[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Crash! kernel 2.4.2 jumps to mon



Benjamin Herrenschmidt wrote:
> 
> >Can you send me your System.map for that kernel?
> >
> >The linuxcare.com kernels that Paul keeps have changes that he's not ready
> >to push into the main tree yet.  It's likely you discovered one of the
> >reasons he holds off on pushing his changes :)  You may want let Paul know,
> >too.
> 
> I would rather suspect a latent bug in the HFS fs code related to dentry
> caching. There is already a report of this bug on the sourceforge site.
> I asked Klaus Halfman to have a look at it (he's working on HFS+ and have
> some knowledge of the VFS and HFS filesystem that I don't have).
> 

Well, considering that the HFS code hasn't been actively maintained since the
2.0 days as far as I can tell, it's no surprise it isn't completely stable. In
fact, I took a look at the code, and it just doesn't look right. I haven't
really pinned it down, but it looks like someone went through and "got it to work"
in 2.4 and didn't have the time to do a better job of it. I seem to recall that
happening in the 2.2 kernel as well. The dcache stuff in HFS is all added after
the fact, and isn't really part of the design as far as I can tell. It really
needs an almost complete rewrite, but I doubt anyone has the time. I know I
don't.

On a somewhat brighter note, there's now a 'linux-hfsplus' project on
sourceforge that everyone can keep an eye on. Hopefully I'll have some
actual code up there in a few weeks.

http://sourceforge.net/projects/linux-hfsplus/

	Brad Boyer
	flar@pants.nu



Reply to: