* [Qemu-devel] block IO thread creation question
@ 2016-03-24 15:02 Chris Friesen
2016-03-25 2:25 ` Fam Zheng
0 siblings, 1 reply; 2+ messages in thread
From: Chris Friesen @ 2016-03-24 15:02 UTC (permalink / raw)
To: qemu-devel
Hi,
Could someone point me at the code for creating threads to handle block IO? I'm
seeing up to 30 threads per virtual disk, which seems high.
In case it's related, the block devices are iSCSI with the host acting as the
initiator and exposing block devices to qemu.
I'm particularly interested in qemu 2.2, just in case things have changed
significantly since then.
Thanks,
Chris
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [Qemu-devel] block IO thread creation question
2016-03-24 15:02 [Qemu-devel] block IO thread creation question Chris Friesen
@ 2016-03-25 2:25 ` Fam Zheng
0 siblings, 0 replies; 2+ messages in thread
From: Fam Zheng @ 2016-03-25 2:25 UTC (permalink / raw)
To: Chris Friesen; +Cc: qemu-devel
On Thu, 03/24 09:02, Chris Friesen wrote:
> Hi,
>
> Could someone point me at the code for creating threads to handle
> block IO? I'm seeing up to 30 threads per virtual disk, which seems
> high.
It is thread-pool.c, assuming you are using "aio=threads" for drive, which is
the default. With the alternative, "aio=native", spawning worker threads will
be much more uncommon.
Fam
>
> In case it's related, the block devices are iSCSI with the host
> acting as the initiator and exposing block devices to qemu.
>
> I'm particularly interested in qemu 2.2, just in case things have
> changed significantly since then.
>
> Thanks,
> Chris
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2016-03-25 2:25 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-03-24 15:02 [Qemu-devel] block IO thread creation question Chris Friesen
2016-03-25 2:25 ` Fam Zheng
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).