Re: [libvirt] [PATCH] migration: Fix possible bug for migrate cancel

-----Original Message----- From: Eric Blake [mailto:eblake@redhat.com] Sent: Tuesday, March 25, 2014 12:01 AM To: Paolo Bonzini; Gonglei (Arei); qemu-devel(a)nongnu.org Cc: quintela(a)redhat.com; owasserm(a)redhat.com; Yanqiangjun; Zhaoyanbin (A); Zengjunliang; libvir-list(a)redhat.com Subject: Re: [PATCH] migration: Fix possible bug for migrate cancel [adding libvirt] On 03/24/2014 09:47 AM, Paolo Bonzini wrote: > Il 24/03/2014 14:04, arei.gonglei(a)huawei.com ha scritto: >> From: zengjunliang <zengjunliang(a)huawei.com&gt; >> >> Return error for migrate cancel, when migration status is not >> MIG_STATE_SETUP or MIG_STATE_ACTIVE. Thus, libvirt can can >> perceive the operation fails. >> >> Signed-off-by: zengjunliang <zengjunliang(a)huawei.com&gt; >> Signed-off-by: Gonglei <arei.gonglei(a)huawei.com&gt; > > I think this is done on purpose, because canceling migration is racy. > Instead, libvirt should do "query-migrate" and check if the migration > was completed or canceled. Can you please give more details at how you are triggering the problem with libvirt? I think Paolo is probably right - the bug is more likely to be in libvirt not expecting the race and not recovering correctly when the race occurs, than it is to be in changing qemu's state algorithm.