)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":1000859,"name":"Karl Palsson","email":"karlp@tweak.au","username":"karlp"},"change_message_id":"a140b326637cf2072d0aa523beeaae0fc51711f6","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"e03a289f_5638fc18","updated":"2021-09-17 19:35:13.000000000","message":"Confirm this fixes SWO configuration, but it was broken for \"new\" commands as well as \"legacy\" commands.\n\nWith _either_ \n```\ntpiu config internal :3443 uart off 32000000 2000000\n```\n_OR_\n```\ntpiu create stm32wbx.cpu.tpiu -dap stm32wbx.dap -ap-num 0\nstm32wbx.cpu.tpiu configure -protocol uart -traceclk 32000000 -output :3443 -formatter off -pin-freq 2000000\nstm32wbx.cpu.tpiu enable\n```\n\ntrace would fail to be initialized, this didn\u0027t only affect legacy commands.","commit_id":"f2f01bcb1f799573da2af45c3a3285c7a92a9f60"},{"author":{"_account_id":1000859,"name":"Karl Palsson","email":"karlp@tweak.au","username":"karlp"},"change_message_id":"d94c21dcb200daf0352f603c2b126a158a1595a7","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"47eec3df_ab1969c0","updated":"2021-09-20 10:07:21.000000000","message":"So, is there any way of configuring things that was working without this patch?  I find the title, \"support for deprecated...\" to be pretty misleading, as the old way was the only way?  Still, far better having this whatever the title, than not :)","commit_id":"f2f01bcb1f799573da2af45c3a3285c7a92a9f60"},{"author":{"_account_id":1000021,"name":"Antonio Borneo","email":"borneo.antonio@gmail.com","username":"borneoa"},"change_message_id":"4fd8c4a95d7216cc975872ad8a588d1f00402ca2","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"8c001066_e1306e6c","in_reply_to":"47eec3df_ab1969c0","updated":"2021-09-20 10:29:38.000000000","message":"The key point is that running tpiu commands before \u0027init\u0027 requires putting them on hold and run them immediately after \u0027init\u0027, when the target is accessible.\nThe handler \u0027trace-config\u0027 was incorrectly executed before \u0027init\u0027, triggering error and not setting the device-specific configuration (clocks, enable, ...).\nThis fix delays the execution of \u0027trace-config\u0027 too.\nTo run without this patch, you need use the new TPIU command syntax AND rename in your config the target\u0027s event \u0027trace-config\u0027 as tpiu\u0027s event \u0027post-enable\u0027.\n\nAny suggest for the commit title?","commit_id":"f2f01bcb1f799573da2af45c3a3285c7a92a9f60"},{"author":{"_account_id":1000021,"name":"Antonio Borneo","email":"borneo.antonio@gmail.com","username":"borneoa"},"change_message_id":"93e9fb43dbc4aa2d690ed903b4bc31023098ccba","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"cdf53d3d_614e4fcc","in_reply_to":"5a0ce2f6_2fbdd11c","updated":"2021-09-20 16:35:11.000000000","message":"Yes, it\u0027s a regression.\nI use to run openocd with one of the upstream script, then type the commands in the telnet interface. In that case (typing after \u0027init\u0027) I missed the issue.\nThe issue was clear by using your script that embeds the tpiu command, thus run before \u0027init\u0027.","commit_id":"f2f01bcb1f799573da2af45c3a3285c7a92a9f60"},{"author":{"_account_id":1000859,"name":"Karl Palsson","email":"karlp@tweak.au","username":"karlp"},"change_message_id":"0cb127bb9aa092311c51d138217b4e7645241cef","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"5a0ce2f6_2fbdd11c","in_reply_to":"899d9a02_3b87926a","updated":"2021-09-20 15:05:18.000000000","message":"Indeed, the upstream files should be fixed.\n\nI guess my problem is that I don\u0027t see how it was possible, _at all_ to have working trace, configured statically in a local project openocd.cfg file.  If I\u0027ve understood, if I manually called tpiu config functions after openocd was running it would work?  That seems like a vastly less common setup, and certainly not one I\u0027ve seen in any example projects.\n\nSo to me, this is just straight up regression.  Anything along the lines of \"oh, fix it when you don\u0027t do it in this other way, that no-one actually did\" feels like just trying to downplay that trace configurations were straight up broken.","commit_id":"f2f01bcb1f799573da2af45c3a3285c7a92a9f60"},{"author":{"_account_id":1000859,"name":"Karl Palsson","email":"karlp@tweak.au","username":"karlp"},"change_message_id":"cad11be24aae7b8979dd99cfb7c3f3648f4c9ab1","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"b75171a8_b0aad81a","in_reply_to":"8c001066_e1306e6c","updated":"2021-09-20 11:59:15.000000000","message":"But those trace-config events are defined in the upstream _target_ config, not end user local config.  As far as I can tell, there was _no way_ to have trace working without using a local _target_ config file, which I would say is a fairly unreasonable expectation.\n\nBy that end, I\u0027d go with a title that\u0027s a lot more descriptive about \"actually enable trace when requested\" or something.  This whole description implies that it is only fixing things for people who were doing things the \"wrong way\" but there is no other way?!","commit_id":"f2f01bcb1f799573da2af45c3a3285c7a92a9f60"},{"author":{"_account_id":1000021,"name":"Antonio Borneo","email":"borneo.antonio@gmail.com","username":"borneoa"},"change_message_id":"488d5d51f9c6fc68469e4a4f1d032c67922254d9","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"899d9a02_3b87926a","in_reply_to":"b75171a8_b0aad81a","updated":"2021-09-20 13:18:37.000000000","message":"actually the upstream files should be reworked, adding the tpiu and renaming the event. The problem is to identify all the files that need change. When I checked in November there were 73 target files with a cortex-m, but only 10 of them have a trace-config event. For the other 63 I need to download the documentation and wish to find info about the TPIU. Cortex-m0 has no TPIU, for the others it is optional.\n\nWhat about:\narm_tpiu_swo: fix execute \u0027trace-config\u0027 event after \u0027init\u0027","commit_id":"f2f01bcb1f799573da2af45c3a3285c7a92a9f60"},{"author":{"_account_id":1000021,"name":"Antonio Borneo","email":"borneo.antonio@gmail.com","username":"borneoa"},"change_message_id":"5815fa4632af005fdb4cd247942c49785dbfccb1","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"149b3e62_a8910db8","in_reply_to":"e03a289f_5638fc18","updated":"2021-09-17 21:15:05.000000000","message":"On stm32wb, the event handler trace-config enables the SWO pin.\nThe \u0027wrapper\u0027 for old command failed to use the event (try to write memory before init) and the new command didn\u0027t call the old event.\nSo, yes, this fixes both cases.","commit_id":"f2f01bcb1f799573da2af45c3a3285c7a92a9f60"}]}