Re: [Xen-devel] [PATCH] xl: Special case tap/aio for disk validation

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Thu, 2011-01-27 at 15:22 +0000, Ian Jackson wrote:
> Stefano Stabellini writes ("Re: [Xen-devel] [PATCH] xl: Special case tap/aio 
> for disk validation"):
> > - A fix for parse_disk_config.
> > The new algorithm should assume format=raw and ignore tap:, tap2:, aio:,
> > tapdisk:, ioemu:, until it gets to a real disk format (qcow:, qcow2:,
> > vhd:) or the file name.
> 
> You can recognise the filename by the fact that it starts with a "/",
> since most of our stuff doesn't work with relative pathnames for block
> devices and we don't want it to try to make it work now.

FWIW xend notices a relative path and wacks a "/dev/" on the front.

> So the specification string should always be either
>    someprefix:[stuff]
> in which case you strip off someprefix: (perhaps making a note of its
> implications) or /somepath
> in which case you've got to the end and know what to do.

Shall we pre-deprecate the someprefix: along with all the
tap:{aio,ioemu,FOO}: stuff while we are at this? IOW the syntax we want
to actually support going forward would be:
        (raw|vhd|qcow2|qcow):/path/to/something
with the toolstack being at liberty to best decide how to achieve the
necessary format support? (optionally raw: as the default). All the rest
of the format is really semantically meaningless and supported only for
compatibility with existing configuration files.

Ian.



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel