When waiting for the return path (RP) thread to finish, there is
really nothing wrong in the RP if the destination end of the migration
stops responding, leaving it stuck.
Stop returning an error at that point and leave it to other parts of
the code to catch. One such part is the very next routine run by
migration_completion() which checks 'to_dst_file' for an error and fails
the migration. Another is the RP thread itself when the recvmsg()
returns an error.
With this we stop marking RP bad from outside of the thread and can
reuse await_return_path_close_on_source() in the next patches to wait
on the thread during a paused migration.
Signed-off-by: Fabiano Rosas <faro...@suse.de>
---
migration/migration.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/migration/migration.c b/migration/migration.c
index 91bba630a8..051067f8c5 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -2049,7 +2049,6 @@ static int
await_return_path_close_on_source(MigrationState *ms)
* waiting for the destination.
*/
qemu_file_shutdown(ms->rp_state.from_dst_file);
- mark_source_rp_bad(ms);
}
trace_await_return_path_close_on_source_joining();
qemu_thread_join(&ms->rp_state.rp_thread);
--
2.35.3
