Re: Don't use 2.1.44 (inode code race)

Mark Hemment (markhe@sco.COM)
Tue, 8 Jul 1997 14:00:32 +0100 (BST)


On 8 Jul 1997, Linus Torvalds wrote:
> Anyway, the jury is still out on 2.1.44, and I'd still like people to
> test it out - but be aware that there is certainly some feature out
> there that people of little faith would call a bug. So test at your own
> risk (but I'd be very happy indeed if people test it out and mayhaps
> help me find the problem).

>From a quick look at fs/inode.c, there appears to be a problem with
This function calls _get_empty_inode(), which may block, but it does not
re-check the inode-cache. (It does check the cache before calling
_get_empty_inode(), which iget() does anyway.)

I couldn't see any semaphore which would lock the inode-cache. vfs_lock()
is empty! Perhaps there is a guard else where, and I missed seeing it.

This is the same in 2.1.43, so I doubt this is causing the latest



Mark Hemment UNIX/C Software Engineer (contractor)
"Success has many fathers. Failure is a b**tard"