linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "J. Bruce Fields" <bfields@fieldses.org>
To: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: linux-nfs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	jlayton@redhat.com
Subject: Re: linux-next NFSD: NULL pointer dereference at nfsd_svc()
Date: Fri, 6 Aug 2010 18:10:00 -0400	[thread overview]
Message-ID: <20100806221000.GF29536@fieldses.org> (raw)
In-Reply-To: <20100806220537.GE29536@fieldses.org>

On Fri, Aug 06, 2010 at 06:05:37PM -0400, J. Bruce Fields wrote:
> On Fri, Aug 06, 2010 at 05:27:28PM -0400, J. Bruce Fields wrote:
> > Bah, so what you were hitting was simple--I just moved the
> > nfsd_reset_versions() call to the wrong place; the below should fix it.
> > 
> > There's also a couple other bugs in the area.

And also there was one more problem with my original "nfsd: fix
startup/shutdown order bug": it doesn't work to use sv_nrthreads
changing from zero to nonzero as the signal for when to do all this
startup, because write_pool_threads() adjusts the number of threads
without calling nfsd_svc().  (Maybe that should be fixed.)

For now, just use the nfsd_up variable to keep track of this (which is a
little closer to Jeff's original solution).

This is a replacement.

--b.

commit 4cd7eb015e92f7cefb43eaab3e111d1b3c7b3cbf
Author: J. Bruce Fields <bfields@redhat.com>
Date:   Mon Aug 2 14:12:44 2010 -0400

    nfsd: fix startup/shutdown order bug
    
    We must create the server before we can call init_socks or check the
    number of threads.
    
    Symptoms were a NULL pointer dereference in nfsd_svc().  Problem
    identified by Jeff Layton.
    
    Also fix a minor cleanup-on-error case in nfsd_startup().
    
    Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
    Signed-off-by: J. Bruce Fields <bfields@redhat.com>

diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c
index 92173bd..2a20f89 100644
--- a/fs/nfsd/nfssvc.c
+++ b/fs/nfsd/nfssvc.c
@@ -204,6 +204,9 @@ static bool nfsd_up = false;
 static int nfsd_startup(unsigned short port, int nrservs)
 {
 	int ret;
+
+	if (nfsd_up)
+		return 0;
 	/*
 	 * Readahead param cache - will no-op if it already exists.
 	 * (Note therefore results will be suboptimal if number of
@@ -217,7 +220,7 @@ static int nfsd_startup(unsigned short port, int nrservs)
 		goto out_racache;
 	ret = lockd_up();
 	if (ret)
-		return ret;
+		goto out_racache;
 	ret = nfs4_state_start();
 	if (ret)
 		goto out_lockd;
@@ -420,7 +423,7 @@ int
 nfsd_svc(unsigned short port, int nrservs)
 {
 	int	error;
-	bool	first_thread;
+	bool	nfsd_up_before;
 
 	mutex_lock(&nfsd_mutex);
 	dprintk("nfsd: creating service\n");
@@ -432,29 +435,29 @@ nfsd_svc(unsigned short port, int nrservs)
 	if (nrservs == 0 && nfsd_serv == NULL)
 		goto out;
 
-	first_thread = (nfsd_serv->sv_nrthreads == 0) && (nrservs != 0);
-
-	if (first_thread) {
-		error = nfsd_startup(port, nrservs);
-		if (error)
-			goto out;
-	}
 	error = nfsd_create_serv();
 	if (error)
-		goto out_shutdown;
-	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
+		goto out;
+
+	nfsd_up_before = nfsd_up;
+
+	error = nfsd_startup(port, nrservs);
 	if (error)
 		goto out_destroy;
+	}
+	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
+	if (error)
+		goto out_shutdown;
 	/* We are holding a reference to nfsd_serv which
 	 * we don't want to count in the return value,
 	 * so subtract 1
 	 */
 	error = nfsd_serv->sv_nrthreads - 1;
-out_destroy:
-	svc_destroy(nfsd_serv);		/* Release server */
 out_shutdown:
-	if (error < 0 && first_thread)
+	if (error < 0 && !nfsd_up_before)
 		nfsd_shutdown();
+out_destroy:
+	svc_destroy(nfsd_serv);		/* Release server */
 out:
 	mutex_unlock(&nfsd_mutex);
 	return error;

  reply	other threads:[~2010-08-06 22:11 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <201008020747.o727lqXs064715@www262.sakura.ne.jp>
2010-08-02 14:32 ` linux-next NFSD: NULL pointer dereference at nfsd_svc() Jeff Layton
2010-08-02 14:36   ` Jeff Layton
2010-08-02 18:16     ` J. Bruce Fields
2010-08-02 18:53       ` Jeff Layton
2010-08-03  1:09       ` Tetsuo Handa
2010-08-03 15:48         ` J. Bruce Fields
2010-08-03 16:24           ` J. Bruce Fields
2010-08-04  0:13           ` Tetsuo Handa
     [not found]             ` <201008040013.o740DmYK024832-etx+eQDEXHD7nzcFbJAaVXf5DAMn2ifp@public.gmane.org>
2010-08-04 19:40               ` J. Bruce Fields
2010-08-05  1:10                 ` Tetsuo Handa
2010-08-05 20:46                   ` J. Bruce Fields
2010-08-05 21:31                     ` J. Bruce Fields
2010-08-06  1:37                       ` Tetsuo Handa
2010-08-06 21:27                       ` J. Bruce Fields
2010-08-06 22:05                         ` J. Bruce Fields
2010-08-06 22:10                           ` J. Bruce Fields [this message]
2010-08-07  1:48                             ` Tetsuo Handa
2010-08-07  2:33                               ` J. Bruce Fields

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=20100806221000.GF29536@fieldses.org \
    --to=bfields@fieldses.org \
    --cc=jlayton@redhat.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=penguin-kernel@I-love.SAKURA.ne.jp \
    /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;
as well as URLs for NNTP newsgroup(s).