This increases performance drematically for local connections, which is the
most likely arrangement for a VPI connection.
Change-Id: Id15b29ae663f5d8100b2175357649bd03d05b7c8
Signed-off-by: Darius Rad <darius@bluespec.com>
Reviewed-on: http://openocd.zylin.com/4549
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Make flash_driver methods protect() and protect_check()
optional.
Remove dummy definitions of these methods from the drivers
which do not implement protection handling.
Some drivers did not define protect method. It raised segfault
before this change and now it is handled properly.
Lot of drivers returned ERROR_OK from dummy protect()
- now flash_driver_protect() returns an error if not handled by the driver.
Change-Id: I2d4a0da316bf03c6379791b1b1c6198fbf22e66c
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4765
Tested-by: jenkins
Reviewed-by: Andreas Bolsch <hyphen0break@gmail.com>
Reviewed-by: Christopher Head <chead@zaber.com>
Since commit bae76053dc
gdb-attach event is defined as halt by default.
Remove useless and in case of bcm281xx wrong definitions of the event.
Change-Id: I8e69780a93722eb9392673303f54d502e71eceb6
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4763
Tested-by: jenkins
Reviewed-by: Steven Stallion <sstallion@gmail.com>
The dual bank option was being incorrectly read and the
bank b base incorrectly set. This is tested with 512kB
dual bank configuration but needs checking with other
configurations (e.g. 256kb).
This fix should remove the need to use a mass_erase command
prior to programming with OpenOCD
Change-Id: I6e920f11b794c4c1fd34c0e44fb8fa01e7fe8f85
Signed-off-by: Alex J Lennon <alex.lennon@s19.tech>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4641
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-by: Thomas Søhus <soehus@gmail.com>
Prepare for additional flash banks not located at address 0
Change-Id: I60b78c917f94fa52bf24df9e3315536f776eec84
Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
Reviewed-on: http://openocd.zylin.com/4440
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
S6 record is equivalent to S5 but it uses 24-bit value (instead of 16-bit)
to store total number of data records. It can be found in big srec files
with total number of data records bigger than 65535
Handle S6 record in the same way as S5 (simply ignore it)
Change-Id: I3e15a4e8f784fd38803c00accf422f803e8469cd
Signed-off-by: Bohdan Tymkiv <bhdt@cypress.com>
Reviewed-on: http://openocd.zylin.com/4645
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The reset_config line in the config file does not actually set
connect_assert_srst (the default is connect_deassert_srst), but it reads
as if it does. Clarify that the target is compatible with
connect_assert_srst, without suggesting that the file actually sets it
to that value.
Change-Id: I14e9445ab282d386b5d0055f6adf03d7c8878a8c
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/4743
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
In previous versions of OpenOCD, it was possible to connect to the TCL
RPC interface without a current target. In `tcl_new_connection`, the
curent target would be queried by number, and the possibility of a null
current target was handled properly.
In commit bb9d9c6026, the
`get_target_by_num` call was replaced by a `get_current_target` call,
without noticing that `get_current_target` aborts if there is no current
target, whereas `tcl_new_connection` is perfectly able to handle that
situation.
Provide a `get_current_target_or_null` function for use by consumers who
are OK with a null current target, and use it in `tcl_new_connection`.
Change-Id: I06f7e1e149f1169e23c73ba328c7ad9f9425cc2a
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/4730
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Franck Jullien <franck.jullien@gmail.com>
The stm32h7x.cfg does not specify connect_assert_srst or
connect_deassert_srst in its reset_config. The comment claims that it
will therefore connect in reset. However, per the manual, the default
configuration is actually connect_deassert_srst, not
connect_assert_srst. In actual fact, connect_assert_srst does not work
on the STM32H7 because, while SRST is asserted, everything on the AXI
bus is inaccessible. The CPU core is accessible, but since the
examine-end event handler also pokes at the DBGMCU peripheral, that will
fail in connect_assert_srst mode. So using connect_deassert_srst is
appropriate, so fix the comment accordingly.
Change-Id: If3e32e871fb19cc61183bdf911b7c5efd80b62e2
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/4741
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The for loop inside bitq_path_move function is not correct, this will
overflow the cmd->path array and produces an unpredictable result.
Change-Id: I81e3bc9ee6d1dd948acd2fe4c667103ac22bb26f
Signed-off-by: xuguangxiao <szgxxu@qq.com>
Reviewed-on: http://openocd.zylin.com/4733
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The functions cortex_a_dap_{read,write}_coreregister_u32() are
duplicate of the functions dpm_{read,write}_reg().
Remove both duplicated functions in cortex_a.c while export only
dpm_read_reg(), since dpm_write_reg() is currently not used.
Rename dpm_read_reg() as arm_dpm_read_reg() to keep uniform the
naming style.
Change-Id: I501bc99dc402039e630c47917a086a0bb382782c
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4747
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Tested-by: jenkins
The name of the function dpm_modeswitch() does not follow the
common style of the other function names in the same file.
Rename it as arm_dpm_modeswitch().
Change-Id: Idebf3c7bbddcd9b3c7b44f8d0dea1e5f7549b0eb
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4756
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The armv7m debug port provides a direct access to the CPU memory
bus, allowing the debugger to bypass the CPU for every memory
operation.
The armv7a debug port doesn't offer the same feature, mainly
because CPU caches and MMU makes the direct memory access more
tricky. Nevertheless most SoC with armv7a provide direct memory
access through an AHB bus available on another DAP access port,
different from the debug port.
