Re: NOT !! 2.3.99-pre9-5 fixed deadlockswith ReiserFS (possible fix to reiserfs)

From: Chris Mason (
Date: Wed May 24 2000 - 22:13:59 EST

On Wed, 24 May 2000, Matt Yourst wrote:

> David Ford wrote:
> >
> > Do you want to make avail the patched patch? :)
> > Chris Mason wrote:
> >
> > > > The -pre9 + reiserfs 3.6.5 combo is still definitely *DANGEROUS*.
> > >
> > > 3.6.5 needed minor patching to work under 2.3.99-pre9-4 and higher. You
> > > need to fix the readpage/writepage, and remove all the d_delete calls from
> > > fs/reiserfs/namei.c
> I tried to apply these changes manually and it appears to fix the problem (i.e.,
> cron no longer hangs on startup, other programs don't deadlock with mmap-related
> things, etc.) I found a *nasty* bug in the parameters to writepage(): the method
> signature changed after pre7-3 such that block_write_full_page was unknowingly
> being passed a struct dentry instead of a struct page! Maybe this was the bug
> Chris was referring to, as it could very well have been responsible for the data
> corruption problem reported earlier.
> Here are the patches that appear to fix this:
These are correct ;-) Sorry for not sending them out with my first mail.

Yet another promise for 3.6.6 real soon now, and against 2.3.99pre9.
Unless yura reports his machine has blown up in overnight testing, it
should go out tomorrow.


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

This archive was generated by hypermail 2b29 : Wed May 31 2000 - 21:00:13 EST