What was left from v4 was to figure out how to pair the yank register/unregister for the QEMUFiles.
I noticed that we don't really need to call unregister_yank under the lock. We just need to make sure the pointer is still valid. So I went with Peter's first suggestion which was to call the function individually for the QIOChannels we know yank is registered. CI run: https://gitlab.com/farosas/qemu/-/pipelines/988224440 v4: https://lore.kernel.org/r/20230816142510.5637-1-faro...@suse.de v3: https://lore.kernel.org/r/20230811150836.2895-1-faro...@suse.de v2: https://lore.kernel.org/r/20230802143644.7534-1-faro...@suse.de v1: https://lore.kernel.org/r/20230728121516.16258-1-faro...@suse.de Fabiano Rosas (8): migration: Fix possible race when setting rp_state.error migration: Fix possible races when shutting down the return path migration: Fix possible race when shutting down to_dst_file migration: Remove redundant cleanup of postcopy_qemufile_src migration: Consolidate return path closing code migration: Replace the return path retry logic migration: Move return path cleanup to main migration thread migration: Add a wrapper to cleanup migration files migration/migration.c | 231 +++++++++++++------------------------ migration/migration.h | 1 - migration/yank_functions.c | 5 + 3 files changed, 88 insertions(+), 149 deletions(-) -- 2.35.3
