Re: [Qemu-devel] [PATCH v5 00/11] qcow2: Fix image repairing

[Max Reitz <mreitz@redhat.com>]

On 29.08.2014 23:40, Max Reitz wrote:
> As can be seen in the final patch of this series, there are certain
> cases where the current repair implementation of qcow2 actually damages
> the image further because it allocates new clusters for the refcount
> structure which overlap with existing but according to the on-disk
> refcounts (which are assumed to be wrong to begin with) unallocated
> clusters.
>
> This series fixes this by completely recreating the refcount structure
> based on the in-memory information calculated during the check operation
> if the possibility of damaging the image while repairing the refcount
> structures in-place exists.
>
>
> v5:
> - Added patch 1 which adds two helper variables to BDRVQcowState
>    reflecting the number of entries per refcount block;
>    in contrast to v4, we don't need to clamp the refcount order against
>    sub-byte widths, because sub-byte widths are actually correct (that
>    means, I dropped the MAX() around refcount_order - 3)
> - Patch 8 (prev. 7):
>    - Use these new variables [Benoît]
>    - Use a struct for rt_offset_and_clusters [Benoît]
>
>
> git-backport-diff against v4:
>
> 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
>
> 001/11:[down] 'qcow2: Calculate refcount block entry count'
> 002/11:[----] [--] 'qcow2: Fix leaks in dirty images'
> 003/11:[----] [--] 'qcow2: Split qcow2_check_refcounts()'
> 004/11:[----] [--] 'qcow2: Pull check_refblocks() up'
> 005/11:[----] [--] 'qcow2: Reuse refcount table in calculate_refcounts()'
> 006/11:[----] [--] 'qcow2: Fix refcount blocks beyond image end'
> 007/11:[----] [--] 'qcow2: Do not perform potentially damaging repairs'
> 008/11:[0025] [FC] 'qcow2: Rebuild refcount structure during check'
> 009/11:[----] [--] 'qcow2: Clean up after refcount rebuild'
> 010/11:[----] [--] 'iotests: Fix test outputs'
> 011/11:[----] [-C] 'iotests: Add test for potentially damaging repairs'
>
>
> Max Reitz (11):
>    qcow2: Calculate refcount block entry count
>    qcow2: Fix leaks in dirty images
>    qcow2: Split qcow2_check_refcounts()
>    qcow2: Pull check_refblocks() up
>    qcow2: Reuse refcount table in calculate_refcounts()
>    qcow2: Fix refcount blocks beyond image end
>    qcow2: Do not perform potentially damaging repairs
>    qcow2: Rebuild refcount structure during check
>    qcow2: Clean up after refcount rebuild
>    iotests: Fix test outputs
>    iotests: Add test for potentially damaging repairs
>
>   block/qcow2-refcount.c     | 677 ++++++++++++++++++++++++++++++++-------------
>   block/qcow2.c              |   4 +-
>   block/qcow2.h              |   2 +
>   tests/qemu-iotests/039.out |  10 +-
>   tests/qemu-iotests/060.out |  10 +-
>   tests/qemu-iotests/061.out |  18 +-
>   tests/qemu-iotests/104     | 141 ++++++++++
>   tests/qemu-iotests/104.out | 110 ++++++++
>   tests/qemu-iotests/group   |   1 +
>   9 files changed, 767 insertions(+), 206 deletions(-)
>   create mode 100755 tests/qemu-iotests/104
>   create mode 100644 tests/qemu-iotests/104.out
>

Ping. Patch 8 still requires a review. I know it's probably the hardest 
patch of the series, but well...

Max