There is a WIP on upstream to address this virtio-scsi request allocation slowness:

http://lists.gnu.org/archive/html/qemu-devel/2014-09/msg02896.html

It should be introduced by the g_malloc -> g_malloc0 change, because the structure we are allocating is nearly 50KB (because of the .elem field).

I've applied the patches onto RHEV 7.1 to get this scratch build:

https://brewweb.devel.redhat.com/taskinfo?taskID=7954656

Xiaomei, could you test it? Thanks,

Fam

(In reply to Fam Zheng from comment #1)
> There is a WIP on upstream to address this virtio-scsi request allocation
> slowness:
> 
> http://lists.gnu.org/archive/html/qemu-devel/2014-09/msg02896.html
> 
> It should be introduced by the g_malloc -> g_malloc0 change, because the
> structure we are allocating is nearly 50KB (because of the .elem field).
> 
> I've applied the patches onto RHEV 7.1 to get this scratch build:
> 
> https://brewweb.devel.redhat.com/taskinfo?taskID=7954656
> 
> Xiaomei, could you test it? Thanks,

Fam, sure. We will test the build at once when the machine is available.

Xiaomei

Fix included in qemu-kvm-rhev-2.1.2-15.el7

Opened another new Bug 1174103 to track ~5%-10% gap between 1.5.3 and 2.1.2, so set the bug as verified.

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHSA-2015-0624.html