Message-ID: <4A65EC3F.4050400@gibraltar.at>
Date:	Tue, 21 Jul 2009 18:26:39 +0200
From:	Rene Mayrhofer <rene.mayrhofer@...raltar.at>
To:	netdev@...r.kernel.org
CC:	Richard Leitner <leitner@...s.at>
Subject: Kernel oops on setting sky2 interfaces down

Hi everybody,

[Please CC me in replies, I am not currently subscribed to this list.]

I have a fully reproducible kernel oops in the sky2 module in kernel
2.6.28.10. The kernel is a vanilla 2.6.28.10 (and I can't switch to
anything newer at this time because of missing squashfs-lzma support),
patched with PaX, netfilter-layer7, squashfs (with LZMA), and IMQ. The
base system is a Debian Lenny with some updates from testing/unstable.

Whenever interfaces using the sky2 module (this box has 8 network
interfaces in a 19" rack appliance) go down, the oops occurs:

[~]# ifdown -a --exclude=lo
[ 1535.000069] sky2 0000:01:00.0: error interrupt status=0xffffffff
[ 1535.006649] sky2 0000:01:00.0: PCI hardware error (0xffff)
[ 1535.012608] sky2 0000:01:00.0: PCI Express error (0xffffffff)
[ 1535.018821] sky2 wan: ram data read parity error
[ 1535.023827] sky2 wan: ram data write parity error
[ 1535.028913] sky2 wan: MAC parity error
[ 1535.032992] sky2 wan: RX parity error
[ 1535.036983] sky2 wan: TCP segmentation error
[ 1535.041655] general protection fault: 0000 [#1] PREEMPT SMP
[ 1535.045601] last sysfs file:
/sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed

[ 1535.045601] Modules linked in: xt_multiport cpufreq_userspace xt_DSCP
xt_length xt_mark xt_dscp xt_MARK xt_CONNMARK xt_comment xt_policy
ipt_REDIRECT ip6t_LOG xt_tcpudp ip6table_mangle iptable_mangle
ip6table_filter ip6_tables sit tunnel4 8021q garp stp llc ipt_LOG
xt_limit xt_state iptable_nat iptable_filter ip_tables x_tables dm_mod
p4_clockmod speedstep_lib freq_table tun imq nf_nat_ftp nf_nat
nf_conntrack_ftp nf_conntrack_ipv6 nf_conntrack_ipv4 nf_conntrack
nf_defrag_ipv4 ipv6 evdev parport_pc parport serio_raw pcspkr i2c_i801
i2c_core iTCO_wdt rng_core intel_agp agpgart squashfs sqlzma unlzma loop
aufs exportfs nls_utf8 nls_cp437 ide_generic sd_mod ide_gd_mod
ata_generic pata_acpi ata_piix piix ide_pci_generic ide_core skge sky2
thermal_sys
[ 1535.045601]

[ 1535.045601] Pid: 9960, comm: mv Not tainted (2.6.28.10 #2)

[ 1535.045601] EIP: 0060:[<f808085a>] EFLAGS: 00010286 CPU: 0

[ 1535.045601] EIP is at sky2_mac_intr+0x22/0x9d [sky2]

[ 1535.045601] EAX: f8090f88 EBX: 00000001 ECX: 00000008 EDX: 000000ff

[ 1535.045601] ESI: 00000000 EDI: f682cb80 EBP: 00000080 ESP: f5f13ed4

[ 1535.045601]  DS: 0068 ES: 0068 FS: 00d8 GS: 0033 SS: 0068

[ 1535.045601] Process mv (pid: 9960, ti=f5f12000 task=f4a961c0
task.ti=f5f12000)

[ 1535.045601] Stack:

[ 1535.045601]  ff08340b f682cb88 ffffffff ffffffff f712b800 f80839d6
00000040 f682cb88

