can you be more specific on swd support? how did you test it? what probes did you use to sample the swo output? where do you get the trace stream? is it raw or parsed?
the openocd manual does not explain swd/swo support, if you have more information, please provide me some links.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Will be very difficult to include support for this in OpenOCD plugin?
I'm not sure what do you mean by 'this'.
J-Link handles automatically all swo/swv issues, I just telnet to a raw socket and forward the ascii stream to the Eclipse console.
if openocd provides such a raw socket, yes, it can be done.
unfortunately experience with openocd was not very good, when I tried to integrate the existing 4444 console I discovered it is not raw, but requires the telnet client protocol, which is way too complicated for the purpose of this plug-in.
please let me know if you succeed in finding a portable method to get the swo stream out of openocd.
Last edit: Liviu Ionescu (ilg) 2014-03-10
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The patch hosted by http://openocd.zylin.com/#/c/1524 was included in OpenOCD 0.8.0 that was released quite recently so somewhat simpler SWO tracing with OpenOCD should be available now. I'm trying to get it running myself, but I have had no luck yet.
I've hacked the stm32f4discovery.cfg file to include "trace 168000000 swo.log" and I get the log file created during my run, but it's always empty. I'm not sure if it is the OpenOCD setup, the ITM code (project default form GNU ARM Eclipse) or something else right now though. It is worth noting that the frequency and output file parameters have changed placed in a later commit so the commit message in the patch is not correct anymore. The documentation reflects this though.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
can you be more specific on swd support? how did you test it? what probes did you use to sample the swo output? where do you get the trace stream? is it raw or parsed?
the openocd manual does not explain swd/swo support, if you have more information, please provide me some links.
SWD in st-link is suportet for long time and I'm using it since 0.6.0
http://openocd.sourceforge.net/2012/09/final-openocd-0-6-0-is-out/
Look for stlink cfg scripts provided with openocd. I'm trying to enable SWO trace but with no succes yet.
Check this:
http://openocd.sourceforge.net/doc/doxygen/html/structswd__driver.html
http://openocd.sourceforge.net/doc/html/Debug-Adapter-Configuration.html
http://false.ekta.is/2013/12/using-swoswv-streaming-data-with-stlink-under-linux-part-2/
http://openocd.zylin.com/#/c/1524/
Will be very difficult to include support for this in OpenOCD plugin?
Last edit: Sebastian Oller 2014-03-09
ok, I'll update the manual page.
I'm not sure what do you mean by 'this'.
J-Link handles automatically all swo/swv issues, I just telnet to a raw socket and forward the ascii stream to the Eclipse console.
if openocd provides such a raw socket, yes, it can be done.
unfortunately experience with openocd was not very good, when I tried to integrate the existing 4444 console I discovered it is not raw, but requires the telnet client protocol, which is way too complicated for the purpose of this plug-in.
please let me know if you succeed in finding a portable method to get the swo stream out of openocd.
Last edit: Liviu Ionescu (ilg) 2014-03-10
Hi,
The patch hosted by http://openocd.zylin.com/#/c/1524 was included in OpenOCD 0.8.0 that was released quite recently so somewhat simpler SWO tracing with OpenOCD should be available now. I'm trying to get it running myself, but I have had no luck yet.
I've hacked the stm32f4discovery.cfg file to include "trace 168000000 swo.log" and I get the log file created during my run, but it's always empty. I'm not sure if it is the OpenOCD setup, the ITM code (project default form GNU ARM Eclipse) or something else right now though. It is worth noting that the frequency and output file parameters have changed placed in a later commit so the commit message in the patch is not correct anymore. The documentation reflects this though.
ok, I'm not sure I understand what your suggestion is, but if you manage to configure it properly, please let me know.
good luck!