[Cluster-devel] [GFS2/DLM] Bug fixes [0/12]

[Cluster-devel] [GFS2/DLM] Bug fixes [0/12]

[Steven Whitehouse]

Hi,

Following this message are 12 bugs fixes and minor clean ups from the
GFS2 -fixes git tree. They are all pretty small, most are just a few
lines long.

I've only just pushed the patches, so it may take a little while for
kernel.org's mirrors to pick them up. They are all in the -nmw tree as
well, of course,

Steve.

[Cluster-devel] [GFS2] fix locking mistake [1/12]

[Steven Whitehouse]

From 2e95b6653bb69c893e6ee1b42b537939c1ea2b9c Mon Sep 17 00:00:00 2001
From: Josef Whiter <jwhiter@redhat.com>
Date: Tue, 20 Feb 2007 00:03:29 -0500
Subject: [PATCH] [GFS2] fix locking mistake

This patch fixes a locking mistake in the quota code, we do a mutex_lock instead
of a mutex_unlock.

Signed-off-by: Josef Whiter <jwhiter@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index d0db881..c186857 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -279,7 +279,7 @@ static int bh_get(struct gfs2_quota_data *qd)
 		(bh->b_data + sizeof(struct gfs2_meta_header) +
 		 offset * sizeof(struct gfs2_quota_change));
 
-	mutex_lock(&sdp->sd_quota_mutex);
+	mutex_unlock(&sdp->sd_quota_mutex);
 
 	return 0;
 
-- 
1.4.4.2

[Cluster-devel] [GFS2] add newline to printk message [2/12]

[Steven Whitehouse]

From d5a6751b32c79680da90eaa76919ffe6e5b8a94f Mon Sep 17 00:00:00 2001
From: Richard Fearn <richardfearn@gmail.com>
Date: Sat, 17 Feb 2007 17:29:15 +0000
Subject: [PATCH] [GFS2] add newline to printk message

Patch for the 2.6.20 stable tree that adds a missing newline to one of
the printk messages in fs/gfs2/ops_fstype.c.

Signed-off-by: Richard Fearn <richardfearn@gmail.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index ee80b8a..ee54cb6 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -840,7 +840,7 @@ static struct super_block* get_gfs2_sb(const char *dev_name)
 	}
 
 	printk(KERN_WARNING "GFS2: Unrecognized block device or "
-	       "mount point %s", dev_name);
+	       "mount point %s\n", dev_name);
 
 free_nd:
 	path_release(&nd);
-- 
1.4.4.2

[Cluster-devel] [GFS2] NFS filehandle check [3/12]

[Steven Whitehouse]

From a7d2b2bdc9a0b55d5b08e15756c7e65c48c4bca5 Mon Sep 17 00:00:00 2001
From: Wendy Cheng <wcheng@redhat.com>
Date: Fri, 23 Feb 2007 00:21:17 -0500
Subject: [PATCH] [GFS2] NFS filehandle check

File handle checking error found in '07 NFS connectathon. The fh_type
and fh_len are not necessarily identical. Some of the client machines
could fail mount with stale filehandle without this patch.

Signed-off-by: S. Wendy Cheng <wcheng@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/ops_export.c b/fs/gfs2/ops_export.c
index 1de05b6..aad9183 100644
--- a/fs/gfs2/ops_export.c
+++ b/fs/gfs2/ops_export.c
@@ -38,14 +38,11 @@ static struct dentry *gfs2_decode_fh(struct super_block *sb,
 	struct gfs2_fh_obj fh_obj;
 	struct gfs2_inum_host *this, parent;
 
-	if (fh_type != fh_len)
-		return NULL;
-
 	this 		= &fh_obj.this;
 	fh_obj.imode 	= DT_UNKNOWN;
 	memset(&parent, 0, sizeof(struct gfs2_inum));
 
-	switch (fh_type) {
+	switch (fh_len) {
 	case GFS2_LARGE_FH_SIZE:
 		parent.no_formal_ino = ((u64)be32_to_cpu(fh[4])) << 32;
 		parent.no_formal_ino |= be32_to_cpu(fh[5]);
-- 
1.4.4.2

[Cluster-devel] [GFS2] fix hangup when multiple processes are trying to write to the same file [4/12]

[Steven Whitehouse]

From a13cbe375303585fec1425135ed54adb62be41fc Mon Sep 17 00:00:00 2001
From: Josef Whiter <jwhiter@redhat.com>
Date: Fri, 23 Feb 2007 12:49:51 -0500
Subject: [PATCH] [GFS2] fix hangup when multiple processes are trying to write to the same file

This fixes a problem I encountered while running bonnie++.  When you have one
thread that opens a file and starts to write to it, and then another thread that
tries to open and write to the same file, the second thread will loop forever
trying to grab the inode lock for that inode.  Basically we come in through
generic_buffered_file_write, which calls gfs2_prepare_write, which then attempts
to grab the glock.  Because we don't own the lock, gfs2_prepare_write gets
GLR_TRYFAILED, which returns AOP_TRUNCATED_PAGE to generic_buffered_file_write.
At this point generic_buffered_file_write loops around again and immediately
retries the prepare_write.  This means that the second process never gets off of
the processor in order to allow the process that holds the lock to finish its
work and let go of the lock.  This patch makes gfs2_glock_nq schedule() if it
gets back a GLR_TRYFAILED, which resolves this problem.

Signed-off-by: Josef Whiter <jwhiter@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c
index 56e3359..b3b7e84 100644
--- a/fs/gfs2/ops_address.c
+++ b/fs/gfs2/ops_address.c
@@ -266,9 +266,11 @@ skip_lock:
 out:
 	return error;
 out_unlock:
-	if (error == GLR_TRYFAILED)
-		error = AOP_TRUNCATED_PAGE;
 	unlock_page(page);
+	if (error == GLR_TRYFAILED) {
+		error = AOP_TRUNCATED_PAGE;
+		yield();
+	}
 	if (do_unlock)
 		gfs2_holder_uninit(&gh);
 	goto out;
@@ -364,6 +366,7 @@ static int gfs2_prepare_write(struct file *file, struct page *page,
 		if (error == GLR_TRYFAILED) {
 			unlock_page(page);
 			error = AOP_TRUNCATED_PAGE;
+			yield();
 		}
 		goto out_uninit;
 	}
-- 
1.4.4.2

[Cluster-devel] [DLM] fs/dlm/user.c should #include "user.h" [5/12]

[Steven Whitehouse]

From 84c6e8cd359adc34d21e40efcafe09297510b3c8 Mon Sep 17 00:00:00 2001
From: Adrian Bunk <bunk@stusta.de>
Date: Mon, 26 Feb 2007 00:18:42 +0100
Subject: [PATCH] [DLM] fs/dlm/user.c should #include "user.h"

Every file should include the headers containing the prototypes for
it's global functions.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/dlm/user.c b/fs/dlm/user.c
index 40db61d..3870150 100644
--- a/fs/dlm/user.c
+++ b/fs/dlm/user.c
@@ -22,6 +22,7 @@
 #include "lockspace.h"
 #include "lock.h"
 #include "lvb_table.h"
+#include "user.h"
 
 static const char *name_prefix="dlm";
 static struct miscdevice ctl_device;
-- 
1.4.4.2

[Cluster-devel] [GFS2] pass formal ino in do_filldir_main [6/12]

[Steven Whitehouse]

From fb0d3bce8e88cca4abb26076f778f64edcaf19aa Mon Sep 17 00:00:00 2001
From: Wendy Cheng <wcheng@redhat.com>
Date: Wed, 28 Feb 2007 11:24:25 -0500
Subject: [PATCH] [GFS2] pass formal ino in do_filldir_main

ok, the following is the minimum changes to get NFSD going before we
settle down this issue .. would appreciate this in the tree so other NFS
related works can get done in parallel.

Signed-off-by: S. Wendy Cheng <wcheng@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 0d6831a..902dd81 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -61,13 +61,13 @@ static int iget_set(struct inode *inode, void *opaque)
 
 struct inode *gfs2_ilookup(struct super_block *sb, struct gfs2_inum_host *inum)
 {
-	return ilookup5(sb, (unsigned long)inum->no_formal_ino,
+	return ilookup5(sb, (unsigned long)inum->no_addr,
 			iget_test, inum);
 }
 
 static struct inode *gfs2_iget(struct super_block *sb, struct gfs2_inum_host *inum)
 {
-	return iget5_locked(sb, (unsigned long)inum->no_formal_ino,
+	return iget5_locked(sb, (unsigned long)inum->no_addr,
 		     iget_test, iget_set, inum);
 }
 
-- 
1.4.4.2

[Cluster-devel] [GFS2] Fix bz 230143, incorrect flushing of rgrps [7/12]

[Steven Whitehouse]

From cad5b9392754910ee7dbe551eb004010a864c882 Mon Sep 17 00:00:00 2001
From: Steven Whitehouse <swhiteho@redhat.com>
Date: Wed, 28 Feb 2007 14:03:00 +0000
Subject: [PATCH] [GFS2] Fix bz 230143, incorrect flushing of rgrps

The below patch fixes a problem where we were not flushing rgrps
correctly. It only occurred in the specific case that a callback was
received for an rgrp which was dirty and when a journal log flush had
not already resulted in the rgrp being flushed anyway. This fixes Red
Hat bz 230143,

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index 46af553..4221110 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -459,6 +459,8 @@ const struct gfs2_glock_operations gfs2_inode_glops = {
 };
 
 const struct gfs2_glock_operations gfs2_rgrp_glops = {
+	.go_xmote_th = meta_go_sync,
+	.go_drop_th = meta_go_sync,
 	.go_inval = meta_go_inval,
 	.go_demote_ok = rgrp_go_demote_ok,
 	.go_lock = rgrp_go_lock,
-- 
1.4.4.2

[Cluster-devel] [GFS2] Fix bz 229831, lookup returns wrong inode [8/12]

[Steven Whitehouse]

From 1be3867955731b5cb2dc14060cc46f0882e87873 Mon Sep 17 00:00:00 2001
From: Steven Whitehouse <swhiteho@redhat.com>
Date: Thu, 1 Mar 2007 10:00:53 +0000
Subject: [PATCH] [GFS2] Fix bz 229831, lookup returns wrong inode

The following patch fixes Red Hat bz 229831. Without this patch its
possible for the wrong inode to be returned in certain cases. It is a
pretty unusual event, so that its taken some time to track down. Thanks
and due to Josef Whiter who did a lot of the testing required to thrack
this down and fix it.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 902dd81..df0b8b3 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -43,7 +43,8 @@ static int iget_test(struct inode *inode, void *opaque)
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_inum_host *inum = opaque;
 
-	if (ip->i_num.no_addr == inum->no_addr)
+	if (ip->i_num.no_addr == inum->no_addr &&
+	    inode->i_private != NULL)
 		return 1;
 
 	return 0;
-- 
1.4.4.2

[Cluster-devel] [GFS2] Remove unused variable [9/12]

[Steven Whitehouse]

From 04b159b132c0d8e92dae8c72f134fd5b13b43deb Mon Sep 17 00:00:00 2001
From: Steven Whitehouse <swhiteho@redhat.com>
Date: Thu, 1 Mar 2007 11:14:44 +0000
Subject: [PATCH] [GFS2] Remove unused variable

Remove an unused variable.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index 12c80fd..0b2b545 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -416,7 +416,6 @@ struct gfs2_tune {
 	unsigned int gt_stall_secs; /* Detects trouble! */
 	unsigned int gt_complain_secs;
 	unsigned int gt_reclaim_limit; /* Max num of glocks in reclaim list */
-	unsigned int gt_entries_per_readdir;
 	unsigned int gt_statfs_quantum;
 	unsigned int gt_statfs_slow;
 };
diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index 70f424f..4fdda97 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -76,7 +76,6 @@ void gfs2_tune_init(struct gfs2_tune *gt)
 	gt->gt_stall_secs = 600;
 	gt->gt_complain_secs = 10;
 	gt->gt_reclaim_limit = 5000;
-	gt->gt_entries_per_readdir = 32;
 	gt->gt_statfs_quantum = 30;
 	gt->gt_statfs_slow = 0;
 }
-- 
1.4.4.2

[Cluster-devel] [GFS2] go_drop_bh is never used, so remove it [10/12]

[Steven Whitehouse]

From 631c42e170564108423fa4073531db159f2523ea Mon Sep 17 00:00:00 2001
From: Steven Whitehouse <swhiteho@redhat.com>
Date: Thu, 1 Mar 2007 10:36:32 +0000
Subject: [PATCH] [GFS2] go_drop_bh is never used, so remove it

The ->go_drop_bh function is never used, so this removes it and the single
caller,

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 6618c11..8597fde 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -953,9 +953,6 @@ static void drop_bh(struct gfs2_glock *gl, unsigned int ret)
 		spin_unlock(&gl->gl_spin);
 	}
 
-	if (glops->go_drop_bh)
-		glops->go_drop_bh(gl);
-
 	spin_lock(&gl->gl_spin);
 	gl->gl_req_gh = NULL;
 	gl->gl_req_bh = NULL;
diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index 0b2b545..49f0dbf 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -104,7 +104,6 @@ struct gfs2_glock_operations {
 	void (*go_xmote_th) (struct gfs2_glock *gl);
 	void (*go_xmote_bh) (struct gfs2_glock *gl);
 	void (*go_drop_th) (struct gfs2_glock *gl);
-	void (*go_drop_bh) (struct gfs2_glock *gl);
 	void (*go_inval) (struct gfs2_glock *gl, int flags);
 	int (*go_demote_ok) (struct gfs2_glock *gl);
 	int (*go_lock) (struct gfs2_holder *gh);
-- 
1.4.4.2

[Cluster-devel] [GFS2] build fix [11/12]

[Steven Whitehouse]

From 95d97b7dd7d7a7a13d11a38b3ecb64849d2e5086 Mon Sep 17 00:00:00 2001
From: akpm@linux-foundation.org <akpm@linux-foundation.org>
Date: Mon, 5 Mar 2007 23:10:39 -0800
Subject: [PATCH] [GFS2] build fix

fs/gfs2/glock.c:2198: error: 'THIS_MODULE' undeclared here (not in a function)

Cc: Steven Whitehouse <swhiteho@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 8597fde..12accb0 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -20,6 +20,7 @@
 #include <linux/list.h>
 #include <linux/lm_interface.h>
 #include <linux/wait.h>
+#include <linux/module.h>
 #include <linux/rwsem.h>
 #include <asm/uaccess.h>
 
-- 
1.4.4.2

[Cluster-devel] [GFS2] Fix bz 229873, alternate test: assertion "!ip->i_inode.i_mapping->nrpages" failed [12/12]

[Steven Whitehouse]

From c3f49bc209b28d2b5f82b78baaa827eb3a4d1891 Mon Sep 17 00:00:00 2001
From: Steven Whitehouse <swhiteho@redhat.com>
Date: Wed, 7 Mar 2007 09:06:48 +0000
Subject: [PATCH] [GFS2] Fix bz 229873, alternate test: assertion "!ip->i_inode.i_mapping->nrpages" failed

The following removes an incorrect assertion from the GFS2 glops code. This
fixes Red Hat bz 229873. Thanks to Abhijith Das for testing the patch
and confirming the fix.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Cc: Abhijith Das <adas@redhat.com>

diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index 4221110..39c8ae2 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -245,7 +245,6 @@ static void inode_go_inval(struct gfs2_glock *gl, int flags)
 
 	if (ip && S_ISREG(ip->i_inode.i_mode)) {
 		truncate_inode_pages(ip->i_inode.i_mapping, 0);
-		gfs2_assert_withdraw(GFS2_SB(&ip->i_inode), !ip->i_inode.i_mapping->nrpages);
 		clear_bit(GIF_PAGED, &ip->i_flags);
 	}
 }
-- 
1.4.4.2

[Cluster-devel] [GFS2/DLM] Pull request

[Steven Whitehouse]

Hi,

Please consider pulling the following GFS2 & DLM bug fixes and trivial clean ups. 
They are all relatively small in size,

Steve.

The following changes since commit 08e15e81a40e3241ce93b4a43886f3abda184aa6:
  Linus Torvalds (1):
        Linux 2.6.21-rc3

are found in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-2.6-fixes.git

Adrian Bunk (1):
      [DLM] fs/dlm/user.c should #include "user.h"

akpm at linux-foundation.org (1):
      [GFS2] build fix

Josef Whiter (2):
      [GFS2] fix locking mistake
      [GFS2] fix hangup when multiple processes are trying to write to the same file

Richard Fearn (1):
      [GFS2] add newline to printk message

Steven Whitehouse (5):
      [GFS2] Fix bz 230143, incorrect flushing of rgrps
      [GFS2] Fix bz 229831, lookup returns wrong inode
      [GFS2] Remove unused variable
      [GFS2] go_drop_bh is never used, so remove it
      [GFS2] Fix bz 229873, alternate test: assertion "!ip->i_inode.i_mapping->nrpages" failed

Wendy Cheng (2):
      [GFS2] NFS filehandle check
      [GFS2] pass formal ino in do_filldir_main

 fs/dlm/user.c         |    1 +
 fs/gfs2/glock.c       |    4 +---
 fs/gfs2/glops.c       |    3 ++-
 fs/gfs2/incore.h      |    2 --
 fs/gfs2/inode.c       |    7 ++++---
 fs/gfs2/ops_address.c |    7 +++++--
 fs/gfs2/ops_export.c  |    5 +----
 fs/gfs2/ops_fstype.c  |    2 +-
 fs/gfs2/quota.c       |    2 +-
 fs/gfs2/super.c       |    1 -
 10 files changed, 16 insertions(+), 18 deletions(-)