Re: 2.3.42 fs corruption

[Horst von Brand]

[Synopsis: I've been seeing ext2 FS corruption with every 2.3.x kernel
 here, seemingly triggered by intense disk activity (like compiling a
 kernel). Andrea Arcangeli came up with a patch, included in 2.3.42aa2,
 that should solve this; a variant of this is supposed to be in 2.3.43]

Jeff Garzik <[e-mail]> said:

[...]

> 2.95.2 should be good enough for a corruption test...  can you try your
> tests replacing egcs, above, with gcc 2.95.2?

OK, again: Built 2.3.43 with gcc-2.95.2, binutils-2.9.5.0.27; on that I ran
6 "rpm -bb" in paralell. No corruption (yet?). Currently running that
kernel.

The other kernels tested (2.3.42aa[23], egcs-1.1.2-24 (RH 6.1) or
gcc-2.95.2) were running some time after similar torture tests; due to a
bug in IDE CD as a module + isofs as a module (try mounting one, the
machine hangs) I had to reset (==> fsck everywhere ;). No problems surfaced
either.

Looks fixed, but I don't know if my torture is a reliable way to trigger
the bug(s). Caveat lector.
-- 
Horst von Brand                             [e-mail]
Casilla 9G, Viņa del Mar, Chile                               [phone]