Re: [PATCH 4/8] autofs4-2.6 - to support autofs 4.1.x
From: Mike Waychison
Date: Wed Jan 28 2004 - 20:29:04 EST
Ian Kent wrote:
On Wed, 28 Jan 2004, Mike Waychison wrote:
raven@xxxxxxxxxx wrote:
Patch:
4-autofs4-2.6.0-test9-waitq2.patch
Adds a spin lock to serialize access to wait queue in the super block info
struct.
A while back I wrote up a patch for autofs3 that serialized waitq access
and changed the waitq counters to atomic_t. I never sent it out because
I had realized that the changes I made weren't needed as all waitq
code-paths were running under the BKL (the big ones were ->lookup and
the ioctls).
My understanding is that this code can be reached at least via lookup,
readdir and revalidate. I thought that in 2.6 none of these held the BKL
on entry (I'll have to check). Certainly this is the case for revalidate.
I just took a look at the autofs4 code and the BKL is explicitly held in
the lookup code. It is also implicitly held in the ioctl code.
My understanding is that when the dcache (and subsequently
->d_revalidate) was introduced into the Linux kernel, autofs3 was
modified to do revalidates, and would have had it's lock_kernel calls
added then. If autofs4 had forked off before the dcache addition, it
wouldn't have picked the lock_kernel in d_revalidate.
Peter/Jeremy, could you confirm this?
--
Mike Waychison
Sun Microsystems, Inc.
1 (650) 352-5299 voice
1 (416) 202-8336 voice
mailto: Michael.Waychison@xxxxxxx
http://www.sun.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
NOTICE: The opinions expressed in this email are held by me,
and may not represent the views of Sun Microsystems, Inc.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/