Re: [Qemu-devel] [PATCHv3 1.8 0/9] qemu-img convert optimizations

[Peter Lieven]

Am 05.12.2013 13:15, schrieb Stefan Hajnoczi:
> On Wed, Nov 27, 2013 at 11:07:00AM +0100, Peter Lieven wrote:
>> this series adds some optimizations for qemu-img during convert which
>> have been developed recently:
>> - skipping input based on get_block_status
>> - variable I/O buffer size
>> - align write requests to cluster_size
>>
>> v2->v3:
>>   - added Paolos comments in Patch 1
>>   - changed the comment in patch 7 [Paolo]
>>   - remove the patch to add sector progress output
>>   - added a new patch to decrease the progress update interval.
>>
>> v1->v2:
>>   - introduce opt_transfer_length in BlockLimits [Paolo]
>>   - remove knobs for iobuffer_size and alignment and
>>     use them unconditionally [Paolo]
>>   - calculate I/O buffer size by BlockLimits information [Paolo]
>>   - change the alignment patch to round down to the
>>     last and not to the next aligned sector [Paolo]
>>   - limit updates in the sector progress output
>>   - new patch to increase the default for min_sparse [Paolo]
>>
>> Peter Lieven (9):
>>   qemu-img: add support for skipping zeroes in input during convert
>>   qemu-img: fix usage instruction for qemu-img convert
>>   block/iscsi: set bdi->cluster_size
>>   block: add opt_transfer_length to BlockLimits
>>   block/iscsi: set bs->bl.opt_transfer_length
>>   qemu-img: dynamically adjust iobuffer size during convert
>>   qemu-img: round down request length to an aligned sector
>>   qemu-img: increase min_sparse to 128 sectors (64kb)
>>   qemu-img: decrease progress update interval on convert
>>
>>  block/iscsi.c             |   10 ++++
>>  include/block/block_int.h |    3 ++
>>  qemu-img.c                |  131 
>> +++++++++++++++++++++++++++------------------
>>  qemu-img.texi             |    2 +-
>>  4 files changed, 93 insertions(+), 53 deletions(-)
> Merged all except patch 8/9.
>
> Thanks, applied to my block tree:
> https://github.com/stefanha/qemu/commits/block
Thank you.

As discussed I've sent the follow-up patch

    [PATCH] qemu-img: make progress output more accurate during convert

to the list some minutes ago.

Peter