* GSO checksum warning debug info
@ 2009-02-04 0:23 Chris Wright
2009-02-04 12:40 ` Herbert Xu
0 siblings, 1 reply; 4+ messages in thread
From: Chris Wright @ 2009-02-04 0:23 UTC (permalink / raw)
To: Herbert Xu; +Cc: netdev
I'm able to trigger the GSO checksum warning, and after 67fd1a73 "net: Add
debug info to track down GSO checksum bug" here's the more verbose output:
e1000e: caps=(0x1119a9, 0x0) len=2960 data_len=0 ip_summed=0
What info would be helpful for you (couple full backtraces below)?
thanks,
-chris
[ 2172.557466] WARNING: at /home/chrisw/git/linux/linus-2.6/net/core/dev.c:1548 skb_gso_segment+0x103/0x23a()
[ 2172.557468] Hardware name: To Be Filled By O.E.M.
[ 2172.557470] e1000e: caps=(0x1119a9, 0x0) len=2960 data_len=0 ip_summed=0Modules linked in: kvm_intel kvm [last unloaded: scsi_wait_scan]
[ 2172.557476] Pid: 4750, comm: qemu-kvm Not tainted 2.6.29-rc3 #13
[ 2172.557478] Call Trace:
[ 2172.557480] <IRQ> [<ffffffff8103acbd>] warn_slowpath+0xb1/0xed
[ 2172.557489] [<ffffffff811b75ba>] ? alloc_iova+0x1e6/0x218
[ 2172.557492] [<ffffffff811a29bd>] ? sprintf+0x40/0x42
[ 2172.557496] [<ffffffff81290217>] ? tun_chr_aio_write+0x0/0x44d
[ 2172.557501] [<ffffffff81267aa5>] ? e1000_get_drvinfo+0xa6/0xcd
[ 2172.557503] [<ffffffff813a38aa>] skb_gso_segment+0x103/0x23a
[ 2172.557505] [<ffffffff813a3b8d>] dev_hard_start_xmit+0x1ac/0x26d
[ 2172.557510] [<ffffffff813b5759>] __qdisc_run+0xe7/0x1ea
[ 2172.557513] [<ffffffff8102e81b>] ? enqueue_task+0x50/0x5b
[ 2172.557516] [<ffffffff813a40b0>] dev_queue_xmit+0x350/0x44d
[ 2172.557520] [<ffffffff8144fe98>] br_dev_queue_push_xmit+0x5c/0x62
[ 2172.557522] [<ffffffff81035264>] ? default_wake_function+0xd/0xf
[ 2172.557524] [<ffffffff8144feed>] br_forward_finish+0x4f/0x53
[ 2172.557526] [<ffffffff8144ff5b>] __br_forward+0x6a/0x6e
[ 2172.557529] [<ffffffff8144ffa8>] br_forward+0x49/0x57
[ 2172.557531] [<ffffffff81450a97>] br_handle_frame_finish+0x12b/0x14f
[ 2172.557534] [<ffffffff81450c73>] br_handle_frame+0x1b8/0x1d5
[ 2172.557536] [<ffffffff813a3195>] netif_receive_skb+0x2ea/0x3bb
[ 2172.557538] [<ffffffff813a3682>] napi_gro_receive+0x47/0x57
[ 2172.557541] [<ffffffff8101bd8f>] ? lapic_next_event+0x15/0x19
[ 2172.557543] [<ffffffff813a370f>] process_backlog+0x7d/0xb2
[ 2172.557545] [<ffffffff813a17e9>] net_rx_action+0xa7/0x17a
[ 2172.557548] [<ffffffff8103fcc3>] __do_softirq+0x83/0x144
[ 2172.557551] [<ffffffff8100d33c>] call_softirq+0x1c/0x28
[ 2172.557552] <EOI> [<ffffffff8100e258>] do_softirq+0x34/0x76
[ 2172.557556] [<ffffffff813a3765>] netif_rx_ni+0x21/0x28
[ 2172.557558] [<ffffffff8129060d>] tun_chr_aio_write+0x3f6/0x44d
[ 2172.557561] [<ffffffff810a0000>] ? sys_swapon+0x2bb/0xa58
[ 2172.557563] [<ffffffff81290217>] ? tun_chr_aio_write+0x0/0x44d
[ 2172.557566] [<ffffffff810b0153>] do_sync_readv_writev+0xe3/0x12b
[ 2172.557569] [<ffffffff8104db04>] ? autoremove_wake_function+0x0/0x38
[ 2172.557574] [<ffffffff8116e3ba>] ? selinux_file_permission+0xbd/0xc6
[ 2172.557578] [<ffffffff811673b0>] ? security_file_permission+0x11/0x13
[ 2172.557580] [<ffffffff810b089d>] do_readv_writev+0xba/0x121
[ 2172.557583] [<ffffffff8116e3ba>] ? selinux_file_permission+0xbd/0xc6
[ 2172.557585] [<ffffffff81035257>] ? default_wake_function+0x0/0xf
[ 2172.557587] [<ffffffff811673b0>] ? security_file_permission+0x11/0x13
[ 2172.557589] [<ffffffff810b0942>] vfs_writev+0x3e/0x49
[ 2172.557591] [<ffffffff810b0994>] sys_writev+0x47/0x94
[ 2172.557593] [<ffffffff8100c25b>] system_call_fastpath+0x16/0x1b
[ 2172.557595] ---[ end trace bbdce316762c5c8a ]---
[ 2172.558366] ------------[ cut here ]------------
[ 2172.558368] WARNING: at /home/chrisw/git/linux/linus-2.6/net/core/dev.c:1548 skb_gso_segment+0x103/0x23a()
[ 2172.558370] Hardware name: To Be Filled By O.E.M.
[ 2172.558371] e1000e: caps=(0x1119a9, 0x0) len=5880 data_len=0 ip_summed=0Modules linked in: kvm_intel kvm [last unloaded: scsi_wait_scan]
[ 2172.558375] Pid: 4750, comm: qemu-kvm Tainted: G W 2.6.29-rc3 #13
[ 2172.558376] Call Trace:
[ 2172.558377] <IRQ> [<ffffffff8103acbd>] warn_slowpath+0xb1/0xed
[ 2172.558382] [<ffffffff810be05c>] ? pollwake+0x47/0x4f
[ 2172.558384] [<ffffffff81035257>] ? default_wake_function+0x0/0xf
[ 2172.558386] [<ffffffff811a29bd>] ? sprintf+0x40/0x42
[ 2172.558389] [<ffffffff81267aa5>] ? e1000_get_drvinfo+0xa6/0xcd
[ 2172.558391] [<ffffffff813a38aa>] skb_gso_segment+0x103/0x23a
[ 2172.558393] [<ffffffff813a3b8d>] dev_hard_start_xmit+0x1ac/0x26d
[ 2172.558396] [<ffffffff813b5759>] __qdisc_run+0xe7/0x1ea
[ 2172.558398] [<ffffffff81030158>] ? check_preempt_wakeup+0x13d/0x18e
[ 2172.558400] [<ffffffff813a40b0>] dev_queue_xmit+0x350/0x44d
[ 2172.558402] [<ffffffff81031850>] ? enqueue_task_fair+0x25/0x92
[ 2172.558405] [<ffffffff8144fe98>] br_dev_queue_push_xmit+0x5c/0x62
[ 2172.558407] [<ffffffff81035264>] ? default_wake_function+0xd/0xf
[ 2172.558409] [<ffffffff8144feed>] br_forward_finish+0x4f/0x53
[ 2172.558411] [<ffffffff8144ff5b>] __br_forward+0x6a/0x6e
[ 2172.558413] [<ffffffff8144ffa8>] br_forward+0x49/0x57
[ 2172.558415] [<ffffffff81450a97>] br_handle_frame_finish+0x12b/0x14f
[ 2172.558418] [<ffffffff81450c73>] br_handle_frame+0x1b8/0x1d5
[ 2172.558420] [<ffffffff813a3195>] netif_receive_skb+0x2ea/0x3bb
[ 2172.558422] [<ffffffff813a3682>] napi_gro_receive+0x47/0x57
[ 2172.558424] [<ffffffff8101bd8f>] ? lapic_next_event+0x15/0x19
[ 2172.558426] [<ffffffff813a370f>] process_backlog+0x7d/0xb2
[ 2172.558427] [<ffffffff813a17e9>] net_rx_action+0xa7/0x17a
[ 2172.558430] [<ffffffff8103fcc3>] __do_softirq+0x83/0x144
[ 2172.558432] [<ffffffff8100d33c>] call_softirq+0x1c/0x28
[ 2172.558433] <EOI> [<ffffffff8100e258>] do_softirq+0x34/0x76
[ 2172.558436] [<ffffffff813a3765>] netif_rx_ni+0x21/0x28
[ 2172.558438] [<ffffffff8129060d>] tun_chr_aio_write+0x3f6/0x44d
[ 2172.558440] [<ffffffff810a0000>] ? sys_swapon+0x2bb/0xa58
[ 2172.558442] [<ffffffff81290217>] ? tun_chr_aio_write+0x0/0x44d
[ 2172.558444] [<ffffffff810b0153>] do_sync_readv_writev+0xe3/0x12b
[ 2172.558447] [<ffffffff8104db04>] ? autoremove_wake_function+0x0/0x38
[ 2172.558449] [<ffffffff8116e3ba>] ? selinux_file_permission+0xbd/0xc6
[ 2172.558451] [<ffffffff811673b0>] ? security_file_permission+0x11/0x13
[ 2172.558454] [<ffffffff810b089d>] do_readv_writev+0xba/0x121
[ 2172.558457] [<ffffffff810bc2fc>] ? do_vfs_ioctl+0x462/0x4a2
[ 2172.558459] [<ffffffff810b0942>] vfs_writev+0x3e/0x49
[ 2172.558461] [<ffffffff810b0994>] sys_writev+0x47/0x94
[ 2172.558463] [<ffffffff8100c25b>] system_call_fastpath+0x16/0x1b
[ 2172.558464] ---[ end trace bbdce316762c5c8b ]---
[ 2172.559059] ------------[ cut here ]------------
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: GSO checksum warning debug info
2009-02-04 0:23 GSO checksum warning debug info Chris Wright
@ 2009-02-04 12:40 ` Herbert Xu
2009-02-04 16:11 ` Chris Wright
0 siblings, 1 reply; 4+ messages in thread
From: Herbert Xu @ 2009-02-04 12:40 UTC (permalink / raw)
To: Chris Wright; +Cc: netdev
On Tue, Feb 03, 2009 at 04:23:45PM -0800, Chris Wright wrote:
> I'm able to trigger the GSO checksum warning, and after 67fd1a73 "net: Add
> debug info to track down GSO checksum bug" here's the more verbose output:
>
> e1000e: caps=(0x1119a9, 0x0) len=2960 data_len=0 ip_summed=0
>
> What info would be helpful for you (couple full backtraces below)?
Interesting. Is this a Linux guest or a Window guest?
If it's a Linux guest, it means that somewhere along the line
the partial checksum got lost. But I can't see anything obvious.
Since you can reproduce this, could you add some printks along
the path shown in the back trace to see where ip_summed gets
zapped? So that would be tun and bridge basically.
Thanks,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: GSO checksum warning debug info
2009-02-04 12:40 ` Herbert Xu
@ 2009-02-04 16:11 ` Chris Wright
2009-02-04 20:52 ` Herbert Xu
0 siblings, 1 reply; 4+ messages in thread
From: Chris Wright @ 2009-02-04 16:11 UTC (permalink / raw)
To: Herbert Xu; +Cc: Chris Wright, netdev
* Herbert Xu (herbert@gondor.apana.org.au) wrote:
> On Tue, Feb 03, 2009 at 04:23:45PM -0800, Chris Wright wrote:
> > I'm able to trigger the GSO checksum warning, and after 67fd1a73 "net: Add
> > debug info to track down GSO checksum bug" here's the more verbose output:
> >
> > e1000e: caps=(0x1119a9, 0x0) len=2960 data_len=0 ip_summed=0
> >
> > What info would be helpful for you (couple full backtraces below)?
>
> Interesting. Is this a Linux guest or a Window guest?
Windows guest.
> If it's a Linux guest, it means that somewhere along the line
> the partial checksum got lost. But I can't see anything obvious.
> Since you can reproduce this, could you add some printks along
> the path shown in the back trace to see where ip_summed gets
> zapped? So that would be tun and bridge basically.
Sure.
thanks,
-chris
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: GSO checksum warning debug info
2009-02-04 16:11 ` Chris Wright
@ 2009-02-04 20:52 ` Herbert Xu
0 siblings, 0 replies; 4+ messages in thread
From: Herbert Xu @ 2009-02-04 20:52 UTC (permalink / raw)
To: Chris Wright; +Cc: netdev
On Wed, Feb 04, 2009 at 08:11:55AM -0800, Chris Wright wrote:
>
> Windows guest.
That makes a lot more sense :) The Windows driver needs to set
a partial checksum in order to use GSO/LSO.
Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2009-02-04 20:52 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-02-04 0:23 GSO checksum warning debug info Chris Wright
2009-02-04 12:40 ` Herbert Xu
2009-02-04 16:11 ` Chris Wright
2009-02-04 20:52 ` Herbert Xu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).