Skip to contentSkip to conversation

Driver core patches for 2.6.20

PATCH New Greg KH 3 participants 31 messages Feb 8, 2007

Subsystems (16)
lore
Share & tools
lore

Greg KH

author
#0 lore raw

PATCH 1/1 Driver core patches for 2.6.20

Here are a bunch of Driver core, sysfs, and driver model patches for 2.6.20-git

They include the conversion of the network drivers to use 'struct device' instead of 'struct class' (the patch has been in -mm for over 6 months and was acked by David Miller) some module sysfs additions to make userspace tools simpler, and lots of cleanups and fixes. See the full log below for details.

All of these have been in the -mm tree for quite a while.

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

 Documentation/HOWTO                       |    1 +
 drivers/base/class.c                      |   21 ++-
 drivers/base/core.c                       |  203 +++++++++++++-------
 drivers/base/dd.c                         |   21 ++-
 drivers/base/firmware_class.c             |    2 +-
 drivers/base/platform.c                   |   11 +-
 drivers/ide/setup-pci.c                   |    7 +-
 drivers/infiniband/ulp/ipoib/ipoib_main.c |   33 ++--
 drivers/infiniband/ulp/ipoib/ipoib_vlan.c |   11 +-
 drivers/input/serio/serio.c               |    6 +-
 drivers/net/arm/at91_ether.c              |    2 +-
 drivers/net/arm/etherh.c                  |    2 +-
 drivers/net/bonding/bond_sysfs.c          |  287 +++++++++++++++++-----------
 drivers/net/iseries_veth.c                |    2 +-
 drivers/net/macb.c                        |   36 ++--
 drivers/net/smc911x.c                     |    2 +-
 drivers/net/smc91x.c                      |    2 +-
 drivers/net/wireless/hostap/hostap_main.c |    2 +-
 drivers/net/wireless/orinoco.c            |    4 +-
 drivers/net/wireless/orinoco_cs.c         |    2 +-
 drivers/net/wireless/spectrum_cs.c        |    2 +-
 drivers/pci/pci-driver.c                  |    4 +-
 drivers/pcmcia/cs.c                       |   34 ++--
 drivers/pcmcia/cs_internal.h              |    4 +-
 drivers/pcmcia/ds.c                       |   14 +-
 drivers/pcmcia/i82092.c                   |    2 +-
 drivers/pcmcia/i82365.c                   |    2 +-
 drivers/pcmcia/pcmcia_ioctl.c             |    1 -
 drivers/pcmcia/pcmcia_resource.c          |    1 -
 drivers/pcmcia/pd6729.c                   |    2 +-
 drivers/pcmcia/rsrc_nonstatic.c           |   56 +++---
 drivers/pcmcia/soc_common.c               |    6 +-
 drivers/pcmcia/socket_sysfs.c             |  104 ++++++-----
 drivers/pcmcia/tcic.c                     |    2 +-
 drivers/pcmcia/yenta_socket.c             |    2 +-
 drivers/spi/pxa2xx_spi.c                  |    2 +-
 drivers/spi/spi.c                         |   32 ++--
 drivers/spi/spi_bitbang.c                 |    6 +-
 drivers/spi/spi_butterfly.c               |    4 +-
 drivers/usb/core/driver.c                 |    4 +-
 drivers/usb/input/hid-lgff.c              |    4 +-
 fs/sysfs/bin.c                            |    6 +-
 fs/sysfs/dir.c                            |  214 ++++++++++++++++++----
 fs/sysfs/file.c                           |   82 ++++++---
 fs/sysfs/group.c                          |    2 +
 fs/sysfs/inode.c                          |   36 ++++-
 fs/sysfs/mount.c                          |   11 +-
 fs/sysfs/symlink.c                        |    1 +
 fs/sysfs/sysfs.h                          |   21 ++
 include/linux/device.h                    |   12 +-
 include/linux/ide.h                       |    4 +-
 include/linux/kobject.h                   |    4 +
 include/linux/module.h                    |    3 +-
 include/linux/netdevice.h                 |    5 +-
 include/linux/pci.h                       |    5 +-
 include/linux/serio.h                     |    5 +
 include/linux/spi/spi.h                   |   10 +-
 include/linux/sysfs.h                     |   24 ++-
 include/linux/usb.h                       |    5 +-
 include/pcmcia/ss.h                       |    2 +-
 kernel/module.c                           |   76 ++++++--
 kernel/params.c                           |   28 ++-
 lib/kobject.c                             |   73 ++++++--
 net/bridge/br_if.c                        |    2 +-
 net/bridge/br_sysfs_br.c                  |  234 +++++++++++++-----------
 net/bridge/br_sysfs_if.c                  |    2 +-
 net/core/dev.c                            |    6 +-
 net/core/net-sysfs.c                      |  175 ++++++++++--------
 net/core/skbuff.c                         |    2 +-
 69 files changed, 1286 insertions(+), 709 deletions(-)

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

Cornelia Huck (6):
driver core: Remove device_is_registered() in device_move(). driver core: Allow device_move(dev, NULL).
driver core fixes: make_class_name() retval checks driver core fixes: device_register() retval check in platform.c driver core: Don't stop probing on ->probe errors. driver core: Change function call order in device_bind_driver().

Dave Jones (1):
Driver Core: Increase the default timeout value of the firmware subsystem

Eric W. Biederman (1):
sysfs: Shadow directory support

Frank Haverkamp (1):
SYSFS: Fix missing include of list.h in sysfs.h

Frederik Deweerdt (1):
sysfs: suppress lockdep warnings

Greg Kroah-Hartman (8):
Kobject: make kobject apis more robust in handling NULL pointers Driver core: convert pcmcia code to use struct device
Driver core: convert SPI code to use struct device
Network: convert network devices to use struct device instead of class_device
Modules: only add drivers/ direcory if needed
PCI: add the sysfs driver name to all modules
SERIO: add the sysfs driver name to all modules
USB: add the sysfs driver name to all modules

Kay Sievers (5):
MODULES: add the module name for built in kernel drivers /sys/modules/*/holders
Driver core: add uevent vars for devices of a class Driver core: add device_type to struct device Driver core: allow to delay the uevent at device creation time

Mariusz Kozlowski (2):
sysfs: kobject_put cleanup
kobject: kobject_put cleanup

Oliver Neukum (2):
Driver core: fix race in sysfs between sysfs_remove_file() and read()/write()
sysfs: error handling in sysfs, fill_read_buffer()

Robert P. J. Day (1):
HOWTO: Add a reference to Harbison and Steele

#1 lore raw

PATCH 1/28 Kobject: make kobject apis more robust in handling NULL pointers

From: Greg Kroah-Hartman <[email protected]>

It should be ok to pass in NULL for some kobject functions, so add error checking for all exported kobject functions to be more robust.

---
 lib/kobject.c |   19 +++++++++++++++++++
 1 files changed, 19 insertions(+), 0 deletions(-)
lib/kobject.c +19 -0
index 7ce6dc1..9aed594 100644
@@ -126,6 +126,8 @@ EXPORT_SYMBOL_GPL(kobject_get_path);
  */
 void kobject_init(struct kobject * kobj)
 {
+	if (!kobj)
+		return;
 	kref_init(&kobj->kref);
 	INIT_LIST_HEAD(&kobj->entry);
 	init_waitqueue_head(&kobj->poll);
@@ -366,6 +368,8 @@ out:
 
 void kobject_del(struct kobject * kobj)
 {
+	if (!kobj)
+		return;
 	sysfs_remove_dir(kobj);
 	unlink(kobj);
 }
@@ -377,6 +381,8 @@ void kobject_del(struct kobject * kobj)
 
 void kobject_unregister(struct kobject * kobj)
 {
+	if (!kobj)
+		return;
 	pr_debug("kobject %s: unregistering\n",kobject_name(kobj));
 	kobject_uevent(kobj, KOBJ_REMOVE);
 	kobject_del(kobj);
@@ -523,6 +529,8 @@ int kset_add(struct kset * k)
 
 int kset_register(struct kset * k)
 {
+	if (!k)
+		return -EINVAL;
 	kset_init(k);
 	return kset_add(k);
 }
@@ -535,6 +543,8 @@ int kset_register(struct kset * k)
 
 void kset_unregister(struct kset * k)
 {
+	if (!k)
+		return;
 	kobject_unregister(&k->kobj);
 }
 
@@ -586,6 +596,9 @@ int subsystem_register(struct subsystem * s)
 {
 	int error;
 
+	if (!s)
+		return -EINVAL;
+
 	subsystem_init(s);
 	pr_debug("subsystem %s: registering\n",s->kset.kobj.name);
 
@@ -598,6 +611,8 @@ int subsystem_register(struct subsystem * s)
 
 void subsystem_unregister(struct subsystem * s)
 {
+	if (!s)
+		return;
 	pr_debug("subsystem %s: unregistering\n",s->kset.kobj.name);
 	kset_unregister(&s->kset);
 }
@@ -612,6 +627,10 @@ void subsystem_unregister(struct subsystem * s)
 int subsys_create_file(struct subsystem * s, struct subsys_attribute * a)
 {
 	int error = 0;
+
+	if (!s || !a)
+		return -EINVAL;
+
 	if (subsys_get(s)) {
 		error = sysfs_create_file(&s->kset.kobj,&a->attr);
 		subsys_put(s);
Signature · 1 line
1.4.4.4
#2 lore raw

PATCH 2/28 Driver core: convert pcmcia code to use struct device

From: Greg Kroah-Hartman <[email protected]>

Converts from using struct "class_device" to "struct device" making everything show up properly in /sys/devices/ with symlinks from the /sys/class directory.

