driver core fixes against 2.6.25-rc3 git

Here are a few driver core fixes against your current git tree that fix some more problems that have cropped up:
- suspend/resume bugfixes
- ksets were not getting properly initialized
- CONFIG_SYSFS_DEPRECATED was catching people with old userspace versions of LVM now that the block core is changed to use struct device.
- other driver core bugs on error paths.

Please pull from:
master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6.git/

Patches will be sent as a follow-on to this message to lkml for people to see.

thanks,

greg k-h

------------

 drivers/base/core.c       |   14 +-----
 drivers/base/power/main.c |  108 +++++++--------------------------------------
 drivers/usb/core/usb.c    |    2 +-
 fs/debugfs/inode.c        |    4 --
 include/linux/debugfs.h   |    5 ++
 init/Kconfig              |   11 +++-
 lib/kobject.c             |   11 ++---
 7 files changed, 39 insertions(+), 116 deletions(-)

---------------

Alan Stern (1):
PM: fix new mutex-locking bug in the PM core

Cornelia Huck (1):
Driver core: Fix cleanup when failing device_add().

Greg Kroah-Hartman (2):
driver core: fix up Kconfig text for CONFIG_SYSFS_DEPRECATED
kobject: properly initialize ksets

Harvey Harrison (1):
debugfs: fix sparse warnings

Ingo Molnar (1):
sysfs: CONFIG_SYSFS_DEPRECATED fix

Rafael J. Wysocki (2):
PM: Do not acquire device semaphores upfront during suspend driver core: Remove dpm_sysfs_remove() from error path of device_add()

As things get moved into this config option, the hard date of 2006 does not work anymore, so update the text to be more descriptive.

---
 init/Kconfig |    7 ++++---
 1 files changed, 4 insertions(+), 3 deletions(-)

index f698a5a..e6606e6 100644
@@ -382,10 +382,11 @@ config SYSFS_DEPRECATED
 
 	  If enabled, this option will also move any device structures
 	  that belong to a class, back into the /sys/class hierarchy, in
-	  order to support older versions of udev.
+	  order to support older versions of udev and some userspace
+	  programs.
 
-	  If you are using a distro that was released in 2006 or later,
-	  it should be safe to say N here.
+	  If you are using a distro with the most recent userspace
+	  packages, it should be safe to say N here.
 
 config CGROUP_MEM_CONT
 	bool "Memory controller for cgroups"

1.5.4.3

udev did not break, and nash is a shell, so it would have a hard time "breaking" :)

The versions "required" did not change if you just enable that config option.

thanks,

greg k-h

The problem with nash was that it has some builtins for creating entries in /dev (basically, a subset of udev).

And you're right, this time was nash, the others were the *last* time we busticated /sys. ;)

From: Ingo Molnar <[email protected]>

CONFIG_SYSFS_DEPRECATED=y changed its meaning recently and causes regressions in working setups that had SYSFS_DEPRECATED disabled.

so rename it to SYSFS_DEPRECATED_V2 so that testers pick up the new default via 'make oldconfig', even if their old .config's disabled CONFIG_SYSFS_DEPRECATED ...

---
 init/Kconfig |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

index e6606e6..98ebf37 100644
@@ -367,9 +367,13 @@ config RESOURCE_COUNTERS
 	depends on CGROUPS
 
 config SYSFS_DEPRECATED
+	bool
+
+config SYSFS_DEPRECATED_V2
 	bool "Create deprecated sysfs files"
 	depends on SYSFS
 	default y
+	select SYSFS_DEPRECATED
 	help
 	  This option creates deprecated symlinks such as the
 	  "device"-link, the <subsystem>:<name>-link, and the

1.5.4.3

kset_initialize was calling kobject_init_internal() which didn't initialize the kobject as well as kobject_init() was. So have kobject_init() call kobject_init_internal() and move the logic to initalize the kobject there.

---
 lib/kobject.c |   11 +++++------
 1 files changed, 5 insertions(+), 6 deletions(-)

index d784dae..0d03252 100644
@@ -153,6 +153,10 @@ static void kobject_init_internal(struct kobject *kobj)
 		return;
 	kref_init(&kobj->kref);
 	INIT_LIST_HEAD(&kobj->entry);
+	kobj->state_in_sysfs = 0;
+	kobj->state_add_uevent_sent = 0;
+	kobj->state_remove_uevent_sent = 0;
+	kobj->state_initialized = 1;
 }
 
 

@@ -289,13 +293,8 @@ void kobject_init(struct kobject *kobj, struct kobj_type *ktype)
 		dump_stack();
 	}
 
-	kref_init(&kobj->kref);
-	INIT_LIST_HEAD(&kobj->entry);
+	kobject_init_internal(kobj);
 	kobj->ktype = ktype;
-	kobj->state_in_sysfs = 0;
-	kobj->state_add_uevent_sent = 0;
-	kobj->state_remove_uevent_sent = 0;
-	kobj->state_initialized = 1;
 	return;
 
 error:

1.5.4.3

From: Rafael J. Wysocki <[email protected]>

Remove the code that acquires all device semaphores from the suspend code path as it causes multiple problems to appear (most notably,
http://bugzilla.kernel.org/show_bug.cgi?id=10030) and revert the change introduced by commit 4145ed6dc597a9bea5f6ae8c574653b2de10620f depending on the code being removed.

Remove pm_sleep_lock()/pm_sleep_unlock() from device_add() to avoid the issue reported at http://bugzilla.kernel.org/show_bug.cgi?id=9874.

It should fix the regreesions reported at:

	http://bugzilla.kernel.org/show_bug.cgi?id=9874
	http://bugzilla.kernel.org/show_bug.cgi?id=10030

---
 drivers/base/core.c       |    8 ---
 drivers/base/power/main.c |  106 +++++++--------------------------------------
 drivers/usb/core/usb.c    |    2 +-
 3 files changed, 17 insertions(+), 99 deletions(-)

index 9c0070b..1e2e0fa 100644
@@ -770,13 +770,6 @@ int device_add(struct device *dev)
 	struct class_interface *class_intf;
 	int error;
 
-	error = pm_sleep_lock();
-	if (error) {
-		dev_warn(dev, "Suspicious %s during suspend\n", __FUNCTION__);
-		dump_stack();
-		return error;
-	}
-
 	dev = get_device(dev);
 	if (!dev || !strlen(dev->bus_id)) {
 		error = -EINVAL;

@@ -843,7 +836,6 @@ int device_add(struct device *dev)
 	}
  Done:
 	put_device(dev);
-	pm_sleep_unlock();
 	return error;
  BusError:
 	device_pm_remove(dev);

index f6f1990..1f0db51 100644
@@ -233,7 +233,7 @@ static int ksuspend_usb_init(void)
 	 * singlethreaded.  Its job doesn't justify running on more
 	 * than one CPU.
 	 */
-	ksuspend_usb_wq = create_singlethread_workqueue("ksuspend_usbd");
+	ksuspend_usb_wq = create_freezeable_workqueue("ksuspend_usbd");
 	if (!ksuspend_usb_wq)
 		return -ENOMEM;
 	return 0;

1.5.4.3

From: Alan Stern <[email protected]>

This patch (as1041) fixes a bug introduced by the acquire-all-device-semaphores reversion. The error pathway of dpm_suspend() fails to reacquire a mutex it should be holding.

---
 drivers/base/power/main.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

index b0c16f6..d887d5c 100644
@@ -428,6 +428,7 @@ static int dpm_suspend(pm_message_t state)
 
 		mutex_unlock(&dpm_list_mtx);
 		error = suspend_device(dev, state);
+		mutex_lock(&dpm_list_mtx);
 		if (error) {
 			printk(KERN_ERR "Could not suspend device %s: "
 					"error %d%s\n",

@@ -438,7 +439,6 @@ static int dpm_suspend(pm_message_t state)
 					""));
 			break;
 		}
-		mutex_lock(&dpm_list_mtx);
 		if (!list_empty(&dev->power.entry))
 			list_move(&dev->power.entry, &dpm_off);
 	}

1.5.4.3

From: Rafael J. Wysocki <[email protected]>

Since device_pm_remove(dev) calls dpm_sysfs_remove(dev), it's incorrect to call the latter after the former in the device_add() error path.

---
 drivers/base/core.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

index 1e2e0fa..7f59c2f 100644
@@ -839,7 +839,6 @@ int device_add(struct device *dev)
 	return error;
  BusError:
 	device_pm_remove(dev);
-	dpm_sysfs_remove(dev);
  PMError:
 	if (dev->bus)
 		blocking_notifier_call_chain(&dev->bus->p->bus_notifier,

1.5.4.3

From: Cornelia Huck <[email protected]>

Driver core: Fix cleanup when failing device_add().

- Don't call cleanup_device_parent() if we didn't call setup_parent().
- dev->kobj.parent may be NULL when cleanup_device_parent() is called, so we need to handle glue_dir == NULL in cleanup_glue_dir().

---
 drivers/base/core.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

index 7f59c2f..7de543d 100644
@@ -621,7 +621,8 @@ static struct kobject *get_device_parent(struct device *dev,
 static void cleanup_glue_dir(struct device *dev, struct kobject *glue_dir)
 {
 	/* see if we live in a "glue" directory */
-	if (!dev->class || glue_dir->kset != &dev->class->class_dirs)
+	if (!glue_dir || !dev->class ||
+	    glue_dir->kset != &dev->class->class_dirs)
 		return;
 
 	kobject_put(glue_dir);

@@ -773,7 +774,7 @@ int device_add(struct device *dev)
 	dev = get_device(dev);
 	if (!dev || !strlen(dev->bus_id)) {
 		error = -EINVAL;
-		goto Error;
+		goto Done;
 	}
 
 	pr_debug("device: '%s': %s\n", dev->bus_id, __FUNCTION__);

1.5.4.3

From: Harvey Harrison <[email protected]>

extern does not belong in C files, move declaration to linux/debugfs.h fs/debugfs/file.c:42:30: warning: symbol 'debugfs_file_operations' was not declared. Should it be static? fs/debugfs/file.c:54:31: warning: symbol 'debugfs_link_operations' was not declared. Should it be static?

---
 fs/debugfs/inode.c      |    4 ----
 include/linux/debugfs.h |    5 +++++
 2 files changed, 5 insertions(+), 4 deletions(-)

index d26e282..e9602d8 100644
@@ -29,10 +29,6 @@
 
 #define DEBUGFS_MAGIC	0x64626720
 
-/* declared over in file.c */
-extern struct file_operations debugfs_file_operations;
-extern struct inode_operations debugfs_link_operations;
-
 static struct vfsmount *debugfs_mount;
 static int debugfs_mount_count;
 

index f592d6d..7266124 100644
@@ -27,6 +27,11 @@ struct debugfs_blob_wrapper {
 };
 
 #if defined(CONFIG_DEBUG_FS)
+
+/* declared over in file.c */
+extern const struct file_operations debugfs_file_operations;
+extern const struct inode_operations debugfs_link_operations;
+
 struct dentry *debugfs_create_file(const char *name, mode_t mode,
 				   struct dentry *parent, void *data,
 				   const struct file_operations *fops);

1.5.4.3