Re: [Qemu-devel] [PATCH v2 0/3] qemu-img convert: Rewrite copying logic

[Kevin Wolf <kwolf@redhat.com>]

Am 19.03.2015 um 13:33 hat Kevin Wolf geschrieben:
> See the commit message of patch 2 for the background.
> 
> v2:
> - Rebased. Commit f1d3cd79 changed many bdrv_* calls to blk_* calls.
>   This version tries to keep those conversions and use BlockBackend
>   where qemu.git master does or where it makes sense. Added patch 1 to
>   get blk_write_zeroes().
> - Renamed s->out_backing as s->target_has_backing [Max]
> - convert_iteration_sectors(): Limit n to BDRV_REQUEST_MAX_SECTORS
>   instead of INT_MAX [Max]
> - convert_read(): Fixed sector number for image concatenation [Max]
> - Fixed -S 0 for full allocation [Max, Peter]
> - Leave has_zero_init = false instead of failing the conversion if
>   bdrv_make_zero() returns an error [Max]
> - Set has_zero_init = true instead of 1 [Max]
> - Don't ignore error returns in a final blk_write_compressed() [Max]
> - Removed TODO comment after deciding that no change is needed
> - Many added tests
> 
> Kevin Wolf (3):
>   block-backend: Expose bdrv_write_zeroes()
>   qemu-img convert: Rewrite copying logic
>   qemu-iotests: Some qemu-img convert tests

Removed the test case from the "quick" group as suggested by Max,
and applied to block-next.

Kevin