See https://github.com/riscv-collab/riscv-openocd/issues/1139
riscv013_invalidate_cached_progbuf() was failing to zeroize the final
buffer array element. Use memset() instead of a manual loop to zeroize
it in order to address this and simplify the code.
(1) Error code and 'skip_reason' string were replaced with memory access
status. It allows to specify whether OpenOCD should exit the access
early.
(2) Slightly refactored 'read_memory' and 'write_memory' functions.
Checkpatch-ignore: MACRO_ARG_PRECEDENCE, MULTISTATEMENT_MACRO_USE_DO_WHILE
Checkpatch-ignore: TRAILING_SEMICOLON
Signed-off-by: Farid Khaydari <f.khaydari@syntacore.com>
Fixes#1043
There were multiple issuese with DMI logging:
1. Address was assumed to be the same (#1043).
2. Reported IDLE count was not affected by a reset of the delays.
3. VLA were used.
These issues are addressed in the commit.
Change-Id: I82f45505e8a62dfdd7dcb418784975fe10180109
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
There was a conflict:
1. commit 2cd8ebf44d ("breakpoints: use 64-bit type for watchpoint mask
and value")
2. commit 0bf3373e80 ("target/breakpoints: Use 'unsigned int' for
length")
The second commit was created erlier, but merged later so the types of
`mask` and `value` became `uint32_t` in `watchpoint_add_internal()`.
This created a bug:
`WATCHPOINT_IGNORE_DATA_VALUE_MASK` is defined as `(~(uint64_t)0)`.
Truncation to uint32_t makes it so the comparisons with the constant
don't work.
Change-Id: I19c414c351f52aff72a60330d83c29db7bbca375
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Use the TRAX interface DEBUGPC if available.
Otherwise use default stop-and-go profiling.
ESP32-S3, before this patch:
Internal: 8 samples/second
FT2232H: 12 samples/second
After this patch:
Internal: 18ksamples/second
FT2232H: 100ksamples/second
Change-Id: I681f0bccf4263c1e24f38be511e3b3aec8bf4d60
Signed-off-by: Richard Allen <rsaxvc@rsaxvc.net>
Reviewed-on: https://review.openocd.org/c/openocd/+/8431
Reviewed-by: Erhan Kurubas <erhan.kurubas@espressif.com>
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Ian Thompson <ianst@cadence.com>
Reviewed-by: Yurii Shutkin <yurii.shutkin@gmail.com>
hide_csrs should not emit warnings on an attempt to hide non-exitents CSR.
hide_csrs funcitonality is intended to be used for scenarios when we don`t
want certain groups of registers to be available in GDB. Typically this is
needed to simplify integration with various IDE. In such scenarious it may
be impractical/unfeseable to figure out which register is present on a
target. So reporting a situation when a user wants to hide a non-existent
register creates way too much noise. This commit reduces severity of
relevant debug message to LOG_TARGET_DEBUG
Change-Id: Icbb982c4bcce7586fe35b6b004d0874d6014d5a7
This reverts commit e56dc61697.
The reverted commit claims to be the same as
b201a5db23, but it's not -- it changes the
warning in `riscv_reg_impl_expose_csrs()` instead of the one in
`riscv_reg_impl_hide_csrs()`.
Before the change, if the user wrote to any `tdata*` register, OpenOCD
would sometimes start to disable all the triggers (by writing zeroes to
`tdata1`) and re-enable them again (by witing all trigger registers to the
values read before for each `tselect` value), e.g. on `step`
(see `disable/enable_triggers()`).
There are a couple of issues with such approach:
1. RISC-V Debug Specification does not require custom register types
to support re-enabling by such sequence of writes (e.g. some custom
trigger type may require writing a custom CSR to enable it).
2. OpenOCD may still overwrite these triggers when a user asks to set a
new WP.
This commit introduces `riscv reserve_trigger ...` command to explicitly
mark the triggers OpenOCD should not touch.
Such approach allows to separate management of custom triggers and
offload it onto the user (e.g. disable/enable such triggers by setting up
an event handler on `step`-related events).
Change-Id: I3339000445185ab221368442a070f412bf44bfab
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Currently, during register file examination:
1. A read of an XPR is attempted via 64-bit abstract access.
2. If such a read fails (e.g. connection unstable) XLEN is assumed to be
32.
3. Then `misa` is read. Since `misa` is a CSR and it may be only
readable via program buffer, `s0` should be readable beforehand (at
least some assumption about `xlen` should be made).
4. Before the commit, the `misa.mxl` field was not checked against
`xlen`, therefore erroneous info may have been reported to the user.
Moreover, the `examine()` would pass indicating no error at all.
5. After the commit, `misa.mxl` is checked against `xlen` value.
Change-Id: I3fe5bd6742e564e6de782aad9ed10e65c0728923
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
* removed `progbuf_size` field from `riscv_info`; added getter
* moved `impebreak` field from `riscv_info` to `riscv013_info`
as implementation dependent field; added getter
Signed-off-by: Farid Khaydari <f.khaydari@syntacore.com>
Open output file in binary mode to disable EOL
conversion on Windows (and sometimes cygwin depending
on installation settings and path).
Change-Id: I38276dd1af011ce5781b0264b7cbb08c32a1a2ad
Signed-off-by: Richard Allen <rsaxvc@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8278
Reviewed-by: Karl Palsson <karlp@tweak.au>
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
This is useful for setting a reset catch on a CPU that is being
brought out of reset.
Change-Id: Id8fe9bc3f75fd170f207f470a9f3b0faba7f24c1
Signed-off-by: Peter Collingbourne <pcc@google.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8422
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reduce the amount of boilerplate by moving cti_regs into its only
user, making it a local variable and removing the now-redundant
p_val pointer.
Change-Id: I778cc1e960532fae1ac1a952c6ff19c54e578a5f
Signed-off-by: Peter Collingbourne <pcc@google.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8421
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Some targets start up with the sticky overrun bit set. On such targets
we need to clear it in order to avoid subsequent incorrect reads.
Change-Id: I3e939a9e092de6fcea9494d3179a3386aa1701d2
Signed-off-by: Peter Collingbourne <pcc@google.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8420
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
While at it, fix some small coding style issues.
Change-Id: Ifb8e78b55d29a06d69a3ce71d12d0040777aef13
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8423
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Commit 53b94fad58 ("binarybuffer: Fix str_to_buf() parsing
function") introduces the helper command_parse_str_to_buf() to
parse as number a string on TCL command-line.
The parameter 'radix' can specify the base (decimal, octal,
hexadecimal, or auto-detected).
TCL is supposed to use decimal numbers by default, while octal and
hexadecimal numbers must be prefixed respectively with '0' and
'0x' (or '0X').
This would require the helper to always run auto-detection of the
base, thus always set the 'radix' parameter to zero. This makes
the parameter useless.
Keeping the 'radix' parameter can open the door to future abuse of
TCL syntax, E.g. a command can require an octal value without the
mandatory TCL '0' prefix; the octal value cannot be the result of
TCL expression.
To prevent any future abuse of the 'radix' parameter, drop it.
Change-Id: I88855bd83b4e08e8fdcf86a2fa5ef3269dd4ad57
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8393
Tested-by: jenkins
Reviewed-by: Jan Matyas <jan.matyas@codasip.com>
- src/jtag/drivers/ftdi.c:
```
++<<<<<<< HEAD
+ int i;
+ static const uint8_t zero;
++=======
+ uint8_t zero = 0;
++>>>>>>> ocd_upstream
```
Decided to choose the latter.
- src/target/riscv/riscv-013.c:
```
++<<<<<<< HEAD
+ int abs_chain_position;
+ /* The base address to access this DM on DMI */
+ uint32_t base;
++=======
+ unsigned int abs_chain_position;
+
++>>>>>>> ocd_upstream
```
Decided to choose the latter (abs_chain_position is unsigned now)
- src/target/riscv/batch.c:
```
++<<<<<<< HEAD
++=======
+ void dump_field(int idle, const struct scan_field *field)
+ {
...
+ }
++>>>>>>> ocd_upstream
```
dump_field function is not needed anymore
Signed-off-by: Parshintsev Anatoly <anatoly.parshintsev@syntacore.com>
Logically, BSCAN tunneling is used to establish a connection, therefore
it should be set up before the communication starts (i.e. before
`init`).
Moreover, current implementation does not support changing
`bscan_tunnel_ir_width` after `init`. This is evident by RISC-V handler
of the `init` itself.
Link: 9a23c9e679/src/target/riscv/riscv.c (L467-L481)
Change-Id: I817c6a996f7f7171b2286e181daf1092bd358f69
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Also avoid receiving data if the value is discarded on the call-site.
Change-Id: Ied87b551536a00d9fad469b9843cccae1976e6b6
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
* Register file examination is separated.
* Allow to access registers through cache as early as possible to re-use
general register access interface and propely track state of the
register.
* Reduces the number of operations: S0 and S1 are saved/restored only
when needed (targets without abstract CSR access).
Change-Id: I2e205ae4e88733a5c792f8a35cf30325c68d96b2
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
The 'orig_instr' information of software breakpoints is incorrect
because buf_to_hex_str() expects the length of the buffer to be
converted in bits and not bytes.
Change-Id: I9a9ed383a8c25200d461b899749d5259ee4c6e3d
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8218
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
This patch modifies as little code as possible in order to simplify the
review. Data types that are affected by these changes will be addresses
in following patches.
While at it, apply coding style fixes if these are not too extensive.
Change-Id: Idcbbbbbea2705512201eb326c3e6cef110dbc674
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8413
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
This patch modifies as little code as possible in order to simplify the
review. Data types that are affected by these changes will be addresses
in following patches.
While at it, apply coding style fixes if these are not too extensive.
Change-Id: Ie048b3d472f546fecb6733f17f9d0f17fda40187
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8404
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
This patch modifies as little code as possible in order to simplify the
review. Data types that are affected by these changes will be modified
in following patches.
Change-Id: I83921d70e017095d63547e0bc9fe61779191d9d0
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8403
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Scan-build is unable to detect that 'target->dbg_msg_enabled' does
not change across the function cortex_m_fast_read_all_regs().
It incorrectly assumes that it can be false at the first check (so
'dcrdr' get not assigned) and it is true later on (when 'dcrdr'
get used).
This triggers a false positive:
src/target/cortex_m.c:338:12: warning:
3rd function call argument is an uninitialized value
[core.CallAndMessage]
retval = mem_ap_write_atomic_u32(armv7m->debug_ap, DCB_DCRDR, dcrdr);
Use a local variable for 'target->dbg_msg_enabled' so scan-build
can track it as not modified.
While there, change the type of 'target->dbg_msg_enabled' to
boolean as there is no reason to use uint32_t.
Change-Id: Icaf1a1b2dea8bc55108182ea440708ab76396cd7
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8391
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
* Eliminates the use of VLA, which is prohibited by `doc/manual
/style.txt`:
Link: c6bb902629/doc/manual/style.txt (L164-L166)
* Unifies DMI access interface.
* Reduces code duplication.
Change-Id: I2d7b0595f171e21062049ff61f76fb5a3c992d11
Signed-off-by: Evgeniy Naydanov <evgeniy.naydanov@syntacore.com>
Use a command group 'gdb' with subcommands instead of individual
commands with 'gdb_' prefix.
The old commands are still available to ensure backwards compatibility,
but are marked as deprecated.
Change-Id: I037dc58554e589d5710cf46924e0a00f863aa300
Signed-off-by: Marc Schink <dev@zapb.de>
Reviewed-on: https://review.openocd.org/c/openocd/+/8336
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
The function str_to_buf() was too benevolent and did
not perform sufficient error checking on the input
string being parsed. Especially:
- Invalid numbers were silently ignored.
- Out-of-range numbers were silently truncated.
The following commands that use str_to_buf()
were affected:
- reg (when writing a register value)
- set_reg
- jtag drscan
This pull request fixes that by:
- Rewriting str_to_buf() to add the missing checks.
- Adding function command_parse_str_to_buf() which can
be used in command handlers. It parses the input
numbers and provides user-readable error messages
in case of parsing errors.
Examples:
jtag drscan 10 huh10
- Old behavior: The string "huh10" is silently
converted to 10 and the command is then executed.
No warning error or warning is shown to the user.
- New behavior: Error message is shown:
"'huh10' is not a valid number"
reg pc 0x123456789
Assuming the "pc" is 32 bits wide:
- Old behavior: The register value is silently
truncated to 0x23456789 and the command is performed.
- New behavior: Error message is shown to the user:
"Number 0x123456789 exceeds 32 bits"
Change-Id: I079e19cd153aec853a3c2eb66953024b8542d0f4
Signed-off-by: Jan Matyas <jan.matyas@codasip.com>
Reviewed-on: https://review.openocd.org/c/openocd/+/8315
Tested-by: jenkins
Reviewed-by: Marek Vrbka <marek.vrbka@codasip.com>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tommy Murphy
Evgeniy Naydanov
Farid Khaydari
Evgeniy Naydanov
Richard Allen
Parshintsev Anatoly
Tim Newsome
Anatoly Parshintsev
Tomas Vanek
Richard Allen
Marc Schink
Peter Collingbourne
Antonio Borneo
Jan Matyas