Patchwork [net,V2,1/2] virtio-net: don't respond to cpu hotplug notifier if we're not ready

login
register
mail settings
Submitter Jason Wang
Date Oct. 15, 2013, 3:18 a.m.
Message ID <1381807139-3450-1-git-send-email-jasowang@redhat.com>
Download mbox | patch
Permalink /patch/283452/
State Accepted
Delegated to: David Miller
Headers show

Comments

Jason Wang - Oct. 15, 2013, 3:18 a.m.
We're trying to re-configure the affinity unconditionally in cpu hotplug
callback. This may lead the issue during resuming from s3/s4 since

- virt queues haven't been allocated at that time.
- it's unnecessary since thaw method will re-configure the affinity.

Fix this issue by checking the config_enable and do nothing is we're not ready.

The bug were introduced by commit 8de4b2f3ae90c8fc0f17eeaab87d5a951b66ee17
(virtio-net: reset virtqueue affinity when doing cpu hotplug).

Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
The patch is need for 3.8 and above.
---
 drivers/net/virtio_net.c | 8 ++++++++
 1 file changed, 8 insertions(+)
Rusty Russell - Oct. 16, 2013, 11:27 p.m.
Jason Wang <jasowang@redhat.com> writes:
> We're trying to re-configure the affinity unconditionally in cpu hotplug
> callback. This may lead the issue during resuming from s3/s4 since
>
> - virt queues haven't been allocated at that time.
> - it's unnecessary since thaw method will re-configure the affinity.
>
> Fix this issue by checking the config_enable and do nothing is we're not ready.
>
> The bug were introduced by commit 8de4b2f3ae90c8fc0f17eeaab87d5a951b66ee17
> (virtio-net: reset virtqueue affinity when doing cpu hotplug).
>
> Cc: Rusty Russell <rusty@rustcorp.com.au>
> Cc: Michael S. Tsirkin <mst@redhat.com>
> Cc: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> Acked-by: Michael S. Tsirkin <mst@redhat.com>
> Reviewed-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
> The patch is need for 3.8 and above.

Please put 'CC: stable@kernel.org # 3.8+' in the commit.

(The specification of the stable line is poor, but that seems to be one
common method).

Cheers,
Rusty.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Michael S. Tsirkin - Oct. 17, 2013, 5:07 a.m.
On Thu, Oct 17, 2013 at 09:57:41AM +1030, Rusty Russell wrote:
> Jason Wang <jasowang@redhat.com> writes:
> > We're trying to re-configure the affinity unconditionally in cpu hotplug
> > callback. This may lead the issue during resuming from s3/s4 since
> >
> > - virt queues haven't been allocated at that time.
> > - it's unnecessary since thaw method will re-configure the affinity.
> >
> > Fix this issue by checking the config_enable and do nothing is we're not ready.
> >
> > The bug were introduced by commit 8de4b2f3ae90c8fc0f17eeaab87d5a951b66ee17
> > (virtio-net: reset virtqueue affinity when doing cpu hotplug).
> >
> > Cc: Rusty Russell <rusty@rustcorp.com.au>
> > Cc: Michael S. Tsirkin <mst@redhat.com>
> > Cc: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> > Acked-by: Michael S. Tsirkin <mst@redhat.com>
> > Reviewed-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > ---
> > The patch is need for 3.8 and above.
> 
> Please put 'CC: stable@kernel.org # 3.8+' in the commit.

Not if this is going in through the net tree.

> 
> (The specification of the stable line is poor, but that seems to be one
> common method).
> 
> Cheers,
> Rusty.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller - Oct. 17, 2013, 7:55 p.m.
From: Jason Wang <jasowang@redhat.com>
Date: Tue, 15 Oct 2013 11:18:58 +0800

> We're trying to re-configure the affinity unconditionally in cpu hotplug
> callback. This may lead the issue during resuming from s3/s4 since
> 
> - virt queues haven't been allocated at that time.
> - it's unnecessary since thaw method will re-configure the affinity.
> 
> Fix this issue by checking the config_enable and do nothing is we're not ready.
> 
> The bug were introduced by commit 8de4b2f3ae90c8fc0f17eeaab87d5a951b66ee17
> (virtio-net: reset virtqueue affinity when doing cpu hotplug).
> 
> Cc: Rusty Russell <rusty@rustcorp.com.au>
> Cc: Michael S. Tsirkin <mst@redhat.com>
> Cc: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> Acked-by: Michael S. Tsirkin <mst@redhat.com>
> Reviewed-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> Signed-off-by: Jason Wang <jasowang@redhat.com>

Applied and queued up for -stable.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Rusty Russell - Oct. 18, 2013, 1 a.m.
"Michael S. Tsirkin" <mst@redhat.com> writes:
> On Thu, Oct 17, 2013 at 09:57:41AM +1030, Rusty Russell wrote:
>> Jason Wang <jasowang@redhat.com> writes:
>> > We're trying to re-configure the affinity unconditionally in cpu hotplug
>> > callback. This may lead the issue during resuming from s3/s4 since
>> >
>> > - virt queues haven't been allocated at that time.
>> > - it's unnecessary since thaw method will re-configure the affinity.
>> >
>> > Fix this issue by checking the config_enable and do nothing is we're not ready.
>> >
>> > The bug were introduced by commit 8de4b2f3ae90c8fc0f17eeaab87d5a951b66ee17
>> > (virtio-net: reset virtqueue affinity when doing cpu hotplug).
>> >
>> > Cc: Rusty Russell <rusty@rustcorp.com.au>
>> > Cc: Michael S. Tsirkin <mst@redhat.com>
>> > Cc: Wanlong Gao <gaowanlong@cn.fujitsu.com>
>> > Acked-by: Michael S. Tsirkin <mst@redhat.com>
>> > Reviewed-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
>> > Signed-off-by: Jason Wang <jasowang@redhat.com>
>> > ---
>> > The patch is need for 3.8 and above.
>> 
>> Please put 'CC: stable@kernel.org # 3.8+' in the commit.
>
> Not if this is going in through the net tree.

WTF?  Wow, there really *is* an FAQ: https://lwn.net/Articles/561669/

DaveM is the best maintainer I've ever known, but I abhor the idea that
every subsystem has its own incompatible variant on workflow and style.

Asking people to express 'CC: stable' in words is error-prone; if Dave
wants to filter it, he's quite capable.

Rusty.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller - Oct. 18, 2013, 3:48 a.m.
From: Rusty Russell <rusty@rustcorp.com.au>
Date: Fri, 18 Oct 2013 11:30:15 +1030

> Asking people to express 'CC: stable' in words is error-prone; if Dave
> wants to filter it, he's quite capable.

Filtering it one time is one thing.

Potentially acting on that filter 100 or so times a day...

That's completely another.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Rusty Russell - Oct. 18, 2013, 6:17 a.m.
David Miller <davem@davemloft.net> writes:
> From: Rusty Russell <rusty@rustcorp.com.au>
> Date: Fri, 18 Oct 2013 11:30:15 +1030
>
>> Asking people to express 'CC: stable' in words is error-prone; if Dave
>> wants to filter it, he's quite capable.
>
> Filtering it one time is one thing.
>
> Potentially acting on that filter 100 or so times a day...
>
> That's completely another.

I don't see the difference between reacting to:

> The patch were need for 3.10 and above.

And:

CC: stable@kernel.org # 3.10+

Except the latter is the standard form which everyone else uses.

Do you want awk script to turn the latter into the former?  Would that
really help?

Confused,
Rusty.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller - Oct. 18, 2013, 7:13 a.m.
From: Rusty Russell <rusty@rustcorp.com.au>
Date: Fri, 18 Oct 2013 16:47:14 +1030

> Do you want awk script to turn the latter into the former?  Would that
> really help?

Rusty in the several years I've been operating this way, you're
the first person who seems to mind it.

To be honest I sometimes just silently deal with the stable CC:,
but it's much easier if things operate as they do now, and the
burdon is on the submitter instead of me.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index defec2b..c4bc1cc 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1116,6 +1116,11 @@  static int virtnet_cpu_callback(struct notifier_block *nfb,
 {
 	struct virtnet_info *vi = container_of(nfb, struct virtnet_info, nb);
 
+	mutex_lock(&vi->config_lock);
+
+	if (!vi->config_enable)
+		goto done;
+
 	switch(action & ~CPU_TASKS_FROZEN) {
 	case CPU_ONLINE:
 	case CPU_DOWN_FAILED:
@@ -1128,6 +1133,9 @@  static int virtnet_cpu_callback(struct notifier_block *nfb,
 	default:
 		break;
 	}
+
+done:
+	mutex_unlock(&vi->config_lock);
 	return NOTIFY_OK;
 }