qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH 0/6] block: Fix slow pre-zeroing in qemu-img con


From: Nir Soffer
Subject: Re: [Qemu-devel] [PATCH 0/6] block: Fix slow pre-zeroing in qemu-img convert
Date: Sun, 24 Mar 2019 01:11:19 +0200

On Fri, Mar 22, 2019 at 10:47 PM Eric Blake <address@hidden> wrote:

> On 3/22/19 3:22 PM, Nir Soffer wrote:
>
> > Thanks for fixing this! This will be a very important performance fix for
> > importing
> > VMs.
> >
> > The change description sounds good to me.
> >
> > Can you make these patches available on some git repo to make it easy to
> > test?
>
> Based on commit titles, this branch looks reasonably close (at the moment):
>
> https://repo.or.cz/qemu/kevin.git/shortlog/refs/heads/qcow2-data-file


I tested the patches and they work correctly with NBD server (tested both
nbdkit
and qemu-nbd), skipping pre zero step, and zeroing only the unallocated
areas.

However zeroing is always done without BDRV_REQ_MAY_UNMAP, creating
fully allocated images when the expected result is sparse image.

I posted a patch with possible fix, explaining this issue with mode details.
I don't see it yest in mail archive, but the patch is available here:
https://github.com/nirs/qemu/commit/5023405ee42dc0d8e9a03ac03d247079f719b455

Nir


reply via email to

[Prev in Thread] Current Thread [Next in Thread]