On Fri, Jan 28, 2000 at 02:54:52AM -0500, der Mouse wrote:
> ...well, no, it's weirder than that.
> 
> I was working with NetBSD/sun3 and somehow managed to create a
> filesystem that fsck thought was damaged but couldn't fix.  It would
> always say
> 
> BLK(S) MISSING IN BIT MAPS
> SALVAGE?
> 
> and then, when I say "y", would write on the disk...but then running
> fsck again produces the same "error".  And I have reason to think that
> the filesystem actually was corrupt; I twice got a panic ("freeing free
> block", IIRC), and when cleaning up afterwards I found this weirdness.

Could you try a 'fsck -d' and post results ?
> 
> Then, before running off at the keyboard about it, I moved the disk to
> a SPARC and tried fsck there.  It produced no complaints at all, and
> had no problem using the filesystem.
> 
> This is just too weird.  I have no idea how reproducible this may prove
> to be...anyone have any ideas?  Does NetBSD/sun3 DTRT with memory
> parity errors (ie, could this be an undetected memory error)?

Weird. Either a compiler bug or a hardware prolem.
Coudl you try to compile a fsck_ffs with different compilers options ?

--
Manuel Bouyer <bouyer@antioche.eu.org>
--