Bug#508151: linux-image-2.6.26-1-sparc64: soft lockup when trying to bring up eth1
Package: linux-image-2.6.26-1-sparc64
Version: 2.6.26-11
Severity: important
Hey,
I have a Sun Netra T1 200 with two ethernet ports on a sungem card.
eth0 has no cable connected, eth1 is connected to a regular 100BaseT
switch.
When I boot up the box, I get flooded with
eth0: switching to forced 100bt
eth0: switching to forced 10bt
But the box is stable and I can do everything with it.
However, when I try to "ifup eth1" (both interfaces have no "auto" lines
in /e/n/interfaces),
the box gets stuck, prints a soft-lockup and isn't responsible anymore.
The lockup looks like this:
[ 361.852604] BUG: soft lockup - CPU#0 stuck for 61s! [ifconfig:1750]
[ 362.062093] Modules linked in: ipv6 dm_snapshot dm_mirror dm_log
dm_mod ext3 jb3 xt3
jbd sd_mod ide_cd_mod cdrom ata_generic libata ide_pci_generic sym53c8xx
scsi_trrransport_spi scsi_tr
ansport_spi alim15x3 ohci_hcd sungem sungem_phy scsi_mod scsi_mod
[ 363.191678] TSTATE: 0000004480009601 TPC: 0000000000485700 TNPC:
000000000048885704 Y: 00000000 Not tainted
5704 Y: 00000000 Not tainted 0009601 TPC: 0000000000485700 TNPC:
000000000048
[ 363.559375] TPC: <__rcu_process_callbacks+0x4c/0x208>
[ 363.739299] g0: 00000000007b5560 g1: 0000000000000000 g2:
0000000000000554 g333: 0000000000000000
: 0000000000000000 00000000007b5560 g1: 0000000000000000 g2:
0000000000000554 g3
[ 364.078589] g4: fffff8000e05a640 g5: 0000006574683000 g6:
fffff8000e2c0000 g777: 00000000007b5560
: 00000000007b5560 fffff8000e05a640 g5: 0000006574683000 g6:
fffff8000e2c0000 g7
[ 364.417878] o0: 00000000007462a8 o1: 00000000007cd3c8 o2:
00000001000010ca o333: 0000000000000000
: 0000000000000000 00000000007462a8 o1: 00000000007cd3c8 o2:
00000001000010ca o3
[ 364.757167] o4: fffff8000c1b2598 o5: fffff8000fe83e60 sp:
fffff8000fe835c1 reeet_pc: 00000000004858dc
t_pc: 00000000004858dc f8000c1b2598 o5: fffff8000fe83e60 sp:
fffff8000fe835c1 re
[ 365.105197] RPC: <rcu_process_callbacks+0x20/0x30>
[ 365.278583] l0: 00000000000000ca l1: fffff8000fe83e60 l2:
00000000007b48c0 l333: 0000000000000400
: 0000000000000400 00000000000000ca l1: fffff8000fe83e60 l2:
00000000007b48c0 l3
[ 365.617871] l4: 0000000000000000 l5: 0000000000000003 l6:
0000000000000000 l777: 0000000000000008
: 0000000000000008 0000000000000000 l5: 0000000000000003 l6:
0000000000000000 l7
[ 365.957160] i0: 00000000007462c8 i1: 00000000007cd438 i2:
0000000000000001 i333: 0000000000000000
: 0000000000000000 00000000007462c8 i1: 00000000007cd438 i2:
0000000000000001 i3
[ 366.296449] i4: 0000000000003fff i5: 00000000007b5980 i6:
fffff8000fe83681 i777: 00000000004571b8
: 00000000004571b8 0000000000003fff i5: 00000000007b5980 i6:
fffff8000fe83681 i7
[ 366.635744] I7: <__do_softirq+0x48/0xb8>
(sorry for the broken lines, my serial console is somehow borked)
This bug is opened upstream:
http://bugzilla.kernel.org/show_bug.cgi?id=10309
And at ubuntu: https://bugs.launchpad.net/linux/+bug/234304/
It does not happen with Etch's 2.6.18, but is still present in
2.6.28-rc7 :/
Bisecting linux-2.6, it looks like the following commit is the bad one:
commit bea3348eef27e6044b6161fd04c3152215f96411
Author: Stephen Hemminger <shemminger@linux-foundation.org>
Date: Wed Oct 3 16:41:36 2007 -0700
[NET]: Make NAPI polling independent of struct net_device objects.
However, I wasn't able to properly revert it to test the bug.
Hope some of you sparc guys have some idea how to fix that, as this is a
imho bad regression from Etch.
Regards
Evgeni
-- System Information:
Debian Release: lenny/sid
APT prefers testing
APT policy: (500, 'testing'), (500, 'stable')
Architecture: sparc (sparc64)
Kernel: Linux 2.6.23-sparc-1
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash
Versions of packages linux-image-2.6.26-1-sparc64 depends on:
ii debconf [debconf-2.0] 1.5.24 Debian configuration management sy
ii initramfs-tools [linux-initra 0.92j tools for generating an initramfs
ii module-init-tools 3.4-1 tools for managing Linux kernel mo
linux-image-2.6.26-1-sparc64 recommends no packages.
Versions of packages linux-image-2.6.26-1-sparc64 suggests:
pn fdutils <none> (no description available)
pn linux-doc-2.6.26 <none> (no description available)
ii silo 1.4.13a+git20070930-3 Sparc Improved LOader
-- debconf information:
shared/kernel-image/really-run-bootloader: true
linux-image-2.6.26-1-sparc64/preinst/lilo-initrd-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/postinst/old-initrd-link-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/postinst/kimage-is-a-directory:
linux-image-2.6.26-1-sparc64/preinst/lilo-has-ramdisk:
linux-image-2.6.26-1-sparc64/preinst/abort-install-2.6.26-1-sparc64:
linux-image-2.6.26-1-sparc64/preinst/elilo-initrd-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/postinst/old-dir-initrd-link-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/prerm/removing-running-kernel-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/postinst/depmod-error-2.6.26-1-sparc64: false
linux-image-2.6.26-1-sparc64/preinst/already-running-this-2.6.26-1-sparc64:
linux-image-2.6.26-1-sparc64/preinst/overwriting-modules-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/preinst/failed-to-move-modules-2.6.26-1-sparc64:
linux-image-2.6.26-1-sparc64/postinst/create-kimage-link-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/postinst/bootloader-test-error-2.6.26-1-sparc64:
linux-image-2.6.26-1-sparc64/postinst/bootloader-error-2.6.26-1-sparc64:
linux-image-2.6.26-1-sparc64/postinst/depmod-error-initrd-2.6.26-1-sparc64: false
linux-image-2.6.26-1-sparc64/preinst/initrd-2.6.26-1-sparc64:
linux-image-2.6.26-1-sparc64/prerm/would-invalidate-boot-loader-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/preinst/abort-overwrite-2.6.26-1-sparc64:
linux-image-2.6.26-1-sparc64/postinst/old-system-map-link-2.6.26-1-sparc64: true
linux-image-2.6.26-1-sparc64/preinst/bootloader-initrd-2.6.26-1-sparc64: true
Reply to: