Unfortunately, IDE is unstable in 2.1.85 and on, at least under SMP.
I managed to capture a failure by logging over my network. This has
been consistent behavior with 2.1.85, 86, 87, 88, and pre-89-3 and 4.
The system behaves normally for a while, then crashes and burns. It's
not necessarily provoked by heavy IDE load either (although that seems
to set things off faster).
I've been done this booting w/o options and with the
"ide0=noautotune" option. Haven't tried a non-SMP kernel yet.
Start with these 6 messages:
Feb 28 23:44:50 peter kernel: hda: write_intr: status=0xd0 { Busy }
Feb 28 23:44:50 peter kernel: ide0: reset: success
Feb 28 23:44:50 peter kernel: hda: recal_intr: status=0xd0 { Busy }
Feb 28 23:44:50 peter kernel: ide0: reset: success
Feb 28 23:44:51 peter kernel: hda: recal_intr: status=0xd0 { Busy }
Feb 28 23:44:51 peter kernel: end_request: I/O error, dev 03:03, sector 3948627
Repeat with different sector numbers. After a few
iterations, I start getting these periodically (each following an
end_request line):
Feb 28 23:44:52 peter kernel: IO error syncing ext2 inode [03:03:00003007]
Then switch "write_intr" above to "read_intr" and repeat the above;
Then start interspersing a few panics, each after a "read_intr" or
"write_intr" line:
Feb 28 23:44:57 peter kernel: Kernel panic: EXT2-fs panic (device 03:03): ext2_write_inode: unable to read i-node block - inode=12295, block=49164
After which "read_intr" and "write_intr" start appearing
interchangably. These continue to be sprinkled with panics after
{read,write}_intr lines:
Feb 28 23:44:57 peter kernel: Kernel panic: EXT2-fs panic (device 03:03): ext2_write_inode: unable to read i-node block - inode=493867, block=1974321
Feb 28 23:45:12 peter kernel: Kernel panic: EXT2-fs panic (device 03:03): ext2_write_inode: unable to read i-node block - inode=493866, block=1974321
Feb 28 23:45:20 peter kernel: Kernel panic: EXT2-fs panic (device 03:03): ext2_write_inode: unable to read i-node block - inode=493798, block=1974312
Feb 28 23:45:21 peter kernel: Kernel panic: EXT2-fs panic (device 03:03): ext2_write_inode: unable to read i-node block - inode=493852, block=1974319
Finally, there are two panic's back-to-back, and the system goes
brain-dead. Interrupt handlers still work, VC's still switch, pings
are still answered, but userland is toast. Filesystem has (so far)
always been easily recoverable with e2fsck, but I don't know how many
more times I want to risk that ;-)
I can provide the complete log, or just sector numbers, or whatever
subset of the data would be useful/informative/interesting.
Pertinent hardware (abbreviated):
Twin Pentium-MMX 166's, Gigabyte MB, 64 MB EDO
peter kernel: PIIX3: IDE controller on PCI bus 0 function 57
peter kernel: PCI->APIC IRQ transform: (B0,I7,P-1) -> -1
peter kernel: PIIX3: not 100% native mode: will probe irqs later
peter kernel: ide0: BM-DMA at 0xf000-0xf007, BIOS settings: hda:pio, hdb:pio
peter kernel: ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:pio, hdd:pio
peter kernel: hda: QUANTUM FIREBALL ST3.2A, ATA DISK drive
peter kernel: hdb: MAXTOR LXT-340A, ATA DISK drive
peter kernel: hdc: TOSHIBA CD-ROM XM-5702B, ATAPI CDROM drive
peter kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
peter kernel: ide1 at 0x170-0x177,0x376 on irq 15
Suggestions, comments, etc?
(P.S. - 2.1.84 was rock-solid-stable, and I didn't see much in 2.1.85
that should've touched IDE directly. IO-APIC changes maybe? Guess I
better hack up a 2.1.89-4 kernel to compile for UP and start hammering
it...)
Adam
-- All my instincts, they return Adam D. Bradley artdodge@cs.bu.edu The grand facade so soon will burn Boston University Computer Science Without a noise, without my pride Grad student and Linux hacker I reach out from the inside ----------> In Your Eyes ---------> <><
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu