becomes quite
complicated, so please consult the following Linux kernel code when
adding new CPUs:
* arch/mips/include/asm/cpu.h
* arch/mips/kernel/cpu-probe.c
Change-Id: I0f413d5096cd43ef346b02cea85024985b7face6
Signed-off-by: Artur Rojek <contact@artur-rojek.eu>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/7859
Tested-by: jenkins
stlink v2 on Nucleo-64 board (e.g. NUCLEO-L476RG) has target SWO signal
connected to STM32F103CB'S PA10, which is UART1_RX. UART1 within this
MCU in theory can be configured to 4.5 Mbps baudrate, which means this
is the upper limit supported by HW. As a confirmation BMP (Black Magic
Probe) project also states in documentation that UART1 can be used with
up to 4.5 Mbps baudrate.
Tests have shown that configuring 4.5 Mbps baudrate on stlink v2
available on NUCLEO-L476RG board results in receiving corrupted data.
Using 2.25 Mbps however allows to successfully receive all data from
SWO. This makes sense in terms of STM32F103CB capabilities, since 2.25
Mbps is the next supported baudrate due to division by 2.
Increase supported stlink v2 SWO speed from 2 to 2.25 Mbps.
Tested with NUCLEO-L476RG:
$ stm32l4x.tpiu configure -protocol uart \
-traceclk 80000000 -pin-freq 2250000 \
-output /dev/stdout
$ stm32l4x.tpiu enable
2.25 Mbps speed confirmed with logic analyzer.
Signed-off-by: Marcin Niestroj <m.niestroj@emb.dev>
Change-Id: Icbec04585664aba8b217e8f9a75458e577f7617f
Reviewed-on: https://review.openocd.org/c/openocd/+/7848
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
This emulation mode supports software translation of an AP request
into an address mapped transaction that does not rely on physical AP
hardware. This is necessary in some hardware such as K3 SoCs since the
hardware architecture anticipates a potential race condition between
AP doing direct memory access generating transactions back to system
bus and firewalls that data path out.
This emulation mode allows direct memory driver to emulate CoreSight
Access Port (AP) and reuse the SoC configuration meant for JTAG
debuggers.
Since the address ranges are flat in nature, the requisite memory base
and size will need to be provided a-priori to the driver for mapping.
The other design alternative would be to map requested memory map for
every register operation, but, that would defeat our intent of getting
max debug performance.
Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Jason Peck <jpeck@ti.com>
Change-Id: I2d3c5f7833f1973e90b4f6b247827f62fc2905d0
Reviewed-on: https://review.openocd.org/c/openocd/+/7089
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Direct memory driver support for CoreSight Access Port(AP).
Even though we emulate SWD (serial wire debug), we aren't actually
using swd. Instead, we are using a direct memory access to get to the
register set. This is similar in approach to other fast access native
drivers such as am335xgpio drivers.
Example operation on Texas Instrument's AM62x K3 SoC:
+-----------+
| OpenOCD | SoC mem map
| on |--------------+
| Cortex-A53| |
+-----------+ |
|
+-----------+ +-----v-----+
|Cortex-M4F |<───────| |
+-----------+ | |
| DebugSS |
+-----------+ | |
|Cortex-M4F |<───────| |
+-----------+ +-----------+
Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Jason Peck <jpeck@ti.com>
Change-Id: I8470cb15348863dd844b2c0e3f63a9063cb032c6
Reviewed-on: https://review.openocd.org/c/openocd/+/7088
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Since zephyrproject-rtos/zephyr@c3eeae8,
Zephyr OS exposes offset of mode_exc_return in the arch struct for ARM.
Accounting for this allows for consistency and enables
logic with further offsets that may be added after this.
Signed-off-by: Bruno Mendes <bd_mendes@outlook.com>
Change-Id: Id53ebd80c5d98a7d94eb6b00ad638ce51e719822
Reviewed-on: https://review.openocd.org/c/openocd/+/7851
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
These cores are advertised as M23 and M33 compatible, but are identified
by the Realtek implementor id. These cores are found on the RTL872xD
family, at least.
Raw CPUIDs:
Real-M200 (KM0): 721cd200
Real-M300 (KM4): 721fd220
Change-Id: I4106ccb7e8c562f98072a71e9e818f57999d664e
Signed-off-by: Karl Palsson <karlp@tweak.au>
Reviewed-on: https://review.openocd.org/c/openocd/+/7846
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Presently, we only look at the Part Number field of the CPUID, and
completely ignore the Implmentor field, simply assuming it to be ARM.
Parts have since been found, with different implementors, that use
overlapping part numbers, causing detection to fail.
Expand the "part number" field to be a full implementor+part number,
excluding the revision/patch fields, to make checking more reliable.
Change-Id: Id81774f829104f57a0c105320d0d2e479fa01522
Signed-off-by: Karl Palsson <karlp@tweak.au>
Reviewed-on: https://review.openocd.org/c/openocd/+/7845
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
There's really no reason to try and add an extra layer of cpu
verification here.
Change-Id: If8c4aa03754607be6c089f514ae300b09b067ffa
Signed-off-by: Karl Palsson <karlp@tweak.au>
Reviewed-on: https://review.openocd.org/c/openocd/+/7844
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
During a previous patch, the ignoring of writes to register zero
was deleted. This patch restores it to the original.
Change-Id: Ieb028a5b2e3f691e4847713c7bc809e10726e18c
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
This also fixes a bug when, after `examine` completion, the target still
has `unknown` status. To reproduce this one spike, it is enough to do
the following:
---
// make sure spike harts are halted
openocd ... -c init -c 'echo "[targets]"'
---
this behavior is quite dangerous and leads to segfaults in some cases
Change-Id: I13915f7038ad6d0251d56d2d519fbad9a2f13c18
Signed-off-by: Parshintsev Anatoly <anatoly.parshintsev@syntacore.com>
This patch improves the following issues:
1. Makes it compatible with targets with progbufsize == 1.
2. Although exceptions don’t update any registers, but do end execution
of the progbuf. This will make fence rw, rw impossible to execute.
Change-Id: I2208fd31ec6a7dae6e61c5952f90901568caada6
Signed-off-by: Xiang W <wxjstz@126.com>
The motivation for this refactor is to fixup error handling for some
corner cases. These functions attempt to cache S0 register and only then
perform a bunch of extra checks to figure out if the requested register
is valid one in this context. The problem is that there are few corner
cases when _*progbuf functions could receive a GPR as an input. For
example, an abstract read could fail (for whatever reason) leading to
infinite recursion:
````
save S0 -> read S0 -> save S0 -> read S0 -> ...
```
The case described above could be fixed by adding extra sanitity checks,
however I decided to make these functions more modular since I find
self-contained functions easier to read.
Change-Id: I01f57bf474ca45ebb67a30cd4d8fdef21f307c7d
Signed-off-by: Parshintsev Anatoly <anatoly.parshintsev@syntacore.com>
Use blocks (64 KiB) instead of sectors (4 KiB) when erasing the zd25Q16
SPI flash memory (thanks to Tomas Vanek!)
Change-Id: I969a69ad35f51b84eb3e11b93f0d79db3e98613a
Signed-off-by: Nikolay Dimitrov <nikolay.dimitrov@retrohub.org>
Reviewed-on: https://review.openocd.org/c/openocd/+/7850
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins
Add missing aarch64_poll() calls to ensure the event
TARGET_EVENT_HALTED is called when necessary.
This is needed with the poller update introduced in commit
95603fae18 ("openocd: revert workarounds for 'expr' syntax change")
Signed-off-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>
Change-Id: I6e91f1b6bc1f0d16e6f0eb76fc67d20111e3afd2
Reviewed-on: https://review.openocd.org/c/openocd/+/7737
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
This is the driver code for NanoXplore's ANGIE USB-JTAG Adapter.
The driver is based on the openULINK project.
This driver communicate with ANGIE's firmware in order to establish
JTAG protocol to debug the target chip.
Since the ANGIE Adapter has a Spartan-6 FPGA in addition to the
FX2 microcontroller, the driver adds two functions, one to download
the firmware (embedded C) to the FX2, and the second to program
the FPGA with its bitstream.
Add ANGIE's configuration file to tcl/interface/
Add the device VID/PID to 60-openocd.rules file.
Add ANGIE to OpenOCD's documentation
Change-Id: Id17111c74073da01450d43d466e11b0cc086691f
Signed-off-by: Ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7702
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
The USB control transfer can be executed without any data.
The libusb API libusb_control_transfer() can thus be called with
zero 'size', thus returning zero byte transferred when succeed.
The OpenOCD API jtag_libusb_control_transfer() returns zero either
in case of transfer error and in case of libusb_control_transfer()
returning zero, making impossible discriminating the two cases.
Extend jtag_libusb_control_transfer() with separate return value
for error code and explicit parameter's pointer for transferred
bytes.
Make the transferred pointer optional, as many callers do not
properly handle the returned value.
Use 'int' type pointer for transferred, instead of the 'uint16_t'
that would have matched the type of 'size'. This can simplify the
caller's code by using a single 'int transferred' variable shared
with other jtag_libusb_bulk_read|write, while keeping possible the
comparison int vs uint16_t without cast.
This change is inspired from commit d612baacaa
("jtag_libusb_bulk_read|write: return error code instead of size")
Change-Id: I14d9bff3e845675be03465c307a136e69eebc317
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7756
Tested-by: jenkins
Reviewed-by: ahmed BOUDJELIDA <aboudjelida@nanoxplore.com>
This patch changes data types of watchpoint value and mask to allow for
64-bit values match that some architectures (like RISCV) allow.
In addition this patch fixes the behavior of watchpoint command to
zero-out mask if only data value is provided.
Change-Id: I3c7ec1630f03ea9534ec34c0ebe99e08ea56e7f0
Signed-off-by: Parshintsev Anatoly <anatoly.parshintsev@syntacore.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7840
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Marek Vrbka <marek.vrbka@codasip.com>
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
register_cache_invalidate() is written a way which uses
pointer arithmetic, which makes it harder to read. This patch
replaces it with more readable way to iterate over array of
structs.
Change-Id: Ia420f70a3bb6998c690c8c600c71301dca9f9dbf
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7735
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tested-by: jenkins
Reviewed-by: Jan Matyas <jan.matyas@codasip.com>
OpenOCD fails in the presence of inactive/unresponsive cores
I faced with case when inactive core returns 0 while reading dtmcontrol.
This leads to failure on assert: "addrbits != 0" in "dbus_scan".
Also change "read_bits","poll_target" funcs to avoid a lot lines in logs
Change-Id: If852126755317789602b7372c5c5732183fff6c5
Signed-off-by: Kirill Radkin <kirill.radkin@syntacore.com>
Lot of messages was logged as LOG_WARNING, but the operation failed
immediately. Sometimes no error message was logged at all.
Add missing messages, change warnings to errors.
Sometimes ERROR_TARGET_INVALID was returned. Some command handlers
returned ERROR_OK! Always return ERROR_TARGET_NOT_HALTED.
While on it use LOG_TARGET_ERROR() whenever possible.
Prefix command_print() message with 'Error:' to get closer
to LOG_TARGET_ERROR() variant.
Error message was not added to get() and set() methods of
struct xxx_reg_type - the return value is properly checked and a message
is logged by the caller in case of ERROR_TARGET_NOT_HALTED.
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Change-Id: I2fe4187c6025f0038956ab387edbf3f461c69398
Reviewed-on: https://review.openocd.org/c/openocd/+/7819
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Support assign DMI address of the debug module by pass
-dbgbase to the target create command
Change-Id: I774c3746567f6e6d77c43a62dea5e9e67bb25770
Signed-off-by: Mark Zhuang <mark.zhuang@spacemit.com>
Compiler would complain that `written` was used without being
initialized.
Simplify the code a little. The number of bytes written is already
checked in usb_write().
Signed-off-by: Tim Newsome <tim@sifive.com>
Change-Id: Ibada85dcccfca6f1269c584cdbc4f2e3b93bb8f3
Reviewed-on: https://review.openocd.org/c/openocd/+/7813
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Jan Matyas <jan.matyas@codasip.com>
This patch adds target logging to logging instances where it makes sense.
This is especially useful when debugging multiple targets at once,
such as multicore systems.
Change-Id: Ia9861f3fa0e6e5908b683c2a8280659c3c264395
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
While not affecting the function's main purpose, an error has
crept into arc_save_context() that results in logging wrong register
values when the debug level is 3 or more. For instance, when debugging a
trivial program and halting at entry to main, the following values are
printed to the log:
Debug: 2915 2020 arc.c:894 arc_save_context(): Get core register regnum=0,
name=r0, value=0x0000000
...
Debug: 2947 2020 arc.c:894 arc_save_context(): Get core register regnum=60,
name=lp_count, value=0x900002d8
Debug: 2948 2020 arc.c:894 arc_save_context(): Get core register regnum=63,
name=pcl, value=0xffffffff
Debug: 2949 2020 arc.c:909 arc_save_context(): Get aux register regnum=64,
name=pc, value=0x900000b4
Debug: 2950 2020 arc.c:909 arc_save_context(): Get aux register regnum=65,
name=lp_start, value=0x900000bc
Debug: 2951 2020 arc.c:909 arc_save_context(): Get aux register regnum=66,
name=lp_end, value=0x00080801
Debug: 2952 2020 arc.c:909 arc_save_context(): Get aux register regnum=67,
name=status32, value=0xffffffff
After the change, the register contents make much more sense:
Debug: 2923 3934 arc.c:889 arc_save_context(): Get core register regnum=0,
name=r0, value=0x00000000
...
Debug: 2955 3934 arc.c:889 arc_save_context(): Get core register regnum=60,
name=lp_count, value=0x00000000
Debug: 2956 3934 arc.c:889 arc_save_context(): Get core register regnum=63,
name=pcl, value=0x900002d8
Debug: 2957 3934 arc.c:903 arc_save_context(): Get aux register regnum=64,
name=pc, value=0x900002da
Debug: 2958 3934 arc.c:903 arc_save_context(): Get aux register regnum=65,
name=lp_start, value=0x900000b4
Debug: 2959 3934 arc.c:903 arc_save_context(): Get aux register regnum=66,
name=lp_end, value=0x900000bc
Debug: 2960 3934 arc.c:903 arc_save_context(): Get aux register regnum=67,
name=status32, value=0x00080801
While at it, simplify a couple of expressions.
Change-Id: I8f2d79404707fbac4503af45b393ea73f91e6beb
Signed-off-by: Artemiy Volkov <artemiy@synopsys.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7765
Tested-by: jenkins
Reviewed-by: Evgeniy Didin <didin@synopsys.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Besides checkpatch, now upstream codes are scanning with
Sparse semantic checker tool.
This commit addresses some Sparse and checkpatch warnings.
Change-Id: I0e3e9f15220d8829c5708897af27aa86a8f90c07
Signed-off-by: Erhan Kurubas <erhan.kurubas@espressif.com>
Below warnings are fixed.
1- A function declaration without a prototype is deprecated in all
versions of C [-Werror,-Wstrict-prototypes]
2- error: variable set but not used [-Werror,-Wunused-but-set-variable]
Signed-off-by: Erhan Kurubas <erhan.kurubas@espressif.com>
Change-Id: I1cf14b8e5e3e732ebc9cacc4b1cb9009276a8ea9
Reviewed-on: https://review.openocd.org/c/openocd/+/7569
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
* Add a warning when eq trigger is setup and it's behavior is different
from other triggers.
* Make eq trigger's behavior consistent with other triggers in case of
length == 1.
* Fix a bug in setting chained triggers (LT, GT case).
* Improve logging.
Change-Id: Id1ed0d11971b8ed875afbb979e6c8a8b51dd3818
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
There were a couple of problems with previous implementation:
* Misalligned read would return ERROR_OK and print all zeroes.
* CMDERR_BUSY for abstract access was improperly handled:
According to the spec, no assumptions can be made about DM_DATA*
contents in such a case, but these were considered valid values from
memory.
* A fallback to one element read was implemented when DMI_STATUS_BUSY
occurred during batch reads, even though this can be accounted for.
Change-Id: I09174c61c951b2bb97a529b7f0aa5afaa995179b
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Memory region addresses are not in use for now.
Signed-off-by: Erhan Kurubas <erhan.kurubas@espressif.com>
Change-Id: I9a2189e956ae59b56245ec914ab16719df857b2d
Reviewed-on: https://review.openocd.org/c/openocd/+/7762
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Right after target halt, some activities needs to be done
such as printing exception reason, disable wdts and reading
debug stubs information.
Missing activities will be submitted in the next patches.
Signed-off-by: Erhan Kurubas <erhan.kurubas@espressif.com>
Change-Id: I27aad5614d903f4bd7c8d6dba6bfb0bdb93ed8dc
Reviewed-on: https://review.openocd.org/c/openocd/+/7757
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
See section 57.6.1 in RM0432.
Change-Id: Ic4977aee74d1838f420c1d9ff19925d09f8f6e2b
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/7763
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
1. update RISCV_MAX_HARTS to 2^20 according to SPEC
2. remove RISCV_MAX_REGISTERS, it's not used anywhere anymore
3. add parentheses
Change-Id: Iadf0fa1ba3bbe5b9420b8430883e140db87f4f9e
Signed-off-by: Mark Zhuang <mark.zhuang@spacemit.com>
If a target where a software breakpoint was set is not currently
available, but there are other targets in the same SMP group that are
available, then we can use those to remove the software breakpoint.
Change-Id: I9faa427c7b3aee31504e6e6599539e6f29b58d8f
Signed-off-by: Tim Newsome <tim@sifive.com>
STLINK-V3PWR is both a standalone debugger probe compatible with
STLINK-V3 and a source measurement unit (SMU).
Link: http://www.st.com/stlink-v3pwr
This code adds support for the debugger probe functionality.
Change-Id: Ib056e55722528f922c5574bb6fbf77e2f2b2b0c1
Signed-off-by: Laurent LEMELE <laurent.lemele@st.com>
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7755
Tested-by: jenkins
To start a ipdbg server one needs to know the tap and the
instruction code to reach the IPDBG-Hub. This instruction is
vendor/family specific. Knowledge which can be provided by the
pld driver.
Change-Id: I13eeb9fee895d65cd48544da4704fcc9b528b869
Signed-off-by: Daniel Anselmi <danselmi@gmx.ch>
Reviewed-on: https://review.openocd.org/c/openocd/+/7369
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
According to section 5.6 in the RISC-V debug specification, the previous
way to set triggers was incorrect, as was discussed as part of
https://github.com/riscv/riscv-openocd/issues/870. This commit fixes the
sequence to be in line with the specification as well as adds some comments
to clarify for any future reader as to what is actually done.
Change-Id: Iffc5cc0f866a466a7aaa72a4c53ee95c9080ac9d
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
this functionality allows to query if a target belongs to some smp group
and to dynamically turn on/off smp-specific behavior
Change-Id: I469453d95e7c1640a91bc60d80c854404e508535
Signed-off-by: Parshintsev Anatoly <anatoly.parshintsev@syntacore.com>
This patch fixes the handling of errno by setting the sys_errn
only if error has actually occurred during the semihosting call.
It also fixes few issues where error was not set in the first place.
Change-Id: I2fbe562f3ec5e6220b800de04cd33aa1f409c7a0
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7730
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Jan Matyas <jan.matyas@codasip.com>
This patch introduces function semihosting_opcode_to_str() which
converts semihosting opcodes to strings. This function is then
used in debug messages to improve log analysis and troubleshooting.
Change-Id: Iffea49dae13d6a626ae0db40d379cba3c9ea5bd3
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7726
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Jan Matyas <jan.matyas@codasip.com>
* rtos/FreeRTOS: solve some conflicting usage of thread id.
1.
There are some RISCV-specific usage of thread_id, which has conflict with upstream.
Some adaptions are made in this patch, to make sure OpenOCD is sending a clear thread list to gdb.
2.
Use freertos_read_struct_value for xSchedulerRunning.
Change-Id: I001a88a0c6d8eac98a389c0217b4897f28124840
Signed-off-by: Chao Du <duchao@eswincomputing.com>
* fix typo.
Change-Id: Id6546cc74de44bbee7e44b7cb29b769a2f35ec4a
* correct the data type.
Change-Id: I28c7e111e569d94ba5f6e1ae21745ddb34d4dd12
* changes as per the review comment.
Change-Id: Ica4c705a8f2657700dc27e24790287ca802480fd
* another macro replacement.
Change-Id: Ia9330fed32d917cf87804051ba1b8d6ac42cfb7b
---------
Signed-off-by: Chao Du <duchao@eswincomputing.com>
With this change, failures to resume a hart due to it not being halted
are more explicitly logged or reported as an error.
Change-Id: Ia55d8df85a908363d0f2140637ce1e47c1ab6251
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
This involves halting the target, which might have unintended side
effects, but when the debugger is connected software breakpoints must
trap to the debugger. Anything else is a terrible user experience.
Change-Id: I1f7bb610eeeb054cc3042dc6bcfc16589ce12a31
Signed-off-by: Tim Newsome <tim@sifive.com>
Intended as a place where we can interact with the target without too
much concern about preserving state and doing exactly the right thing
while poll() is going on.
Change-Id: Ic9bd441caae85901a131fd45e742599803df89b5
Signed-off-by: Tim Newsome <tim@sifive.com>
Specifically, call into the RISC-V version when target becomes halted,
running, or unavailable.
I'll be using unavailable shortly.
Change-Id: I9ffffdccbf22e053fe6390d656b362bf9ab9559a
Signed-off-by: Tim Newsome <tim@sifive.com>
Will be used later when we want to do a quick halt/resume.
Change-Id: Ib80166234c4c277b7d9ce26b7566ac0f93017e64
Signed-off-by: Tim Newsome <tim@sifive.com>
* Only set ebreak bits that might be supported based on misa.
* Don't write dcsr if its value wouldn't change.
Change-Id: I7087af0b0df0fbdbf994373b5c887b9b389df872
Signed-off-by: Tim Newsome <tim@sifive.com>
Make it callable earlier, handle `supported` being NULL, and make enum
names more clear.
Change-Id: If4d286b54ccfc01eb5de5a57eb18f748c920e979
Signed-off-by: Tim Newsome <tim@sifive.com>
The riscv013_on_halt function was being called but its implementation was
empty, providing no additional functionality. Removed the function declaration,
calls to it, and its implementation since it is not required.
Change-Id: I425ea890deadeec945f0a47af247f3f99172e801
Signed-off-by: Tim Newsome <tim@sifive.com>
Otherwise I get a compiler warning, which fails the build.
Change-Id: Ib7d4ab85160b537d07c74f8651ac42906fd661ed
Signed-off-by: Tim Newsome <tim@sifive.com>
When halted we don't need to read all 3 instructions before deciding the
sequence doesn't match.
Change-Id: I9f8345960ce27e859265af901a368166a70b9fde
Signed-off-by: Tim Newsome <tim@sifive.com>
* rtos/FreeRTOS: pxCurrentTCB should be used for judgment.
The current TCB is stored in pxCurrentTCB, which is somehow RISC-V-specific, should not be overwritten from upstream (#816).
* fix the code style check.
Signed-off-by: Chao Du <duchao@eswincomputing.com>
Change-Id: I9ffa8947f0cb9e93c7d96866882a5a1e8e69afad
* revert some over-changes in last commit.
Change-Id: Ie88bd75b59190503db11ee4538281bd13b554e50
Signed-off-by: Chao Du <duchao@eswincomputing.com>
---------
Signed-off-by: Chao Du <duchao@eswincomputing.com>
The original document from Jedec does not report the entry for
"21 NXP (Philips)", replaced by "c". It's clearly a typo.
Keep the line from JEP106BF.01 for "NXP (Philips)".
Change-Id: I30215c4ff08d5f112305cde6ab7a3176cdcef948
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7727
Tested-by: jenkins
On Windows, isatty() can return any non-zero value if it's an interactive
device. Which diverges from the ARM semihosting specification. This patch
introduces a fix to make the SYS_ISTTY operation conform to spec.
Change-Id: I9bc4f3cb82370812825d52419851910b3e3f35cc
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7725
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Reviewed-by: Jan Matyas <jan.matyas@codasip.com>
The old implementation of gdb socket error handling
in the gdb_get_char_inner() differs between Windows and *nix
platforms. This patch simplifies it by using an existing
function log_socket_error() which handles most of the platform
specific things. It also provides better error messages.
Change-Id: Iec871c4965b116dc7cfb03c3565bab66c8b41958
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7724
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Added debug prints to show what is the target debug reason. Also added
debug print for Ctrl-C response. This is useful for troubleshooting and
log analysis.
Change-Id: I055936257d989efe7255656198a8d73a367fcd15
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7720
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
This patch adds support for the NXP QN908x family of Bluetooth
microcontrollers, such as the QN9080. This chip features a Cortex-M4F
with 512 KiB of flash on all the available versions, although the
documentation suggests that there might be 256 kB versions as well.
The initial support allows to read, erase and write the whole user flash
area. Three new sub-commands under the new "qn908x" command are added
in this patch as well: disable_wdog to disabled the watchdog,
mass_erase to perform a mass erase and allow_brick to allow programming
images that disable the SWD interface.
Disabling the watchdog is required after a "reset halt" in order to run
the CRC algorithm from RAM when verifying the chip. However, this is not
done automatically on probing or other initialization since disabling
the watchdog might interfere with debugging real applications.
The "mass_erase" command allows to erase the whole flash without
probing it, since in some scenarios the chip can be locked such that no
flash or ram can be accessed from the SWD interface, allowing only to
run a mass_erase to be able to flash the program.
The flashing process allows to compute a checksum, similar to the
lpc2000 driver "calc_checksum" but done over a different region of the
memory. This checksum is required to be present for the QN908x
bootloader ROM to boot, and otherwise is useless. As with the lpc2000
design, verification when using "calc_checksum" is expected to fail if
the checksum was not valid in the image being verified.
This was manually tested on a QN9080, including the scan-view,
AddressSanitizer/UBSan and test coverage configurations.
Change-Id: Ibd6d8f3608654294795085fcaaffb448b77cc58b
Co-developed-by: Marian Buschsieweke <marian.buschsieweke@ovgu.de>
Signed-off-by: Marian Buschsieweke <marian.buschsieweke@ovgu.de>
Signed-off-by: iosabi <iosabi@protonmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/5584
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
These functions used to exist but don't anymore. (Pointed out in #863)
Change-Id: Iac6b5edd320bdff7628a788861e332f956dcd93d
Signed-off-by: Tim Newsome <tim@sifive.com>
Previously, progbuf execution did not flush or invalidate the register cache which could lead to incorrect behavior. This patch fixes it as well as refactors few sore points in the code related to it.
Change-Id: I353b931ca70a1828d4a9cc512aead00441730875
Signed-off-by: Marek Vrbka <marek.vrbka@codasip.com>
Otherwise we may end up modifying DCSR of a different hart than
intended.
Change-Id: I39bde21a1444623ed150f2b3d504b9318b9d6191
Signed-off-by: Tim Newsome <tim@sifive.com>
This change implements the predefined type tap_state instead of generic
uint8_t in the driver
Change-Id: I3478e8d7b40b961f3ba77711179016cdcc35cd32
Signed-off-by: Jacek Wuwer <jacekmw8@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7722
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
This change fixes endianness support in the driver.
Change-Id: Ida360bb58e988cea0a66fdc79e1610b528846fc4
Signed-off-by: Jacek Wuwer <jacekmw8@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7721
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Clang reports that 3rd function call argument is an uninitialized value
file esp32_apptrace.c line:1270
Signed-off-by: Erhan Kurubas <erhan.kurubas@espressif.com>
Change-Id: I73e254d4eb0c6b3152229717d8827d334784ab92
Reviewed-on: https://review.openocd.org/c/openocd/+/7719
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Clang is unable to fully track the content of the array
write_buffer[] and incorrectly complains that it could contain
some uninitialized value.
To help clang to track the execution flow, rewrite the handling of
the buffer by using simpler indexing and by moving away cmd_byte
from the first buffer's element to the variable cmd_byte.
While there:
- fix the error codes returned while parsing the command line and
- use directly command_print_sameline() instead of passing through
intermediate buffers.
Change-Id: I1969e896887ea3a4abebee057cc04c03005fa57c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7718
Tested-by: jenkins
Clang 15.0.7 complains about snprintf output truncation due to
output between 13 and 22 bytes into a destination of size 20.
Increase the size of the buffer.
Change-Id: I0369255ca1bc02a0cf494f765e91a608c960a0d6
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7717
Tested-by: jenkins
Value in the 'dscsr' variable is garbage until the DAP queue is run.
Postpone evaluation of the 'secure_state' variable. Reading the
core registers in between will execute the DAP queue.
Change-Id: I44959e882dbafb1b9779e813c3d13f3b3dbcd47f
Signed-off-by: Bohdan Tymkiv <bohdan200@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7693
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The registers pauth_dmask and pauth_cmask are not accessible in
AARCH32 mode. Tagging them as 'hidden' is not enough and triggers
error:
Failed to read pauth_dmask register
while halting the core.
Tag the pauth registers as not existing, unless required by user.
Note: for non existing registers there should be no need to
allocate their register cache. Let's keep this for a further
improvement.
Change-Id: Iaa0d006a3d8ee611ee93333ed49a8615a6c94276
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Fixes: d0436b0cda ("armv8: Add support of pointer authentication")
Reviewed-on: https://review.openocd.org/c/openocd/+/7712
Tested-by: jenkins
Reviewed-by: Koudai Iwahori <koudai@google.com>
Register access on running target should fail if mstatus needs to be
modified.
Change-Id: Iec8e8d514ef2f5ca42606a5534cce55aaaa99180
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
- Manual integration of File-IO support from xt0.2 release
- Verified with applications linked using gdbio LSP
- No new clang static analysis warnings
Signed-off-by: Ian Thompson <ianst@cadence.com>
Change-Id: Iedc5f885b2548097ef4f11ae1a675b5944f5fdf0
Reviewed-on: https://review.openocd.org/c/openocd/+/7550
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Removed an unused include from src/jtag/adapter.c.
Signed-off-by: Jan Matyas <jan.matyas@codasip.com>
Change-Id: Ia5ea0acdfa1c011d7c88decd0f63e8032aafd699
Reviewed-on: https://review.openocd.org/c/openocd/+/7687
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
The function is used for commands:
- jtag tapisenabled
- jtag tapenable
- jtag tapdisable
While there, add the missing .help and .usage fields and fix the
incorrect check in jtag_tap_enable() and jtag_tap_disable().
Change-Id: I0e1c9f0b8d9fbad19d09610a97498bec8003c27e
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7554
Tested-by: jenkins
While there:
- fix memory leak in case of error on values tap->chip,
tap->tapname, tap->expected_ids;
- check for out of memory error;
- fix minor coding style issue;
- add the missing .usage field;
- remove functions not in use anymore.
Change-Id: I1c8c3ffeb324e9eacb919c7e0d94fd72122c9a81
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7431
Tested-by: jenkins
Change the prototype of functions:
- target_run_algorithm()
- target_wait_algorithm()
- target_wait_state()
- struct target_type::run_algorithm()
- struct target_type::wait_algorithm()
to use unsigned int for timeout_ms instead of int.
Change accordingly the variables passed as parameter.
Change-Id: I0b8d6e691bb3c749eeb2911dc5a86c38cc0cb65d
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7562
Tested-by: jenkins
While there, add the missing .usage field and remove the now
unused function jim_target_tap_disabled().
Change-Id: I79afcc5097643fc264354c6c3957786a55f40498
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7561
Tested-by: jenkins
The function esp32_cmd_apptrace_generic() is not used outside the
file.
Declare it as static.
Detected through 'sparse' tool.
Change-Id: I08c6b92fb01594320bc3ae6b16067ac4eb51ca12
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/7676
Tested-by: jenkins
This reverts commit 047b1a8fc2.
Commit 047b1a8fc2 ("target/image: zero-initialize ELF segments
up to p_memsz") breaks the backward compatibility introducing some
problem:
- an empty bss segment with paddr in SRAM gets zero filled at load
but does not survive after a reset, causing verify to fail;
- an empty bss segment with paddr in FLASH causes excessive flash
usage, which can exceed flash size (causing error) and makes
flash aging faster.
Revert it while looking for a better implementation.
Change-Id: Iaaf926dafce46a220a5bbe20c8576eb449996d76
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reported-by: Marian Buschsieweke <marian.buschsieweke@ovgu.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/7658
Reviewed-by: Bohdan Tymkiv <bohdan200@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Nemui Trinomius <nemuisan_kawausogasuki@live.jp>
Reviewed-by: Peter Collingbourne <pcc@google.com>
Tested-by: jenkins
This way if you connect to a running target, before it's hit a breakpoint,
then when it does hit the breakpoint OpenOCD will catch it.
Change-Id: I6f1e5f169fa385f46759015786e664693c3872e4
Signed-off-by: Tim Newsome <tim@sifive.com>
Poll failure just means poll failed. It's safer to assume the target is
still running, because then if it is running and subsequently halts we can
relay this to gdb correctly. We can't do the other way around, because once
gdb thinks the target has halted, it can't deal with it spontaneously
running.
Change-Id: Idb56137f1d6baa9afc1b0e55e4a48f407b8ebe83
Signed-off-by: Tim Newsome <tim@sifive.com>
When a DM was powered down, we end up in examine() again, and clearly if
the DM was powered down we need to invalidate that cache.
Change-Id: I5eb6a289939f313e06c09cac22245db083026aa3
Signed-off-by: Tim Newsome <tim@sifive.com>
The error state is sticky, so this has to be done to recover.
Change-Id: I589f3cdab0f2351fd25f89951830cbc16c39bd93
Signed-off-by: Tim Newsome <tim@sifive.com>
Trying to disable OTP write protection by running e.g. `flash protect
1 0 1 off` would already be rejected with an error code, but that would
result in a generic "failed setting protection for blocks 0 to 1"
message. Now a more specific error message is also printed, telling the
user why it failed.
Change-Id: I6d4974eb0bcd23a0a6cf68ff955d9e59b8b1b06a
Signed-off-by: Matthijs Kooijman <matthijs@stdin.nl>
Reviewed-on: https://review.openocd.org/c/openocd/+/7615
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
When a DMI operation does not succeed (either because of a timeout
or an error), the specification says that the error in the `op`
field is sticky and needs to cleared by writing `dmireset` in `dtmcs`.
This is already done for timeouts in increase_dmi_busy_delay
but not for errors.
Change-Id: I7c5f27a5cf145511a1a8b64a45a586521e1cbe41
Signed-off-by: Amaury Pouly <amaury.pouly@lowrisc.org>
Reviewed-on: https://review.openocd.org/c/openocd/+/7688
Tested-by: jenkins
Reviewed-by: Tim Newsome <tim@sifive.com>
Reviewed-by: Jan Matyas <jan.matyas@codasip.com>
this new STM32 series family introduces 2 devices:
STM32C011xx (0x443) and STM32C031xx (0x453)
both devices have 32 Kbytes single flash bank.
Change-Id: I4e890789e44e3b174c0e9c0e1068383ecdbb865f
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/6874
Reviewed-by: Nemui Trinomius <nemuisan_kawausogasuki@live.jp>
Tested-by: jenkins
Reviewed-by: zapb <dev@zapb.de>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The bcm2835gpio driver preferred /dev/gpiomem for access to
memory mapped GPIO control and used /dev/mem as a fallback
only if it couldn't open /dev/gpiomem.
/dev/mem usually requires elevated rights or specific capabilities
of the opening process, so the fallback failed anyway.
Although /dev/gpiomem is the strongly preferred option with respect
to security, there could be also use cases which require /dev/mem
even if /dev/gpiomem is available (e.g. changing the GPIO pad
settings is necessary or testing/debugging OpenOCD).
It was difficult to handle such cases because they required
to block globally the system device /dev/gpiomem
(remove, rename or chmod).
Drop the fallback feature and select the memory device
by 'bcm2835gpio peripheral_mem_dev' configuration command.
Use /dev/gpiomem as a default.
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Change-Id: I60e427bda795d7a13d55d61443590dd31d694832
Reviewed-on: https://review.openocd.org/c/openocd/+/7350
Tested-by: jenkins
Reviewed-by: Jonathan Bell <jonathan@raspberrypi.com>
Raspberry Pi 4 with 64-bit kernel and arm_peri_high=1 config.txt
parameter needs peripheral_base 0x47e000000, uint32_t is not enough.
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Change-Id: Icedd084e2916657fa4478d452a5eb1e84a45c281
Reviewed-on: https://review.openocd.org/c/openocd/+/7685
Tested-by: jenkins
Reviewed-by: Jonathan Bell <jonathan@raspberrypi.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
The pads were configured at a wrong memory address
if /dev/gpiomem was mapped.
The pad setting registers are not accessible in mapped /dev/gpiomem,
disable the pads setting if the driver doesn't open /dev/mem.
While on it, do not fail the driver initialization if pad mapping fails
- just emit a warning and work with unchanged pad setting.
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Change-Id: I0bce76cade8f7efd75efd9087a7d9ba6511a6239
Reviewed-on: https://review.openocd.org/c/openocd/+/7684
Tested-by: jenkins
Reviewed-by: Jonathan Bell <jonathan@raspberrypi.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
This commit introduces a new function, which can be used to reduce number
of register accesses.
Change-Id: I125809726eb7797b11121175c3ad66bedb66dd0d
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Since writing a register can make some GPRs dirty (e.g. S0, S1), registers
should be flushed in reverse order, so GPRs are flushed last.
Change-Id: Ice352a4df4ae064619c0f9905db634a7b57e4711
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Kirill Radkin
Tim Newsome
Evgeniy Naydanov
Marek Vrbka
Brandon Pupp
Ahmed BOUDJELIDA
Artur Rojek
Marcin Niestroj
Jason Peck
Nishanth Menon
Bruno Mendes
Karl Palsson
Parshintsev Anatoly
Xiang W
Nikolay Dimitrov
Daniel Anselmi
Daniel Goehring
Antonio Borneo
Tomas Vanek
Mark Zhuang
eolson
Artemiy Volkov
Erhan Kurubas
Marc Schink
Laurent LEMELE
Chao Du
iosabi
Jacek Wuwer
Bohdan Tymkiv
Ian Thompson
Jan Matyas
Matthijs Kooijman
Amaury Pouly
Tarek BOCHKATI