netdev - Re: [RFC PATCH] Regression in linux 2.6.32 virtio

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 15 Dec 2009 08:26:20 -0800
From:	Sridhar Samudrala <sri@...ibm.com>
To:	Herbert Xu <herbert@...dor.apana.org.au>
Cc:	"Michael S. Tsirkin" <mst@...hat.com>, rusty@...tcorp.com.au,
	netdev@...r.kernel.org
Subject: Re: [RFC PATCH] Regression in linux 2.6.32 virtio_net seen with
 vhost-net

On Tue, 2009-12-15 at 22:42 +0800, Herbert Xu wrote:
> On Mon, Dec 14, 2009 at 01:40:55AM +0200, Michael S. Tsirkin wrote:
> >
> > An interesting thing is that
> > 48925e372f04f5e35fec6269127c62b2c71ab794
> > was supposed to do this already.
> 
> But Rusty missed the queue wake call.  That too needs to check
> that the required amount of space exists.

Yes. I noticied that. skb_xmit_done() doesn't check for MAX_SKB_FRAGS+2
descriptors before doing a netif_wake_queue(). We don't have any easy way 
to get the number of free desciptors in the virtqueue and we may need a 
vring_get_num_free() interface.

Also, can we do free_old_xmit_skbs() to freeup any used descriptors in
the callback(skb_xmit_done) context? I tried this, but hit a panic in
free_old_xmit_skbs(__skb_unlink() trying to remove the skb from the vi->send queue).
Looks like we need to add the skb to the vi->send queue before doing a kick.

Thanks
Sridhar

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html