)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"f10ecb240f710db29ac5c4ab70e5583c56c85af0","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"55ef4230_a84fff9b","updated":"2026-04-30 12:53:38.000000000","message":"Samuel, thanks for the fix.\n\nWhen working on #9180 I missed this direct setting of examine flag because\nall other code uses setters `target_set_examined()` and `target_reset_examined()`. We should avoid direct write to examine flag.\nEither make both `target_reset_examined()` and `target_reset_active_polled()`\nexposed in targed.h or introduce something like `target_reset_examined_stop_polling()` which resets both flags.\n\nPlease tell me when an xtensa targed keeps indicating DM is not powered? Should not we keep polling anyway? Or stop polling only if `-defer-examine` is set?","commit_id":"0f1b5378cbf9935eb9441a0285dbf8dcedc72d05"},{"author":{"_account_id":1001803,"name":"Samuel Obuch","email":"samuel.obuch@espressif.com","username":"sobuch"},"change_message_id":"e9d70cd7fae30ec5c91b00015c8eeafc1cc13719","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"e2c301ab_06f0aa49","updated":"2026-05-04 14:10:58.000000000","message":"I believe these checks are mainly not to spam the user with error messages - with esp32, cpu1 can be stalled and then fails examine.\n\nWe had some failed tests, but that seemed to be due to polling interval growing too large.","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"},{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"ac70b59bb99da2af2fea1d279ac73cd4bc7ab50b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"20c0f1f0_1a1b6ffe","updated":"2026-05-06 07:04:35.000000000","message":"Jenkins build result looks strange:\n`Build timed out (after 45 minutes). Marking the build as aborted.`\n\nPaul, could you please look at the logs. BTW it\u0027s not the first time I\u0027ve seen such timeout https://review.openocd.org/c/openocd/+/6964/comments/4509eb0e_37dee8b1","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"},{"author":{"_account_id":1001982,"name":"Ian Thompson","email":"ianst@cadence.com","username":"ianstcdns"},"change_message_id":"c5246690534c341416a6b12252348004fe8eb2a5","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"e695b5f8_27a96111","updated":"2026-05-04 13:59:47.000000000","message":"Thanks, Samuel.  Looks good.","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"},{"author":{"_account_id":1000160,"name":"Paul Fertser","email":"fercerpav@gmail.com","username":"pfertser"},"change_message_id":"5934cf8efb0d2ff10ac3cc63e6271ff775e4c93f","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":2,"id":"847d25c3_9a45550e","in_reply_to":"20c0f1f0_1a1b6ffe","updated":"2026-05-06 13:41:13.000000000","message":"\u003e Paul, could you please look at the logs. BTW it\u0027s not the first time I\u0027ve seen such timeout https://review.openocd.org/c/openocd/+/6964/comments/4509eb0e_37dee8b1\n\nI noticed it before and tried to check everything related. It looked like the server was just slower than usual for no apparent reason at that moment... Might be Gerrit getting hammered by some bots and since it takes priority it slows the build to a halt, not sure, it doesn\u0027t look like that.\n\nGuess we should ask for VPS upgrade soon as it\u0027s really underpowered (wrt RAM for sure, I had to hand tune swapping behaviour for Gerrit/Jenkins balancing) for today\u0027s software.","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"},{"author":{"_account_id":1001803,"name":"Samuel Obuch","email":"samuel.obuch@espressif.com","username":"sobuch"},"change_message_id":"18e46326a92498f3c86dbc7b6926f067dcbe1bb2","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":2,"id":"f998ecc8_a6089b76","in_reply_to":"44c01a49_aae2395b","updated":"2026-05-11 09:53:31.000000000","message":"Thanks for looking into this.\n\nWe support similar functionality also for riscv targets, however there we simply rely on `target-\u003estate` instead of resetting `target-\u003eexamined`. Should be a better solution, as the current one can break even if the user just holds the reset button a bit too long.