On 16.03.21 17:30, Stefan Reiter wrote: > Saving dirty bitmaps from our savevm-async code didn't work, since we > use a coroutine which holds the iothread mutex already (upstream savevm > is sync, migration uses a thread). Release the mutex before calling the > one function that (according to it's documentation) requires the lock to > *not* be held: qemu_savevm_state_pending. > > Additionally, loading dirty bitmaps requires a call to > dirty_bitmap_mig_before_vm_start after "loadvm", which the upstream > savevm does explicitly afterwards - do that too. > > This is exposed via the query-proxmox-support property > "pbs-dirty-bitmap-savevm". > > Signed-off-by: Stefan Reiter <s.rei...@proxmox.com> > --- > > This fixes a bug reported in the forum[0] as well as enterprise support. > > This patch contains the actual fix, i.e. what makes it not crash, but we can > actually go one step further, and by adding the query-proxmox-support info we > can avoid crashes even with older versions (see patch 2/2). > > As a neat side-effect, this also allows us to fix dirty bitmaps for snapshots > and hibernate/resume, so that's a plus :) > > [0] https://forum.proxmox.com/threads/query-savevm.85719/ > > ...async-for-background-state-snapshots.patch | 15 +++++++++++---- > ...add-optional-buffer-size-to-QEMUFile.patch | 6 +++--- > ...dd-query_proxmox_support-QMP-command.patch | 19 ++++++++++++------- > ...E-add-query-pbs-bitmap-info-QMP-call.patch | 17 ++++++++++------- > ...-transaction-to-synchronize-job-stat.patch | 2 +- > ...-block-on-finishing-and-cleanup-crea.patch | 4 ++-- > ...igrate-dirty-bitmap-state-via-savevm.patch | 18 ++++++++++-------- > ...routine-QMP-for-backup-cancel_backup.patch | 4 ++-- > .../pve/0043-PBS-add-master-key-support.patch | 16 ++++++++-------- > 9 files changed, 59 insertions(+), 42 deletions(-) > >
applied, thanks! _______________________________________________ pve-devel mailing list pve-devel@lists.proxmox.com https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
