2.4.0-test12 not liking high disk i/o
Quoted 8 lines · Niels Kristian Bech Jensen
Quoted 3 lines · Mohammad A. Haque
Any one else experiencing problems when they do lots of disk activity in test12?
Yes, I've had some complete freezes (nothing working at all) in test12-pre8 and test12. They can be triggered by e.g. Netscape. test12-pre7 seems to be stable.
test12-pre8 deadlocked (but probably livelocked, as NMI watchdog was not trigerred? or maybe NMI is not trigerred when spinning with interrupts enabled) on me. One CPU was spinning in drivers/char/tty_io.c:do_tty_write:716 in lock_kernel(), another one in fs/fat/inode.c:fat_write_inode:863 in spin_lock(&fat_inode_lock). But as I did not found how this could happen, I did not report it (afaik if you hold fat_inode_lock, you never sleep, not even talking about invoking fat_write_inode... But maybe I overlooked some codepath).
Of course, I do not have full stack traces of these two offenders, as I do not have kdb in kernel yet...
It happened when I decided to copy old 18GB IDE disk to new 40GB IDE one (both UDMA33, one (18GB src) as primary master, one (40GB dst) as secondary master; i440BX).
During copy of 8GB ext2 partition I decided that copying 200MB of data from VFAT in parallel (and leaving room for a while) with ext2 copy should not kill system, so I did it. Unfortunately, it killed it :-( ext2 was copied with tar clf - /mnt2 | (cd /new/mnt2 ; tar xf -), vfat was copied with midnight commander. From status on screen of mc it looks like that deadlock occured after copy, but before screen was refreshed - mc showed status 100%, copying last file in root directory of vfat. vfat contents was copied to same partition as ext2, if that matters.
BTW, what's correct way of loading .o kernel module into gdb? I noticed that if I create shared module from it (I was doing this up to now), (mine) ld drops .text.lock section. Have I to use (modified) vmlinux.lds for linking? (if loaded as relocatable module, all sections begins at zero, overlapping one over another)
Best regards,
Petr Vandrovec [email protected]
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/
well, i hate to be piling on here, but i just encountered this (i think it's this) this morning. i was printing a 145+m file (to /dev/lp0) from an ide drive and it locked up. just before the lockup, i noticed it was very sluggish, as if it were under very heavy load (which is really wasn't). this is on an smp-enabled machine (noapic at lilo prompt because of the usb/pirq(?) problem). i'm using 2.4.0-test12 on a tyan tiger 133 (via apollo 133a chipset) mobo. the machine has 512m memory and another 512m in swap (didn't notice much swap activity, but i could have missed it). there were no messages in the logs.
if there's any info i can provide or tests i can run, just let me know.
pete
Quoted 9 lines · Petr Vandrovec
Quoted 8 lines · Niels Kristian Bech Jensen
Quoted 3 lines · Mohammad A. Haque
Any one else experiencing problems when they do lots of disk activity in test12?
Yes, I've had some complete freezes (nothing working at all) in test12-pre8 and test12. They can be triggered by e.g. Netscape. test12-pre7 seems to be stable.
Signature · 2 lines
Pete Toscano p:[email protected] w:[email protected] GPG fingerprint: D8F5 A087 9A4C 56BB 8F78 B29C 1FF0 1BA7 9008 2736
i440BX is consistent with mine as is running the drive at UDMA33.
Quoted 3 lines
It happened when I decided to copy old 18GB IDE disk to new 40GB IDE one (both UDMA33, one (18GB src) as primary master, one (40GB dst) as secondary master; i440BX).
Signature · 13 lines
=====================================================================
Mohammad A. Haque http://www.haque.net/
[email protected]
"Alcohol and calculus don't mix. Project Lead
Don't drink and derive." --Unknown http://wm.themes.org/
[email protected]
=====================================================================
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/Quoted 6 lines · Mohammad A. Haque
i440BX is consistent with mine as is running the drive at UDMA33.
Quoted 3 lines
It happened when I decided to copy old 18GB IDE disk to new 40GB IDE one (both UDMA33, one (18GB src) as primary master, one (40GB dst) as secondary master; i440BX).
My system is an old 486DX4-100MHz (AMD processor), SiS 85C496 chipset, and no UDMA33.
Signature · 8 lines
Niels Kristian Bech Jensen -- [email protected] -- http://www.image.dk/~nkbj/ ----------->> Stop software piracy --- use free software! <<----------- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] Please read the FAQ at http://www.tux.org/lkml/