* add opcode for csrrsi and csrrci
* enable MMU while reading/writing memory using progbuf
* fix style issues
* keep old behavior for progbufsize<4, perform r/w/csr only when necessary
* do not pass progbufsize, only write mstatus if changed
* add config option to enable virtualization feature
* throw error if virt enabled but unavaliable, outsource modify_privilege
* support virtualization for read_memory_progbuf_one
gdb has developed a nasty habit of very often reading 30-some
half-words. This change speeds that up significantly.
Change-Id: Iab1b7575bec5c57051c6e630ae292dddf8fe6350
* Make resume order configurable.
This is a customer requirement. Using this option is discouraged.
Change-Id: I520ec19cc23d7837cb8576f69dadf2b922fa2628
* Fix style.
Change-Id: If8e515984c92ce8df52aa69e87abde023897409f
* Make mingw32-gcc happy.
Change-Id: I39852aedec293294b2b2638ab2cc45494fe77beb
* WIP, rewrite of flash algorithm.
Just put all the flashing logic into the algorithm, instead of using an
intermediate format. This should reduce total data written while
flashing by about 9%, and also makes the code much simpler.
Change-Id: I807e60c8ab4f9f376cceaecdbbd10a2326be1c79
* New algorithm works.
Speeds up Arty flashing another 9%.
wrote 2228224 bytes from file /media/sf_tnewsome/SiFive/arty_images/arty.E21TraceFPGAEvaluationConfig.mcs in 86.784538s (25.074 KiB/s)
verified 2192012 bytes in 6.693336s (319.816 KiB/s)
8.66user 13.03system 1:33.91elapsed 23%CPU (0avgtext+0avgdata 12272maxresident)k
Change-Id: Ie55c5250d667251be141cb32b144bbcf3713fce4
* Fix whitespace.
Change-Id: I338d518fa11a108efb530ffe75a2030619457a0b
* Don't reserve so much stack space.
Also properly check XLEN in riscv_wrapper.S.
Change-Id: Ifa0301f3ea80f648fb8a6d6b6c8bf39f386fe4a6
* In theory support RV32E.
Change-Id: Icfe2a40976ae3161f2324e5bb586915aa4c4c7ee
* In theory support RV32E.
At least very basic tests pass.
Change-Id: Ia42e28a3fa020b3e52c92109392c46d009330355
* Fix cut and paste bug.
Change-Id: Ibfea68b39d706f59a8c3aa8153bb4db9803958c6
* Add hacks to make RV32E work with gdb.
gdb currently requires all 32 GPRs to be present, even on RV32E targets.
Once gdb is fixed these hacks can be removed.
Change-Id: Idcde648de2ca1a3f5b31315aab35fac86580af2c
The ST-Link driver supports not only JTAG but also SWD and SWIM.
Change-Id: I9f0e7b018cae54ed8e73a724151647e050e7bb49
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5247
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
The command "soft_reset_halt" is deprecated since mid 2013 with
the commit 146dfe3295 ("cortex_m: deprecate soft_reset_halt").
Nevertheless it is still extremely useful with multicore chips
where it allows to reset only one of the cores, option not
available through asserting the chip-wide srst.
Without a better replacement of the command, it's worth fixing it.
Accordingly to ARM DDI 0403E.d, chapter C1.4.1 "Entering Debug
state on leaving reset state", to halt the core at reset both bits
DHCSR.C_DEBUGEN and DEMCR.VC_CORERESET must be set.
Current code only sets the latter bit, relying on having C_DEBUGEN
already set through other commands, e.g. "halt". This prevents the
command "soft_reset_halt" to work if issued as very first command.
Set the bit C_DEBUGEN in command "soft_reset_halt".
Change-Id: I66bfd6a0da1fca5049dea037b4d258cf6f842966
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4987
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Thanks to command_run_linef() there is no need to pre-build the
command using alloc_printf().
Change-Id: Iccfebd6063d1ac162f090fe2309b1f51bebf0214
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5226
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The `command_run_line` function contains a comment saying it should be
reentrant. However, it isn’t: it NULLs out `current_target_override` and
doesn’t restore it before returning, and it changes the `context`
associated data of the `interp` object and then deletes that associated
data before returning rather than restoring it to its previous value.
