Currently, bdrv_close_all() force-closes all BDSs with a BlockBackend, which can lead to data corruption (see the iotest added in the final patch of this series) and is most certainly very ugly.
This series reworks bdrv_close_all() to instead eject the BDS trees from all BlockBackends and then close the monitor-owned BDS trees, which are the only BDSs without a BB. In effect, all BDSs are closed just by getting closed automatically due to their reference count becoming 0. Note that the approach taken here leaks all BlockBackends. This does not really matter, however, since qemu is about to exit anyway. v5 is here (yes, it has been a while): http://lists.nongnu.org/archive/html/qemu-block/2015-03/msg00044.html v6: - Patch 1: Added (leakage of a BDS in a drive-backup error case, becomes apparent with this series) - Patch 2: Added (leakage of a BDS when a block job failed to be created) - Patch 3: Added (we hacked our way around this so far, but now we have to fix it, see the commit message) - Patch 5: Rebase conflicts - Patch 6: Renamed test from 096 to 140 (rebase conflict) - Patch 7: - Contextual rebase conflicts - Rebase conflict in a comment being removed - Dropping the reference to exp->blk has been moved from nbd_export_close() to nbd_export_put(), so the notifier_remove() call must be moved there, too - Patch 8: blk_remove_bs() no longer silently ignores blk->bs being NULL, therefore we have to keep it wrapped in if (blk->bs) - Patch 9: - The bdrv_drain_all() and bdrv_flush() calls have been removed from hmp_drive_del() already, so we don't need to do it anymore. - As in patch 8, blk_remove_bs() must stay in the if (blk->bs) block - Patch 11: Do not move the list entry in all_bdrv_states in bdrv_move_feature_fields(); every BDS is part of that list, so there is absolutely no point in swapping their positions here - Patch 12: - Do not move the list entry in the list of monitor-owned BDS in bdrv_move_feature_fields(); if a BDS is owned by the monitor, that will not change by putting another BDS on top of it or by putting it on top of another BDS - Add test to qmp_x_blockdev_del() whether the BDS to be deleted (if a node is to be deleted) is actually monitor-owned - If qmp_x_blockdev_del() unrefs a node, it must be dropped from the list of monitor-owned BDS - Patch 13: Enclose blk_remove_bs() in an if (blk->bs) block (just like in patches 8 and 9) - Patch 14: - One semi-contextual conflict in which an empty line addition is dropped, because now there already is an empty line - Changes to bdrv_close_all(): - Add a bdrv_drain_all() at the beginning; it may not do anything, but it looked like a reasonable thing to do to me - Add a comment what the loop is for - A block job will not immediately release its BDS reference once it is canceled (maybe it did back in March, but now it doesn't), so we need to call aio_poll() repeatedly until the reference is dropped git-backport-diff against v5: Key: [----] : patches are identical [####] : number of functional differences between upstream/downstream patch [down] : patch is downstream-only The flags [FC] indicate (F)unctional and (C)ontextual differences, respectively 001/15:[down] 'blockdev: Add missing bdrv_unref() in drive-backup' 002/15:[down] 'blockjob: Call bdrv_unref() on creation error' 003/15:[down] 'block: Release dirty bitmaps in bdrv_close()' 004/15:[----] [--] 'iotests: Move _filter_nbd into common.filter' 005/15:[0009] [FC] 'iotests: Make redirecting qemu's stderr optional' 006/15:[0004] [FC] 'iotests: Add test for eject under NBD server' 007/15:[0007] [FC] 'block: Move BDS close notifiers into BB' 008/15:[0004] [FC] 'block: Use blk_remove_bs() in blk_delete()' 009/15:[0006] [FC] 'blockdev: Use blk_remove_bs() in do_drive_del()' 010/15:[----] [-C] 'block: Make bdrv_close() static' 011/15:[0003] [FC] 'block: Add list of all BlockDriverStates' 012/15:[0009] [FC] 'blockdev: Keep track of monitor-owned BDS' 013/15:[0004] [FC] 'block: Add blk_remove_all_bs()' 014/15:[0025] [FC] 'block: Rewrite bdrv_close_all()' 015/15:[----] [-C] 'iotests: Add test for multiple BB on BDS tree' Max Reitz (15): blockdev: Add missing bdrv_unref() in drive-backup blockjob: Call bdrv_unref() on creation error block: Release dirty bitmaps in bdrv_close() iotests: Move _filter_nbd into common.filter iotests: Make redirecting qemu's stderr optional iotests: Add test for eject under NBD server block: Move BDS close notifiers into BB block: Use blk_remove_bs() in blk_delete() blockdev: Use blk_remove_bs() in do_drive_del() block: Make bdrv_close() static block: Add list of all BlockDriverStates blockdev: Keep track of monitor-owned BDS block: Add blk_remove_all_bs() block: Rewrite bdrv_close_all() iotests: Add test for multiple BB on BDS tree block.c | 83 ++++++++++++++++++++++++------- block/block-backend.c | 41 +++++++++++++--- blockdev-nbd.c | 37 +------------- blockdev.c | 28 ++++++++++- blockjob.c | 1 + hw/block/dataplane/virtio-blk.c | 77 ++++++++++++++++++++++------- hw/scsi/virtio-scsi.c | 59 ++++++++++++++++++++++ include/block/block.h | 2 - include/block/block_int.h | 8 ++- include/hw/virtio/virtio-scsi.h | 10 ++++ include/sysemu/block-backend.h | 4 +- nbd.c | 13 +++++ stubs/Makefile.objs | 1 + stubs/blockdev-close-all-bdrv-states.c | 5 ++ tests/qemu-iotests/083 | 13 +---- tests/qemu-iotests/083.out | 10 ---- tests/qemu-iotests/117 | 86 ++++++++++++++++++++++++++++++++ tests/qemu-iotests/117.out | 14 ++++++ tests/qemu-iotests/140 | 90 ++++++++++++++++++++++++++++++++++ tests/qemu-iotests/140.out | 16 ++++++ tests/qemu-iotests/common.filter | 12 +++++ tests/qemu-iotests/common.qemu | 15 +++++- tests/qemu-iotests/group | 2 + 23 files changed, 518 insertions(+), 109 deletions(-) create mode 100644 stubs/blockdev-close-all-bdrv-states.c create mode 100755 tests/qemu-iotests/117 create mode 100644 tests/qemu-iotests/117.out create mode 100755 tests/qemu-iotests/140 create mode 100644 tests/qemu-iotests/140.out -- 2.6.2
