Re: cvs commit: src/sys/fs/hpfs hpfs_vnops.c src/sys/fs/msdosfs msdosfs_denode.c src/sys/fs/ntfs ntfs_vnops.c src/sys/fs/nwfs nwfs_node.c src/sys/fs/smbfs smbfs_node.c src/sys/fs/udf udf_vnops.c src/sys/isofs/cd9660 cd9660_node.c src/sys/nfsclient ...

From: Frank Mayhar <frank_at_exit.com>
Date: Tue, 17 Jan 2006 14:55:40 -0800
On Tue, 2006-01-17 at 22:50 +0100, Alexander Leidinger wrote:
> On Tue, 17 Jan 2006 17:29:03 +0000 (UTC)
> Alfred Perlstein <alfred_at_FreeBSD.org> wrote:
> >   to be identical, as if they were just clones of one original routine.
> "Clones" as in identical and could be refactored into one function?

As "refactored" is just a pretentious term for "rewritten," yes, pretty
much.  All of the xxx_reclaim() functions that I fixed were essentially
identical.

> Do you think this panic can be triggered by e.g. letting rhytmbox run
> over a lot of music files on 2 msdosfs partitions to update the music
> library?

Yes, if it's dirtying pages on msdosfs (by, say, changing access times)
and msdosfs has a putpages routine, _and_ that routine is called as a
result of the destroy_object call.  Otherwise no.
-- 
Frank Mayhar frank_at_exit.com     http://www.exit.com/
Exit Consulting                 http://www.gpsclock.com/
                                http://www.exit.com/blog/frank/
Received on Tue Jan 17 2006 - 22:55:47 UTC