[ 1535.045601]  00000000 00000001 f682cb80 c082111a 00000000 00000000
00000003 f7014b80
[ 1535.045601]  c0a604e8 00000246 f7014b80 c0838f21 00000000 c0a604e8
00000101 c1d10124
[ 1535.045601] Call Trace:
[ 1535.045601]  [<f80839d6>] sky2_poll+0x1cb/0xbed [sky2]
[ 1535.045601]  [<c082111a>] __wake_up+0x29/0x39
[ 1535.045601]  [<c0a604e8>] _spin_unlock_irqrestore+0x22/0x39
[ 1535.045601]  [<c0838f21>] __queue_work+0x4d/0x5a
[ 1535.045601]  [<c0a604e8>] _spin_unlock_irqrestore+0x22/0x39
[ 1535.045601]  [<c09eda45>] net_rx_action+0xb8/0x1f6
[ 1535.045601]  [<c082f954>] __do_softirq+0x95/0x142
[ 1535.045601]  [<c082fa49>] do_softirq+0x48/0x57
[ 1535.045601]  [<c082fbc9>] irq_exit+0x3b/0x78
[ 1535.045601]  [<c081218f>] smp_apic_timer_interrupt+0x75/0x7f
[ 1535.045601]  [<c0804f48>] apic_timer_interrupt+0x28/0x30
[ 1535.045601]  [<c0a60000>] rwsem_down_failed_common+0xa4/0x175
[ 1535.045601] Code: c0 83 c4 14 5b 5e 5f 5d c3 55 89 d5 57 89 c7 56 53
89 d3 c1 e5 07 83 ec 04 8b 74 90 30 8d 85 08 0f 00 00 03 07 8a 10 88 54
24 03 <f6> 86 0d 05 00 00 02 74 12 0f b6 c2 50 56 68 84 5b 08 f8 e8 cd
[ 1535.045601] EIP: [<f808085a>] sky2_mac_intr+0x22/0x9d [sky2] SS:ESP
0068:f5f13ed4
[ 1535.302490] Kernel panic - not syncing: Fatal exception in interrupt
[ 1535.309412] Rebooting in 30 seconds..


Or even when doing it more slowly, interface by interface:

[~]# ifdown tun6to4; cat /proc/net/dev | cut -d: -f1 | grep -v Inter |
grep -v face | sort -u | while read iface; do echo $iface; ifdown
$iface; sleep 3s; done
hb

lo

dmz

lan
[ 1127.000261] sky2 0000:04:00.0: error interrupt status=0xffffffff
[ 1127.007348] sky2 0000:04:00.0: PCI hardware error (0xffff)
[ 1127.013745] sky2 0000:04:00.0: PCI Express error (0xffffffff)
[ 1127.020468] sky2 lan: ram data read parity error
[ 1127.025834] sky2 lan: ram data write parity error
[ 1127.031302] sky2 lan: MAC parity error
[ 1127.035671] sky2 lan: RX parity error
[ 1127.039910] sky2 lan: TCP segmentation error
[ 1127.045079] general protection fault: 0000 [#1] PREEMPT SMP
[ 1127.048879] last sysfs file:
/sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed

[ 1127.048879] Modules linked in: xt_multiport cpufreq_userspace xt_DSCP
xt_length xt_mark xt_dscp xt_MARK xt_CONNMARK xt_comment xt_policy
ipt_REDIRECT ip6t_LOG xt_tcpudp ip6table_mangle iptable_mangle
ip6table_filter ip6_tables sit tunnel4 8021q garp stp llc ipt_LOG
xt_limit xt_state iptable_nat iptable_filter ip_tables x_tables dm_mod
p4_clockmod speedstep_lib freq_table tun imq nf_nat_ftp nf_nat
nf_conntrack_ftp nf_conntrack_ipv6 nf_conntrack_ipv4 nf_conntrack
nf_defrag_ipv4 ipv6 evdev parport_pc parport pcspkr serio_raw i2c_i801
i2c_core iTCO_wdt rng_core intel_agp agpgart squashfs sqlzma unlzma loop
aufs exportfs nls_utf8 nls_cp437 ide_generic sd_mod ide_gd_mod
ata_generic pata_acpi ata_piix piix ide_pci_generic ide_core skge sky2
thermal_sys
[ 1127.048879]

[ 1127.048879] Pid: 20150, comm: rndc Not tainted (2.6.28.10 #2)

[ 1127.048879] EIP: 0060:[<f808085a>] EFLAGS: 00010286 CPU: 0

[ 1127.048879] EIP is at sky2_mac_intr+0x22/0x9d [sky2]

[ 1127.048879] EAX: f80d8f88 EBX: 00000001 ECX: 00000008 EDX: 000000ff

[ 1127.048879] ESI: 00000000 EDI: f68c2a80 EBP: 00000080 ESP: eb83fb38

[ 1127.048879]  DS: 0068 ES: 0068 FS: 00d8 GS: 0000 SS: 0068

[ 1127.048879] Process rndc (pid: 20150, ti=eb83e000 task=f695bb00
task.ti=eb83e000)

[ 1127.048879] Stack:

[ 1127.048879]  ff08340b f68c2a88 ffffffff ffffffff f712c000 f80839d6
00000040 f68c2a88

[ 1127.048879]  c0a78d54 f70344e0 f68c2a80 f695bb00 c0a78d54 c0a604e8
c1d10980 c0a78d54

[ 1127.048879]  c0827013 00000000 0000000f 00000246 f70344e0 00000102
c0be5180 c0832dc6

[ 1127.048879] Call Trace:

[ 1127.048879]  [<f80839d6>] sky2_poll+0x1cb/0xbed [sky2]

[ 1127.048879]  [<c0a604e8>] _spin_unlock_irqrestore+0x22/0x39

[ 1127.048879]  [<c0827013>] try_to_wake_up+0x158/0x162

[ 1127.048879]  [<c0832dc6>] process_timeout+0x0/0x5

[ 1127.048879]  [<c09eda45>] net_rx_action+0xb8/0x1f6

[ 1127.048879]  [<c082f954>] __do_softirq+0x95/0x142

[ 1127.048879]  [<c082fa49>] do_softirq+0x48/0x57

[ 1127.048879]  [<c082fbc9>] irq_exit+0x3b/0x78

[ 1127.048879]  [<c081218f>] smp_apic_timer_interrupt+0x75/0x7f

[ 1127.048879]  [<c0804f48>] apic_timer_interrupt+0x28/0x30

[ 1127.048879]  [<c0867764>] get_page_from_freelist+0x2b8/0x3df

[ 1127.048879]  [<c0867ae0>] __alloc_pages_internal+0x98/0x37f

[ 1127.048879]  [<c0862ee0>] find_lock_page+0x10/0x43

[ 1127.048879]  [<c0a60555>] _spin_unlock+0x10/0x23

[ 1127.048879]  [<c086fb70>] __do_fault+0xaa/0x3bc

[ 1127.048879]  [<c08718e1>] handle_mm_fault+0x54a/0xbfa

[ 1127.048879]  [<c0a60555>] _spin_unlock+0x10/0x23

[ 1127.048879]  [<c089442d>] __d_lookup+0xfa/0x116

[ 1127.048879]  [<c088cb78>] do_lookup+0x53/0x153

[ 1127.048879]  [<c0893375>] dput+0x16/0xfc

[ 1127.048879]  [<c088eb25>] __link_path_walk+0xb01/0xbfb

[ 1127.048879]  [<c0a60555>] _spin_unlock+0x10/0x23

[ 1127.048879]  [<c086f05e>] kmap_high+0x17c/0x186

[ 1127.048879]  [<c0819b76>] default_spin_lock_flags+0x5/0x7

[ 1127.048879]  [<c081a64b>] do_page_fault+0x335/0x86e

[ 1127.048879]  [<c0a60555>] _spin_unlock+0x10/0x23

[ 1127.048879]  [<c0870a91>] unmap_vmas+0x498/0x6ab

[ 1127.048879]  [<c087321e>] free_pgtables+0x7d/0x93

[ 1127.048879]  [<c086d42e>] vma_prio_tree_insert+0x17/0x7f

[ 1127.048879]  [<c0874a45>] vma_link+0x51/0x73

[ 1127.048879]  [<c0a60555>] _spin_unlock+0x10/0x23

[ 1127.048879]  [<c0874a5f>] vma_link+0x6b/0x73

[ 1127.048879]  [<c08763b8>] mmap_region+0x475/0x58c

[ 1127.048879]  [<c08767a4>] do_mmap_pgoff+0x2d5/0x326

[ 1127.048879]  [<c08081db>] sys_mmap2+0x62/0x77

[ 1127.048879]  [<c08081e9>] sys_mmap2+0x70/0x77

[ 1127.048879]  [<c081a316>] do_page_fault+0x0/0x86e

[ 1127.048879]  [<c0a60805>] error_code+0x75/0x80

[ 1127.048879] Code: c0 83 c4 14 5b 5e 5f 5d c3 55 89 d5 57 89 c7 56 53
89 d3 c1 e5 07 83 ec 04 8b 74 90 30 8d 85 08 0f 00 00 03 07 8a 10 88 54
24 03 <f6> 86 0d 05 00 00 02 74 12 0f b6 c2 50 56 68 84 5b 08 f8 e8 cd

[ 1127.048879] EIP: [<f808085a>] sky2_mac_intr+0x22/0x9d [sky2] SS:ESP
0068:eb83fb38

[ 1127.470534] Kernel panic - not syncing: Fatal exception in interrupt

[ 1127.478035] Rebooting in 30 seconds..



It seems that the oops occurs when the last network interface using the
sky2 module goes down, although I am not completely certain about this.
I am also fairly sure that the other patches applied to 2.6.28.10 are
not at fault, as the same kernel works perfectly well on different
hardware (which is not using the sky2 NIC module).

Attached are the lspci -v output and the kernel config.

Any hints on what may be wrong would be highly appreciated. I am able to
try patches to sky2 and/or give remote ssh access to the box (although
it will be offline for 5 minutes after triggering the oops...).

best regards,
Rene

Download attachment "lspci.lst.gz" of type "application/gzip" (1595 bytes)

Download attachment "config-2.6.28.10.gz" of type "application/gzip" (19547 bytes)

Download attachment "signature.asc" of type "application/pgp-signature" (261 bytes)

Powered by blists - more mailing lists