This is used to compute memory block read alignment, and specifically
allows 64-bit targets to ensure that memory block reads are only
requested on 64-bit boundaries.
Signed-off-by: Tim Newsome <tim@sifive.com>
Change-Id: Idb1a27b9fc02c46245556bb0f3d6d94b368c4817
Reviewed-on: http://openocd.zylin.com/6249
Reviewed-by: Marc Schink <dev@zapb.de>
Tested-by: jenkins
Reviewed-by: Jan Matyas <matyas@codasip.com>
Reviewed-by: Xiang W <wxjstz@126.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
With the API fixed to comply with OpenOCD coding style, fix all
the references in the code.
Patch generated automatically with the script below.
The list is in reverse order to replace a common prefix after the
replacement of the symbols with the same prefix.
%<---%<---%<---%<---%<---%<---%<---%<---%<---%<---%<---%<---%<---
(cat << EOF
Jim_SetResult_NvpUnknown jim_set_result_nvp_unknown
Jim_Nvp_value2name_simple jim_nvp_value2name_simple
Jim_Nvp_value2name_obj jim_nvp_value2name_obj
Jim_Nvp_value2name jim_nvp_value2name
Jim_Nvp_name2value_simple jim_nvp_name2value_simple
Jim_Nvp_name2value_obj_nocase jim_nvp_name2value_obj_nocase
Jim_Nvp_name2value_obj jim_nvp_name2value_obj
Jim_Nvp_name2value_nocase_simple jim_nvp_name2value_nocase_simple
Jim_Nvp_name2value_nocase jim_nvp_name2value_nocase
Jim_Nvp_name2value jim_nvp_name2value
Jim_Nvp struct jim_nvp
Jim_GetOpt_Wide jim_getopt_wide
Jim_GetOpt_String jim_getopt_string
Jim_GetOpt_Setup jim_getopt_setup
Jim_GetOpt_Obj jim_getopt_obj
Jim_GetOpt_NvpUnknown jim_getopt_nvp_unknown
Jim_GetOpt_Nvp jim_getopt_nvp
Jim_GetOpt_Enum jim_getopt_enum
Jim_GetOpt_Double jim_getopt_double
Jim_GetOpt_Debug jim_getopt_debug
Jim_GetOptInfo struct jim_getopt_info
Jim_GetNvp jim_get_nvp
Jim_Debug_ArgvString jim_debug_argv_string
EOF
) | while read a b; do
sed -i "s/$a/$b/g" $(find src -type f ! -name jim-nvp.\? )
done
%<---%<---%<---%<---%<---%<---%<---%<---%<---%<---%<---%<---%<---
Change-Id: I10a12bd64bb8b17575fd9150482c989c92b298a2
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6184
Reviewed-by: Marc Schink <dev@zapb.de>
Tested-by: jenkins
The memory allocated to hold the watchpoints is not freed at
OpenOCD exit.
Free the watchpoint memory at OpenOCD exit.
Change-Id: I518c9ce0dc901cde2913d752e3154734f878b854
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6210
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Use byte address select for 1 and 2 bytes length.
Use normal mode for 4 bytes length.
Change-Id: I28d182f25145d0635de64d0361d456f1ad96640e
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6197
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Decrement the available watchpoints only when succeed setting it.
Initialize the available watchpoint with the correct value.
Change-Id: I0f93b347300b8ebedbcd9e718d4ba32b26cf6846
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6196
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The current implementation of OpenOCD does not support watchpoints for
cortex_a architecture. So, I replaced the add_watchpoint and
remove_watchpoint with the specific implementation for the
cortex a and using the breakpoint implementation and the arm
documentation [1] as reference. In particular, I have made the
following changes:
* added the following functions
- cortex_a_add_watchpoint
This wrapper function check whether there are any watchpoint
available on the target device by calling cortex_a_set_watchpoint.
- cortex_a_set_watchpoint
This function is responsible for settings the watchpoint register
pair. In particular, it sets the WVR and the WCR registers with
the cortex_a_dap command.
- cortex_a_remove_watchpoint
This wrapper function the selected watchpoint on the target device
by calling cortex_a_unset_watchpoint.
- cortex_a_unset_watchpoint
This function sets both the WVR and the WCR registers to zero, thus
unsetting the watchpoint.
[1]
http://infocenter.arm.com/help/topic/com.arm.doc.ddi0464f/BCGDHIEJ.html
Change-Id: I86611dab474cb84836662af572b17636dc68e282
Signed-off-by: Chengyu Zheng <chengyu.zheng@polimi.it>
Reviewed-on: http://openocd.zylin.com/3913
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Marc Schink <dev@zapb.de>
Tested-by: jenkins
When a WAIT occurs the commands after the WAIT are replayed and the
STICKYORUN is cleared. However if another WAIT occurs during the
command replay, the command itself is resent but the STICKYORUN bit
shall also be cleared. If this is not done, the MEM-AP hangs.
Change-Id: I14e8340cd5d8f58f4de31509da96cfa2ecb630d1
Signed-off-by: micbis <michele.bisogno.ct@renesas.com>
Reviewed-on: http://openocd.zylin.com/6278
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Some help text end with a useless space character.
Remove it.
Change-Id: I397e1194fac8042f0fab694222f925f906716de3
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6222
Tested-by: jenkins
OpenOCD already defines the macro ARRAY_SIZE, while riscv code
uses a local macro DIM.
Prefer using the macro ARRAY_SIZE() instead of DIM().
Not all the riscv code has been upstreamed, yes; this patch only
covers the code already upstreamed.
Change-Id: I89a58a6d91916d85c53ba5e4091b558271f8d618
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6258
Reviewed-by: Xiang W <wxjstz@126.com>
Tested-by: jenkins
Reviewed-by: Tim Newsome <tim@sifive.com>
The array newly_halted[] is assigned but its value is never used.
Drop it!
Change-Id: I678812a31c45a3ec03716e3eee6a30b8e8947926
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6257
Tested-by: jenkins
Reviewed-by: Xiang W <wxjstz@126.com>
Reviewed-by: Tim Newsome <tim@sifive.com>
The target mem_ap appears as an ARM target, thus it allows the
execution of ARM specific commands causing the crash of OpenOCD.
E.g. 'arm mrc ...' can be executed and segfaults.
Replace the incorrect ARM magic number with a dedicated one.
While there, remove the 'struct arm', that is now holding only the
mem_ap's dap, and replace it with a pointer to the dap.
Change-Id: I881332d3fdf8d8f8271b8711607737b052a5699b
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6213
Tested-by: jenkins
Minor typos found by the new checkpatch boosted by the dictionary
provided by 'codespell'.
Change-Id: I548581247db72e683249749d1b8725035530b06e
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6217
Tested-by: jenkins
The early documentation for armv8a report the debug register WFAR
as containing the address of the instruction that triggered the
watchpoint. More recent documentation report the register EDWAR as
containing the data memory address that triggered the watchpoint.
The name of macros CPUV8_DBG_WFAR0 and CPUV8_DBG_WFAR1 is not
correct as they point to the debug register EDWAR, so reading such
register returns directly the data memory address that triggered
the watchpoint. The code incorrectly passes this address value to
the function armv8_dpm_report_wfar(); this function is supposed to
adjust the PC value, decrementing it to remove the effects of the
CPU pipeline. This pipeline offset, that has no meaning on the
value in EDWAR, caused commit 651b861d5d ("target/aarch64: Add
watchpoint support") to add back the offset while comparing the
address with the watchpoint enabled.
The upper 32 bits of EDWAR are not valid in aarch32 mode and have
to be ignored.
Rename CPUV8_DBG_WFAR0/1 as CPUV8_DBG_EDWAR0/1.
Remove the function armv8_dpm_report_wfar().
Remove the offset while searching the matching watchpoint.
Ignore the upper 32 bits of EDWAR in aarch32 mode.
Fix a comment and the LOG text.
Change-Id: I7cbdbeb766fa18e31cc72be098ca2bc501877ed1
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6205
Tested-by: jenkins
Reviewed-by: Liming Sun <limings@nvidia.com>
The field 'wp_pc' was originally introduced in commit 55eeea7fce
("ARMv7a/Cortex-A8: report watchpoint trigger insn") in end 2009
to contain the address of the instruction which triggered a
watchpoint. Later on with commit 651b861d5d ("target/aarch64:
Add watchpoint support") it has been reused in to hold directly
the memory address that triggered a watchpoint.
Rename 'wp_pc' as 'wp_addr' and change its doxygen description.
While there, fix the format string to print the field.
Change-Id: I2e5ced1497e4a6fb6b38f91e881807512e8d8c47
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6204
Tested-by: jenkins
Reviewed-by: Liming Sun <limings@nvidia.com>
VECTRESET check should be done after verifying if the core is an ARMv6-M core,
and not before that.
Fixes: 2dc9c1df81 ("cortex_m: [FIX] ARMv8-M does not support VECTRESET")
Change-Id: I8306affd332b3a35cea69bba39ef24ca71244273
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/6232
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Despite of assert(is_packed) clang static analyser complains on use
of the uninitialized offset variable.
Cross compiling with latest x86_64-w64-mingw32-gcc hits warnings
src/target/armv7m.c: In function ‘armv7m_read_core_reg’:
src/target/armv7m.c:337:54: error: ‘reg32_id’ may be used
uninitialized in this function [-Werror=maybe-uninitialized]
It happens because mingw32 defines assert() without the attribute
"noreturn", whatever NDEBUG is defined or not.
Replace assert(is_packed) by if (is_packed) conditional and call
assert(false) in the else branch.
Change-Id: Id3c7dcccb65106e28be200b9a4d2b642f4d31019
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/6256
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-by: Andrzej Sierżęga <asier70@gmail.com>
Use a constant output length and remove the line break to make the
authentication data easier to parse.
Change-Id: Iebbf1f171947ef89b0f360a2cb286a4ea15c6ba5
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: http://openocd.zylin.com/6199
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Tim Newsome <tim@sifive.com>
Jimtcl commit 1843b79a03dd ("expr: TIP 526, only support a single
arg") drops the support for multi-argument syntax for the TCL
command 'expr'.
Modify the script startup.tcl compiled-in OpenOCD binary to comply
with the new jimtcl.
Change-Id: I520dcafacadaa289a815035f93f250447ca66ea0
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6158
Tested-by: jenkins
Reviewed-by: Oleksij Rempel <linux@rempel-privat.de>
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
For both:
- TCL proc that redirect deprecated commands to the new commands,
- TCL proc used internally and not supposed to be exposed to user,
add their name to the list of commands that should be hide by the
telnet auto-complete.
Change-Id: I05237c6a79334b7d2b151dfb129fb57b2f40bba6
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6195
Tested-by: jenkins
Reviewed-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
this change aims to provide a better gdb debugging experience,
by making gdb understand what's really happening.
before this change when hitting a watchpoint
- openocd reports "T05" to gdb
- gdb displays: Program received signal SIGTRAP, Trace/breakpoint trap.
after the change
- openocd reports "T05watch:20000000;" to gdb
- gdb displays:
Hardware watchpoint 1: *0x20000000
Old value = 16000000
New value = 170000000
...
Change-Id: Iac3a85eadd86663617889001dd04513a4211ced9
Signed-off-by: Tarek BOCHKATI <tarek.bouchkati@gmail.com>
Reviewed-on: http://openocd.zylin.com/6181
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Remove a use of AH_BOTTOM from configure.ac. This macro is used by
autoheader to add '#include' of some include file to the end of
config.h.in and then to config.h. OpenOCD can be built with a custom
config.h, so it's preferable to move these '#include' statement directly
in the C files that need them dropping this unneeded dependency.
It also causes problems when I want to use the gnulib library (which
comes with its own Makefile, and does not have the same include path as
the top-level Makefile).
So this change touches a lot of files, but is actually really simple. It
does not affect functionality at all.
