Was rebooting an office machine after a few package upgrades, and right at the tail end of the reboot it appears that the kernel Oopsed right at the remount, just before reboot. I can attach my config and the kernel patchset provided with my distro, just ask. Below is my retyped Oops printout (photo of monitor is URL of bug): EIP is at dentry_iput+0x25/0xb0 EAX: f057983c EBX: f05797f8 ECX: f05cad1c EDX: b05cad1c ESI: f05cad04 EDI: f0579824 EBP: f05797f8 ESP: f69a8ef4 DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 Process umount (pid: 6490, ti=f69a8000 task=f72cc6e0 task.ti=f69a8000) Stack: 00000282 f0579770 f05787f8 f05797f8 f69a8f2c c10940c2 f6897a00 f69a8f2c <0> c10942fd f69a8f34 00000000 f6897a84 00000000 00000000 eafac2d4 f05798ac <0> f69a8f34 f69a8f34 f6897a00 00000000 00000000 f69a8000 c1086775 00000001 Call Trace: [<c10940c2>] ? d_kill+0x32/0x60 [<c10942fd>] ? __shrink_dcache_sb+0x20d/0x290 [<c1086775>] ? do_remount_sb+0x35/0xe0 [<c1099fa1>] ? sys_umount+0x321/0x350 [<c106ef77>] ? remove_vma+0x37/0x50 [<c1070502>] ? do_munmap+0x212/0x270 [<c1099fe7>] ? sys_oldumount+0x17/0x20 [<c1002e08>] ? sysenter_do_call+0x12/0x26 Code: 0f 0b eb fe 66 90 56 53 89 c3 83 ec 0c 8b 70 10 85 f6 0f 84 8e 00 00 00 c7 40 10 00 00 00 00 8d 40 44 8b 4b 44 8b 53 48 89 51 04 <89> 0a 89 43 44 89 43 48 fe 43 08 fe 05 c0 63 3f c1 8b 46 28 85 EIP: [<c1093f85>] dentry_iput+0x25/0xb0 SS:ESP 0068:f69a8ef4 CR2: 00000000b05cad1c ---[ end trace 29d86a99dda601c9 ]---
Oh, it might be helpful to add that this is a pure 64-bit machine. chris@chariot ~ $ uname -a Linux chariot 2.6.32-gentoo-r5 #2 SMP Mon Feb 15 12:51:56 CST 2010 i686 Intel(R) Core(TM)2 Duo CPU E8200 @ 2.66GHz GenuineIntel GNU/Linux
(In reply to comment #1) > Oh, it might be helpful to add that this is a pure 64-bit machine. More bugspam. I meant "this is *not* a pure 64-bit machine" but of course I can't get my booleans straight.
Thanks for the report. I'm sorry but this looks like a single-bit error. From the oops, it looks like the kernel tried to dereference address 0xb05cad1c while the correct address would be f05cad1c (as you can see, lots of addresses in this range are in registers so that's probably where kernel has it's data on your machine). So I'd blame either cosmic rays or faulty HW. Maybe try running memtest, or try replacing memory if it repeats. I'll close the bug as invalid...