AM13E23019: OpenOCD patch for AM13E230x issue

Hi Aafiya Asad,

Could you share the launchpad revision you are using?

I applied the patch to openocd commit e4c49d8 to match your build, but could not reproduce the error on both launchpad revisions. Could you share what you see in the success case?

I'm sure you would have considered it by now, but it's worth trying to re-seat your cables

Thanks,
Sameer Srivastava 

hi Sameer Srivastava we had a look on our issue. so far we found out that the issue is related to a test application which utilizes some jtag pins as we expected that only the swd pins are used. it look the integrated flashing of the ccs is using swd instead of jtag, thats why we could not see the issue locally compared to ci/cd which is using openocd. can you confirm this? for now we excluded that pins from the test as a workaround. is there another option to overcome that issue, like connect under reset (to ensure default pin configuration)? what command would we have to use to avoid that problem in case pins are using alternative functionality?

br

norbert

Hi Norbert Daser , yes, our OpenOCD config for this device currently only uses JTAG, so the workaround you mentioned is correct, and it could also explain the earlier behavior.

You can try editing tcl/target/ti/am13e230x.cfg to use SWD to see if it solves the problem:

line 21:
-transport select jtag
+transport select swd
 
line 25:
-jtag newtap $_CHIPNAME cpu -irlen 4 -expected-id $_DAP_TAPID
+swd newdap $_CHIPNAME cpu -enable

We'll evaluate if SWD support is on par with JTAG and update the patch.

Thanks,
Sameer Srivastava 

thanks Sameer Srivastava for this snippet, we will give it later a try. in generall we can also run into such a problem when using swd pins during run of the application. is there as asked in the previous chat a option to connect under reset like i know it from uvision or other ide's in openocd?