Change-Id: I52c70bf15eb2edc1dd10e0fde23b2bcd4caec000
Signed-off-by: Tim Newsome <tim@sifive.com>
Reviewed-on: http://openocd.zylin.com/6171
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
The existing code asserted in that case, which is not correct. This
would allow the user to crash OpenOCD with a bad ELF file, which is
not what we want. A proper error should be reported in that case and
OpenOCD should not crash.
Change-Id: Ied5a6a6fd4ee0fd163f3fe850d304a121ecbe33a
Signed-off-by: Christian Hoff <christian.hoff@advantest.com>
Reviewed-on: http://openocd.zylin.com/6172
Reviewed-by: Jonathan McDowell <noodles-openocd@earth.li>
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Commit 246782229f ("smp: replace commands smp_on/smp_off with
"smp [on|off]"") deprecates some multi-word comments, when openocd
was unable to properly handle mixes of multi-word commands and tcl
procedures having a common initial word.
This limitation is over, so move in startup.tcl the multi-word
commands deprecation, making it easy to remove them after a grace
period.
Change-Id: Icb550d1fa7559b95692d2a1244880da6c90ec0b2
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5677
Tested-by: jenkins
While still keeping the tree of struct command, stop registering
commands in jim by the root "word" only.
Register the full-name of the command and pass as private data the
struct command of the command itself.
Still use the tree of struct command to un-register the commands.
Some "native" commands (.jim_handler) share the same handler, then
the handler checks the command name to run the right code.
Now argv[0] returns the full-name of the command, so check the
name by looking in the struct command passed as private data.
Change-Id: I5623c61cceee8a75f5d5a551ef3fbf5a303af6be
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5671
Tested-by: jenkins
Reviewed-by: Oleksij Rempel <linux@rempel-privat.de>
Replace the "struct command *parent" parameter with a string that
contains the command prefix.
This abstracts the openocd code from the knowledge of the tree of
struct command.
This also makes unused the function command_find_in_context(), so
remove it.
Change-Id: I598d60719cfdc1811ee6f6edfff8a116f82c7ed6
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5668
Tested-by: jenkins
Reviewed-by: Oleksij Rempel <linux@rempel-privat.de>
In current code the current target is overridden whenever
jim_handler_data is not NULL. This happens not only with target
prefixed commands, but also with cti, dap and swo/tpiu prefixed
commands.
While this is not causing any run-time issue, by now, the
behaviour is tricky and makes the code cryptic.
Add a specific field to struct command for the target override so
the content of jim_handler_data can be restricted to command
specific data only (today only cti, dap and swo/tpiu).
Extend the API register_commands() to specify the presence of
either the command data or the override target.
The new API makes obsolete calling command_set_handler_data() to
set jim_handler_data, so remove it.
Change-Id: Icc323faf754b0546a72208f90abd9e68ff2ef52f
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5667
Tested-by: jenkins
Reviewed-by: Oleksij Rempel <linux@rempel-privat.de>
Now that target override is uniformly implemented for all types of
commands, there is no need for target-prefixed "native" commands
(.jim_handler) to sneakily extract the overridden target from the
struct command.
Modify the commands to use the standard API get_current_target().
Change-Id: I732a09c3261e56524edd5217634fa409eb97a8c6
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5666
Tested-by: jenkins
Reviewed-by: Oleksij Rempel <linux@rempel-privat.de>
While registering a new command, jim accepts a pointer to command's
private data that will be accessible during the command execution.
Today openocd is not consistent and passes different private data
depending on the command, and then even overwrites it:
- "simple" commands (.handler) are registered with their own
struct command pointer as command private data;
- "native" commands (.jim_handler) at root level are registered
with NULL command private data;
- "native" commands (.jim_handler) not at root level are
registered with the struct command pointer of their root command
as command private data but, when executed, the command private
data is overwritten by the value in field jim_handler_data taken
from their struct command.
Uniform the usage of command private data by always set it to the
struct command pointer while registering the new commands.
Note: for multi-word commands only the root command is registered,
so command private data will be set to the struct command of the
root command. This will change later in this series when the full-
name of the command will be registered.
Don't overwrite the command private data, but let the commands that
needs jim_handler_data to get it directly through struct command.
For sake of uniformity, let function command_set_handler_data() to
set the field jim_handler_data also for "group" commands, even if
such value will not be used.
Now Jim_CmdPrivData() always returns a struct command pointer, so
wrap it in the inline function jim_to_command() to gain compile
time check on the returned type.
While there, uniform the code to use the macro Jim_CmdPrivData()
to access the command's private data pointer.
Change-Id: Idba16242ba1f6769341b4030a49cdf35a5278695
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/5664
Tested-by: jenkins
Reviewed-by: Oleksij Rempel <linux@rempel-privat.de>
If current core is in EL0, it cannot use 'msr sctlr_el1, x0'
Change-Id: I04e60e39e4c84f9d9de7cc87a8e438f5d2737dc3
Signed-off-by: Cheng-Shiun Tsai <cheng.shiun.tsai@gmail.com>
Reviewed-on: http://openocd.zylin.com/6051
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The target mem_ap is a convenient way to access buses, memory and
peripherals behind an ARM AP.
The current implementation provides only access through OpenOCD
commands, because GDB remote protocol has to interact with a CPU
and has to operate on CPU states and registers.
Using GDB to access the memory is welcome, because GDB can resolve
the symbol's address from an ELF file and can nicely display the
content of complex struct and data types.
Extend mem_ap target with the bare minimal support for a remote
GDB connection, by emulating a fake basic ARM core. It means that
only a GDB that has support for ARM can be used (either 'aarch64',
'arm' or 'multiarch' GDB). This is not seen as a big limitation,
because the mem_ap target is mainly used on ARM based devices.
Add a minimalist register description for the fake CPU.
Fill the field 'debug_reason' as expected by GDB server.
Call the target halted event to reply to GDB halt requests.
For backward compatibility, don't open the GDB port by default. If
needed, it has to be specified at 'target create' or 'configure'
with the flag '-gdb-port'.
Change-Id: I5a1b7adb749746516f5d4ffc6193c47b70132364
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/6034
Tested-by: jenkins
There are some breakpoint/watchpoint related code in armv8_dpm.c,
but seems not working for aarch64. Target aarch64 has its own
breakpoint implementation in aarch64.c. This commit follows the
same logic to add watchpoint support for target aarch64.
This commit also increases the size of stop_reason[] in function
gdb_signal_reply() since the old size is too small to fit in a
64-bit address, such as ffff8000115e6980.
Change-Id: I907dc0e648130e36b434220f570c37d0e8eb5ce1
Signed-off-by: Liming Sun <lsun@mellanox.com>
Signed-off-by: Daniel Goehring <dgoehrin@os.amperecomputing.com>
Reviewed-on: http://openocd.zylin.com/4761
Tested-by: jenkins
Reviewed-by: Liming Sun <limings@nvidia.com>
Reviewed-by: Kevin Burke <kevinb@os.amperecomputing.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
For Zephyr rtos support it is necessary to define general register
numbers for architecture. There were some already in arc.h file.
Let's define ARC registers numbers as a set instead of separate defines.
Change-Id: I63742b8608f9556c2ec9bd2661a0fd9cf88e9b74
Signed-off-by: Evgeniy Didin <didin@synopsys.com>
Reviewed-on: http://openocd.zylin.com/6105
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
This removes some warnings which prevent a successful build with -Werror
which is enabled by default. I'm using gcc 11, so maybe others are not
getting this warnings yet.
In src/flash/nor/numicro.c the debug messages were misleadingly indented.
In src/target/arm920t.c the array size where smaller than expected from
the receiving function.
Change-Id: I66f5c6a63beb9f9416e73b726299297476c884d8
Signed-off-by: Rene Kita <git@rkta.de>
Reviewed-on: http://openocd.zylin.com/6104
Reviewed-by: Jonathan McDowell <noodles-openocd@earth.li>
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tarek BOCHKATI
Marc Schink
R. Diez
Tim Newsome
Antonio Borneo
Chengyu Zheng
micbis
Tomas Vanek
Christian Hoff
Florian Meister
Cheng-Shiun Tsai
Liming Sun
Evgeniy Didin
Rene Kita