public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [patch 1/1] devtmpfs: Convert dirlock to a mutex
@ 2009-12-16 21:31 Thomas Gleixner
  2009-12-16 22:15 ` Greg KH
  2009-12-17 14:55 ` patch devtmpfs-convert-dirlock-to-a-mutex.patch added to gregkh-2.6 tree gregkh
  0 siblings, 2 replies; 3+ messages in thread
From: Thomas Gleixner @ 2009-12-16 21:31 UTC (permalink / raw)
  To: LKML; +Cc: Kay Sievers, Greg Kroah-Hartman

[-- Attachment #1: devfs-fix-locking-madness.patch --]
[-- Type: text/plain, Size: 1959 bytes --]

devtmpfs has a rw_lock dirlock which serializes delete_path and
create_path.

This code was obviously never tested with the usual set of debugging
facilities enabled. In the dirlock held sections the code calls:

 - vfs functions which take mutexes
 - kmalloc(, GFP_KERNEL)

In both code pathes the might sleep warning triggers and spams dmesg.

Convert the rw_lock to a mutex. There is no reason why this needs to
be a rwlock.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Greg Kroah-Hartman <gregkh@suse.de>
---
 drivers/base/devtmpfs.c |   12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

Index: linux-2.6/drivers/base/devtmpfs.c
===================================================================
--- linux-2.6.orig/drivers/base/devtmpfs.c
+++ linux-2.6/drivers/base/devtmpfs.c
@@ -32,7 +32,7 @@ static int dev_mount = 1;
 static int dev_mount;
 #endif
 
-static rwlock_t dirlock;
+static DEFINE_MUTEX(dirlock);
 
 static int __init mount_param(char *str)
 {
@@ -93,7 +93,7 @@ static int create_path(const char *nodep
 {
 	int err;
 
-	read_lock(&dirlock);
+	mutex_lock(&dirlock);
 	err = dev_mkdir(nodepath, 0755);
 	if (err == -ENOENT) {
 		char *path;
@@ -117,7 +117,7 @@ static int create_path(const char *nodep
 		}
 		kfree(path);
 	}
-	read_unlock(&dirlock);
+	mutex_unlock(&dirlock);
 	return err;
 }
 
@@ -229,7 +229,7 @@ static int delete_path(const char *nodep
 	if (!path)
 		return -ENOMEM;
 
-	write_lock(&dirlock);
+	mutex_lock(&dirlock);
 	for (;;) {
 		char *base;
 
@@ -241,7 +241,7 @@ static int delete_path(const char *nodep
 		if (err)
 			break;
 	}
-	write_unlock(&dirlock);
+	mutex_unlock(&dirlock);
 
 	kfree(path);
 	return err;
@@ -352,8 +352,6 @@ int __init devtmpfs_init(void)
 	int err;
 	struct vfsmount *mnt;
 
-	rwlock_init(&dirlock);
-
 	err = register_filesystem(&dev_fs_type);
 	if (err) {
 		printk(KERN_ERR "devtmpfs: unable to register devtmpfs "



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [patch 1/1] devtmpfs: Convert dirlock to a mutex
  2009-12-16 21:31 [patch 1/1] devtmpfs: Convert dirlock to a mutex Thomas Gleixner
@ 2009-12-16 22:15 ` Greg KH
  2009-12-17 14:55 ` patch devtmpfs-convert-dirlock-to-a-mutex.patch added to gregkh-2.6 tree gregkh
  1 sibling, 0 replies; 3+ messages in thread
From: Greg KH @ 2009-12-16 22:15 UTC (permalink / raw)
  To: Thomas Gleixner; +Cc: LKML, Kay Sievers

On Wed, Dec 16, 2009 at 09:31:33PM -0000, Thomas Gleixner wrote:
> devtmpfs has a rw_lock dirlock which serializes delete_path and
> create_path.
> 
> This code was obviously never tested with the usual set of debugging
> facilities enabled. In the dirlock held sections the code calls:
> 
>  - vfs functions which take mutexes
>  - kmalloc(, GFP_KERNEL)
> 
> In both code pathes the might sleep warning triggers and spams dmesg.
> 
> Convert the rw_lock to a mutex. There is no reason why this needs to
> be a rwlock.

