Kernel Oops/BUG during network operation with kernel version 3.0.101-0.18.1

This document (7014823) is provided subject to the disclaimer at the end of this document.

Environment

SUSE Linux Enterprise Server 11 SP3
SUSE Linux Enterprise Desktop 11 SP3

Situation

Some network operations cause a kernel oop/Bug when running the 3.0.101-0.18.1 version of the kernel.

[ 2453.277767] kernel BUG at /usr/src/packages/BUILD/kernel-default-3.0.101/linux-3.0/include/net/inetpeer.h:71!
[ 2453.277767] invalid opcode: 0000 [#28] SMP 
[ 2453.277767] CPU 4 
[ 2453.277767] Modules linked in: coretemp edd mperf microcode loop dm_multipath dm_mod shpchp pci_hotplug 
i5k_amb ipv6 ipv6_lib i2c_i801 e1000e sg i5000_edac sr_mod iTCO_wdt iTCO_vendor_support ioatdma cdrom ses 
edac_core dca floppy joydev pcspkr enclosure button rtc_cmos serio_raw container ext3 jbd mbcache btrfs 
zlib_deflate crc32c libcrc32c radeon ttm drm_kms_helper drm i2c_algo_bit i2c_core usbhid hid uhci_hcd 
qla2xxx ehci_hcd usbcore scsi_transport_fc sd_mod scsi_tgt crc_t10dif usb_common processor thermal_sys
hwmon scsi_dh_rdac scsi_dh_hp_sw scsi_dh_alua scsi_dh_emc scsi_dh 3w_9xxx ata_generic ata_piix ahci libahci 
libata scsi_mod
[ 2453.277767] Supported: Yes
[ 2453.277767] 
[ 2453.277767] Pid: 3832, comm: bs_sched Tainted: G      D      
3.0.101-0.18-default #1 Supermicro X7DB8/X7DB8
[ 2453.277767] RIP: 0010:[<ffffffff813d17e5>]  [<ffffffff813d17e5>] ipv4_dst_check+0x65/0xc0
[ 2453.277767] RSP: 0018:ffff8803176ffd48  EFLAGS: 00010202
[ 2453.277767] RAX: 000000001c38eec1 RBX: ffff88031c38eec1 RCX: 0000000000000000
[ 2453.277767] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000002
[ 2453.277767] RBP: ffff88031a002bc0 R08: 000000000000001a R09: 0000000000000000
[ 2453.277767] R10: 0000000000000028 R11: 0000000000000001 R12: 0000000000000000
[ 2453.277767] R13: ffff8801fbc35328 R14: 0000000000000000 R15: 0000000000000028
[ 2453.277767] FS:  00007fe319d1a700(0000) GS:ffff88032fd00000(0000) knlGS:0000000000000000
[ 2453.277767] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 
[ 2453.277767] CR2: 00007fbb7859e168 CR3: 0000000312749000 CR4: 00000000000007e0
[ 2453.277767] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 2453.277767] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 2453.277767] Process bs_sched (pid: 3832, threadinfo ffff8803176fe000, task ffff8803176fc140)
[ 2453.277767] Stack:
[ 2453.277767]  000000000000b757 ffff88031a002bc0 ffff8801fbc35040 ffffffff81391d7a
[ 2453.277767]  000102d000000001 ffff880302adbdb0 ffff8801fbc35040 ffffffff813db23a
[ 2453.277767]  ffff8801fbc35040 ffff880317534908 ffff880302adbdb0 00000000000000d0
[ 2453.277767] Call Trace:
[ 2453.277767]  [<ffffffff81391d7a>] __sk_dst_check+0x4a/0x70
[ 2453.277767]  [<ffffffff813db23a>] ip_queue_xmit+0x28a/0x3e0
[ 2453.277767]  [<ffffffff813f1480>] tcp_transmit_skb+0x510/0x710
[ 2453.277767]  [<ffffffff813f3547>] tcp_connect+0x187/0x240
[ 2453.277767]  [<ffffffff813f77e8>] tcp_v4_connect+0x3d8/0x4e0
[ 2453.277767]  [<ffffffff81408cf8>] inet_stream_connect+0x1b8/0x1e0
[ 2453.277767]  [<ffffffff8138cfc7>] sys_connect+0xe7/0xf0
[ 2453.277767]  [<ffffffff81468d12>] system_call_fastpath+0x16/0x1b
[ 2453.277767]  [<00007fe319243ad0>] 0x7fe319243acf
[ 2453.277767] Code: 00 00 74 e6 48 8b 9f d8 00 00 00 89 d8 a8 01 74 16 8b b7
b4 00 00 00 31 d2 e8 78 fd ff ff 48 8b 9d d8 00 00 00 89 d8 a8 01 74 0b <0f> 0b
eb fe 0f 1f 80 00 00 00 00 48 85 db 74 1d 48 89 de 48 89 
[ 2453.277767] RIP  [<ffffffff813d17e5>] ipv4_dst_check+0x65/0xc0
[ 2453.277767]  RSP <ffff8803176ffd48>
[ 2454.407168] ---[ end trace ceb5e9eda14179b5 ]---

Resolution

On April 8, 2014, an updated SLES kernel was released that addressed this issue:

# uname -r
3.0.101-0.21-default

This kernel version is available through the normal update channels.

Prior to the release of this updated kernel, to resolve this problem the recommendation was to backrev to the previous kernel. Those instructions are kept in this document (as seen below), but do not represent the suggested solution for this problem.

Backrev to the previous kernel, until and when a new kernel is available.

# uname -r
3.0.101-0.18-default

# zypper se -s kernel-default
This will list the available kernel versions in the update channel.

# zypper in --no-refresh --oldpackage kernel-default=3.0.101-0.15.1
Installs the older kernel. You will be prompted to backrev other dependent kernel packages. Select Solution 1 to downgrade your kernel.

# zypper in --no-refresh --oldpackage kernel-source=3.0.101-0.15.1
You will also need to downgrade the kernel source so applications that depend on the kernel can properly compile their kernel drivers. Again, select Solution 1 to downgrade dependent packages.

# reboot
The server needs to be rebooted to activate the downgraded kernel.

Additional Information

The kernel has been removed from the update channel and patch areas.

Disclaimer

This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.