On Thu, Jun 7, 2012 at 4:47 PM, Paolo Bonzini wrote:
> Il 07/06/2012 03:38, Asias He ha scritto:
>>> > I never looked at the kvmtool code, but I think Asias has a point. If
>>> > the guest submits I/O to lots of devices, they would contend on the
>>> > single thread. There was a similar proof of
Il 07/06/2012 03:38, Asias He ha scritto:
>> > I never looked at the kvmtool code, but I think Asias has a point. If
>> > the guest submits I/O to lots of devices, they would contend on the
>> > single thread. There was a similar proof of concept patch for QEMU that
>> > provided substantial bene
On Wed, Jun 6, 2012 at 8:35 PM, Paolo Bonzini wrote:
> Il 04/06/2012 18:07, Sasha Levin ha scritto:
>>> > All blk requests are processed in notify_vq() which is in the context of
>>> > ioeventfd thread: ioeventfd__thread(). The processing in notify_vq() may
>>> > take a long time to complete and a
Il 04/06/2012 18:07, Sasha Levin ha scritto:
>> > All blk requests are processed in notify_vq() which is in the context of
>> > ioeventfd thread: ioeventfd__thread(). The processing in notify_vq() may
>> > take a long time to complete and all devices share the single ioeventfd
>> > thead, so this m
On Mon, Jun 4, 2012 at 6:40 PM, Asias He wrote:
> All blk requests are processed in notify_vq() which is in the context of
> ioeventfd thread: ioeventfd__thread(). The processing in notify_vq() may
> take a long time to complete and all devices share the single ioeventfd
> thead, so this might blo
On Tue, Jun 5, 2012 at 4:03 AM, Asias He wrote:
> On Tue, Jun 5, 2012 at 12:07 AM, Sasha Levin wrote:
>> On Mon, 2012-06-04 at 23:40 +0800, Asias He wrote:
>>> All blk requests are processed in notify_vq() which is in the context of
>>> ioeventfd thread: ioeventfd__thread(). The processing in not
On Tue, Jun 05, 2012 at 08:47:17AM +0800, Asias He wrote:
> > I must admit I don't understand this code ;) The data get read into
> > stack variable forever?
>
> The data we read itself is not interesting at all. virtio_blk_thread()
> sleeps on the eventfd bdev->io_efd until notify_vq() writes som
On Tue, Jun 5, 2012 at 12:07 AM, Sasha Levin wrote:
> On Mon, 2012-06-04 at 23:40 +0800, Asias He wrote:
>> All blk requests are processed in notify_vq() which is in the context of
>> ioeventfd thread: ioeventfd__thread(). The processing in notify_vq() may
>> take a long time to complete and all d
On Mon, Jun 4, 2012 at 11:48 PM, Cyrill Gorcunov wrote:
> On Mon, Jun 04, 2012 at 11:40:53PM +0800, Asias He wrote:
>>
>> +static void *virtio_blk_thread(void *dev)
>> +{
>> + struct blk_dev *bdev = dev;
>> + u64 data;
>> +
>> + while (1) {
>> + read(bdev->io_efd, &data, si
On Mon, 2012-06-04 at 23:40 +0800, Asias He wrote:
> All blk requests are processed in notify_vq() which is in the context of
> ioeventfd thread: ioeventfd__thread(). The processing in notify_vq() may
> take a long time to complete and all devices share the single ioeventfd
> thead, so this might b
On Mon, Jun 04, 2012 at 11:40:53PM +0800, Asias He wrote:
>
> +static void *virtio_blk_thread(void *dev)
> +{
> + struct blk_dev *bdev = dev;
> + u64 data;
> +
> + while (1) {
> + read(bdev->io_efd, &data, sizeof(u64));
> + virtio_blk_do_io(bdev->kvm, &bdev->vq
All blk requests are processed in notify_vq() which is in the context of
ioeventfd thread: ioeventfd__thread(). The processing in notify_vq() may
take a long time to complete and all devices share the single ioeventfd
thead, so this might block other device's notify_vq() being called and
starve oth
On Thu, Dec 1, 2011 at 10:58 AM, Sasha Levin wrote:
> I was looking into the concept of adding 'dedicated' threads to the
> threadpool, since when the threadpool was added originally one of the
> purposes was to have all worker threads in a single place.
>
> This way we could still have threads in
On Thu, 2011-12-01 at 10:48 +0200, Pekka Enberg wrote:
> On Wed, 30 Nov 2011, Asias He wrote:
> >>> In virtio net's notify_vq(), we simply signal the tx/rx handle thread
> >>> and return.
> >>
> >> Why not use the threadpool?
> >
> > No.
>
> Sasha?
I was looking into the concept of adding 'dedica
On Wed, 30 Nov 2011, Asias He wrote:
In virtio net's notify_vq(), we simply signal the tx/rx handle thread
and return.
Why not use the threadpool?
No.
Sasha?
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordom
On 11/29/2011 10:36 PM, Sasha Levin wrote:
> On Tue, 2011-11-29 at 22:28 +0800, Asias He wrote:
>> Currently, all blk requests are processed in notify_vq() which is in
>> the context of ioeventfd thread: ioeventfd__thread(). The processing
>> in notify_vq() may take a long time to complete.
>>
>> W
On Tue, 2011-11-29 at 22:28 +0800, Asias He wrote:
> Currently, all blk requests are processed in notify_vq() which is in
> the context of ioeventfd thread: ioeventfd__thread(). The processing
> in notify_vq() may take a long time to complete.
>
> We should make notify_vq() return as soon as possi
Currently, all blk requests are processed in notify_vq() which is in
the context of ioeventfd thread: ioeventfd__thread(). The processing
in notify_vq() may take a long time to complete.
We should make notify_vq() return as soon as possible, since all devices
are sharing the single ioeventfd threa
18 matches
Mail list logo