[PATCH net 1/1] hyperv: Fix a bug in netvsc_send()

[Long Li]

Thanks Sitsofe. Can you provide more details on the test setup?

The kernel trace shows that skb->mac_header=0xffff (which means not yet set, it's in RCX: 000000000000ffff).


-----Original Message-----
From: devel [mailto:driverdev-devel-bounces at linuxdriverproject.org] On Behalf Of Sitsofe Wheeler
Sent: Thursday, October 09, 2014 6:32 AM
To: David Miller
Cc: olaf at aepfle.de; netdev at vger.kernel.org; jasowang at redhat.com; linux-kernel at vger.kernel.org; apw at canonical.com; devel at linuxdriverproject.org
Subject: Re: [PATCH net 1/1] hyperv: Fix a bug in netvsc_send()

On Sun, Oct 05, 2014 at 09:11:29PM -0400, David Miller wrote:
> From: "K. Y. Srinivasan" <kys at microsoft.com>
> Date: Sun,  5 Oct 2014 10:42:51 -0700
> 
> > After the packet is successfully sent, we should not touch the 
> > packet as it may have been freed. This patch is based on the work 
> > done by Long Li <longli at microsoft.com>.
> > 
> > David, please queue this up for stable.

With 3.17.0 g782d59c (which should include this patch) I'm still seeing the following:

Oct 09 13:14:51 a network[428]: Bringing up interface eth0:
Oct 09 13:14:51 a dhclient[538]: DHCPREQUEST on eth0 to 255.255.255.255 port 67 (xid=0x1dd33078) Oct 09 13:14:51 a dhclient[538]: DHCPACK from 10.x.x.x (xid=0x1dd33078) Oct 09 13:14:55 a kernel: BUG: unable to handle kernel paging request at ffff8800ed2e72e3 Oct 09 13:14:55 a kernel: IP: [<ffffffff814ede1d>] netvsc_select_queue+0x3d/0x150 Oct 09 13:14:55 a kernel: PGD 2db5067 PUD 2075be067 PMD 207454067 PTE 80000000ed2e7060 Oct 09 13:14:55 a kernel: Oops: 0000 [#1] SMP DEBUG_PAGEALLOC Oct 09 13:14:55 a kernel: CPU: 6 PID: 566 Comm: arping Not tainted 3.17.0.x86_64-05585-g782d59c #147 Oct 09 13:14:55 a kernel: Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS 090006  05/23/2012 Oct 09 13:14:55 a kernel: task: ffff8801f978b9f0 ti: ffff8801f3b84000 task.ti: ffff8801f3b84000 Oct 09 13:14:55 a kernel: RIP: 0010:[<ffffffff814ede1d>]  [<ffffffff814ede1d>] netvsc_select_queue+0x3d/0x150 Oct 09 13:14:55 a kernel: RSP: 0018:ffff8801f3b87c60  EFLAGS: 00010202 Oct 09 13:14:55 a kernel: RAX: 0000000000000000 RBX: ffff8800f13e8000 RCX: 000000000000ffff Oct 09 13:14:55 a kernel: RDX: ffff8800ed2d72d8 RSI: ffff8801fabca1c0 RDI: ffff8800f13e8000 Oct 09 13:14:55 a kernel: RBP: ffff8801f3b87c88 R08: 000000000000002a R09: 0000000000000000 Oct 09 13:14:55 a kernel: R10: ffff8801f83b3f60 R11: 0000000000000008 R12: ffff8801fabca1c0 Oct 09 13:14:55 a kernel: R13: 0000000000000000 R14: ffff8800ed359bd8 R15: ffff8801fabca1c0 Oct 09 13:14:55 a kernel: FS:  00007f943a5c9740(0000) GS:ffff880206cc0000(0000) knlGS:0000000000000000 Oct 09 13:14:55 a kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Oct 09 13:14:55 a kernel: CR2: ffff8800ed2e72e3 CR3: 00000001f3957000 CR4: 00000000000406e0 Oct 09 13:14:55 a kernel: Stack:
Oct 09 13:14:55 a kernel:  ffffffff816a0221 ffff8800f13e8000 000000000000001c 0000000000000000 Oct 09 13:14:55 a kernel:  ffff8800ed359bd8 ffff8801f3b87d48 ffffffff816a3fce ffff8801f3b87cb0 Oct 09 13:14:55 a kernel:  ffffffff816c34a7 0000000000000001 ffff8801f3b87db8 000000000000001c Oct 09 13:14:55 a kernel: Call Trace:
Oct 09 13:14:55 a kernel:  [<ffffffff816a0221>] ? packet_pick_tx_queue+0x31/0xa0 Oct 09 13:14:55 a kernel:  [<ffffffff816a3fce>] packet_sendmsg+0xc6e/0xe30 Oct 09 13:14:55 a kernel:  [<ffffffff816c34a7>] ? _raw_spin_unlock+0x27/0x40 Oct 09 13:14:55 a kernel:  [<ffffffff81091bba>] ? prepare_creds+0x3a/0x170 Oct 09 13:14:55 a kernel:  [<ffffffff815d2e08>] sock_sendmsg+0x88/0xb0 Oct 09 13:14:55 a kernel:  [<ffffffff81188f83>] ? might_fault+0xa3/0xb0 Oct 09 13:14:55 a kernel:  [<ffffffff81188f3a>] ? might_fault+0x5a/0xb0 Oct 09 13:14:55 a kernel:  [<ffffffff815d2f3e>] SYSC_sendto+0x10e/0x150 Oct 09 13:14:55 a kernel:  [<ffffffff81188f3a>] ? might_fault+0x5a/0xb0 Oct 09 13:14:55 a kernel:  [<ffffffff816c41d5>] ? sysret_check+0x22/0x5d Oct 09 13:14:55 a kernel:  [<ffffffff810ba3fd>] ? trace_hardirqs_on_caller+0x17d/0x210
Oct 09 13:14:55 a kernel:  [<ffffffff813a20ee>] ? trace_hardirqs_on_thunk+0x3a/0x3f Oct 09 13:14:55 a kernel:  [<ffffffff815d3f1e>] SyS_sendto+0xe/0x10 Oct 09 13:14:55 a kernel:  [<ffffffff816c41a9>] system_call_fastpath+0x16/0x1b Oct 09 13:14:55 a kernel: Code: 00 4d 85 d2 0f 84 1c 01 00 00 44 8b 9f 8c 03 00 00 31 c0 41 83 fb 01 0f 86 1b 01 00 00 0f b7 8e b6 00 00 00 Oct 09 13:14:55 a kernel: RIP  [<ffffffff814ede1d>] netvsc_select_queue+0x3d/0x150 Oct 09 13:14:55 a kernel:  RSP <ffff8801f3b87c60> Oct 09 13:14:55 a kernel: CR2: ffff8800ed2e72e3 Oct 09 13:14:55 a kernel: ---[ end trace e52f922dd7435e0d ]---

Was the above meant to have been fixed by the patch "[PATCH 1/1]
Drivers: net: hyperv: Cleanup  netvsc_change_mtu ()" from
https://lkml.org/lkml/2014/8/29/369 ? If so will that patch be resent?

--
Sitsofe | http://sucs.org/~sits/
_______________________________________________
devel mailing list
devel at linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel