Re: [PATCH 2/2] libxl: switch NetBSD image file handling to Qemu

[Roger Pau Monné <roger.pau%citrix.com@localhost>]

On 28/11/12 14:26, Thor Lancelot Simon wrote:
> On Wed, Nov 28, 2012 at 02:20:51PM +0100, Roger Pau Monne wrote:
>> This forces libxl to use Qemu when a raw image file is used as a disk
>> backend. NetBSD currently only supports qemu-traditional, so
> 
> Wait -- do I grasp correctly that this patch _forces_ NetBSD as a dom0
> to handle file-backed I/O requests in a way that includes additional
> context switches to userspace, and which your previous email said
> resulted in a 30% reduction in throughput compared to dom0 I/O?
> 
> I am assuming there's some reason we want this, rather than handling
> such requests in-kernel.  But what is it?

I should have marked this as "experimental", or something like this. I'm
not sure if switching to Qdisk for all image file backends is needed.
From the Xen wiki (http://wiki.xen.org/wiki/Blktap) I've found that:

"loop device had problems with flushing dirty pages (specifically, doing
a large number of writes to an NFS-backed image don't result in the OOM
killer going berserk)."

I'm not sure if NetBSD is in the same situation, but if I remember
correctly (haven't tried that in a long time), trying to use a disk file
on a NFS share caused the NetBSD Dom0 kernel to crash. The PR for this
issue is: http://gnats.netbsd.org/40726.

If this is solved, and the vnd device is solid in NetBSD we could only
use Qdisk when the image format is different than raw (qcow, qcow2,
vmdk, vdi...).