Many thanks, this is correct.  I'll go queue it up.

greg k-h

^ permalink raw reply	[flat|nested] 3+ messages in thread

* patch devtmpfs-convert-dirlock-to-a-mutex.patch added to gregkh-2.6 tree
  2009-12-16 21:31 [patch 1/1] devtmpfs: Convert dirlock to a mutex Thomas Gleixner
  2009-12-16 22:15 ` Greg KH
@ 2009-12-17 14:55 ` gregkh
  1 sibling, 0 replies; 3+ messages in thread
From: gregkh @ 2009-12-17 14:55 UTC (permalink / raw)
  To: tglx, gregkh, kay.sievers, linux-kernel, stable


This is a note to let you know that I've just added the patch titled

    Subject: devtmpfs: Convert dirlock to a mutex

to my gregkh-2.6 tree.  Its filename is

    devtmpfs-convert-dirlock-to-a-mutex.patch

This tree can be found at 
    http://www.kernel.org/pub/linux/kernel/people/gregkh/gregkh-2.6/patches/


>From tglx@linutronix.de  Thu Dec 17 06:44:07 2009
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 16 Dec 2009 21:31:33 -0000
Subject: devtmpfs: Convert dirlock to a mutex
To: LKML <linux-kernel@vger.kernel.org>
Cc: Kay Sievers <kay.sievers@vrfy.org>, Greg Kroah-Hartman <gregkh@suse.de>
Message-ID: <20091216212730.308081108@linutronix.de>


devtmpfs has a rw_lock dirlock which serializes delete_path and
create_path.

This code was obviously never tested with the usual set of debugging
facilities enabled. In the dirlock held sections the code calls:

 - vfs functions which take mutexes
 - kmalloc(, GFP_KERNEL)

In both code pathes the might sleep warning triggers and spams dmesg.

Convert the rw_lock to a mutex. There is no reason why this needs to
be a rwlock.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

---
 drivers/base/devtmpfs.c |   12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

--- a/drivers/base/devtmpfs.c
+++ b/drivers/base/devtmpfs.c
@@ -32,7 +32,7 @@ static int dev_mount = 1;
 static int dev_mount;
 #endif
 
-static rwlock_t dirlock;
+static DEFINE_MUTEX(dirlock);
 
 static int __init mount_param(char *str)
 {
@@ -93,7 +93,7 @@ static int create_path(const char *nodep
 {
 	int err;
 
-	read_lock(&dirlock);
+	mutex_lock(&dirlock);
 	err = dev_mkdir(nodepath, 0755);
 	if (err == -ENOENT) {
 		char *path;
@@ -117,7 +117,7 @@ static int create_path(const char *nodep
 		}
 		kfree(path);
 	}
-	read_unlock(&dirlock);
+	mutex_unlock(&dirlock);
 	return err;
 }
 
@@ -229,7 +229,7 @@ static int delete_path(const char *nodep
 	if (!path)
 		return -ENOMEM;
 
-	write_lock(&dirlock);
+	mutex_lock(&dirlock);
 	for (;;) {
 		char *base;
 
@@ -241,7 +241,7 @@ static int delete_path(const char *nodep
 		if (err)
 			break;
 	}
-	write_unlock(&dirlock);
+	mutex_unlock(&dirlock);
 
 	kfree(path);
 	return err;
@@ -352,8 +352,6 @@ int __init devtmpfs_init(void)
 	int err;
 	struct vfsmount *mnt;
 
-	rwlock_init(&dirlock);
-
 	err = register_filesystem(&dev_fs_type);
 	if (err) {
 		printk(KERN_ERR "devtmpfs: unable to register devtmpfs "


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2009-12-17 14:55 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-12-16 21:31 [patch 1/1] devtmpfs: Convert dirlock to a mutex Thomas Gleixner
2009-12-16 22:15 ` Greg KH
2009-12-17 14:55 ` patch devtmpfs-convert-dirlock-to-a-mutex.patch added to gregkh-2.6 tree gregkh

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox