)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":1001803,"name":"Samuel Obuch","email":"samuel.obuch@espressif.com","username":"sobuch"},"change_message_id":"b6ced60dcd2db5087de2e8af3a75e5c6df8d5805","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":3,"id":"acd49acf_6d1af77b","updated":"2025-09-16 07:52:23.000000000","message":"Please note this will conflict with https://review.openocd.org/c/openocd/+/9120\n\nGiven that software breakpoints are assumed to work for the whole SMP group, could we find a better place to keep them than the `t-\u003ebreakpoints` list for whichever target was active/available at the time the breakpoint was added?","commit_id":"34ecaf8dd7bcbc21fe2b4edc8e4b2785a6cbad56"},{"author":{"_account_id":1001803,"name":"Samuel Obuch","email":"samuel.obuch@espressif.com","username":"sobuch"},"change_message_id":"7394673dabac58b923cddf55a81c86ca948d7649","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":3,"id":"47ec86c6_88fdb531","in_reply_to":"0e41584c_69777590","updated":"2025-09-22 08:38:54.000000000","message":"yes, seems best to adjust https://review.openocd.org/c/openocd/+/9120 for now. \n\nThe note about `smp on|off` reminded me of one more aspect to consider - for targets with core-local memories (tightly/closely coupled memories) it could be useful to have an option to only set for a single target not the whole SMP group.\n\nI guess `smp on|off` can be used now as a workaround, but it doesnt seem practical. E.g. when using GDB, it will remove/reinsert breakpoints on each halt/resume, so that would require configuring some complicated target event handlers.","commit_id":"34ecaf8dd7bcbc21fe2b4edc8e4b2785a6cbad56"},{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"8cfdfd60d80961fd76719662499f172bde2757a6","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":3,"id":"7722eae8_8bdba6a4","in_reply_to":"47ec86c6_88fdb531","updated":"2025-09-23 15:39:12.000000000","message":"\u003e targets with core-local memories (tightly/closely coupled memories) it could be useful to have an option to only set for a single target not the whole SMP group.\n\nGood point. To be precise we should configure a group of targets and memory range they share. Not necessarily same group as the SMP group. Quite complicated...","commit_id":"34ecaf8dd7bcbc21fe2b4edc8e4b2785a6cbad56"},{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"ac375dbb1c535a7b38c9c285fdf66f33e1bd82eb","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":3,"id":"ff2acf3a_98e32fb3","in_reply_to":"acd49acf_6d1af77b","updated":"2025-09-17 10:25:25.000000000","message":"Yes, there is, I already tested the patches together.\n\nNot only yours #9120, also the old code prefers core 0 over the selected code for working with soft breakpoints. It makes sense as core 0 is usually the controlling one and stays always available.\nOn the other hand there as some devices where this assumption is not true\n(e.g. RP2350 with switchable Cortex-M33/RV32 in two core slots).\nAnd moreover, I don\u0027t like that selected current target is ignored - this may complicate testing.\n\nI also thought about introducing a semi-target for SMP group with some methods and with the list of shared soft breakpoints. Honestly I\u0027m not happy to make such changes quickly just before OpenOCD release.\n\nThe second possibility could be to adjust #9120 (and probably other parts of original code) to loop over all soft bkpts in all targets of SMP list. This would be kind of messy as a soft bkpt could end in the list of any core from the SMP group. However it looks to me much easier to implement. What do you think?","commit_id":"34ecaf8dd7bcbc21fe2b4edc8e4b2785a6cbad56"},{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"03a47748ed79d8ba5b6f85c4544a388cff7fbe2e","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":3,"id":"0e41584c_69777590","in_reply_to":"ff2acf3a_98e32fb3","updated":"2025-09-17 16:14:54.000000000","message":"There is another aspect which haven\u0027t been addressed: command `smp on|off`.\nFortunately the command is rarely used. Manual reads\n`Display, enable or disable SMP handling mode. This command is needed only if user wants to temporary disable SMP handling for an existing SMP group`\n\nTaking this command into account, we should set soft bkpt on one core and create breakpoint record in a shared list or in all cores lists in a SMP group (if such exists), regardless of target-\u003esmp state. If the target is later subject to *temporary disable SMP handling* then it would keep the correct list of breakpoints.","commit_id":"34ecaf8dd7bcbc21fe2b4edc8e4b2785a6cbad56"},{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"77d53c5e28946c3b4822baac57883f24405872f0","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":5,"id":"b06d6872_b620c2ca","updated":"2025-10-14 14:03:27.000000000","message":"TODO: Implement breakpoint_remove() counterpart\n\nriscv-collab implementation\nhttps://github.com/riscv-collab/riscv-openocd/commit/39a4f37f84d52aa38ffa1a7db703ad57deec4fdc\ntarget/breakpoints: Clear software breakpoints from available targets\n\npresumes the target must be halted and available, so it\u0027s not generally usable.","commit_id":"b13cf415b5346e7f571981ab5efc946616719bc5"}]}