Re: [Qemu-devel] [PATCH] RFC: Add new block driver for the VDI format

[Anthony Liguori]

Stefan Weil wrote:
> Hello,
>
> this mail will be followed by two patches which
> allow QEMU to create, read and write VDI images.
> VDI is the native image format of SUN's VirtualBox emulator.
>
> The code was written from scratch for QEMU, while earlier patches
> sent to Qemu-devel were wrappers for code from VirtualBox. See
> http://lists.gnu.org/archive/html/qemu-devel/2008-07/msg00366.html
> for those patches.
>
> Patch 1 adds uuid support to QEMU and can be used independent of
> the second patch (look for uuid in vl.c, for example).
>
> Patch 2 adds the VDI block driver. It only needs uuid support
> when a new image is created (without uuid, it will create an image
> with zeroed "uuid" values), so this second patch is semi-independent
> of the first.
>   

I'd really like to get rid of synchronous IO functions in the block 
layer.  One way to do this is to insist that all new block drivers only 
implement the AIO functions.

I think we should make this decree but I'd like to know if other people 
think this is unreasonable first.  One potential model of block drivers 
would involve synchronous IO and threads.  I'm not a big fan of that 
model and I don't think it's an easy conversion from today's synchronous 
IO drivers to that model because the locking and re-entrance needs 
careful consideration.

Since it looks like you're caching the full offset table in memory, and 
it's a single level, making it asynchronous should be very easy to do.

Regards,

Anthony Liguori