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>
This patch adds support for p packet responses by targets configured
with RTOS support. This change required moving to a rtos_reg struct,
which is similar to struct reg used by targets, which resulted in
needing to update each stacking with register numbers. This patch also
allows targets with non-linear register numbers to function with RTOSes
as well.
Change-Id: I5b189d74110d6b6f2fa851a67ab0762ae6b1832f
Signed-off-by: Steven Stallion <stallion@squareup.com>
Reviewed-on: http://openocd.zylin.com/4121
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This patch fixes a number of bugs caused by incomplete support for
non-existent registers. This is needed for targets that provide optional
registers or non-linear register numbers.
Change-Id: I216196e0051f28887a2c3da410959382369eed80
Signed-off-by: Steven Stallion <stallion@squareup.com>
Reviewed-on: http://openocd.zylin.com/4113
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Virtual targets, like mem_ap, do not or cannot implement the required
functionality to accept a GDB connection. In the case of mem_ap, the
method get_gdb_reg_list() is missing and a following connection from
gdb causes OpenOCD to segfault.
OpenOCD opens a GDB port for each target; it's always possible to
connect, by mistake, GDB to one virtual target.
Add a method to check if the target supports GDB connections (for the
moment just checking if get_gdb_reg_list is implemented).
Skip opening a gdb server for every targets that don't support GDB
connections.
Change-Id: Ia439a43efe1a9adbb1771cd9d252db8ffa32eb9d
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4676
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The argument passed to global config command "gdb_port" is usually,
but not always, a TCP port number. In case of multiple targets, this
numeric value is used as the first port of a set of consecutive TCP
ports assigned one per target.
If the argument is not a numeric value (e.g. "pipe", "disabled", ...)
then incrementing it for the next target has no sense.
Add the option "-gdb-port number" to the commands "target create" and
"$target_name configure" to override, for the specific target, the
general global configuration.
This permits to use a per target "-gdb-port disabled", when no gdb
port is required for that specific target.
It also makes possible to choose a custom TCP port number for each
target, overriding the usual sequence of consecutive port numbers.
Change-Id: I3b9a1910b28ab4bc757e839d0e5d08ffc29f7ab4
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4530
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Every TCL command can be renamed (or deleted) and then replaced by
a TCL proc that has the same name of the original TCL command.
This can be used either to completely replace an existing command
or to wrap the original command to extend its functionality.
This applies also to the OpenOCD command "shutdown" and can be
useful, for example, to set back some default value to the target
before quitting OpenOCD.
E.g. (TCL code):
rename shutdown original_shutdown
proc shutdown {} {
puts "This is my implementation of shutdown"
# my own stuff before exit OpenOCD
original_shutdown
}
Unfortunately, sending a signal (or pressing CTRL-C) to terminate
OpenOCD doesn't trigger calling the original "shutdown" command
nor its (eventual) replacement.
Detect if the main loop is terminated by an external signal and
in such case execute explicitly the command "shutdown".
Replace with enum the magic numbers assumed by "shutdown_openocd".
Please notice that it's possible to write a custom "shutdown" TCL
proc that does not call the original "shutdown" command. This is
useful, for example, to prevent the user to quit OpenOCD by typing
"shutdown" in the telnet session.
Such case will not prevent OpenOCD to terminate when receiving a
signal; OpenOCD will quit after executing the custom "shutdown"
command.
Change-Id: I86b8f9eab8dbd7a28dad58b8cafd97caa7a82f43
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4551
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Commit 5087a955 added custom signal handlers for the openocd
server process.
Before this commit, when openocd is run as a background process
having the same controlling terminal as gdb, Control-C would be
handled by gdb to stop target execution and return to the gdb prompt.
However, after commit 5087a955, the SIGINT caused by pressing
Control-C also terminates openocd, effectively crashing the
debugging session. The only way to avoid this is run openocd in
a different controling terminal or to detach openocd from its
controlling terminal,
thus losing all job control for the openocd process.
This patch improves the server's handling of POSIX signals:
1) Keyboard generated signals (INT and QUIT) are ignored
when server process has is no controlling terminal.
2) SIGHUP and SIGPIPE are handled to ensure that .quit functions
for each interface are called if user's logs out of X
session or there is a network failure.
SIG_INT & SIG_QUIT still stop openocd
when it is running in the foreground.
Change-Id: I03ad645e62408fdaf4edc49a3550b89b287eda10
Signed-off-by: Brent Roman <genosensor@gmail.com>
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/3963
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
In a multi-target environment we are supposed to have a single
gdb server for each target (or for each group of targets within
a SMP node).
By default, the gdb attached to a server sends its command to
the target (or to the SMP node targets) linked to that server.
This is working fine for the normal gdb commands, but it is
broken for the native OpenOCD commands executed through gdb
"monitor" command. In the latter case, gdb "monitor" commands
will be executed on the current target of OpenOCD configuration
script (that is either the last target created or the target
specified in a "targets" command).
Fixed in gdb_new_connection() by replacing the current target
in the connection's copy of command context.
Change-Id: If7c8f2dce4a3138f0907d3000dd0b15e670cfa80
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4586
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Christopher Head <chead@zaber.com>
When GDB quits (e.g. with "quit" command) we first execute
gdb_detach() to reply "OK" then, at GDB disconnect (either TCP
or pipe connection type), we execute gdb_connection_closed().
In case GDB is killed or it crashes, OpenOCD only executes the
latter when detects the disconnection.
Both gdb_detach() and gdb_connection_closed() trigger the event
TARGET_EVENT_GDB_DETACH thus getting it triggered twice on clean
GDB quit.
Do not trigger the event TARGET_EVENT_GDB_DETACH in gdb_detach()
and let only gdb_connection_closed() to handle it.
Change-Id: Iacf035c855b8b3e2239c1c0e259c279688b418ee
Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-on: http://openocd.zylin.com/4585
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
This patch fixes a bug where target fails to resume after completing GDB FileIO.
We need to update target last run control information to decide resumption. This
was not being done for vcont packets.
Change-Id: I44bea31720f8b877dba97d77a202303d546ea5bd
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/4539
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
In 2016, ARM released the second edition of the semihosting specs
("Semihosting for AArch32 and AArch64"), adding support for 64-bits.
To ease the reuse of the semihosting logic for other platforms
(like RISC-V), the semihosting code was isolated from the ARM
target and updated to the latest specs.
The new code is already in use since January (in GNU MCU Eclipse
OpenOCD) and no problems were reported, neither for ARM nor for
RISC-V targets, after more than 7K downloads.
The 2 new files were formatted with uncrustify.
Change-Id: Ie84dbd86a547323bb8a5d24eab68fc7dad013d96
Signed-off-by: Liviu Ionescu <ilg@livius.net>
Reviewed-on: http://openocd.zylin.com/4518
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Use sector sizes instead of bank size.
Detect a gap between sectors and emit xml blocks accordingly.
Detect sector overflow over the bank size.
Change-Id: If0e0e44b0c3b93067b4d717c9c7b07c08582e57b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4436
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
With this option enabled (it's disabled by default) errors accessing
registers are returned to gdb. Otherwise they are ignored and success is
reported to gdb. (This is the current behavior.)
We want this for RISC-V, but there's still some cleanup that needs to be
done before that can be upstreamed.
Signed-off-by: Tim Newsome <tim@sifive.com>
Change-Id: I7e56109ea52d18b780c14a07fb35f9e6e8979da4
Reviewed-on: http://openocd.zylin.com/4452
Reviewed-by: Steven Stallion <sstallion@gmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Tested-by: jenkins
Since OpenOCD basically allows to perform arbitrary actions on behalf of
the running user, it makes sense to restrict the exposure by default.
If you need network connectivity and your environment is safe enough,
use "bindto 0.0.0.0" to switch to the old behaviour.
Change-Id: I4a4044b90d0ecb30118cea96fc92a7bcff0924e0
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/4331
Reviewed-by: Jonathan McDowell <noodles-openocd@earth.li>
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Tested-by: jenkins
gdb assumes that a rtos can make any thread active at will in response
to a 'Hg' packet. It further assumes that it needs to step-over after
setting a breakpoint on frame #0 of any non-current thread. Both
assumptions are not valid for an actual rtos. We fake the step-over to
not trigger an internal error in gdb. See
https://sourceware.org/bugzilla/show_bug.cgi?id=22925 for details.
Change-Id: Ida60cd134033c1d58ada77b87fe664a58f61e2c0
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4448
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Normally, when a ctrl-c is received from gdb, a SIGINT is reported back
unconditionally to tell gdb that the target has stopped in response.
However when a rtos support was configured, the rtos awareness overwrote
the signal with an actual thread state, which gdb then ignored and got
stuck without the user able to interrupt.
Change-Id: I40fd62333e020a8c4d9df0079270e84df9c77f88
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4445
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Although the leak is negligible, the clean heap on exit will ease
valgrind testing.
Change-Id: I3a7a9c8e8dc7557aa51d0b9caa244537e5e7007d
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/4410
Tested-by: jenkins
This patch adds support to generate multiple nested architecture defined
data types in gdb target xml generated by openOCD. Architecture defined
structs, unions, vectors nested in one or more architecture defined
types can be generated now.
Example:
<vector id="v2d" type="ieee_double" count="2"/>
<vector id="v2u" type="uint64" count="2"/>
<vector id="v2i" type="int64" count="2"/>
<union id="vnd">
<field name="f" type="v2d"/>
<field name="u" type="v2u"/>
<field name="s" type="v2i"/>
</union>
Change-Id: I0f3c5c6daf3d22cde7e4b7b4165d2e97e25872f7
Signed-off-by: Omair Javaid <omair.javaid@linaro.org>
Reviewed-on: http://openocd.zylin.com/4372
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This patch fixes a memory leak in the internal server. Steps for
reproduction:
* valgrind --leak-check=full --show-reachable=yes ./build/src/openocd
* Establish more than one connection to OpenOCD (e.g. telnet)
* Shutdown OpenOCD
* Check for memory leaks in add_connection()
Change-Id: I0ae6fcf2918fd9bdec350446d3e26742d08ff698
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/4053
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
