Re: [PATCH v4 24/30] qcow2: Clear the L2 bitmap when allocating a compressed cluster

[Vladimir Sementsov-Ogievskiy]

24.04.2020 21:41, Alberto Garcia wrote:
> On Fri 24 Apr 2020 08:15:04 PM CEST, Vladimir Sementsov-Ogievskiy 
> <address@hidden> wrote:
> > AFAIK, now compressed clusters can't be used in scenarios with guest,
> > as qcow2 driver doesn't support rewriting them.
>
> You can write to those images just fine, it's just not efficient because
> you have to COW the compressed clusters.

No, rewriting doesn't work:

[root@kvm master]# ./qemu-img create -f qcow2 x 10M
Formatting 'x', fmt=qcow2 size=10485760 cluster_size=65536 lazy_refcounts=off 
refcount_bits=16
[root@kvm master]# ./qemu-io -c 'write -c 0 64K' x
wrote 65536/65536 bytes at offset 0
64 KiB, 1 ops; 00.23 sec (278.708 KiB/sec and 4.3548 ops/sec)
[root@kvm master]# ./qemu-io -c 'write -c 0 64K' x
write failed: Input/output error


> > Or am I wrong? And we normally don't combine normal and compressed
> > clusters together in one image.
>
> As soon as you start writing to an image with compressed clusters you'll
> have a combination of both.

Ah, you mean, rewriting compressed clusters by normal.. So the use-case is just 
take compressed backup image and use it for the guest, instead of converting 
first.. It makes sense.

> But it's true that you don't have an image with compressed clusters if
> what you're looking for is performance. So I wouldn't add support for
> this if it complicates things too much.
>
> Berto


--
Best regards,
Vladimir