Change-Id: I84fd46ef7173f08cf7c57b9a5b76e4986a60816f
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/5223
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Extra whitespace in file creating build errors with makeinfo.
Change-Id: Ib764850c1c8ff596d3c753eadd8e27f8c5982d20
Signed-off-by: Kevin Gillespie <kgills@gmail.com>
Reviewed-on: http://openocd.zylin.com/5229
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
* Cache program buffer writes.
Speeds up flash program by 3%, flash verify by 2%.
Change-Id: I19f8f44f560a1111fa8f4e4fc04ce6de3c94999a
* Remove nop from batch reads.
program @ 22.123 KiB/s, verify @ 47.654 KiB/s (up from program @ 20.287
KiB/s, verify @ 23.148 KiB/s originally).
Change-Id: I7ee19d967b1080336b0088d20e1fc30828afd935
* Use "algorithm" to compute CRC on RISC-V targets.
Use the C compiler to generate the algorithm code. It's better at
assembly than I am. We need separate RV32 and RV64 binaries to handle
shift instructions. I used the code from gdb (libiberty really) because
it returns the correct result. I'm not sure if the table is worth it
since we do have to save/download/restore more bytes now.
riscv_run_algorithm() now properly saves and reads back all registers
used for parameters. It also doesn't check final_pc if exit_point is 0.
Using gdb means I don't know the exact address where the code will end.
Small target.[ch] change to be able to run algorithms at 64-bit
addresses.
Flashing an arty board now:
```
wrote 2228224 bytes from file /media/sf_tnewsome/SiFive/arty_images/arty.E21TraceFPGAEvaluationConfig.mcs in 105.589180s (20.608 KiB/s)
verified 2192012 bytes in 7.037476s (304.177 KiB/s)
9.87user 16.16system 1:53.16elapsed 23%CPU (0avgtext+0avgdata 24768maxresident)k
```
Change-Id: I6696bd4cda7c89ac5ccd21b2ff3aa1663d7d7190
* Clean up formatting.
Change-Id: I7f2d792a2b9432a04209272abb00d8136ee01025
There is not flash bank support at the moment.
Change-Id: I833c009d9d21cdeb70b57d67eb557d50ed0fb4de
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5205
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The AHB5-AP is implemented in Cortex-M23/33 based devices.
Change-Id: I505954a2e2c6462ce0aa96eba1d55b016c5028b9
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5232
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Tested-by: jenkins
This access port type comes with the AMBA 5 protocol specification,
see 'C1.3 AP' in ARM IHI 0031D.
Change-Id: I3c4f0a69230daaf4f5f979de6213fe3c025a089a
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5231
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Muhammad Omair Javaid <omair.javaid@linaro.org>
This patch adds configuration files for ARM CoreLink SSE-200 SoCs. Also
adds configuration file for SSE-200 based Musca A board. Flash programming
support for Musca A QSPI flash is still not functional. This configuration
will be updated once that support lands into OpenOCD.
Please refer to ARM documentation for more information about SSE-200 and
Musca A.
Change-Id: Id3783c34d6e2609d659ef91c0bf7252c39439874
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/5006
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This patch adds ARMv8-M CPUs detection logic in ARMv7m target specific code.
Also adds a slightly different watchpoint manipulation logic for ARMv8-M.
This is based on ARMv8-M architecture reference manual.
Tested on ARM Musca A board.
Change-Id: I0652560954ef02c378a7067fab586edf39d3e9cc
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/4997
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Replace jtag specific API jtag_add_reset() with transport
independent API adapter_{de}assert_reset().
Change-Id: I1b917a4c1205115c4e0315373d81a9305e931258
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4944
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Replace jtag specific API jtag_add_reset() with transport
independent API adapter_{de}assert_reset().
Change-Id: I32c43e2e47366363521fa3f387de9e2fb1c20852
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4943
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The TCL command "return" always returns error code JIM_RETURN, to
indicate that the effective error code and message are elsewhere.
In the current implementation, the caller of jtag's event only
checks for return code JIM_OK and considers any other value,
including JIM_RETURN, as an error condition.
It can be tested running openocd on a jtag target and adding a
jtag event "setup" with a single line "return", e.g.
openocd -f board/ti_cc3200_launchxl.cfg \
-c 'jtag configure cc32xx.cpu -event setup return'
to get the message:
../src/jtag/core.c:1599: Error:
in procedure 'jtag_init' called at file "../src/jtag/core.c",
line 1599
Modify jtag_tap_handle_event() to detect the specific return value
of the "return" command and to test the real error code that is,
eventually, specified to the TCL "return" command.
Change-Id: I6d6febc15ef169638afffbffc1810e0b84fcf5c8
Reported-by: Tomas Vanek <vanekt@fbl.cz>
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5199
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The TCL command "return" always returns error code JIM_RETURN, to
indicate that the effective error code and message are elsewhere.
In the current implementation, the caller of target's event only
checks for return code JIM_OK and considers any other value,
including JIM_RETURN, as an error condition, thus dumping the
call-trace. The execution is not stopped because the error is not
further propagated, but the error message is annoying and
misleading.
It can be tested running
openocd -f ./test.cfg
using the following script "test.cfg". You can replace the board
file in line 1, to use a board available in your lab:
1 source [find board/st_nucleo_f4.cfg]
2 [target current] configure -event reset-start {}
3 [target current] configure -event reset-end {return}
4 init
5 proc a {} {[target current] invoke-event reset-start}
6 proc b {} {[target current] invoke-event reset-end}
7 proc c {} {a;b;echo "arrived at the end"}
8 c
9 shutdown
The execution produces:
./test.cfg:7: Error:
in procedure 'c' called at file "./test.cfg", line 8
in procedure 'b' called at file "./test.cfg", line 7
arrived at the end
that shows the call-trace but does not halt the execution.
The developer can avoid using the "return" command in the event
body by defining a TCL procedure that implements the handler and
that contains the "return" command, reducing the handler body to
a simple call to the procedure above. But this approach is either
not documented nor always intuitive while writing the handler,
causing waste of time to look for the false error.
Modify target_handle_event() to detect the specific return value
of the "return" command and to test the real error code that is,
eventually, specified to the TCL "return" command.
Change-Id: I2b860bab7233c6ed13ee4098e348d7533e1c4626
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4974
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Commit c2e18bfaea changed the size of the JTAG-to-SWD sequence
from 15 bytes to 17 bytes. This broke SWD switch sequence transfer
for buspirate, since buspirate packets can only hold a payload of up
to 16 bytes and we tried to fit the whole sequence in a single packet.
Splitting up the sequence transfer in appropriately sized packets
makes buspirate SWD work again (successfully tested with buspirate
firmwares v6.1 and v7.0).
Change-Id: Ib5b412b9e77287d705d2762e31c16d30318b50e3
Signed-off-by: Tilman Sauerbeck <tilman@code-monkey.de>
Reviewed-on: http://openocd.zylin.com/5200
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Change-Id: If1f56fd5d610b993a4ecbc900fac9f90638037c9
Signed-off-by: Christopher Head <chead@zaber.com>
Reviewed-on: http://openocd.zylin.com/5202
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Before commit 877cec20dc
("command: check command mode for native jim commands") all the jim commands
were erroneously treated as they had mode COMMAND_ANY.
The command '$_TARGET configure -xxx' was therefore applicable on running
OpenOCD to change the target configuration. It is handy e.g. for changing
an event handler or changes of the work area.
Change 'configure' command .mode to COMMAND_ANY to make it possible again.
The only parameter which cannot be re-configured after init is -gdb-port.
Test the command mode and refuse setting of gdb port after init.
Change-Id: I88493ac10a46647dc52a88fbc9f8ce6b5ba3bcd0
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/5214
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
since b5964191f0, all XScale-specific registers were missing, breaking
config scripts.
Change-Id: Ia56f3ca17500ba54bd08f417e9a5aaaa8a1be8c4
Signed-off-by: Michael Schwingen <spam-openocd@discworld.dascon.de>
Reviewed-on: http://openocd.zylin.com/5136
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The macro TARGET_ADDR_FMT, defined in helper/types.h, already
includes the prefix "0x" in front of the hexadecimal number,
being defined as:
#define TARGET_ADDR_FMT "0x%8.8" TARGET_PRIxADDR
An additional "0x" is present in mem_ap; it prints debug messages
with a double "0x" before the address:
Debug: 2921 34180 mem_ap.c:153 mem_ap_write_memory():
Writing memory at physical address 0x0x5000000c; size 4;
count 1
Remove the incorrect hexadecimal prefix.
Change-Id: I38f19ed2a2f542bd5df53e947a2604f1cbe80e08
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5222
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
To support 128 bit registers, the rtos_reg structure value
array needs to be updated from 8 to 16 bytes.
Tested by reading ARMv8 NEON FP regs on an Ampere eMAG 8180 with GDB.
Change-Id: I7f3fe1a5b2def599d021787fbe9cdd51f92859a4
Signed-off-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>
Reviewed-on: http://openocd.zylin.com/5209
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The ARMv7-A architecture supports super sections which allows mapping
physical addresses up to 40-bit into a 32-bit virtual address using the
short descriptor format (see ARM DDI 0406C.c section B4.1.112 for
details).
Change-Id: I8e64d0e93e36ae7a7da7b7bf2a8342856bb044f1
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5212
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
In preparation for adding super section decoding, do not restrict
armv7a_mmu_translate_va_pa() to 32-bit virtual addresses since ARMv7-A
processors with VMSA extensions (including LPAE) can issue wider
physical addresses. Update casting to uint32_t where necessary.
Change-Id: Id1c3d0d5ac324cbdc334259d9ea75fe4981671a1
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5211
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This function is not used anywhere in the tree, remove it, such that we
only have a single function moving forward that might need to deal with
short vs. long format specifics.
Change-Id: I80e81cd7eba1e028d1afaeaedb675b46c0ca6fa1
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5213
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Check earlier that the read of the PAR register was successful instead
of starting the decoding and then checking for an error reading that
register.
Change-Id: Id96c2b2f76d2d1c745fcfa55ad4c1e6db92106f9
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-on: http://openocd.zylin.com/5215
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
* Inverted Frame to Pseudo Tap for Simpler Hardware to Decode
Given the variable supported message length , a prefix decoding approach is significantly simpler for a pseudo tap architecture with a shift reg of len = max len of packet. This prefix coding packet also makes the message len field redundant , as that is implict in ir_len and the ir selected.
* style patch
* non-conflict with original
* style patch
* style patch
* requested changes
* style-patch
The prefixed commands has been removed in commit 0840414f0e
("helper/command: do not replace new commands with ocd_ prefix").
Change-Id: I9f101beb85533973041386896bbb215bb141962f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5191
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
The prefix is not necessary anymore.
Change-Id: Ie0df06a70ff51e6719d7564396739d28618b0196
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/5190
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
armv7m_get_core_reg() calls arm->read_core_reg()
arm->read_core_reg() expects the register number as an index
to core reglist, not an ARMv7M specific register code.
Use reg->number instead of armv7m_reg->num.
The change solves assert
src/target/armv7m.c:222: armv7m_read_core_reg: Assertion
`num < (int)armv7m->arm.core_cache->num_regs' failed.
when gdb 'info reg' is issued on a Cortex-M target and
no cortex_m_debug_entry() has been called since OpenOCD start
(target was halted before OpenOCD start).
Change-Id: I32a2294693ef979b613be93aeceb3b0eb06ee6df
Ticket: https://sourceforge.net/p/openocd/tickets/216/
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/5203
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
For legacy support, drivers that do not define a list of
transports get identified as jtag_only.
Cleanup this old crust and initialize properly the transports
field in the jtag_interface for all the drivers.
Change-Id: I9c86064e5d05bd0212bc18f4424414e615e617fe
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4893
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
In the initial commit 6c9b804d61
in 2007's svn://svn.berlios.de/openocd/trunk@246
a target script gdb_program_config was called before gdb flash
programming. To guarantee the script does not left any pending
command in the jtag queue, a call to jtag_execute_queue() was
inserted after the execution of the script.
In following commit ef1cfb2394
in 2008's svn://svn.berlios.de/openocd/trunk@975
the script was replaced by the event "old-gdb_program_config" and
the call to jtag_execute_queue() get executed in every case, even
if the event handler was not present.
At last, commit bb3793c9a4
("target: remove legacy target events") stripped away the
obsolete event but left the call to jtag_execute_queue(), now
completely useless.
Remove the call to jtag_execute_queue() and clean-up the code
around it.
Change-Id: I284f54d656d431ad6cdc25ca18218c09db31bd25
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4911
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Since the first commit 09883194f8 that introduced cortex_m, the
code has a delay of 50ms after srst has been asserted.
The specific delay is implemented through the JTAG_SLEEP command
sent in the jtag queue.
To remove the dependency from the jtag queue, replace the delay
with a transport independent function.
In case of jtag transport, this change keeps the same behaviour
only if the jtag queue has been flushed before the delay. This
does not happen if the call to dap_dp_init(), few lines above,
fails while calling a dap_queue_dp_{read,write}(); in this case
the jtag queue will be flushed later, after the delay, while in
the original code the delay would follow the flushing of the
commands already queued. Anyway, this different behavior would
only happen in case of DAP already not responsive so anticipating
the delay in such error condition is not supposed to add further
problems.
Change-Id: If15978246764e4266b10e707d86c03e5ed907de7
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4912
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
There is just a single entry point for the jtag API .khz(),
.speed_div(), .power_dropout(), .srst_asserted().
Simplify the code by in-lining the default handler.
The overall code behaviour is not changed.
This change prevents modifying at run-time the content of
struct jtag_interface.
Inspired from change http://openocd.zylin.com/943
by Evan Hunter <ehunter@broadcom.com>
Change-Id: I09aeb76d614db57b1884ac7ee9f00c152cd77849
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4892
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
The specific SWD API to change/query the adapter speed is never
called because the equivalent JTAG API is used in place of it.
In the restructure of struct jtag_interface, the JTAG API is
promoted as global adapter API, thus a specific SWD one is not
anymore required.
Change-Id: I1e810d255b4dfcd5791b4fac8ae1260c31a057fd
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4891
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Currently size of the GDB buffer is 16384 bytes but it is treated as
nul-terminated string in most of the code, so effective size of the
buffer is actually 16383 bytes. OpenOCD responds with `PacketSize=3fff`
to qSupported request. Result of GDB's `m` command is encoded in hex so
each data byte uses two bytes in the buffer. As a result GDB will split
bulk read requests into chunks 0x1fff bytes each. This causes troubles
on targets (or memory regions) which support only aligned, word-sized
access (such as MMIO buffers).
Steps to reproduce (psoc6 target):
gdb> dump binary memory dump.bin 0x040320000 (0x040320000 + 65536)
OpenOCD:
Error: Failed to read memory at 0x40321ffe
Error: Failed to read memory at 0x40321000
Error: Failed to read memory at 0x40323000
Error: Failed to read memory at 0x40325ffe
Error: Failed to read memory at 0x40329ffa
Error: Failed to read memory at 0x40329ffc
Error: Failed to read memory at 0x4032bffc
Error: Failed to read memory at 0x4032dffa
Consolidate GDB_BUFFER_SIZE usage: ensure size of each buffer is
(GDB_BUFFER_SIZE + 1), add explicit comment that additional byte is used
for nul-termination. Report correct size of the buffer to GDB (0x4000)
as recommended in GDB's docummentation: `if the stub stores packets in a
NUL-terminated format, it should allow an extra byte in its buffer for
the NUL`
Checked with clang-asan, clang-analyzer, valgrind - no new errors.
Change-Id: I909e8a2c6b010c5d4a304641808d4a807a4ec18d
Signed-off-by: Bohdan Tymkiv <bhdt@cypress.com>
Reviewed-on: http://openocd.zylin.com/5109
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Tim Newsome
Nils Wistoff
Marc Schink
Antonio Borneo
Christopher Head
Kevin Gillespie
jhjung81
Omair Javaid
Tarek BOCHKATI
Tilman Sauerbeck
Tomas Vanek
Michael Schwingen
Daniel Goehring
Florian Fainelli
Paul George
Andreas Fritiofson
Bohdan Tymkiv
Oleksij Rempel