]> git.hungrycats.org Git - linux/commit
projects / linux / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 90b163a) | patch
[PATCH] dentry d_bucket fix
authorAndrew Morton <akpm@osdl.org>
Fri, 14 May 2004 12:47:06 +0000 (05:47 -0700)
committerLinus Torvalds <torvalds@ppc970.osdl.org>
Fri, 14 May 2004 12:47:06 +0000 (05:47 -0700)
commitb7b5563e74bbeff253ea454ff301fa6e6712a84c
tree238902c8d6860a131c1253df4c38df4641d5922ftree | snapshot
parent90b163a4b5dd909f8a6f9c61c5bdce17130350a0commit | diff
[PATCH] dentry d_bucket fix

The gap between checking d_bucket and sampling d_move_count looks like a bug
to me.

It feels safer to be checking d_bucket after taking the lock, when we know
that it is stable.

And it's a little faster to check d_bucket after having checked the hash
rather than before.
fs/dcache.c diff | blob | history
Linux kernels and configurations used by Zygo.