search for: busyloop

...018?06?30? 00:38, Michael S. Tsirkin wrote: > On Fri, Jun 29, 2018 at 05:09:50PM +0900, Toshiaki Makita wrote: >> Under heavy load vhost busypoll may run without suppressing >> notification. For example tx zerocopy callback can push tx work while >> handle_tx() is running, then busyloop exits due to vhost_has_work() >> condition and enables notification but immediately reenters handle_tx() >> because the pushed work was tx. In this case handle_tx() tries to >> disable notification again, but when using event_idx it by design >> cannot. Then busyloop will ru...

...018?06?30? 00:38, Michael S. Tsirkin wrote: > On Fri, Jun 29, 2018 at 05:09:50PM +0900, Toshiaki Makita wrote: >> Under heavy load vhost busypoll may run without suppressing >> notification. For example tx zerocopy callback can push tx work while >> handle_tx() is running, then busyloop exits due to vhost_has_work() >> condition and enables notification but immediately reenters handle_tx() >> because the pushed work was tx. In this case handle_tx() tries to >> disable notification again, but when using event_idx it by design >> cannot. Then busyloop will ru...

Under heavy load vhost busypoll may run without suppressing notification. For example tx zerocopy callback can push tx work while handle_tx() is running, then busyloop exits due to vhost_has_work() condition and enables notification but immediately reenters handle_tx() because the pushed work was tx. In this case handle_tx() tries to disable notification again, but when using event_idx it by design cannot. Then busyloop will run without suppressing notification....

Under heavy load vhost busypoll may run without suppressing notification. For example tx zerocopy callback can push tx work while handle_tx() is running, then busyloop exits due to vhost_has_work() condition and enables notification but immediately reenters handle_tx() because the pushed work was tx. In this case handle_tx() tries to disable notification again, but when using event_idx it by design cannot. Then busyloop will run without suppressing notification....

Under heavy load vhost busypoll may run without suppressing notification. For example tx zerocopy callback can push tx work while handle_tx() is running, then busyloop exits due to vhost_has_work() condition and enables notification but immediately reenters handle_tx() because the pushed work was tx. In this case handle_tx() tries to disable notification again, but when using event_idx it by design cannot. Then busyloop will run without suppressing notification....

...- Split patches into 3 parts (renaming variables, tx-kick fix, rx-wakeup fix). - Avoid rx-kicks too (patch 4). - Don't memorize endtime as it is not needed for now. Toshiaki Makita (4): vhost_net: Rename local variables in vhost_net_rx_peek_head_len vhost_net: Avoid tx vring kicks during busyloop vhost_net: Avoid rx queue wake-ups during busypoll vhost_net: Avoid rx vring kicks during busyloop drivers/vhost/net.c | 95 +++++++++++++++++++++++++++++++++-------------------- 1 file changed, 60 insertions(+), 35 deletions(-) -- 1.8.3.1

...e_tx side, we poll the sock receive > queue at the same time. handle_rx do that in the same way. > > For avoiding deadlock, change the code to lock the vq one > by one and use the VHOST_NET_VQ_XX as a subclass for > mutex_lock_nested. With the patch, qemu can set differently > the busyloop_timeout for rx or tx queue. > > We set the poll-us=100us and use the iperf3 to test > its throughput. The iperf3 command is shown as below. > > on the guest: > iperf3 -s -D > > on the host: > iperf3 -c 192.168.1.100 -i 1 -P 10 -t 10 -M 1400 > > * With the patch:...

..., Michael S. Tsirkin wrote: >> On Fri, Jun 29, 2018 at 05:09:50PM +0900, Toshiaki Makita wrote: >>> Under heavy load vhost busypoll may run without suppressing >>> notification. For example tx zerocopy callback can push tx work while >>> handle_tx() is running, then busyloop exits due to vhost_has_work() >>> condition and enables notification but immediately reenters handle_tx() >>> because the pushed work was tx. In this case handle_tx() tries to >>> disable notification again, but when using event_idx it by design >>> cannot. Then...

...ST_NET_VQ_TX]; + struct vhost_virtqueue *rvq = &rnvq->vq; struct vhost_virtqueue *tvq = &tnvq->vq; unsigned long uninitialized_var(endtime); int len = peek_head_len(rnvq, sk); @@ -677,7 +678,8 @@ static int vhost_net_rx_peek_head_len(struct vhost_net *net, struct sock *sk, *busyloop_intr = true; break; } - if (sk_has_rx_data(sk) || + if ((sk_has_rx_data(sk) && + !vhost_vq_avail_empty(&net->dev, rvq)) || !vhost_vq_avail_empty(&net->dev, tvq)) break; cpu_relax(); @@ -827,7 +829,6 @@ static void handle_rx(struct vhost_net...

On 2018?07?04? 12:31, xiangxia.m.yue at gmail.com wrote: > From: Tonghao Zhang <xiangxia.m.yue at gmail.com> > > This patches improve the guest receive and transmit performance. > On the handle_tx side, we poll the sock receive queue at the same time. > handle_rx do that in the same way. > > For more performance report, see patch 4. > > v4 -> v5: > fix some

On 2018?07?11? 11:49, Tonghao Zhang wrote: > On Wed, Jul 11, 2018 at 10:56 AM Jason Wang <jasowang at redhat.com> wrote: >> >> >> On 2018?07?04? 12:31, xiangxia.m.yue at gmail.com wrote: >>> From: Tonghao Zhang <xiangxia.m.yue at gmail.com> >>> >>> This patches improve the guest receive and transmit performance. >>> On the

On 2018?07?11? 19:59, Michael S. Tsirkin wrote: > On Wed, Jul 11, 2018 at 01:12:59PM +0800, Jason Wang wrote: >> >> On 2018?07?11? 11:49, Tonghao Zhang wrote: >>> On Wed, Jul 11, 2018 at 10:56 AM Jason Wang <jasowang at redhat.com> wrote: >>>> >>>> On 2018?07?04? 12:31, xiangxia.m.yue at gmail.com wrote: >>>>> From: Tonghao Zhang

...net: vhost: lock the vqs one by one > net: vhost: replace magic number of lock annotation > net: vhost: factor out busy polling logic to vhost_net_busy_poll() > net: vhost: add rx busy polling in tx path > net: vhost: disable rx wakeup during tx busypoll > net: vhost: make busyloop_intr more accurate > > drivers/vhost/net.c | 163 +++++++++++++++++++++++++++++++------------------- > drivers/vhost/vhost.c | 24 +++----- > 2 files changed, 108 insertions(+), 79 deletions(-) > > -- > 1.8.3.1 > ping ?

...one by one >> net: vhost: replace magic number of lock annotation >> net: vhost: factor out busy polling logic to vhost_net_busy_poll() >> net: vhost: add rx busy polling in tx path >> net: vhost: disable rx wakeup during tx busypoll >> net: vhost: make busyloop_intr more accurate >> >> drivers/vhost/net.c | 163 +++++++++++++++++++++++++++++++------------------- >> drivers/vhost/vhost.c | 24 +++----- >> 2 files changed, 108 insertions(+), 79 deletions(-) >> >> -- >> 1.8.3.1 >> > ping ? HI Tonghao...

SunOS pop01.unix 5.10 Generic_118844-26 i86pc i386 i86pc dovecot-1.0.beta3 Dovecot itself runs well, and was easy to confirgure with LDAP. However, I am seeing a cpu busy-loop in dovecot-auth. Is this a known issue, or do you know why it happens? If needed I can go deeper if so required. PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND 26440 root 1 10 0 4232K

...struct vhost_virtqueue *tvq = &tnvq->vq; >> ????? unsigned long uninitialized_var(endtime); >> ????? int len = peek_head_len(rnvq, sk); >> @@ -677,7 +678,8 @@ static int vhost_net_rx_peek_head_len(struct >> vhost_net *net, struct sock *sk, >> ????????????????? *busyloop_intr = true; >> ????????????????? break; >> ????????????? } >> -??????????? if (sk_has_rx_data(sk) || >> +??????????? if ((sk_has_rx_data(sk) && >> +???????????????? !vhost_vq_avail_empty(&net->dev, rvq)) || >> ????????????????? !vhost_vq_avail_emp...

On 2018?07?12? 11:34, Michael S. Tsirkin wrote: > On Thu, Jul 12, 2018 at 11:26:12AM +0800, Jason Wang wrote: >> >> On 2018?07?11? 19:59, Michael S. Tsirkin wrote: >>> On Wed, Jul 11, 2018 at 01:12:59PM +0800, Jason Wang wrote: >>>> On 2018?07?11? 11:49, Tonghao Zhang wrote: >>>>> On Wed, Jul 11, 2018 at 10:56 AM Jason Wang <jasowang at

On 2018?06?29? 16:09, Toshiaki Makita wrote: > Under heavy load vhost busypoll may run without suppressing > notification. For example tx zerocopy callback can push tx work while > handle_tx() is running, then busyloop exits due to vhost_has_work() > condition and enables notification but immediately reenters handle_tx() > because the pushed work was tx. In this case handle_tx() tries to > disable notification again, but when using event_idx it by design > cannot. Then busyloop will run without suppre...

On Fri, Jun 29, 2018 at 05:09:50PM +0900, Toshiaki Makita wrote: > Under heavy load vhost busypoll may run without suppressing > notification. For example tx zerocopy callback can push tx work while > handle_tx() is running, then busyloop exits due to vhost_has_work() > condition and enables notification but immediately reenters handle_tx() > because the pushed work was tx. In this case handle_tx() tries to > disable notification again, but when using event_idx it by design > cannot. Then busyloop will run without suppre...

http://bugs.freedesktop.org/show_bug.cgi?id=15845 Summary: X busyloops when huge image in firefox Product: xorg Version: git Platform: x86-64 (AMD64) OS/Version: Linux (All) Status: NEW Severity: normal Priority: medium Component: Driver/nouveau AssignedTo: nouveau at lists.freed...