[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 1/2] block: qemu-iotests - add basic ability to
From: |
Jeff Cody |
Subject: |
Re: [Qemu-devel] [PATCH 1/2] block: qemu-iotests - add basic ability to use binary sample images |
Date: |
Fri, 20 Sep 2013 09:23:44 -0400 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
On Fri, Sep 20, 2013 at 06:46:56AM -0600, Eric Blake wrote:
> On 09/19/2013 09:48 PM, Jeff Cody wrote:
> > For image formats that are not "QEMU native", but supported for
> > compatibility, it is useful to verify that an image created with
> > the 'gold standard' native tool can be read / written to successfully
> > by QEMU.
> >
> > In addition to testing non-native images, this could also be useful to
> > test against image files created by older versions of QEMU.
> >
> > This provides a directory to store small sample images, for use by
> > scripts in tests/qemu-iotests.
> >
> > The MANIFEST file should be updated when a new image is added to the
> > directory, to give some idea about the nature of the image and the
> > data store therein.
> >
> > Image files should be compressed with bzip2.
>
> Why bzip2? xz beats bzip2 in both compression ratio and decompression
> speed; or if you are worried about portability, gzip is present on more
> machines. bzip2 will probably be around for some time, but it is no
> longer the compression engine of choice.
>
iotest-dynamic-1G.vhdx compressed with gzip, xv, and bzip2:
gzip: 102548 bytes
xv: 15892 bytes
bzip2: 874 bytes
I think bzip2 is particularly well suited for image files with
hyper-repetitive consecutive bytes due to its use of RLE. I don't
know that all sample images will have data like that, but chances are
good many will - and even if they don't, bzip2 is not horrible
compared to xz even on normal data.
-Jeff