From: Dipankar Sarma <dipankar@in.ibm.com>
To: John M Flinchbaugh <glynis@butterfly.hjsoft.com>
Cc: linux-kernel@vger.kernel.org,
Trond Myklebust <trond.myklebust@fys.uio.no>,
Linus Torvalds <torvalds@transmeta.com>,
Maneesh Soni <maneesh@in.ibm.com>
Subject: Re: 2.5.70-bk16: nfs crash
Date: Thu, 12 Jun 2003 21:03:55 +0530 [thread overview]
Message-ID: <20030612153355.GA1438@in.ibm.com> (raw)
In-Reply-To: <20030612135254.GA2482@in.ibm.com>
On Thu, Jun 12, 2003 at 07:22:54PM +0530, Dipankar Sarma wrote:
> I am not supprised at all by this, I can see two csets in Linus' tree
> that will definitely break dcache -
>
> 1. http://linux.bkbits.net:8080/linux-2.5/cset@1.1215.104.2?nav=index.html|ChangeSet@-2d
>
> __d_drop() *must not* initialize d_hash fields. Lockfree lookup depends on
> that. If __d_drop() needs to be allowed on an unhashed dentry, the right
> thing to do would be to check for DCACHE_UNHASHED before unhashing. I will
> submit a patch a little later to do this.
>
>
> 2. http://linux.bkbits.net:8080/linux-2.5/cset@1.1215.104.1?nav=index.html|ChangeSet@-2d
>
> hlist poison patch is broken. list_del_rcu() and hlist_del_rcu()
> *must not* re-initialize the pointers. Maneesh submitted a patch
> earlier today that corrects this -
>
> http://marc.theaimsgroup.com/?l=linux-kernel&m=105541206017154&w=2
John,
Can you try the patch below along with Maneesh's patch mentioned
above to see if they fix your problem ?
Thanks
Dipankar
Fix __d_drop() to be allowed on unhashed dentries correctly. Don't
re-initialize the pointers, instead check for DCACHE_UNHASHED
before really unhashing it.
include/linux/dcache.h | 6 ++++--
1 files changed, 4 insertions(+), 2 deletions(-)
diff -puN include/linux/dcache.h~dentry-hash-init-fix include/linux/dcache.h
--- linux-2.5.70-bk16/include/linux/dcache.h~dentry-hash-init-fix 2003-06-12 20:43:32.000000000 +0530
+++ linux-2.5.70-bk16-dipankar/include/linux/dcache.h 2003-06-12 20:47:10.000000000 +0530
@@ -174,8 +174,10 @@ extern spinlock_t dcache_lock;
static inline void __d_drop(struct dentry *dentry)
{
- dentry->d_vfs_flags |= DCACHE_UNHASHED;
- hlist_del_rcu_init(&dentry->d_hash);
+ if (!(dentry->d_vfs_flags & DCACHE_UNHASHED)) {
+ dentry->d_vfs_flags |= DCACHE_UNHASHED;
+ hlist_del_rcu(&dentry->d_hash);
+ }
}
static inline void d_drop(struct dentry *dentry)
_
next prev parent reply other threads:[~2003-06-12 15:17 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2003-06-12 12:56 2.5.70-bk16: nfs crash John M Flinchbaugh
2003-06-12 13:52 ` Dipankar Sarma
2003-06-12 15:33 ` Dipankar Sarma [this message]
2003-06-12 15:35 ` Trond Myklebust
2003-06-12 15:53 ` Dipankar Sarma
2003-06-12 16:26 ` Trond Myklebust
2003-06-12 16:49 ` Linus Torvalds
2003-06-12 16:55 ` Linus Torvalds
2003-06-12 19:53 ` Dipankar Sarma
2003-06-13 5:24 ` Trond Myklebust
2003-06-13 5:50 ` Dipankar Sarma
2003-06-13 6:13 ` Trond Myklebust
2003-06-13 6:54 ` Dipankar Sarma
2003-06-13 6:06 ` Dipankar Sarma
2003-06-12 16:30 ` viro
2003-06-12 16:55 ` Dipankar Sarma
2003-06-12 15:49 ` Linus Torvalds
2003-06-12 16:05 ` Dipankar Sarma
2003-06-12 16:18 ` Linus Torvalds
2003-06-12 16:35 ` Dipankar Sarma
2003-06-12 16:47 ` Linus Torvalds
2003-06-13 12:48 ` Maneesh Soni
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20030612153355.GA1438@in.ibm.com \
--to=dipankar@in.ibm.com \
--cc=glynis@butterfly.hjsoft.com \
--cc=linux-kernel@vger.kernel.org \
--cc=maneesh@in.ibm.com \
--cc=torvalds@transmeta.com \
--cc=trond.myklebust@fys.uio.no \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox