)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":1002047,"name":"Evgeniy Naydanov","email":"eugnay@gmail.com","username":"en-sc"},"change_message_id":"299329892ba22a6407ba9052c6f9c2c281db86bd","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"1b9d6782_aed433b8","updated":"2026-07-23 08:52:56.000000000","message":"Thank you for the patch!\n\nWere you able to observe this leak or is it a result of some static code analysis?\n\nThe issue is -- AFAIU, these resources should be de-allocated in `free_custom_register_names()`, that is called by `free_reg_names()`, called by `riscv_reg_free_all()`.\n\n`init_custom_register_names()` is only called by `riscv_reg_impl_init_cache()` (with a call to `riscv_reg_free_all()` before that, therefore preventing memory leak on examination re-entry).\n\nWhen the target is de-initialized, `riscv_reg_free_all()` is also called.\n\nThis makes me wonder how this leak can occur.\n\nI\u0027m not strictly opposed to making the code more robust by applying this patch regardless -- I don\u0027t like this approach of double-checking, but I do understand the appeal.\n\nAnyway, I believe if the patch does not fix a leak but rather is intended to improve code \"robustness\" by inserting an extra check, the commit message should be updated.","commit_id":"3dc538fff08d13df378129239fb83d0fb3b804a3"},{"author":{"_account_id":1000021,"name":"Antonio Borneo","email":"borneo.antonio@gmail.com","username":"borneoa"},"change_message_id":"c29aad1b9398d7ffd65954b04f91a77986188410","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"e489d794_ffdc1fbc","updated":"2026-07-23 08:28:33.000000000","message":"Thanks. Looks ok to me","commit_id":"3dc538fff08d13df378129239fb83d0fb3b804a3"},{"author":{"_account_id":1000021,"name":"Antonio Borneo","email":"borneo.antonio@gmail.com","username":"borneoa"},"change_message_id":"75770b90378bb9e7f38a8733dd3a80f57bb0ab8f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"6554f0d8_3c3a9ae4","updated":"2026-07-23 07:57:56.000000000","message":"When you push a new version of a patch, please keep the same line\n`Change-Id: ...`\nin the commit message so Gerrit understand it is a new version and tracks the evolution.","commit_id":"3dc538fff08d13df378129239fb83d0fb3b804a3"},{"author":{"_account_id":1002548,"name":"mengjunjun","display_name":"mengjunjun","email":"mengjunjun@eswincomputing.com","username":"mengjunjun1"},"change_message_id":"df2732ff81a6ca89b55a6c54cb951c934423956d","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"edd43700_c971339c","in_reply_to":"1b9d6782_aed433b8","updated":"2026-07-24 01:46:20.000000000","message":"Thanks for the review.\nI see now that there isn\u0027t actually a memory leak here. I had just come across this function and was concerned because there was no free operation before returning ERROR. However, if this function were to be used independently in other patches later on, there might be an issue. I agree with you—this patch can improves code robustness—so I’ve updated the commit message.","commit_id":"3dc538fff08d13df378129239fb83d0fb3b804a3"}]}