Thanks for the fix! Looks good.  Also verified it fixes the crash reported.

Acked-by: Andy Zhou <az...@nicira.com>

On Sat, Sep 20, 2014 at 9:10 PM, Pravin B Shelar <pshe...@nicira.com> wrote:
> OVS needs to segments large skb before sending it for miss
> packet handling to userspace. but skb_gso_segment uses
> skb->cb. This corrupted OVS_CB which result in following panic.
>
> [  735.419921] BUG: unable to handle kernel paging request at 00000014000001b2
> [  735.423168] Oops: 0000 [#1] SMP DEBUG_PAGEALLOC
> [  735.445097] RIP: 0010:[<ffffffffa05df0d7>]  [<ffffffffa05df0d7>] 
> ovs_nla_put_flow+0x37/0x7c0 [openvswitch]
> [  735.468858] Call Trace:
> [  735.470384]  [<ffffffffa05d7ec2>] queue_userspace_packet+0x332/0x4d0 
> [openvswitch]
> [  735.471741]  [<ffffffffa05d8155>] queue_gso_packets+0xf5/0x180 
> [openvswitch]
> [  735.481862]  [<ffffffffa05da9f5>] ovs_dp_upcall+0x65/0x70 [openvswitch]
> [  735.483031]  [<ffffffffa05dab81>] ovs_dp_process_packet+0x181/0x1b0 
> [openvswitch]
> [  735.484391]  [<ffffffffa05e2f55>] ovs_vport_receive+0x65/0x90 [openvswitch]
> [  735.492638]  [<ffffffffa05e5738>] internal_dev_xmit+0x68/0x110 
> [openvswitch]
> [  735.495334]  [<ffffffff81588eb6>] dev_hard_start_xmit+0x2e6/0x8b0
> [  735.496503]  [<ffffffff81589847>] __dev_queue_xmit+0x3c7/0x920
> [  735.499827]  [<ffffffff81589db0>] dev_queue_xmit+0x10/0x20
> [  735.500798]  [<ffffffff815d3b60>] ip_finish_output+0x6a0/0x950
> [  735.502818]  [<ffffffff815d55f8>] ip_output+0x68/0x110
> [  735.503835]  [<ffffffff815d4979>] ip_local_out+0x29/0x90
> [  735.504801]  [<ffffffff815d4e46>] ip_queue_xmit+0x1d6/0x640
> [  735.507015]  [<ffffffff815ee0d7>] tcp_transmit_skb+0x477/0xac0
> [  735.508260]  [<ffffffff815ee856>] tcp_write_xmit+0x136/0xba0
> [  735.510829]  [<ffffffff815ef56e>] __tcp_push_pending_frames+0x2e/0xc0
> [  735.512296]  [<ffffffff815e0593>] tcp_sendmsg+0xa63/0xd50
> [  735.513526]  [<ffffffff81612c2c>] inet_sendmsg+0x10c/0x220
> [  735.516025]  [<ffffffff81566b8c>] sock_sendmsg+0x9c/0xe0
> [  735.518066]  [<ffffffff81566d41>] SYSC_sendto+0x121/0x1c0
> [  735.521398]  [<ffffffff8156801e>] SyS_sendto+0xe/0x10
> [  735.522473]  [<ffffffff816df5e9>] system_call_fastpath+0x16/0x1b
>
> Reported-by: Andy Zhou <az...@nicira.com>
> Signed-off-by: Pravin B Shelar <pshe...@nicira.com>
> ---
>  datapath/datapath.c | 4 ++++
>  1 file changed, 4 insertions(+)
>
> diff --git a/datapath/datapath.c b/datapath/datapath.c
> index 9850130..ed40c6a 100644
> --- a/datapath/datapath.c
> +++ b/datapath/datapath.c
> @@ -337,9 +337,12 @@ static int queue_gso_packets(struct datapath *dp, struct 
> sk_buff *skb,
>         unsigned short gso_type = skb_shinfo(skb)->gso_type;
>         struct sw_flow_key later_key;
>         struct sk_buff *segs, *nskb;
> +       struct ovs_skb_cb ovs_cb;
>         int err;
>
> +       ovs_cb = *OVS_CB(skb);
>         segs = __skb_gso_segment(skb, NETIF_F_SG, false);
> +       *OVS_CB(skb) = ovs_cb;
>         if (IS_ERR(segs))
>                 return PTR_ERR(segs);
>
> @@ -355,6 +358,7 @@ static int queue_gso_packets(struct datapath *dp, struct 
> sk_buff *skb,
>         /* Queue all of the segments. */
>         skb = segs;
>         do {
> +               *OVS_CB(skb) = ovs_cb;
>                 if (gso_type & SKB_GSO_UDP && skb != segs)
>                         key = &later_key;
>
> --
> 1.8.3.1
>
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev

Reply via email to