\n\nI checked that this also works for the xtensa esp32 target, but it will require bunch of checks for `target-\u003estate`. For example we have this patch to handle breakpoints for SMP targets when one of them is `unavailable` - https://github.com/espressif/openocd-esp32/commit/b51fb6f25bfaae49a8e09354e96113518bc4b515\n\nSo I think best to merge this as a bugfix, and then improve this logic later (probably after the release)","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"},{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"9ee186a3cada2ac29f441e9cb138d20f5e68a06e","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":2,"id":"44c01a49_aae2395b","in_reply_to":"8afd3bad_addd2bf5","updated":"2026-05-06 08:58:22.000000000","message":"I changed my mind: using `target-\u003edefer_examine` here is not a good idea as option `-defer-examine` would block examination. Anyway we should at least inform user similarly as we do in case of `-defer-examine`.","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"},{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"29793e390b60139a3460348736c026687e760c11","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":2,"id":"8afd3bad_addd2bf5","in_reply_to":"e2c301ab_06f0aa49","updated":"2026-05-06 07:27:46.000000000","message":"\u003e I believe these checks are mainly not to spam the user with error messages - with esp32, cpu1 can be stalled and then fails examine.\n\nIf I understand correctly this scenario is covered in 3 counts of `come_online_probes_num` and `target_reset_active_polled()` is not called.\n\nBut if the outage is longer and `target_reset_active_polled()` gets called, the target is abandoned without polling and no warning issued about this.\nTo revive the target operation OpenOCD needs at least `arp_examine` or some `reset`. No other target can similarly disconnect of its own will. The only mechanism which resembles this behaviour is a target with `-defer-examine` option. That\u0027s why I would recommend to call `target_reset_active_polled()`\nonly if `target-\u003edefer_examine` is set.","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"}],"src/target/target.c":[{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"9ee186a3cada2ac29f441e9cb138d20f5e68a06e","unresolved":false,"context_lines":[{"line_number":5253,"context_line":""},{"line_number":5254,"context_line":"\tif (allow_defer \u0026\u0026 target-\u003edefer_examine) {"},{"line_number":5255,"context_line":"\t\tLOG_INFO(\"Deferring arp_examine of %s\", target_name(target));"},{"line_number":5256,"context_line":"\t\tLOG_INFO(\"Use arp_examine command to examine it manually!\");"},{"line_number":5257,"context_line":"\t\treturn ERROR_OK;"},{"line_number":5258,"context_line":"\t}"},{"line_number":5259,"context_line":""}],"source_content_type":"text/x-csrc","patch_set":2,"id":"f2a639e5_e08fa68a","line":5256,"updated":"2026-05-06 08:58:22.000000000","message":"Issue similar info.","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"},{"author":{"_account_id":1001803,"name":"Samuel Obuch","email":"samuel.obuch@espressif.com","username":"sobuch"},"change_message_id":"18e46326a92498f3c86dbc7b6926f067dcbe1bb2","unresolved":false,"context_lines":[{"line_number":5253,"context_line":""},{"line_number":5254,"context_line":"\tif (allow_defer \u0026\u0026 target-\u003edefer_examine) {"},{"line_number":5255,"context_line":"\t\tLOG_INFO(\"Deferring arp_examine of %s\", target_name(target));"},{"line_number":5256,"context_line":"\t\tLOG_INFO(\"Use arp_examine command to examine it manually!\");"},{"line_number":5257,"context_line":"\t\treturn ERROR_OK;"},{"line_number":5258,"context_line":"\t}"},{"line_number":5259,"context_line":""}],"source_content_type":"text/x-csrc","patch_set":2,"id":"adca616a_2c418bea","line":5256,"in_reply_to":"f2a639e5_e08fa68a","updated":"2026-05-11 09:53:31.000000000","message":"Added printouts, explicitly suggesting `reset halt` as that is the only way this works reliably.","commit_id":"c257c0c1a80e5c5b5752754eb70681220a911d7e"}],"src/target/xtensa/xtensa.c":[{"author":{"_account_id":1001982,"name":"Ian Thompson","email":"ianst@cadence.com","username":"ianstcdns"},"change_message_id":"948c3575c7eb46df46b82c620d67392fc2ae2f0e","unresolved":true,"context_lines":[{"line_number":2367,"context_line":"\t\ttarget-\u003estate \u003d TARGET_UNKNOWN;"},{"line_number":2368,"context_line":"\t\tif (xtensa-\u003ecome_online_probes_num \u003d\u003d 0) {"},{"line_number":2369,"context_line":"\t\t\ttarget-\u003eexamined \u003d false;"},{"line_number":2370,"context_line":"\t\t\ttarget-\u003eactive_polled \u003d false;"},{"line_number":2371,"context_line":"\t\t} else {"},{"line_number":2372,"context_line":"\t\t\txtensa-\u003ecome_online_probes_num--;"},{"line_number":2373,"context_line":"\t\t}"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"a7b8c105_25da2c7a","line":2370,"updated":"2026-05-01 14:47:05.000000000","message":"Thanks, Samuel.  Your changes look correct to me.  I agree with Tomas that exposing the reset functions and using them would be cleaner.  This is the only location the Xtensa target clears the active_polled or examined fields, so exposing those APIs in this review would remain small in scope (assuming Tomas supports that).\n\nRe: DM not powered indication--this could happen briefly during Xtensa reset.  Some Espressif chips have multiple Xtensa targets; Erhan might be able to comment on whether they rely on this behavior for longer power-down support.","commit_id":"0f1b5378cbf9935eb9441a0285dbf8dcedc72d05"}]}