The original port of cortex_a in OpenOCD was inspired from the
working cortex_m code, and provided optional memory access
through the AHB, if present.
The code for AHB access is problematic and partially buggy due
to incomplete management of cache coherency and missing check of
page boundary during virtual address operations.
With the commit 5d458cf727
("target/mem_ap: generic mem-ap target") we have a clean support
for memory access through system buses connected to DAP AP, which
obsoletes the buggy memory AP hack in cortex_a.
Remove any code that uses the memory AP accesses in cortex_a.
Change-Id: I7cd1f94885e5817448058953e043d8da90dea3cc
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4748
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The variable fast_reg_read is always zero, causing some code to
never be executed.
Such code try to read the target registers by dumping them in
memory and then reading back the memory through the debugger. But
it is broken due to lack of cache and MMU management.
This code also uses the broken memory_ap access that is going to
be removed soon.
Remove all the code that depends on fast_reg_read not zero.
Add a missing check on arm_dpm_read_current_registers() return.
Keep the unused function cortex_a_dap_write_coreregister_u32()
to balance the used "read" version.
Change-Id: If2ff28a8c49eb0a87dc85207f5431978efd158db
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4746
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Handle write protection status in blocks instead of sectors, removing
unnecessary complexity in the process. Now closer to stm32f2x.
Support sequential modification of option bytes by read/modify/write
directly to option bytes area instead of always starting with the
currently loaded bytes from FLASH_OBR/WRPR registers.
Added new command 'options_load' to force re-load of option bytes w/o
having to power cycle target.
Change-Id: I5c76191e29c17a1e11482df06379d10ca8d6d04d
Signed-off-by: Dominik Peklo <dom.peklo@gmail.com>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4576
Tested-by: jenkins
Reviewed-by: Jan Vojtěch <honza.vojtech@gmail.com>
At OpenOCD start-up the operation of clearing the sticky error in
CTRL/STAT register ignores the current value of the power domains
bits CDBGPWRUPREQ and CSYSPWRUPREQ in the same register and
incorrectly set them to zero.
This abrupt disable does not follow the requirement in IHI0031 to
wait for the acknowledgment of power disabled before continuing.
The power domains are then re-enabled immediately after; it is
possible that such short disable period has passed undetected or
has been tested only on devices that do not implement the power
domains.
Anyway, this sequence is incorrect and can generate unexpected
and hard-to-debug issues while OpenOCD attaches to a running
target that implements power domains.
Anticipate the initialization of dap->dp_ctrl_stat and use it
while clearing the sticky bit. This has the additional effect of
avoiding a power disable in the error recovery part of the
function dap_dp_read_atomic().
Keep the same sequence of read/write in dap_dp_init() to avoid
breaking the initialization of some problematic target.
Add comments to document these choices.
Change-Id: I8d6da788f2dd11909792b5d6b69bc90fbe4df25d
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4677
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
While on it correct RAM amount of SAMR21x16A devices
Change-Id: Ie9ab9de1551bdceff17af7597a9a2ee41f5aebe0
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4734
Reviewed-by: Eduardo Montoya
Tested-by: jenkins
When using stlink for CM7 targets we have to rely on its firmware
to do the right thing as direct DAP access is not possible.
Change-Id: Ieee69f4eeea5c911f89f060f31ce86ed043bdfd0
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/4732
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The performance of CMSIS-DAP in long data transfers was improved substantially in
ef02b69b14. But it not as good as some
other USB/MCU based adapters. Using HID and therefore interrupt endpoint
is slower than USB bulk transfer.
CMSIS-DAP adapter implements multiple HID buffer handling and OpenOCD already
reads number of buffers from info command.
This change adds capability to sumbit more than one HID requests before
driver waits for a HID response. This scenario is used for long transfers only.
Results show about double speed on USB FS and ~140% speed on USB HS:
| w/o this change | with multi HIDrq
-----------------------------------------+-----------------+-----------------
Open source CMSIS-DAP, USB FS, adapter_khz 1000
dump_image ram32k.bin 0x1fffe000 0x8000 | 23.225 KiB/s | 45.901 KiB/s
load_image ram32k.bin 0x1fffe000 | 23.324 KiB/s | 46.552 KiB/s
Cypress' Kitprog in CMSIS-DAP mode, USB FS, adapter_khz 1000 (over firmware limit)
dump_image ram64k.bin 0x20000000 0x10000 | 15.537 KiB/s | 42.558 KiB/s
load_image ram64k.bin 0x20000000 | 15.605 KiB/s | 43.291 KiB/s
Atmel's EDBG, USB HS, adapter_khz 10000 (#3945 applied)
dump_image ram384k.bin 0x20400000 0x6000 | 248.402 KiB/s | 345.250 KiB/s
load_image ram384k.bin 0x20400000 | 256.039 KiB/s | 365.945 KiB/s
Change-Id: I9edbe018086176d357c6aaba5d6b657a5e5e1c64
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4080
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Adding the call to keep_alive() to suppress warnings when
running the async flash algorithm. Issue observed when
loading large pieces of code on slower debuggers.
Change-Id: I7660fa05f68ebd7be07b2ca0a55b0f3b6ae718f3
Signed-off-by: Kevin Gillespie <kgills@gmail.com>
Reviewed-on: http://openocd.zylin.com/4686
Tested-by: jenkins
Reviewed-by: Jesse Marroquin
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
