Description of problem: I have an issue about Kernel 2.6.23.1-21.fc7. The system has an md device made of four partitions (raid5). Ontop is lvm where I created a logical volume and put an ext3 filesystem on. This raid has been created about one year ago. When I read (cp, cat) a certain file in that filesystem the kernel triggers an kernel bug. If I access the same file with 2.6.23.1-10.fc7 everythins works fine. Version-Release number of selected component (if applicable): 2.6.23.1-21.fc7 How reproducible: read file from filesystem Steps to Reproduce: 1. cp fileA fileB Actual results: <0>------------[ cut here ]------------ <2>kernel BUG at drivers/md/raid5.c:144! <0>invalid opcode: 0000 [#1] <0>SMP <4>Modules linked in: nfsd exportfs lockd nfs_acl auth_rpcgss autofs4 hidp l2cap bluetooth sunrpc ipv6 raid1 video output sbs battery ac via_ircc irda button crc_ccitt i2c_viapro e1000 i2c_core sg dm_snapshot dm_zero dm_mirror dm_mod ata_generic pata_via libata sd_mod scsi_mod raid456 async_xor async_memcpy async_tx xor ext3 jbd mbcache ehci_hcd ohci_hcd uhci_hcd <0>CPU: 0 <0>EIP: 0060:[<df8c08c9>] Not tainted VLI <0>EFLAGS: 00010002 (2.6.23.1-21.fc7 #1) <0>EIP is at __release_stripe+0x96/0x117 [raid456] <0>eax: 00000000 ebx: de70f840 ecx: 00000296 edx: de70f840 <0>esi: deeb4400 edi: de70f848 ebp: deeb44d4 esp: ded5af3c <0>ds: 007b es: 007b fs: 00d8 gs: 0000 ss: 0068 <0>Process md3_raid5 (pid: 433, ti=ded5a000 task=def2d840 task.ti=ded5a000) <0>Stack: 00000296 de70f840 deeb4400 00000296 df8c096b deeb44d4 deeb4400 de70f840 <0> 00000000 df8c59d2 7fffffff ded5afcc 00000000 c061b8a0 dee81000 00000000 <0> dee6afa0 00000246 00000200 7fffffff ded5afcc 00000000 ded5af94 ded5afb8 <0>Call Trace: <0> [<df8c096b>] release_stripe+0x21/0x2e [raid456] <0> [<df8c59d2>] raid5d+0x397/0x39e [raid456] <0> [phys_startup_32+-1084114784/-1073741824] schedule_timeout+0x13/0x8f <0> [phys_startup_32+-1084639073/-1073741824] md_thread+0xb9/0xcf <0> [phys_startup_32+-1086073915/-1073741824] autoremove_wake_function+0x0/0x35 <0> [phys_startup_32+-1084639258/-1073741824] md_thread+0x0/0xcf <0> [phys_startup_32+-1086074114/-1073741824] kthread+0x38/0x5e <0> [phys_startup_32+-1086074170/-1073741824] kthread+0x0/0x5e <0> [phys_startup_32+-1086300741/-1073741824] kernel_thread_helper+0x7/0x10 <0> ======================= <0>Code: 80 c4 01 00 00 e8 f5 6f c2 e0 e9 86 00 00 00 90 0f ba 73 20 08 8d 4e 38 8b 51 04 8d 43 08 e8 d0 64 c3 e0 eb 70 83 7a 34 00 74 04 <0f> 0b eb fe 90 0f ba 72 20 05 19 c0 85 c0 74 18 90 ff 4e 58 83 <0>EIP: [<df8c08c9>] __release_stripe+0x96/0x117 [raid456] SS:ESP 0068:ded5af3c Expected results: copy of file Additional info: If I access the same file with 2.6.23.1-10.fc7 everythins works fine.
This is probably due to a patch that got misapplied upstream, then copied into Fedora. Can you try 2.6.23.1-28 (or later) when it is available?
2.6.23.1-28 submitted to updates-testing.