Re: [PATCH] one of $BIGNUM devfs races

[Richard Gooch]

Alexander Viro writes:
> 	OK, folks - that's it.  By all reasonable standards a year
> _is_ sufficient time to fix an obvious race.  One in
> devfs/base.c::create_entry() had been described to Richard more than
> a year ago.  While I respect the "I'll do it myself, don't spoil the
> fun" stance, it's clearly over the bleedin' top.  Patch for that one
> is in the end of posting.  Linus, see if it looks sane for you.

Linus: please don't apply.
Alan: I notice you've put Al's patch into 2.4.7-ac8. Please remove it.

This patch has the following ugly construct:

> +	/*  Ensure table size is enough  */
> +	while (fs_info.num_inodes >= fs_info.table_size) {

Putting the allocation inside a while loop is horrible, and isn't a
perfect solution anyway. I'm fixing this (and other races) with proper
locking. If you went to the trouble to start patching, why at least
didn't you do it cleanly with a lock?

Furthermore, the patch makes gratuitous formatting changes (which made
it harder to see what it actually *changed*).

				Regards,

					Richard....
Permanent: [e-mail]
Current:   [e-mail]