diff mbox series

[v4] virtio_ring: interrupt disable flag updated to vq even with event_triggered is set

Message ID 20230329102300.61000-1-huangjie.albert@bytedance.com
State New
Headers
Series [v4] virtio_ring: interrupt disable flag updated to vq even with event_triggered is set |

Commit Message

黄杰 March 29, 2023, 10:23 a.m. UTC 
  From: "huangjie.albert" <huangjie.albert@bytedance.com>

in virtio_net, if we disable the napi_tx, when we triger a tx interrupt,
the vq->event_triggered will be set to true. It will no longer be set to
false. Unless we explicitly call virtqueue_enable_cb_delayed or
virtqueue_enable_cb_prepare.

If we disable the napi_tx, it will only be called when the tx ring
buffer is relatively small.

Because event_triggered is true. Therefore, VRING_AVAIL_F_NO_INTERRUPT or
VRING_PACKED_EVENT_FLAG_DISABLE will not be set. So we update
vring_used_event(&vq->split.vring) or vq->packed.vring.driver->off_wrap
every time we call virtqueue_get_buf_ctx.This bring more interruptions.

To summarize:
1) event_triggered was set to true in vring_interrupt()
2) after this nothing will happen for virtqueue_disable_cb() so
   VRING_AVAIL_F_NO_INTERRUPT is not set in avail_flags_shadow
3) virtqueue_get_buf_ctx_split() will still think the cb is enabled
   then it tries to publish new event

To fix:
update VRING_AVAIL_F_NO_INTERRUPT or VRING_PACKED_EVENT_FLAG_DISABLE to vq
when we call virtqueue_disable_cb even the event_triggered is set to true.

Tested with iperf:
iperf3 tcp stream:
vm1 -----------------> vm2
vm2 just receives tcp data stream from vm1, and sends the ack to vm1,
there are many tx interrupts in vm2.
but without event_triggered there are just a few tx interrupts.

v2->v3:
-update the interrupt disable flag even with the event_triggered is set,
-instead of checking whether event_triggered is set in
-virtqueue_get_buf_ctx_{packed/split}, will cause the drivers  which have
-not called virtqueue_{enable/disable}_cb to miss notifications.

v3->v4:
-remove change for
-"if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE)"
-in virtqueue_disable_cb_packed

Fixes: 8d622d21d248 ("virtio: fix up virtio_disable_cb")
Signed-off-by: huangjie.albert <huangjie.albert@bytedance.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 drivers/virtio/virtio_ring.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)

Comments

Xuan Zhuo March 29, 2023, 11:45 a.m. UTC | #1 
On Wed, 29 Mar 2023 18:23:00 +0800, Albert Huang <huangjie.albert@bytedance.com> wrote:
> From: "huangjie.albert" <huangjie.albert@bytedance.com>
>
> in virtio_net, if we disable the napi_tx, when we triger a tx interrupt,
> the vq->event_triggered will be set to true. It will no longer be set to
> false. Unless we explicitly call virtqueue_enable_cb_delayed or
> virtqueue_enable_cb_prepare.
>
> If we disable the napi_tx, it will only be called when the tx ring
> buffer is relatively small.
>
> Because event_triggered is true. Therefore, VRING_AVAIL_F_NO_INTERRUPT or
> VRING_PACKED_EVENT_FLAG_DISABLE will not be set. So we update
> vring_used_event(&vq->split.vring) or vq->packed.vring.driver->off_wrap
> every time we call virtqueue_get_buf_ctx.This bring more interruptions.
>
> To summarize:
> 1) event_triggered was set to true in vring_interrupt()
> 2) after this nothing will happen for virtqueue_disable_cb() so
>    VRING_AVAIL_F_NO_INTERRUPT is not set in avail_flags_shadow
> 3) virtqueue_get_buf_ctx_split() will still think the cb is enabled
>    then it tries to publish new event
>
> To fix:
> update VRING_AVAIL_F_NO_INTERRUPT or VRING_PACKED_EVENT_FLAG_DISABLE to vq
> when we call virtqueue_disable_cb even the event_triggered is set to true.
>
> Tested with iperf:
> iperf3 tcp stream:
> vm1 -----------------> vm2
> vm2 just receives tcp data stream from vm1, and sends the ack to vm1,
> there are many tx interrupts in vm2.
> but without event_triggered there are just a few tx interrupts.
>
> v2->v3:
> -update the interrupt disable flag even with the event_triggered is set,
> -instead of checking whether event_triggered is set in
> -virtqueue_get_buf_ctx_{packed/split}, will cause the drivers  which have
> -not called virtqueue_{enable/disable}_cb to miss notifications.
>
> v3->v4:
> -remove change for
> -"if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE)"
> -in virtqueue_disable_cb_packed
>
> Fixes: 8d622d21d248 ("virtio: fix up virtio_disable_cb")

Hi Michael,

I want to know what the purpose of this patch("virtio: fix up virtio_disable_cb")
is. I re-review this patch. I didn't understand what the purpose of this
patches. Does it reduce one write to vring_used_event(&vq->split.vring) ?

Thanks.


> Signed-off-by: huangjie.albert <huangjie.albert@bytedance.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
>  drivers/virtio/virtio_ring.c | 22 ++++++++++++++++------
>  1 file changed, 16 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> index cbeeea1b0439..ec7ab8e04846 100644
> --- a/drivers/virtio/virtio_ring.c
> +++ b/drivers/virtio/virtio_ring.c
> @@ -931,6 +931,14 @@ static void virtqueue_disable_cb_split(struct virtqueue *_vq)
>
>  	if (!(vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT)) {
>  		vq->split.avail_flags_shadow |= VRING_AVAIL_F_NO_INTERRUPT;
> +
> +		/*
> +		 * If device triggered an event already it won't trigger one again:
> +		 * no need to disable.
> +		 */
> +		if (vq->event_triggered)
> +			return;
> +
>  		if (vq->event)
>  			/* TODO: this is a hack. Figure out a cleaner value to write. */
>  			vring_used_event(&vq->split.vring) = 0x0;
> @@ -1761,6 +1769,14 @@ static void virtqueue_disable_cb_packed(struct virtqueue *_vq)
>
>  	if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE) {
>  		vq->packed.event_flags_shadow = VRING_PACKED_EVENT_FLAG_DISABLE;
> +
> +		/*
> +		 * If device triggered an event already it won't trigger one again:
> +		 * no need to disable.
> +		 */
> +		if (vq->event_triggered)
> +			return;
> +
>  		vq->packed.vring.driver->flags =
>  			cpu_to_le16(vq->packed.event_flags_shadow);
>  	}
> @@ -2462,12 +2478,6 @@ void virtqueue_disable_cb(struct virtqueue *_vq)
>  {
>  	struct vring_virtqueue *vq = to_vvq(_vq);
>
> -	/* If device triggered an event already it won't trigger one again:
> -	 * no need to disable.
> -	 */
> -	if (vq->event_triggered)
> -		return;
> -
>  	if (vq->packed_ring)
>  		virtqueue_disable_cb_packed(_vq);
>  	else
> --
> 2.20.1
>
Michael S. Tsirkin March 29, 2023, 4:34 p.m. UTC | #2 
On Wed, Mar 29, 2023 at 07:45:01PM +0800, Xuan Zhuo wrote:
> On Wed, 29 Mar 2023 18:23:00 +0800, Albert Huang <huangjie.albert@bytedance.com> wrote:
> > From: "huangjie.albert" <huangjie.albert@bytedance.com>
> >
> > in virtio_net, if we disable the napi_tx, when we triger a tx interrupt,
> > the vq->event_triggered will be set to true. It will no longer be set to
> > false. Unless we explicitly call virtqueue_enable_cb_delayed or
> > virtqueue_enable_cb_prepare.
> >
> > If we disable the napi_tx, it will only be called when the tx ring
> > buffer is relatively small.
> >
> > Because event_triggered is true. Therefore, VRING_AVAIL_F_NO_INTERRUPT or
> > VRING_PACKED_EVENT_FLAG_DISABLE will not be set. So we update
> > vring_used_event(&vq->split.vring) or vq->packed.vring.driver->off_wrap
> > every time we call virtqueue_get_buf_ctx.This bring more interruptions.
> >
> > To summarize:
> > 1) event_triggered was set to true in vring_interrupt()
> > 2) after this nothing will happen for virtqueue_disable_cb() so
> >    VRING_AVAIL_F_NO_INTERRUPT is not set in avail_flags_shadow
> > 3) virtqueue_get_buf_ctx_split() will still think the cb is enabled
> >    then it tries to publish new event
> >
> > To fix:
> > update VRING_AVAIL_F_NO_INTERRUPT or VRING_PACKED_EVENT_FLAG_DISABLE to vq
> > when we call virtqueue_disable_cb even the event_triggered is set to true.
> >
> > Tested with iperf:
> > iperf3 tcp stream:
> > vm1 -----------------> vm2
> > vm2 just receives tcp data stream from vm1, and sends the ack to vm1,
> > there are many tx interrupts in vm2.
> > but without event_triggered there are just a few tx interrupts.
> >
> > v2->v3:
> > -update the interrupt disable flag even with the event_triggered is set,
> > -instead of checking whether event_triggered is set in
> > -virtqueue_get_buf_ctx_{packed/split}, will cause the drivers  which have
> > -not called virtqueue_{enable/disable}_cb to miss notifications.
> >
> > v3->v4:
> > -remove change for
> > -"if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE)"
> > -in virtqueue_disable_cb_packed
> >
> > Fixes: 8d622d21d248 ("virtio: fix up virtio_disable_cb")
> 
> Hi Michael,
> 
> I want to know what the purpose of this patch("virtio: fix up virtio_disable_cb")
> is. I re-review this patch. I didn't understand what the purpose of this
> patches. Does it reduce one write to vring_used_event(&vq->split.vring) ?
> 
> Thanks.

Are you asking why I applied 8d622d21d248 ("virtio: fix up
virtio_disable_cb")?

It was a prerequisite to fixing interrupt storms we saw in the field previously.

See Message-ID: <20210413011508-mutt-send-email-mst@kernel.org>
for the bug report and Message-ID: <20210413054733.36363-1-mst@redhat.com>
for the original patchset fixing it.


> 
> > Signed-off-by: huangjie.albert <huangjie.albert@bytedance.com>
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > ---
> >  drivers/virtio/virtio_ring.c | 22 ++++++++++++++++------
> >  1 file changed, 16 insertions(+), 6 deletions(-)
> >
> > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> > index cbeeea1b0439..ec7ab8e04846 100644
> > --- a/drivers/virtio/virtio_ring.c
> > +++ b/drivers/virtio/virtio_ring.c
> > @@ -931,6 +931,14 @@ static void virtqueue_disable_cb_split(struct virtqueue *_vq)
> >
> >  	if (!(vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT)) {
> >  		vq->split.avail_flags_shadow |= VRING_AVAIL_F_NO_INTERRUPT;
> > +
> > +		/*
> > +		 * If device triggered an event already it won't trigger one again:
> > +		 * no need to disable.
> > +		 */
> > +		if (vq->event_triggered)
> > +			return;
> > +
> >  		if (vq->event)
> >  			/* TODO: this is a hack. Figure out a cleaner value to write. */
> >  			vring_used_event(&vq->split.vring) = 0x0;
> > @@ -1761,6 +1769,14 @@ static void virtqueue_disable_cb_packed(struct virtqueue *_vq)
> >
> >  	if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE) {
> >  		vq->packed.event_flags_shadow = VRING_PACKED_EVENT_FLAG_DISABLE;
> > +
> > +		/*
> > +		 * If device triggered an event already it won't trigger one again:
> > +		 * no need to disable.
> > +		 */
> > +		if (vq->event_triggered)
> > +			return;
> > +
> >  		vq->packed.vring.driver->flags =
> >  			cpu_to_le16(vq->packed.event_flags_shadow);
> >  	}
> > @@ -2462,12 +2478,6 @@ void virtqueue_disable_cb(struct virtqueue *_vq)
> >  {
> >  	struct vring_virtqueue *vq = to_vvq(_vq);
> >
> > -	/* If device triggered an event already it won't trigger one again:
> > -	 * no need to disable.
> > -	 */
> > -	if (vq->event_triggered)
> > -		return;
> > -
> >  	if (vq->packed_ring)
> >  		virtqueue_disable_cb_packed(_vq);
> >  	else
> > --
> > 2.20.1
> >
Michael S. Tsirkin March 29, 2023, 4:36 p.m. UTC | #3 
On Wed, Mar 29, 2023 at 06:23:00PM +0800, Albert Huang wrote:
> From: "huangjie.albert" <huangjie.albert@bytedance.com>
> 
> in virtio_net, if we disable the napi_tx, when we triger a tx interrupt,
> the vq->event_triggered will be set to true. It will no longer be set to
> false. Unless we explicitly call virtqueue_enable_cb_delayed or
> virtqueue_enable_cb_prepare.
> 
> If we disable the napi_tx, it will only be called when the tx ring
> buffer is relatively small.
> 
> Because event_triggered is true. Therefore, VRING_AVAIL_F_NO_INTERRUPT or
> VRING_PACKED_EVENT_FLAG_DISABLE will not be set. So we update
> vring_used_event(&vq->split.vring) or vq->packed.vring.driver->off_wrap
> every time we call virtqueue_get_buf_ctx.This bring more interruptions.
> 
> To summarize:
> 1) event_triggered was set to true in vring_interrupt()
> 2) after this nothing will happen for virtqueue_disable_cb() so
>    VRING_AVAIL_F_NO_INTERRUPT is not set in avail_flags_shadow
> 3) virtqueue_get_buf_ctx_split() will still think the cb is enabled
>    then it tries to publish new event
> 
> To fix:
> update VRING_AVAIL_F_NO_INTERRUPT or VRING_PACKED_EVENT_FLAG_DISABLE to vq
> when we call virtqueue_disable_cb even the event_triggered is set to true.
> 
> Tested with iperf:
> iperf3 tcp stream:
> vm1 -----------------> vm2
> vm2 just receives tcp data stream from vm1, and sends the ack to vm1,
> there are many tx interrupts in vm2.
> but without event_triggered there are just a few tx interrupts.
> 
> v2->v3:
> -update the interrupt disable flag even with the event_triggered is set,
> -instead of checking whether event_triggered is set in
> -virtqueue_get_buf_ctx_{packed/split}, will cause the drivers  which have
> -not called virtqueue_{enable/disable}_cb to miss notifications.
> 
> v3->v4:
> -remove change for
> -"if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE)"
> -in virtqueue_disable_cb_packed
> 
> Fixes: 8d622d21d248 ("virtio: fix up virtio_disable_cb")
> Signed-off-by: huangjie.albert <huangjie.albert@bytedance.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> Signed-off-by: Jason Wang <jasowang@redhat.com>

Patch is ok so I rewrote the commit log and applied.
Please do not add signed-off-by tags by others -
read up what this tag means.
I also rewrote your Signed-off-by to have your name
formatted same as in email - hope that is ok.


> ---
>  drivers/virtio/virtio_ring.c | 22 ++++++++++++++++------
>  1 file changed, 16 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> index cbeeea1b0439..ec7ab8e04846 100644
> --- a/drivers/virtio/virtio_ring.c
> +++ b/drivers/virtio/virtio_ring.c
> @@ -931,6 +931,14 @@ static void virtqueue_disable_cb_split(struct virtqueue *_vq)
>  
>  	if (!(vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT)) {
>  		vq->split.avail_flags_shadow |= VRING_AVAIL_F_NO_INTERRUPT;
> +
> +		/*
> +		 * If device triggered an event already it won't trigger one again:
> +		 * no need to disable.
> +		 */
> +		if (vq->event_triggered)
> +			return;
> +
>  		if (vq->event)
>  			/* TODO: this is a hack. Figure out a cleaner value to write. */
>  			vring_used_event(&vq->split.vring) = 0x0;
> @@ -1761,6 +1769,14 @@ static void virtqueue_disable_cb_packed(struct virtqueue *_vq)
>  
>  	if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE) {
>  		vq->packed.event_flags_shadow = VRING_PACKED_EVENT_FLAG_DISABLE;
> +
> +		/*
> +		 * If device triggered an event already it won't trigger one again:
> +		 * no need to disable.
> +		 */
> +		if (vq->event_triggered)
> +			return;
> +
>  		vq->packed.vring.driver->flags =
>  			cpu_to_le16(vq->packed.event_flags_shadow);
>  	}
> @@ -2462,12 +2478,6 @@ void virtqueue_disable_cb(struct virtqueue *_vq)
>  {
>  	struct vring_virtqueue *vq = to_vvq(_vq);
>  
> -	/* If device triggered an event already it won't trigger one again:
> -	 * no need to disable.
> -	 */
> -	if (vq->event_triggered)
> -		return;
> -
>  	if (vq->packed_ring)
>  		virtqueue_disable_cb_packed(_vq);
>  	else
> -- 
> 2.20.1
Michael S. Tsirkin March 29, 2023, 4:40 p.m. UTC | #4 
On Wed, Mar 29, 2023 at 06:23:00PM +0800, Albert Huang wrote:
> From: "huangjie.albert" <huangjie.albert@bytedance.com>
> 
> in virtio_net, if we disable the napi_tx, when we triger a tx interrupt,
> the vq->event_triggered will be set to true. It will no longer be set to
> false. Unless we explicitly call virtqueue_enable_cb_delayed or
> virtqueue_enable_cb_prepare.
> 
> If we disable the napi_tx, it will only be called when the tx ring
> buffer is relatively small.
> 
> Because event_triggered is true. Therefore, VRING_AVAIL_F_NO_INTERRUPT or
> VRING_PACKED_EVENT_FLAG_DISABLE will not be set. So we update
> vring_used_event(&vq->split.vring) or vq->packed.vring.driver->off_wrap
> every time we call virtqueue_get_buf_ctx.This bring more interruptions.
> 
> To summarize:
> 1) event_triggered was set to true in vring_interrupt()
> 2) after this nothing will happen for virtqueue_disable_cb() so
>    VRING_AVAIL_F_NO_INTERRUPT is not set in avail_flags_shadow
> 3) virtqueue_get_buf_ctx_split() will still think the cb is enabled
>    then it tries to publish new event
> 
> To fix:
> update VRING_AVAIL_F_NO_INTERRUPT or VRING_PACKED_EVENT_FLAG_DISABLE to vq
> when we call virtqueue_disable_cb even the event_triggered is set to true.
> 
> Tested with iperf:
> iperf3 tcp stream:
> vm1 -----------------> vm2
> vm2 just receives tcp data stream from vm1, and sends the ack to vm1,
> there are many tx interrupts in vm2.
> but without event_triggered there are just a few tx interrupts.
> 
> v2->v3:
> -update the interrupt disable flag even with the event_triggered is set,
> -instead of checking whether event_triggered is set in
> -virtqueue_get_buf_ctx_{packed/split}, will cause the drivers  which have
> -not called virtqueue_{enable/disable}_cb to miss notifications.
> 
> v3->v4:
> -remove change for
> -"if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE)"
> -in virtqueue_disable_cb_packed
> 
> Fixes: 8d622d21d248 ("virtio: fix up virtio_disable_cb")
> Signed-off-by: huangjie.albert <huangjie.albert@bytedance.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> Signed-off-by: Jason Wang <jasowang@redhat.com>

Can you confirm you tested 9p and it still works fine?

> ---
>  drivers/virtio/virtio_ring.c | 22 ++++++++++++++++------
>  1 file changed, 16 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> index cbeeea1b0439..ec7ab8e04846 100644
> --- a/drivers/virtio/virtio_ring.c
> +++ b/drivers/virtio/virtio_ring.c
> @@ -931,6 +931,14 @@ static void virtqueue_disable_cb_split(struct virtqueue *_vq)
>  
>  	if (!(vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT)) {
>  		vq->split.avail_flags_shadow |= VRING_AVAIL_F_NO_INTERRUPT;
> +
> +		/*
> +		 * If device triggered an event already it won't trigger one again:
> +		 * no need to disable.
> +		 */
> +		if (vq->event_triggered)
> +			return;
> +
>  		if (vq->event)
>  			/* TODO: this is a hack. Figure out a cleaner value to write. */
>  			vring_used_event(&vq->split.vring) = 0x0;
> @@ -1761,6 +1769,14 @@ static void virtqueue_disable_cb_packed(struct virtqueue *_vq)
>  
>  	if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE) {
>  		vq->packed.event_flags_shadow = VRING_PACKED_EVENT_FLAG_DISABLE;
> +
> +		/*
> +		 * If device triggered an event already it won't trigger one again:
> +		 * no need to disable.
> +		 */
> +		if (vq->event_triggered)
> +			return;
> +
>  		vq->packed.vring.driver->flags =
>  			cpu_to_le16(vq->packed.event_flags_shadow);
>  	}
> @@ -2462,12 +2478,6 @@ void virtqueue_disable_cb(struct virtqueue *_vq)
>  {
>  	struct vring_virtqueue *vq = to_vvq(_vq);
>  
> -	/* If device triggered an event already it won't trigger one again:
> -	 * no need to disable.
> -	 */
> -	if (vq->event_triggered)
> -		return;
> -
>  	if (vq->packed_ring)
>  		virtqueue_disable_cb_packed(_vq);
>  	else
> -- 
> 2.20.1
黄杰 March 30, 2023, 2:15 a.m. UTC | #5 
Michael S. Tsirkin <mst@redhat.com> 于2023年3月30日周四 00:40写道：
>
> On Wed, Mar 29, 2023 at 06:23:00PM +0800, Albert Huang wrote:
> > From: "huangjie.albert" <huangjie.albert@bytedance.com>
> >
> > in virtio_net, if we disable the napi_tx, when we triger a tx interrupt,
> > the vq->event_triggered will be set to true. It will no longer be set to
> > false. Unless we explicitly call virtqueue_enable_cb_delayed or
> > virtqueue_enable_cb_prepare.
> >
> > If we disable the napi_tx, it will only be called when the tx ring
> > buffer is relatively small.
> >
> > Because event_triggered is true. Therefore, VRING_AVAIL_F_NO_INTERRUPT or
> > VRING_PACKED_EVENT_FLAG_DISABLE will not be set. So we update
> > vring_used_event(&vq->split.vring) or vq->packed.vring.driver->off_wrap
> > every time we call virtqueue_get_buf_ctx.This bring more interruptions.
> >
> > To summarize:
> > 1) event_triggered was set to true in vring_interrupt()
> > 2) after this nothing will happen for virtqueue_disable_cb() so
> >    VRING_AVAIL_F_NO_INTERRUPT is not set in avail_flags_shadow
> > 3) virtqueue_get_buf_ctx_split() will still think the cb is enabled
> >    then it tries to publish new event
> >
> > To fix:
> > update VRING_AVAIL_F_NO_INTERRUPT or VRING_PACKED_EVENT_FLAG_DISABLE to vq
> > when we call virtqueue_disable_cb even the event_triggered is set to true.
> >
> > Tested with iperf:
> > iperf3 tcp stream:
> > vm1 -----------------> vm2
> > vm2 just receives tcp data stream from vm1, and sends the ack to vm1,
> > there are many tx interrupts in vm2.
> > but without event_triggered there are just a few tx interrupts.
> >
> > v2->v3:
> > -update the interrupt disable flag even with the event_triggered is set,
> > -instead of checking whether event_triggered is set in
> > -virtqueue_get_buf_ctx_{packed/split}, will cause the drivers  which have
> > -not called virtqueue_{enable/disable}_cb to miss notifications.
> >
> > v3->v4:
> > -remove change for
> > -"if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE)"
> > -in virtqueue_disable_cb_packed
> >
> > Fixes: 8d622d21d248 ("virtio: fix up virtio_disable_cb")
> > Signed-off-by: huangjie.albert <huangjie.albert@bytedance.com>
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
>
> Can you confirm you tested 9p and it still works fine?
>

Yes, I tested it in my local environment. After applying this patch,
9p works fine.

Thanks.
> > ---
> >  drivers/virtio/virtio_ring.c | 22 ++++++++++++++++------
> >  1 file changed, 16 insertions(+), 6 deletions(-)
> >
> > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> > index cbeeea1b0439..ec7ab8e04846 100644
> > --- a/drivers/virtio/virtio_ring.c
> > +++ b/drivers/virtio/virtio_ring.c
> > @@ -931,6 +931,14 @@ static void virtqueue_disable_cb_split(struct virtqueue *_vq)
> >
> >       if (!(vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT)) {
> >               vq->split.avail_flags_shadow |= VRING_AVAIL_F_NO_INTERRUPT;
> > +
> > +             /*
> > +              * If device triggered an event already it won't trigger one again:
> > +              * no need to disable.
> > +              */
> > +             if (vq->event_triggered)
> > +                     return;
> > +
> >               if (vq->event)
> >                       /* TODO: this is a hack. Figure out a cleaner value to write. */
> >                       vring_used_event(&vq->split.vring) = 0x0;
> > @@ -1761,6 +1769,14 @@ static void virtqueue_disable_cb_packed(struct virtqueue *_vq)
> >
> >       if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE) {
> >               vq->packed.event_flags_shadow = VRING_PACKED_EVENT_FLAG_DISABLE;
> > +
> > +             /*
> > +              * If device triggered an event already it won't trigger one again:
> > +              * no need to disable.
> > +              */
> > +             if (vq->event_triggered)
> > +                     return;
> > +
> >               vq->packed.vring.driver->flags =
> >                       cpu_to_le16(vq->packed.event_flags_shadow);
> >       }
> > @@ -2462,12 +2478,6 @@ void virtqueue_disable_cb(struct virtqueue *_vq)
> >  {
> >       struct vring_virtqueue *vq = to_vvq(_vq);
> >
> > -     /* If device triggered an event already it won't trigger one again:
> > -      * no need to disable.
> > -      */
> > -     if (vq->event_triggered)
> > -             return;
> > -
> >       if (vq->packed_ring)
> >               virtqueue_disable_cb_packed(_vq);
> >       else
> > --
> > 2.20.1
>
diff mbox series

Patch

diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
index cbeeea1b0439..ec7ab8e04846 100644
--- a/drivers/virtio/virtio_ring.c
+++ b/drivers/virtio/virtio_ring.c
@@ -931,6 +931,14 @@  static void virtqueue_disable_cb_split(struct virtqueue *_vq)
 
 	if (!(vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT)) {
 		vq->split.avail_flags_shadow |= VRING_AVAIL_F_NO_INTERRUPT;
+
+		/*
+		 * If device triggered an event already it won't trigger one again:
+		 * no need to disable.
+		 */
+		if (vq->event_triggered)
+			return;
+
 		if (vq->event)
 			/* TODO: this is a hack. Figure out a cleaner value to write. */
 			vring_used_event(&vq->split.vring) = 0x0;
@@ -1761,6 +1769,14 @@  static void virtqueue_disable_cb_packed(struct virtqueue *_vq)
 
 	if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE) {
 		vq->packed.event_flags_shadow = VRING_PACKED_EVENT_FLAG_DISABLE;
+
+		/*
+		 * If device triggered an event already it won't trigger one again:
+		 * no need to disable.
+		 */
+		if (vq->event_triggered)
+			return;
+
 		vq->packed.vring.driver->flags =
 			cpu_to_le16(vq->packed.event_flags_shadow);
 	}
@@ -2462,12 +2478,6 @@  void virtqueue_disable_cb(struct virtqueue *_vq)
 {
 	struct vring_virtqueue *vq = to_vvq(_vq);
 
-	/* If device triggered an event already it won't trigger one again:
-	 * no need to disable.
-	 */
-	if (vq->event_triggered)
-		return;
-
 	if (vq->packed_ring)
 		virtqueue_disable_cb_packed(_vq);
 	else