Stefano Garzarella
2022-Jan-14 13:38 UTC
[PATCH v1] vhost: cache avail index in vhost_enable_notify()
On Fri, Jan 14, 2022 at 07:45:35AM -0500, Michael S. Tsirkin wrote:>On Fri, Jan 14, 2022 at 10:05:08AM +0100, Stefano Garzarella wrote: >> In vhost_enable_notify() we enable the notifications and we read >> the avail index to check if new buffers have become available in >> the meantime. >> >> We are not caching the avail index, so when the device will call >> vhost_get_vq_desc(), it will find the old value in the cache and >> it will read the avail index again. >> >> It would be better to refresh the cache every time we read avail >> index, so let's change vhost_enable_notify() caching the value in >> `avail_idx` and compare it with `last_avail_idx` to check if there >> are new buffers available. >> >> Anyway, we don't expect a significant performance boost because >> the above path is not very common, indeed vhost_enable_notify() >> is often called with unlikely(), expecting that avail index has >> not been updated. >> >> Signed-off-by: Stefano Garzarella <sgarzare at redhat.com> > >... and can in theory even hurt due to an extra memory write. >So ... performance test restults pls?Right, could be. I'll run some perf test with vsock, about net, do you have a test suite or common step to follow to test it? Thanks, Stefano
Michael S. Tsirkin
2022-Jan-14 13:40 UTC
[PATCH v1] vhost: cache avail index in vhost_enable_notify()
On Fri, Jan 14, 2022 at 02:38:16PM +0100, Stefano Garzarella wrote:> On Fri, Jan 14, 2022 at 07:45:35AM -0500, Michael S. Tsirkin wrote: > > On Fri, Jan 14, 2022 at 10:05:08AM +0100, Stefano Garzarella wrote: > > > In vhost_enable_notify() we enable the notifications and we read > > > the avail index to check if new buffers have become available in > > > the meantime. > > > > > > We are not caching the avail index, so when the device will call > > > vhost_get_vq_desc(), it will find the old value in the cache and > > > it will read the avail index again. > > > > > > It would be better to refresh the cache every time we read avail > > > index, so let's change vhost_enable_notify() caching the value in > > > `avail_idx` and compare it with `last_avail_idx` to check if there > > > are new buffers available. > > > > > > Anyway, we don't expect a significant performance boost because > > > the above path is not very common, indeed vhost_enable_notify() > > > is often called with unlikely(), expecting that avail index has > > > not been updated. > > > > > > Signed-off-by: Stefano Garzarella <sgarzare at redhat.com> > > > > ... and can in theory even hurt due to an extra memory write. > > So ... performance test restults pls? > > Right, could be. > > I'll run some perf test with vsock, about net, do you have a test suite or > common step to follow to test it? > > Thanks, > StefanoYou can use the vhost test as a unit test as well. -- MST