---
 drivers/pcmcia/cs.c              |   34 ++++++------
 drivers/pcmcia/cs_internal.h     |    4 +-
 drivers/pcmcia/ds.c              |   14 +++---
 drivers/pcmcia/i82092.c          |    2 +-
 drivers/pcmcia/i82365.c          |    2 +-
 drivers/pcmcia/pcmcia_ioctl.c    |    1 -
 drivers/pcmcia/pcmcia_resource.c |    1 -
 drivers/pcmcia/pd6729.c          |    2 +-
 drivers/pcmcia/rsrc_nonstatic.c  |   56 +++++++++++---------
 drivers/pcmcia/soc_common.c      |    6 +-
 drivers/pcmcia/socket_sysfs.c    |  104 ++++++++++++++++++++++----------------
 drivers/pcmcia/tcic.c            |    2 +-
 drivers/pcmcia/yenta_socket.c    |    2 +-
 include/pcmcia/ss.h              |    2 +-
 14 files changed, 127 insertions(+), 105 deletions(-)
drivers/pcmcia/cs.c +17 -17
index 606a467..ac00424 100644
@@ -110,7 +110,7 @@ int pcmcia_socket_dev_suspend(struct device *dev, pm_message_t state)
 
 	down_read(&pcmcia_socket_list_rwsem);
 	list_for_each_entry(socket, &pcmcia_socket_list, socket_list) {
-		if (socket->dev.dev != dev)
+		if (socket->dev.parent != dev)
 			continue;
 		mutex_lock(&socket->skt_mutex);
 		socket_suspend(socket);
@@ -128,7 +128,7 @@ int pcmcia_socket_dev_resume(struct device *dev)
 
 	down_read(&pcmcia_socket_list_rwsem);
 	list_for_each_entry(socket, &pcmcia_socket_list, socket_list) {
-		if (socket->dev.dev != dev)
+		if (socket->dev.parent != dev)
 			continue;
 		mutex_lock(&socket->skt_mutex);
 		socket_resume(socket);
@@ -143,12 +143,12 @@ EXPORT_SYMBOL(pcmcia_socket_dev_resume);
 
 struct pcmcia_socket * pcmcia_get_socket(struct pcmcia_socket *skt)
 {
-	struct class_device *cl_dev = class_device_get(&skt->dev);
-	if (!cl_dev)
+	struct device *dev = get_device(&skt->dev);
+	if (!dev)
 		return NULL;
-	skt = class_get_devdata(cl_dev);
+	skt = dev_get_drvdata(dev);
 	if (!try_module_get(skt->owner)) {
-		class_device_put(&skt->dev);
+		put_device(&skt->dev);
 		return NULL;
 	}
 	return (skt);
@@ -159,14 +159,14 @@ EXPORT_SYMBOL(pcmcia_get_socket);
 void pcmcia_put_socket(struct pcmcia_socket *skt)
 {
 	module_put(skt->owner);
-	class_device_put(&skt->dev);
+	put_device(&skt->dev);
 }
 EXPORT_SYMBOL(pcmcia_put_socket);
 
 
-static void pcmcia_release_socket(struct class_device *class_dev)
+static void pcmcia_release_socket(struct device *dev)
 {
-	struct pcmcia_socket *socket = class_get_devdata(class_dev);
+	struct pcmcia_socket *socket = dev_get_drvdata(dev);
 
 	complete(&socket->socket_released);
 }
@@ -181,7 +181,7 @@ int pcmcia_register_socket(struct pcmcia_socket *socket)
 	struct task_struct *tsk;
 	int ret;
 
-	if (!socket || !socket->ops || !socket->dev.dev || !socket->resource_ops)
+	if (!socket || !socket->ops || !socket->dev.parent || !socket->resource_ops)
 		return -EINVAL;
 
 	cs_dbg(socket, 0, "pcmcia_register_socket(0x%p)\n", socket->ops);
@@ -226,9 +226,9 @@ int pcmcia_register_socket(struct pcmcia_socket *socket)
 #endif
 
 	/* set proper values in socket->dev */
-	socket->dev.class_data = socket;
+	dev_set_drvdata(&socket->dev, socket);
 	socket->dev.class = &pcmcia_socket_class;
-	snprintf(socket->dev.class_id, BUS_ID_SIZE, "pcmcia_socket%u", socket->sock);
+	snprintf(socket->dev.bus_id, BUS_ID_SIZE, "pcmcia_socket%u", socket->sock);
 
 	/* base address = 0, map = 0 */
 	socket->cis_mem.flags = 0;
@@ -640,7 +640,7 @@ static int pccardd(void *__skt)
 	skt->ops->set_socket(skt, &skt->socket);
 
 	/* register with the device core */
-	ret = class_device_register(&skt->dev);
+	ret = device_register(&skt->dev);
 	if (ret) {
 		printk(KERN_WARNING "PCMCIA: unable to register socket 0x%p\n",
 			skt);
@@ -689,7 +689,7 @@ static int pccardd(void *__skt)
 	remove_wait_queue(&skt->thread_wait, &wait);
 
 	/* remove from the device core */
-	class_device_unregister(&skt->dev);
+	device_unregister(&skt->dev);
 
 	return 0;
 }
@@ -904,7 +904,7 @@ int pcmcia_insert_card(struct pcmcia_socket *skt)
 EXPORT_SYMBOL(pcmcia_insert_card);
 
 
-static int pcmcia_socket_uevent(struct class_device *dev, char **envp,
+static int pcmcia_socket_uevent(struct device *dev, char **envp,
 			        int num_envp, char *buffer, int buffer_size)
 {
 	struct pcmcia_socket *s = container_of(dev, struct pcmcia_socket, dev);
@@ -930,8 +930,8 @@ static void pcmcia_release_socket_class(struct class *data)
 
 struct class pcmcia_socket_class = {
 	.name = "pcmcia_socket",
-	.uevent = pcmcia_socket_uevent,
-	.release = pcmcia_release_socket,
+	.dev_uevent = pcmcia_socket_uevent,
+	.dev_release = pcmcia_release_socket,
 	.class_release = pcmcia_release_socket_class,
 };
 EXPORT_SYMBOL(pcmcia_socket_class);
drivers/pcmcia/cs_internal.h +2 -2
index f573ea0..9fa207e 100644
@@ -142,7 +142,7 @@ struct pcmcia_callback{
 
 int pccard_register_pcmcia(struct pcmcia_socket *s, struct pcmcia_callback *c);
 
-#define cs_socket_name(skt)	((skt)->dev.class_id)
+#define cs_socket_name(skt)	((skt)->dev.bus_id)
 
 #ifdef DEBUG
 extern int cs_debug_level(int);
@@ -158,6 +158,6 @@ extern int cs_debug_level(int);
 #endif
 
 #define cs_err(skt, fmt, arg...) \
-	printk(KERN_ERR "cs: %s: " fmt, (skt)->dev.class_id , ## arg)
+	printk(KERN_ERR "cs: %s: " fmt, (skt)->dev.bus_id , ## arg)
 
 #endif /* _LINUX_CS_INTERNAL_H */
drivers/pcmcia/ds.c +7 -7
index 7355eb4..18e111e 100644
@@ -572,7 +572,7 @@ struct pcmcia_device * pcmcia_device_add(struct pcmcia_socket *s, unsigned int f
 	p_dev->func   = function;
 
 	p_dev->dev.bus = &pcmcia_bus_type;
-	p_dev->dev.parent = s->dev.dev;
+	p_dev->dev.parent = s->dev.parent;
 	p_dev->dev.release = pcmcia_release_dev;
 	bus_id_len = sprintf (p_dev->dev.bus_id, "%d.%d", p_dev->socket->sock, p_dev->device_no);
 
@@ -1328,10 +1328,10 @@ static struct pcmcia_callback pcmcia_bus_callback = {
 	.resume = pcmcia_bus_resume,
 };
 
-static int __devinit pcmcia_bus_add_socket(struct class_device *class_dev,
+static int __devinit pcmcia_bus_add_socket(struct device *dev,
 					   struct class_interface *class_intf)
 {
-	struct pcmcia_socket *socket = class_get_devdata(class_dev);
+	struct pcmcia_socket *socket = dev_get_drvdata(dev);
 	int ret;
 
 	socket = pcmcia_get_socket(socket);
@@ -1364,10 +1364,10 @@ static int __devinit pcmcia_bus_add_socket(struct class_device *class_dev,
 	return 0;
 }
 
-static void pcmcia_bus_remove_socket(struct class_device *class_dev,
+static void pcmcia_bus_remove_socket(struct device *dev,
 				     struct class_interface *class_intf)
 {
-	struct pcmcia_socket *socket = class_get_devdata(class_dev);
+	struct pcmcia_socket *socket = dev_get_drvdata(dev);
 
 	if (!socket)
 		return;
@@ -1389,8 +1389,8 @@ static void pcmcia_bus_remove_socket(struct class_device *class_dev,
 /* the pcmcia_bus_interface is used to handle pcmcia socket devices */
 static struct class_interface pcmcia_bus_interface = {
 	.class = &pcmcia_socket_class,
-	.add = &pcmcia_bus_add_socket,
-	.remove = &pcmcia_bus_remove_socket,
+	.add_dev = &pcmcia_bus_add_socket,
+	.remove_dev = &pcmcia_bus_remove_socket,
 };
 
 
drivers/pcmcia/i82092.c +1 -1
index c2ea07a..df21e2d 100644
@@ -161,7 +161,7 @@ static int __devinit i82092aa_pci_probe(struct pci_dev *dev, const struct pci_de
 	pci_set_drvdata(dev, &sockets[i].socket);
 
 	for (i = 0; i<socket_count; i++) {
-		sockets[i].socket.dev.dev = &dev->dev;
+		sockets[i].socket.dev.parent = &dev->dev;
 		sockets[i].socket.ops = &i82092aa_operations;
 		sockets[i].socket.resource_ops = &pccard_nonstatic_ops;
 		ret = pcmcia_register_socket(&sockets[i].socket);
drivers/pcmcia/i82365.c +1 -1
index ea74f98..72ff2f6 100644
@@ -1298,7 +1298,7 @@ static int __init init_i82365(void)
     
     /* register sockets with the pcmcia core */
     for (i = 0; i < sockets; i++) {
-	    socket[i].socket.dev.dev = &i82365_device->dev;
+	    socket[i].socket.dev.parent = &i82365_device->dev;
 	    socket[i].socket.ops = &pcic_operations;
 	    socket[i].socket.resource_ops = &pccard_nonstatic_ops;
 	    socket[i].socket.owner = THIS_MODULE;
drivers/pcmcia/pcmcia_ioctl.c +0 -1
index 327372b..8849414 100644
@@ -59,7 +59,6 @@ typedef struct user_info_t {
 
 #ifdef DEBUG
 extern int ds_pc_debug;
-#define cs_socket_name(skt)    ((skt)->dev.class_id)
 
 #define ds_dbg(lvl, fmt, arg...) do {		\
 	if (ds_pc_debug >= lvl)				\
drivers/pcmcia/pcmcia_resource.c +0 -1
index b9201c2..0ce39de 100644
@@ -48,7 +48,6 @@ static u8 pcmcia_used_irq[NR_IRQS];
 
 #ifdef DEBUG
 extern int ds_pc_debug;
-#define cs_socket_name(skt)    ((skt)->dev.class_id)
 
 #define ds_dbg(skt, lvl, fmt, arg...) do {			\
 	if (ds_pc_debug >= lvl)					\
drivers/pcmcia/pd6729.c +1 -1
index 360c248..dd0ddf1 100644
@@ -682,7 +682,7 @@ static int __devinit pd6729_pci_probe(struct pci_dev *dev,
 
 		socket[i].socket.ops = &pd6729_operations;
 		socket[i].socket.resource_ops = &pccard_nonstatic_ops;
-		socket[i].socket.dev.dev = &dev->dev;
+		socket[i].socket.dev.parent = &dev->dev;
 		socket[i].socket.driver_data = &socket[i];
 	}
 
drivers/pcmcia/rsrc_nonstatic.c +31 -25

131 lines not on this page · raw message

drivers/pcmcia/soc_common.c +3 -3
index e433704..d2a3bea 100644
@@ -478,10 +478,10 @@ dump_bits(char **p, const char *prefix, unsigned int val, struct bittbl *bits, i
  *
  * Returns: the number of characters added to the buffer
  */
-static ssize_t show_status(struct class_device *class_dev, char *buf)
+static ssize_t show_status(struct device *dev, char *buf)
 {
 	struct soc_pcmcia_socket *skt =
-		container_of(class_dev, struct soc_pcmcia_socket, socket.dev);
+		container_of(dev, struct soc_pcmcia_socket, socket.dev);
 	char *p = buf;
 
 	p+=sprintf(p, "slot     : %d\n", skt->nr);
@@ -747,7 +747,7 @@ int soc_common_drv_pcmcia_probe(struct device *dev, struct pcmcia_low_level *ops
 
 		add_timer(&skt->poll_timer);
 
-		class_device_create_file(&skt->socket.dev, &class_device_attr_status);
+		device_create_file(&skt->socket.dev, &device_attr_status);
 	}
 
 	dev_set_drvdata(dev, sinfo);
drivers/pcmcia/socket_sysfs.c +61 -43

234 lines not on this page · raw message

drivers/pcmcia/tcic.c +1 -1
index 2d2f415..c158cf3 100644
@@ -512,7 +512,7 @@ static int __init init_tcic(void)
     for (i = 0; i < sockets; i++) {
 	    socket_table[i].socket.ops = &tcic_operations;
 	    socket_table[i].socket.resource_ops = &pccard_nonstatic_ops;
-	    socket_table[i].socket.dev.dev = &tcic_device.dev;
+	    socket_table[i].socket.dev.parent = &tcic_device.dev;
 	    ret = pcmcia_register_socket(&socket_table[i].socket);
 	    if (ret && i)
 		    pcmcia_unregister_socket(&socket_table[0].socket);
drivers/pcmcia/yenta_socket.c +1 -1
index da471bd..a61d768 100644
@@ -1104,7 +1104,7 @@ static int __devinit yenta_probe (struct pci_dev *dev, const struct pci_device_i
 	/* prepare pcmcia_socket */
 	socket->socket.ops = &yenta_socket_operations;
 	socket->socket.resource_ops = &pccard_nonstatic_ops;
-	socket->socket.dev.dev = &dev->dev;
+	socket->socket.dev.parent = &dev->dev;
 	socket->socket.driver_data = socket;
 	socket->socket.owner = THIS_MODULE;
 	socket->socket.features = SS_CAP_PAGE_REGS | SS_CAP_PCCARD;
include/pcmcia/ss.h +1 -1
index 623a0fc..6e84258 100644
@@ -284,7 +284,7 @@ struct pcmcia_socket {
 #endif
 
 	/* socket device */
-	struct class_device		dev;
+	struct device			dev;
 	void				*driver_data;	/* data internal to the socket driver */
 
 };
Signature · 1 line
1.4.4.4
#3 lore raw

PATCH 3/28 Driver core: convert SPI code to use struct device

From: Greg Kroah-Hartman <[email protected]>

Converts from using struct "class_device" to "struct device" making everything show up properly in /sys/devices/ with symlinks from the /sys/class directory.

---
 drivers/spi/pxa2xx_spi.c    |    2 +-
 drivers/spi/spi.c           |   32 ++++++++++++++++----------------
 drivers/spi/spi_bitbang.c   |    6 +++---
 drivers/spi/spi_butterfly.c |    4 ++--
 include/linux/spi/spi.h     |   10 +++++-----
 5 files changed, 27 insertions(+), 27 deletions(-)
drivers/spi/pxa2xx_spi.c +1 -1
index 8b41f9c..dccdc50 100644
@@ -1234,7 +1234,7 @@ static int init_queue(struct driver_data *drv_data)
 
 	INIT_WORK(&drv_data->pump_messages, pump_messages);
 	drv_data->workqueue = create_singlethread_workqueue(
-					drv_data->master->cdev.dev->bus_id);
+					drv_data->master->dev.parent->bus_id);
 	if (drv_data->workqueue == NULL)
 		return -EBUSY;
 
drivers/spi/spi.c +16 -16
index 6307428..35d8c01 100644
@@ -193,7 +193,7 @@ struct spi_device *__init_or_module
 spi_new_device(struct spi_master *master, struct spi_board_info *chip)
 {
 	struct spi_device	*proxy;
-	struct device		*dev = master->cdev.dev;
+	struct device		*dev = &master->dev;
 	int			status;
 
 	/* NOTE:  caller did any chip->bus_num checks necessary */
@@ -215,7 +215,7 @@ spi_new_device(struct spi_master *master, struct spi_board_info *chip)
 	proxy->modalias = chip->modalias;
 
 	snprintf(proxy->dev.bus_id, sizeof proxy->dev.bus_id,
-			"%s.%u", master->cdev.class_id,
+			"%s.%u", master->dev.bus_id,
 			chip->chip_select);
 	proxy->dev.parent = dev;
 	proxy->dev.bus = &spi_bus_type;
@@ -290,7 +290,7 @@ static void __init_or_module
 scan_boardinfo(struct spi_master *master)
 {
 	struct boardinfo	*bi;
-	struct device		*dev = master->cdev.dev;
+	struct device		*dev = master->dev.parent;
 
 	down(&board_lock);
 	list_for_each_entry(bi, &board_list, list) {
@@ -319,18 +319,18 @@ scan_boardinfo(struct spi_master *master)
 
 /*-------------------------------------------------------------------------*/
 
-static void spi_master_release(struct class_device *cdev)
+static void spi_master_release(struct device *dev)
 {
 	struct spi_master *master;
 
-	master = container_of(cdev, struct spi_master, cdev);
+	master = container_of(dev, struct spi_master, dev);
 	kfree(master);
 }
 
 static struct class spi_master_class = {
 	.name		= "spi_master",
 	.owner		= THIS_MODULE,
-	.release	= spi_master_release,
+	.dev_release	= spi_master_release,
 };
 
 
@@ -364,9 +364,9 @@ spi_alloc_master(struct device *dev, unsigned size)
 	if (!master)
 		return NULL;
 
-	class_device_initialize(&master->cdev);
-	master->cdev.class = &spi_master_class;
-	master->cdev.dev = get_device(dev);
+	device_initialize(&master->dev);
+	master->dev.class = &spi_master_class;
+	master->dev.parent = get_device(dev);
 	spi_master_set_devdata(master, &master[1]);
 
 	return master;
@@ -396,7 +396,7 @@ int __init_or_module
 spi_register_master(struct spi_master *master)
 {
 	static atomic_t		dyn_bus_id = ATOMIC_INIT((1<<16) - 1);
-	struct device		*dev = master->cdev.dev;
+	struct device		*dev = master->dev.parent;
 	int			status = -ENODEV;
 	int			dynamic = 0;
 
@@ -412,12 +412,12 @@ spi_register_master(struct spi_master *master)
 	/* register the device, then userspace will see it.
 	 * registration fails if the bus ID is in use.
 	 */
-	snprintf(master->cdev.class_id, sizeof master->cdev.class_id,
+	snprintf(master->dev.bus_id, sizeof master->dev.bus_id,
 		"spi%u", master->bus_num);
-	status = class_device_add(&master->cdev);
+	status = device_add(&master->dev);
 	if (status < 0)
 		goto done;
-	dev_dbg(dev, "registered master %s%s\n", master->cdev.class_id,
+	dev_dbg(dev, "registered master %s%s\n", master->dev.bus_id,
 			dynamic ? " (dynamic)" : "");
 
 	/* populate children from any spi device tables */
@@ -449,8 +449,8 @@ void spi_unregister_master(struct spi_master *master)
 {
 	int dummy;
 
-	dummy = device_for_each_child(master->cdev.dev, NULL, __unregister);
-	class_device_unregister(&master->cdev);
+	dummy = device_for_each_child(&master->dev, NULL, __unregister);
+	device_unregister(&master->dev);
 }
 EXPORT_SYMBOL_GPL(spi_unregister_master);
 
@@ -471,7 +471,7 @@ struct spi_master *spi_busnum_to_master(u16 bus_num)
 
 	down(&spi_master_class.sem);
 	list_for_each_entry(cdev, &spi_master_class.children, node) {
-		m = container_of(cdev, struct spi_master, cdev);
+		m = container_of(cdev, struct spi_master, dev.kobj);
 		if (m->bus_num == bus_num) {
 			master = spi_master_get(m);
 			break;
drivers/spi/spi_bitbang.c +3 -3
index 57289b6..4638e6c 100644
@@ -479,7 +479,7 @@ int spi_bitbang_start(struct spi_bitbang *bitbang)
 	/* this task is the only thing to touch the SPI bits */
 	bitbang->busy = 0;
 	bitbang->workqueue = create_singlethread_workqueue(
-			bitbang->master->cdev.dev->bus_id);
+			bitbang->master->dev.parent->bus_id);
 	if (bitbang->workqueue == NULL) {
 		status = -EBUSY;
 		goto err1;
@@ -513,14 +513,14 @@ int spi_bitbang_stop(struct spi_bitbang *bitbang)
 	while (!list_empty(&bitbang->queue) && limit--) {
 		spin_unlock_irq(&bitbang->lock);
 
-		dev_dbg(bitbang->master->cdev.dev, "wait for queue\n");
+		dev_dbg(&bitbang->master->dev, "wait for queue\n");
 		msleep(10);
 
 		spin_lock_irq(&bitbang->lock);
 	}
 	spin_unlock_irq(&bitbang->lock);
 	if (!list_empty(&bitbang->queue)) {
-		dev_err(bitbang->master->cdev.dev, "queue didn't empty\n");
+		dev_err(&bitbang->master->dev, "queue didn't empty\n");
 		return -EBUSY;
 	}
 
drivers/spi/spi_butterfly.c +2 -2
index 312987a..31b7970 100644
@@ -246,7 +246,7 @@ static void butterfly_attach(struct parport *p)
 	 * and no way to be selective about what it binds to.
 	 */
 
-	/* FIXME where should master->cdev.dev come from?
+	/* FIXME where should master->dev.parent come from?
 	 * e.g. /sys/bus/pnp0/00:0b, some PCI thing, etc
 	 * setting up a platform device like this is an ugly kluge...
 	 */
@@ -386,7 +386,7 @@ static void butterfly_detach(struct parport *p)
 	butterfly = NULL;
 
 	/* stop() unregisters child devices too */
-	pdev = to_platform_device(pp->bitbang.master->cdev.dev);
+	pdev = to_platform_device(pp->bitbang.master->dev.parent);
 	status = spi_bitbang_stop(&pp->bitbang);
 
 	/* turn off VCC */
include/linux/spi/spi.h +5 -5
index 176f6e3..8c2edd8 100644
@@ -170,7 +170,7 @@ static inline void spi_unregister_driver(struct spi_driver *sdrv)
  * message's completion function when the transaction completes.
  */
 struct spi_master {
-	struct class_device	cdev;
+	struct device		dev;
 
 	/* other than negative (== assign one dynamically), bus_num is fully
 	 * board-specific.  usually that simplifies to being SOC-specific.
@@ -216,17 +216,17 @@ struct spi_master {
 
 static inline void *spi_master_get_devdata(struct spi_master *master)
 {
-	return class_get_devdata(&master->cdev);
+	return dev_get_drvdata(&master->dev);
 }
 
 static inline void spi_master_set_devdata(struct spi_master *master, void *data)
 {
-	class_set_devdata(&master->cdev, data);
+	dev_set_drvdata(&master->dev, data);
 }
 
 static inline struct spi_master *spi_master_get(struct spi_master *master)
 {
-	if (!master || !class_device_get(&master->cdev))
+	if (!master || !get_device(&master->dev))
 		return NULL;
 	return master;
 }
@@ -234,7 +234,7 @@ static inline struct spi_master *spi_master_get(struct spi_master *master)
 static inline void spi_master_put(struct spi_master *master)
 {
 	if (master)
-		class_device_put(&master->cdev);
+		put_device(&master->dev);
 }
 
 
Signature · 1 line
1.4.4.4
#4 lore raw

PATCH 4/28 Network: convert network devices to use struct device instead of class_device

From: Greg Kroah-Hartman <[email protected]>

This lets the network core have the ability to handle suspend/resume issues, if it wants to.

Thanks to Frederik Deweerdt <[email protected]> for the arm driver fixes.

---
 drivers/infiniband/ulp/ipoib/ipoib_main.c |   33 ++--
 drivers/infiniband/ulp/ipoib/ipoib_vlan.c |   11 +-
 drivers/net/arm/at91_ether.c              |    2 +-
 drivers/net/arm/etherh.c                  |    2 +-
 drivers/net/bonding/bond_sysfs.c          |  287 +++++++++++++++++-----------
 drivers/net/iseries_veth.c                |    2 +-
 drivers/net/macb.c                        |   36 ++--
 drivers/net/smc911x.c                     |    2 +-
 drivers/net/smc91x.c                      |    2 +-
 drivers/net/wireless/hostap/hostap_main.c |    2 +-
 drivers/net/wireless/orinoco.c            |    4 +-
 drivers/net/wireless/orinoco_cs.c         |    2 +-
 drivers/net/wireless/spectrum_cs.c        |    2 +-
 include/linux/netdevice.h                 |    5 +-
 net/bridge/br_if.c                        |    2 +-
 net/bridge/br_sysfs_br.c                  |  234 +++++++++++++-----------
 net/bridge/br_sysfs_if.c                  |    2 +-
 net/core/dev.c                            |    6 +-
 net/core/net-sysfs.c                      |  175 ++++++++++--------
 net/core/skbuff.c                         |    2 +-
 20 files changed, 451 insertions(+), 362 deletions(-)
drivers/infiniband/ulp/ipoib/ipoib_main.c +15 -18
index 705eb1d..af5ee2e 100644
@@ -958,16 +958,17 @@ struct ipoib_dev_priv *ipoib_intf_alloc(const char *name)
 	return netdev_priv(dev);
 }
 
-static ssize_t show_pkey(struct class_device *cdev, char *buf)
+static ssize_t show_pkey(struct device *dev,
+			 struct device_attribute *attr, char *buf)
 {
-	struct ipoib_dev_priv *priv =
-		netdev_priv(container_of(cdev, struct net_device, class_dev));
+	struct ipoib_dev_priv *priv = netdev_priv(to_net_dev(dev));
 
 	return sprintf(buf, "0x%04x\n", priv->pkey);
 }
-static CLASS_DEVICE_ATTR(pkey, S_IRUGO, show_pkey, NULL);
+static DEVICE_ATTR(pkey, S_IRUGO, show_pkey, NULL);
 
-static ssize_t create_child(struct class_device *cdev,
+static ssize_t create_child(struct device *dev,
+			    struct device_attribute *attr,
 			    const char *buf, size_t count)
 {
 	int pkey;
@@ -985,14 +986,14 @@ static ssize_t create_child(struct class_device *cdev,
 	 */
 	pkey |= 0x8000;
 
-	ret = ipoib_vlan_add(container_of(cdev, struct net_device, class_dev),
-			     pkey);
+	ret = ipoib_vlan_add(to_net_dev(dev), pkey);
 
 	return ret ? ret : count;
 }
-static CLASS_DEVICE_ATTR(create_child, S_IWUGO, NULL, create_child);
+static DEVICE_ATTR(create_child, S_IWUGO, NULL, create_child);
 
-static ssize_t delete_child(struct class_device *cdev,
+static ssize_t delete_child(struct device *dev,
+			    struct device_attribute *attr,
 			    const char *buf, size_t count)
 {
 	int pkey;
@@ -1004,18 +1005,16 @@ static ssize_t delete_child(struct class_device *cdev,
 	if (pkey < 0 || pkey > 0xffff)
 		return -EINVAL;
 
-	ret = ipoib_vlan_delete(container_of(cdev, struct net_device, class_dev),
-				pkey);
+	ret = ipoib_vlan_delete(to_net_dev(dev), pkey);
 
 	return ret ? ret : count;
 
 }
-static CLASS_DEVICE_ATTR(delete_child, S_IWUGO, NULL, delete_child);
+static DEVICE_ATTR(delete_child, S_IWUGO, NULL, delete_child);
 
 int ipoib_add_pkey_attr(struct net_device *dev)
 {
-	return class_device_create_file(&dev->class_dev,
-					&class_device_attr_pkey);
+	return device_create_file(&dev->dev, &dev_attr_pkey);
 }
 
 static struct net_device *ipoib_add_port(const char *format,
@@ -1083,11 +1082,9 @@ static struct net_device *ipoib_add_port(const char *format,
 
 	if (ipoib_add_pkey_attr(priv->dev))
 		goto sysfs_failed;
-	if (class_device_create_file(&priv->dev->class_dev,
-				     &class_device_attr_create_child))
+	if (device_create_file(&priv->dev->dev, &dev_attr_create_child))
 		goto sysfs_failed;
-	if (class_device_create_file(&priv->dev->class_dev,
-				     &class_device_attr_delete_child))
+	if (device_create_file(&priv->dev->dev, &dev_attr_delete_child))
 		goto sysfs_failed;
 
 	return priv->dev;
drivers/infiniband/ulp/ipoib/ipoib_vlan.c +5 -6
index f887780..085eafe 100644
@@ -42,15 +42,15 @@
 
 #include "ipoib.h"
 
-static ssize_t show_parent(struct class_device *class_dev, char *buf)
+static ssize_t show_parent(struct device *d, struct device_attribute *attr,
+			   char *buf)
 {
-	struct net_device *dev =
-		container_of(class_dev, struct net_device, class_dev);
+	struct net_device *dev = to_net_dev(d);
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 
 	return sprintf(buf, "%s\n", priv->parent->name);
 }
-static CLASS_DEVICE_ATTR(parent, S_IRUGO, show_parent, NULL);
+static DEVICE_ATTR(parent, S_IRUGO, show_parent, NULL);
 
 int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey)
 {
@@ -118,8 +118,7 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey)
 	if (ipoib_add_pkey_attr(priv->dev))
 		goto sysfs_failed;
 
-	if (class_device_create_file(&priv->dev->class_dev,
-				     &class_device_attr_parent))
+	if (device_create_file(&priv->dev->dev, &dev_attr_parent))
 		goto sysfs_failed;
 
 	list_add_tail(&priv->list, &ppriv->child_intfs);
drivers/net/arm/at91_ether.c +1 -1
index fada15d..1621b8f 100644
@@ -641,7 +641,7 @@ static void at91ether_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo
 {
 	strlcpy(info->driver, DRV_NAME, sizeof(info->driver));
 	strlcpy(info->version, DRV_VERSION, sizeof(info->version));
-	strlcpy(info->bus_info, dev->class_dev.dev->bus_id, sizeof(info->bus_info));
+	strlcpy(info->bus_info, dev->dev.parent->bus_id, sizeof(info->bus_info));
 }
 
 static const struct ethtool_ops at91ether_ethtool_ops = {
drivers/net/arm/etherh.c +1 -1
index f3faa4f..72c41f5 100644
@@ -587,7 +587,7 @@ static void etherh_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *i
 {
 	strlcpy(info->driver, DRV_NAME, sizeof(info->driver));
 	strlcpy(info->version, DRV_VERSION, sizeof(info->version));
-	strlcpy(info->bus_info, dev->class_dev.dev->bus_id,
+	strlcpy(info->bus_info, dev->dev.parent->bus_id,
 		sizeof(info->bus_info));
 }
 
drivers/net/bonding/bond_sysfs.c +174 -113

685 lines not on this page · raw message

drivers/net/iseries_veth.c +1 -1
index 2194b56..0e9ba3c 100644
@@ -1102,7 +1102,7 @@ static struct net_device * __init veth_probe_one(int vlan,
 	}
 
 	kobject_init(&port->kobject);
-	port->kobject.parent = &dev->class_dev.kobj;
+	port->kobject.parent = &dev->dev.kobj;
 	port->kobject.ktype  = &veth_port_ktype;
 	kobject_set_name(&port->kobject, "veth_port");
 	if (0 != kobject_add(&port->kobject))
drivers/net/macb.c +18 -18
index 25b559b..2af2045 100644
@@ -27,8 +27,6 @@
 
 #include "macb.h"
 
-#define to_net_dev(class) container_of(class, struct net_device, class_dev)
-
 #define RX_BUFFER_SIZE		128
 #define RX_RING_SIZE		512
 #define RX_RING_BYTES		(sizeof(struct dma_desc) * RX_RING_SIZE)
@@ -945,10 +943,10 @@ static int macb_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
 	return ret;
 }
 
-static ssize_t macb_mii_show(const struct class_device *cd, char *buf,
+static ssize_t macb_mii_show(const struct device *_dev, char *buf,
 			unsigned long addr)
 {
-	struct net_device *dev = to_net_dev(cd);
+	struct net_device *dev = to_net_dev(_dev);
 	struct macb *bp = netdev_priv(dev);
 	ssize_t ret = -EINVAL;
 
@@ -962,11 +960,13 @@ static ssize_t macb_mii_show(const struct class_device *cd, char *buf,
 }
 
 #define MII_ENTRY(name, addr)					\
-static ssize_t show_##name(struct class_device *cd, char *buf)	\
+static ssize_t show_##name(struct device *_dev,			\
+			   struct device_attribute *attr,	\
+			   char *buf)				\
 {								\
-	return macb_mii_show(cd, buf, addr);			\
+	return macb_mii_show(_dev, buf, addr);			\
 }								\
-static CLASS_DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
+static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
 
 MII_ENTRY(bmcr, MII_BMCR);
 MII_ENTRY(bmsr, MII_BMSR);
@@ -977,13 +977,13 @@ MII_ENTRY(lpa, MII_LPA);
 MII_ENTRY(expansion, MII_EXPANSION);
 
 static struct attribute *macb_mii_attrs[] = {
-	&class_device_attr_bmcr.attr,
-	&class_device_attr_bmsr.attr,
-	&class_device_attr_physid1.attr,
-	&class_device_attr_physid2.attr,
-	&class_device_attr_advertise.attr,
-	&class_device_attr_lpa.attr,
-	&class_device_attr_expansion.attr,
+	&dev_attr_bmcr.attr,
+	&dev_attr_bmsr.attr,
+	&dev_attr_physid1.attr,
+	&dev_attr_physid2.attr,
+	&dev_attr_advertise.attr,
+	&dev_attr_lpa.attr,
+	&dev_attr_expansion.attr,
 	NULL,
 };
 
@@ -994,17 +994,17 @@ static struct attribute_group macb_mii_group = {
 
 static void macb_unregister_sysfs(struct net_device *net)
 {
-	struct class_device *class_dev = &net->class_dev;
+	struct device *_dev = &net->dev;
 
-	sysfs_remove_group(&class_dev->kobj, &macb_mii_group);
+	sysfs_remove_group(&_dev->kobj, &macb_mii_group);
 }
 
 static int macb_register_sysfs(struct net_device *net)
 {
-	struct class_device *class_dev = &net->class_dev;
+	struct device *_dev = &net->dev;
 	int ret;
 
-	ret = sysfs_create_group(&class_dev->kobj, &macb_mii_group);
+	ret = sysfs_create_group(&_dev->kobj, &macb_mii_group);
 	if (ret)
 		printk(KERN_WARNING
 		       "%s: sysfs mii attribute registration failed: %d\n",
drivers/net/smc911x.c +1 -1
index 43af614..c956141 100644
@@ -1659,7 +1659,7 @@ smc911x_ethtool_getdrvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
 {
 	strncpy(info->driver, CARDNAME, sizeof(info->driver));
 	strncpy(info->version, version, sizeof(info->version));
-	strncpy(info->bus_info, dev->class_dev.dev->bus_id, sizeof(info->bus_info));
+	strncpy(info->bus_info, dev->dev.parent->bus_id, sizeof(info->bus_info));
 }
 
 static int smc911x_ethtool_nwayreset(struct net_device *dev)
drivers/net/smc91x.c +1 -1
index e62a958..49f4b77 100644
@@ -1712,7 +1712,7 @@ smc_ethtool_getdrvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
 {
 	strncpy(info->driver, CARDNAME, sizeof(info->driver));
 	strncpy(info->version, version, sizeof(info->version));
-	strncpy(info->bus_info, dev->class_dev.dev->bus_id, sizeof(info->bus_info));
+	strncpy(info->bus_info, dev->dev.parent->bus_id, sizeof(info->bus_info));
 }
 
 static int smc_ethtool_nwayreset(struct net_device *dev)
drivers/net/wireless/hostap/hostap_main.c +1 -1
index 04c19ce..9077e6e 100644
@@ -84,7 +84,7 @@ struct net_device * hostap_add_interface(struct local_info *local,
 	if (strchr(dev->name, '%'))
 		ret = dev_alloc_name(dev, dev->name);
 
-	SET_NETDEV_DEV(dev, mdev->class_dev.dev);
+	SET_NETDEV_DEV(dev, mdev->dev.parent);
 	if (ret >= 0)
 		ret = register_netdevice(dev);
 
drivers/net/wireless/orinoco.c +2 -2
index 936c888..656f216 100644
@@ -4293,8 +4293,8 @@ static void orinoco_get_drvinfo(struct net_device *dev,
 	strncpy(info->driver, DRIVER_NAME, sizeof(info->driver) - 1);
 	strncpy(info->version, DRIVER_VERSION, sizeof(info->version) - 1);
 	strncpy(info->fw_version, priv->fw_name, sizeof(info->fw_version) - 1);
-	if (dev->class_dev.dev)
-		strncpy(info->bus_info, dev->class_dev.dev->bus_id,
+	if (dev->dev.parent)
+		strncpy(info->bus_info, dev->dev.parent->bus_id,
 			sizeof(info->bus_info) - 1);
 	else
 		snprintf(info->bus_info, sizeof(info->bus_info) - 1,
drivers/net/wireless/orinoco_cs.c +1 -1
index d08ae8d..d1e5022 100644
@@ -332,7 +332,7 @@ orinoco_cs_config(struct pcmcia_device *link)
 
 	/* Finally, report what we've done */
 	printk(KERN_DEBUG "%s: " DRIVER_NAME " at %s, irq %d, io "
-	       "0x%04x-0x%04x\n", dev->name, dev->class_dev.dev->bus_id,
+	       "0x%04x-0x%04x\n", dev->name, dev->dev.parent->bus_id,
 	       link->irq.AssignedIRQ, link->io.BasePort1,
 	       link->io.BasePort1 + link->io.NumPorts1 - 1);
 
drivers/net/wireless/spectrum_cs.c +1 -1
index cf2d148..af70460 100644
@@ -806,7 +806,7 @@ spectrum_cs_config(struct pcmcia_device *link)
 
 	/* Finally, report what we've done */
 	printk(KERN_DEBUG "%s: " DRIVER_NAME " at %s, irq %d, io "
-	       "0x%04x-0x%04x\n", dev->name, dev->class_dev.dev->bus_id,
+	       "0x%04x-0x%04x\n", dev->name, dev->dev.parent->bus_id,
 	       link->irq.AssignedIRQ, link->io.BasePort1,
 	       link->io.BasePort1 + link->io.NumPorts1 - 1);
 
include/linux/netdevice.h +3 -2
index fea0d9d..2e37f50 100644
@@ -529,10 +529,11 @@ struct net_device
 	struct net_bridge_port	*br_port;
 
 	/* class/net/name entry */
-	struct class_device	class_dev;
+	struct device		dev;
 	/* space for optional statistics and wireless sysfs groups */
 	struct attribute_group  *sysfs_groups[3];
 };
+#define to_net_dev(d) container_of(d, struct net_device, dev)
 
 #define	NETDEV_ALIGN		32
 #define	NETDEV_ALIGN_CONST	(NETDEV_ALIGN - 1)
@@ -548,7 +549,7 @@ static inline void *netdev_priv(struct net_device *dev)
 /* Set the sysfs physical device reference for the network logical device
  * if set prior to registration will cause a symlink during initialization.
  */
-#define SET_NETDEV_DEV(net, pdev)	((net)->class_dev.dev = (pdev))
+#define SET_NETDEV_DEV(net, pdev)	((net)->dev.parent = (pdev))
 
 struct packet_type {
 	__be16			type;	/* This is really htons(ether_type). */
net/bridge/br_if.c +1 -1
index 55bb263..2b7c2c7 100644
@@ -286,7 +286,7 @@ static struct net_bridge_port *new_nbp(struct net_bridge *br,
 	kobject_init(&p->kobj);
 	kobject_set_name(&p->kobj, SYSFS_BRIDGE_PORT_ATTR);
 	p->kobj.ktype = &brport_ktype;
-	p->kobj.parent = &(dev->class_dev.kobj);
+	p->kobj.parent = &(dev->dev.kobj);
 	p->kobj.kset = NULL;
 
 	return p;
net/bridge/br_sysfs_br.c +127 -107

395 lines not on this page · raw message

net/bridge/br_sysfs_if.c +1 -1
index c51c9e4..0bc2aef 100644
@@ -211,7 +211,7 @@ int br_sysfs_addif(struct net_bridge_port *p)
 	struct brport_attribute **a;
 	int err;
 
-	err = sysfs_create_link(&p->kobj, &br->dev->class_dev.kobj, 
+	err = sysfs_create_link(&p->kobj, &br->dev->dev.kobj,
 				SYSFS_BRIDGE_PORT_LINK);
 	if (err)
 		goto out2;
net/core/dev.c +3 -3
index e660cb5..455d589 100644
@@ -751,7 +751,7 @@ int dev_change_name(struct net_device *dev, char *newname)
 	else
 		strlcpy(dev->name, newname, IFNAMSIZ);
 
-	err = class_device_rename(&dev->class_dev, dev->name);
+	err = device_rename(&dev->dev, dev->name);
 	if (!err) {
 		hlist_del(&dev->name_hlist);
 		hlist_add_head(&dev->name_hlist, dev_name_hash(dev->name));
@@ -3221,8 +3221,8 @@ void free_netdev(struct net_device *dev)
 	BUG_ON(dev->reg_state != NETREG_UNREGISTERED);
 	dev->reg_state = NETREG_RELEASED;
 
-	/* will free via class release */
-	class_device_put(&dev->class_dev);
+	/* will free via device release */
+	put_device(&dev->dev);
 #else
 	kfree((char *)dev - dev->padded);
 #endif
net/core/net-sysfs.c +93 -82

349 lines not on this page · raw message

net/core/skbuff.c +1 -1
index de7801d..f3404ae 100644
@@ -268,7 +268,7 @@ nodata:
 struct sk_buff *__netdev_alloc_skb(struct net_device *dev,
 		unsigned int length, gfp_t gfp_mask)
 {
-	int node = dev->class_dev.dev ? dev_to_node(dev->class_dev.dev) : -1;
+	int node = dev->dev.parent ? dev_to_node(dev->dev.parent) : -1;
 	struct sk_buff *skb;
 
  	skb = __alloc_skb(length + NET_SKB_PAD, gfp_mask, 0, node);
Signature · 1 line
1.4.4.4
#5 lore raw

PATCH 5/28 driver core: Remove device_is_registered() in device_move().

From: Cornelia Huck <[email protected]>

device_is_registered() will always be false for a device with no bus. Remove this check and trust the caller to know what they're doing.

---
 drivers/base/core.c |    4 ----
 1 files changed, 0 insertions(+), 4 deletions(-)
drivers/base/core.c +0 -4
index 67b79a7..36eedfd 100644
@@ -1004,10 +1004,6 @@ int device_move(struct device *dev, struct device *new_parent)
 	if (!dev)
 		return -EINVAL;
 
-	if (!device_is_registered(dev)) {
-		error = -EINVAL;
-		goto out;
-	}
 	new_parent = get_device(new_parent);
 	if (!new_parent) {
 		error = -EINVAL;
Signature · 1 line
1.4.4.4
#6 lore raw

PATCH 6/28 driver core: Allow device_move(dev, NULL).

From: Cornelia Huck <[email protected]>

If we allow NULL as the new parent in device_move(), we need to make sure that the device is placed into the same place as it would if it was newly registered:

- Consider the device virtual tree. In order to be able to reuse code, setup_parent() has been tweaked a bit.
- kobject_move() can fall back to the kset's kobject.
- sysfs_move_dir() uses the sysfs root dir as fallback.

---
 drivers/base/core.c |   76 +++++++++++++++++++++++++++++++-------------------
 fs/sysfs/dir.c      |    6 +---
 lib/kobject.c       |    6 ++--
 3 files changed, 52 insertions(+), 36 deletions(-)
drivers/base/core.c +47 -29
index 36eedfd..c60114d 100644
@@ -390,22 +390,23 @@ void device_initialize(struct device *dev)
 }
 
 #ifdef CONFIG_SYSFS_DEPRECATED
-static int setup_parent(struct device *dev, struct device *parent)
+static struct kobject * get_device_parent(struct device *dev,
+					  struct device *parent)
 {
 	/* Set the parent to the class, not the parent device */
 	/* this keeps sysfs from having a symlink to make old udevs happy */
 	if (dev->class)
-		dev->kobj.parent = &dev->class->subsys.kset.kobj;
+		return &dev->class->subsys.kset.kobj;
 	else if (parent)
-		dev->kobj.parent = &parent->kobj;
+		return &parent->kobj;
 
-	return 0;
+	return NULL;
 }
 #else
-static int virtual_device_parent(struct device *dev)
+static struct kobject * virtual_device_parent(struct device *dev)
 {
 	if (!dev->class)
-		return -ENODEV;
+		return ERR_PTR(-ENODEV);
 
 	if (!dev->class->virtual_dir) {
 		static struct kobject *virtual_dir = NULL;
@@ -415,25 +416,31 @@ static int virtual_device_parent(struct device *dev)
 		dev->class->virtual_dir = kobject_add_dir(virtual_dir, dev->class->name);
 	}
 
-	dev->kobj.parent = dev->class->virtual_dir;
-	return 0;
+	return dev->class->virtual_dir;
 }
 
-static int setup_parent(struct device *dev, struct device *parent)
+static struct kobject * get_device_parent(struct device *dev,
+					  struct device *parent)
 {
-	int error;
-
 	/* if this is a class device, and has no parent, create one */
 	if ((dev->class) && (parent == NULL)) {
-		error = virtual_device_parent(dev);
-		if (error)
-			return error;
+		return virtual_device_parent(dev);
 	} else if (parent)
-		dev->kobj.parent = &parent->kobj;
+		return &parent->kobj;
+	return NULL;
+}
 
+#endif
+static int setup_parent(struct device *dev, struct device *parent)
+{
+	struct kobject *kobj;
+	kobj = get_device_parent(dev, parent);
+	if (IS_ERR(kobj))
+		return PTR_ERR(kobj);
+	if (kobj)
+		dev->kobj.parent = kobj;
 	return 0;
 }
-#endif
 
 /**
  *	device_add - add device to device hierarchy.
@@ -976,12 +983,18 @@ static int device_move_class_links(struct device *dev,
 		sysfs_remove_link(&dev->kobj, "device");
 		sysfs_remove_link(&old_parent->kobj, class_name);
 	}
-	error = sysfs_create_link(&dev->kobj, &new_parent->kobj, "device");
-	if (error)
-		goto out;
-	error = sysfs_create_link(&new_parent->kobj, &dev->kobj, class_name);
-	if (error)
-		sysfs_remove_link(&dev->kobj, "device");
+	if (new_parent) {
+		error = sysfs_create_link(&dev->kobj, &new_parent->kobj,
+					  "device");
+		if (error)
+			goto out;
+		error = sysfs_create_link(&new_parent->kobj, &dev->kobj,
+					  class_name);
+		if (error)
+			sysfs_remove_link(&dev->kobj, "device");
+	}
+	else
+		error = 0;
 out:
 	kfree(class_name);
 	return error;
@@ -993,25 +1006,28 @@ out:
 /**
  * device_move - moves a device to a new parent
  * @dev: the pointer to the struct device to be moved
- * @new_parent: the new parent of the device
+ * @new_parent: the new parent of the device (can by NULL)
  */
 int device_move(struct device *dev, struct device *new_parent)
 {
 	int error;
 	struct device *old_parent;
+	struct kobject *new_parent_kobj;
 
 	dev = get_device(dev);
 	if (!dev)
 		return -EINVAL;
 
 	new_parent = get_device(new_parent);
-	if (!new_parent) {
-		error = -EINVAL;
+	new_parent_kobj = get_device_parent (dev, new_parent);
+	if (IS_ERR(new_parent_kobj)) {
+		error = PTR_ERR(new_parent_kobj);
+		put_device(new_parent);
 		goto out;
 	}
 	pr_debug("DEVICE: moving '%s' to '%s'\n", dev->bus_id,
-		new_parent->bus_id);
-	error = kobject_move(&dev->kobj, &new_parent->kobj);
+		 new_parent ? new_parent->bus_id : "<NULL>");
+	error = kobject_move(&dev->kobj, new_parent_kobj);
 	if (error) {
 		put_device(new_parent);
 		goto out;
@@ -1020,7 +1036,8 @@ int device_move(struct device *dev, struct device *new_parent)
 	dev->parent = new_parent;
 	if (old_parent)
 		klist_remove(&dev->knode_parent);
-	klist_add_tail(&dev->knode_parent, &new_parent->klist_children);
+	if (new_parent)
+		klist_add_tail(&dev->knode_parent, &new_parent->klist_children);
 	if (!dev->class)
 		goto out_put;
 	error = device_move_class_links(dev, old_parent, new_parent);
@@ -1028,7 +1045,8 @@ int device_move(struct device *dev, struct device *new_parent)
 		/* We ignore errors on cleanup since we're hosed anyway... */
 		device_move_class_links(dev, new_parent, old_parent);
 		if (!kobject_move(&dev->kobj, &old_parent->kobj)) {
-			klist_remove(&dev->knode_parent);
+			if (new_parent)
+				klist_remove(&dev->knode_parent);
 			if (old_parent)
 				klist_add_tail(&dev->knode_parent,
 					       &old_parent->klist_children);
fs/sysfs/dir.c +2 -4
index 511edef..2bab1b4 100644
@@ -378,12 +378,10 @@ int sysfs_move_dir(struct kobject *kobj, struct kobject *new_parent)
 	struct sysfs_dirent *new_parent_sd, *sd;
 	int error;
 
-	if (!new_parent)
-		return -EINVAL;
-
 	old_parent_dentry = kobj->parent ?
 		kobj->parent->dentry : sysfs_mount->mnt_sb->s_root;
-	new_parent_dentry = new_parent->dentry;
+	new_parent_dentry = new_parent ?
+		new_parent->dentry : sysfs_mount->mnt_sb->s_root;
 
 again:
 	mutex_lock(&old_parent_dentry->d_inode->i_mutex);
lib/kobject.c +3 -3
index 9aed594..c033dc8 100644
@@ -314,7 +314,7 @@ int kobject_rename(struct kobject * kobj, const char *new_name)
 /**
  *	kobject_move - move object to another parent
  *	@kobj:	object in question.
- *	@new_parent: object's new parent
+ *	@new_parent: object's new parent (can be NULL)
  */
 
 int kobject_move(struct kobject *kobj, struct kobject *new_parent)
@@ -330,8 +330,8 @@ int kobject_move(struct kobject *kobj, struct kobject *new_parent)
 		return -EINVAL;
 	new_parent = kobject_get(new_parent);
 	if (!new_parent) {
-		error = -EINVAL;
-		goto out;
+		if (kobj->kset)
+			new_parent = kobject_get(&kobj->kset->kobj);
 	}
 	/* old object path */
 	devpath = kobject_get_path(kobj, GFP_KERNEL);
Signature · 1 line
1.4.4.4
#7 lore raw

PATCH 7/28 MODULES: add the module name for built in kernel drivers

From: Kay Sievers <[email protected]>

---
 include/linux/device.h |    1 +
 include/linux/module.h |    2 +-
 kernel/module.c        |   33 ++++++++++++++++++++++++---------
 kernel/params.c        |   12 +++++-------
 4 files changed, 31 insertions(+), 17 deletions(-)
include/linux/device.h +1 -0
index f44247f..da72219 100644
@@ -126,6 +126,7 @@ struct device_driver {
 	struct klist_node	knode_bus;
 
 	struct module		* owner;
+	const char 		* mod_name;	/* used for built-in modules */
 
 	int	(*probe)	(struct device * dev);
 	int	(*remove)	(struct device * dev);
include/linux/module.h +1 -1
index 10f771a..90dc254 100644
@@ -58,6 +58,7 @@ struct module_kobject
 {
 	struct kobject kobj;
 	struct module *mod;
+	struct kobject *drivers_dir;
 };
 
 /* These are either module local, or the kernel's dummy ones. */
@@ -263,7 +264,6 @@ struct module
 	struct module_attribute *modinfo_attrs;
 	const char *version;
 	const char *srcversion;
-	struct kobject *drivers_dir;
 
 	/* Exported symbols */
 	const struct kernel_symbol *syms;
kernel/module.c +24 -9
index d0f2260..0f4489a 100644
@@ -1131,8 +1131,8 @@ static int mod_sysfs_setup(struct module *mod,
 	if (err)
 		goto out;
 
-	mod->drivers_dir = kobject_add_dir(&mod->mkobj.kobj, "drivers");
-	if (!mod->drivers_dir) {
+	mod->mkobj.drivers_dir = kobject_add_dir(&mod->mkobj.kobj, "drivers");
+	if (!mod->mkobj.drivers_dir) {
 		err = -ENOMEM;
 		goto out_unreg;
 	}
@@ -1151,7 +1151,7 @@ static int mod_sysfs_setup(struct module *mod,
 out_unreg_param:
 	module_param_sysfs_remove(mod);
 out_unreg_drivers:
-	kobject_unregister(mod->drivers_dir);
+	kobject_unregister(mod->mkobj.drivers_dir);
 out_unreg:
 	kobject_del(&mod->mkobj.kobj);
 	kobject_put(&mod->mkobj.kobj);
@@ -1163,7 +1163,7 @@ static void mod_kobject_remove(struct module *mod)
 {
 	module_remove_modinfo_attrs(mod);
 	module_param_sysfs_remove(mod);
-	kobject_unregister(mod->drivers_dir);
+	kobject_unregister(mod->mkobj.drivers_dir);
 
 	kobject_unregister(&mod->mkobj.kobj);
 }
@@ -2344,15 +2344,30 @@ void module_add_driver(struct module *mod, struct device_driver *drv)
 {
 	char *driver_name;
 	int no_warn;
+	struct module_kobject *mk = NULL;
 
-	if (!mod || !drv)
+	if (!drv)
+		return;
+
+	if (mod)
+		mk = &mod->mkobj;
+	else if (drv->mod_name) {
+		struct kobject *mkobj;
+
+		/* Lookup built-in module entry in /sys/modules */
+		mkobj = kset_find_obj(&module_subsys.kset, drv->mod_name);
+		if (mkobj)
+			mk = container_of(mkobj, struct module_kobject, kobj);
+	}
+
+	if (!mk)
 		return;
 
 	/* Don't check return codes; these calls are idempotent */
-	no_warn = sysfs_create_link(&drv->kobj, &mod->mkobj.kobj, "module");
+	no_warn = sysfs_create_link(&drv->kobj, &mk->kobj, "module");
 	driver_name = make_driver_name(drv);
 	if (driver_name) {
-		no_warn = sysfs_create_link(mod->drivers_dir, &drv->kobj,
+		no_warn = sysfs_create_link(mk->drivers_dir, &drv->kobj,
 					    driver_name);
 		kfree(driver_name);
 	}
@@ -2367,10 +2382,10 @@ void module_remove_driver(struct device_driver *drv)
 		return;
 
 	sysfs_remove_link(&drv->kobj, "module");
-	if (drv->owner && drv->owner->drivers_dir) {
+	if (drv->owner && drv->owner->mkobj.drivers_dir) {
 		driver_name = make_driver_name(drv);
 		if (driver_name) {
-			sysfs_remove_link(drv->owner->drivers_dir,
+			sysfs_remove_link(drv->owner->mkobj.drivers_dir,
 					  driver_name);
 			kfree(driver_name);
 		}
kernel/params.c +5 -7
index 718945d..737b7c5 100644
@@ -561,14 +561,12 @@ static void __init kernel_param_sysfs_setup(const char *name,
 	mk->mod = THIS_MODULE;
 	kobj_set_kset_s(mk, module_subsys);
 	kobject_set_name(&mk->kobj, name);
-	ret = kobject_register(&mk->kobj);
+	kobject_init(&mk->kobj);
+	ret = kobject_add(&mk->kobj);
 	BUG_ON(ret < 0);
-
-	/* no need to keep the kobject if no parameter is exported */
-	if (!param_sysfs_setup(mk, kparam, num_params, name_skip)) {
-		kobject_unregister(&mk->kobj);
-		kfree(mk);
-	}
+	param_sysfs_setup(mk, kparam, num_params, name_skip);
+	mk->drivers_dir = kobject_add_dir(&mk->kobj, "drivers");
+	kobject_uevent(&mk->kobj, KOBJ_ADD);
 }
 
 /*
Signature · 1 line
1.4.4.4
#8 lore raw

PATCH 8/28 Modules: only add drivers/ direcory if needed

From: Greg Kroah-Hartman <[email protected]>

This changes the module core to only create the drivers/ directory if we are going to put something in it.

---
 kernel/module.c |   20 +++++++++++---------
 kernel/params.c |    1 -
 2 files changed, 11 insertions(+), 10 deletions(-)
kernel/module.c +11 -9
index 0f4489a..9de4209 100644
@@ -1131,12 +1131,6 @@ static int mod_sysfs_setup(struct module *mod,
 	if (err)
 		goto out;
 
-	mod->mkobj.drivers_dir = kobject_add_dir(&mod->mkobj.kobj, "drivers");
-	if (!mod->mkobj.drivers_dir) {
-		err = -ENOMEM;
-		goto out_unreg;
-	}
-
 	err = module_param_sysfs_setup(mod, kparam, num_params);
 	if (err)
 		goto out_unreg_drivers;
@@ -1151,8 +1145,6 @@ static int mod_sysfs_setup(struct module *mod,
 out_unreg_param:
 	module_param_sysfs_remove(mod);
 out_unreg_drivers:
-	kobject_unregister(mod->mkobj.drivers_dir);
-out_unreg:
 	kobject_del(&mod->mkobj.kobj);
 	kobject_put(&mod->mkobj.kobj);
 out:
@@ -1163,7 +1155,8 @@ static void mod_kobject_remove(struct module *mod)
 {
 	module_remove_modinfo_attrs(mod);
 	module_param_sysfs_remove(mod);
-	kobject_unregister(mod->mkobj.drivers_dir);
+	if (mod->mkobj.drivers_dir)
+		kobject_unregister(mod->mkobj.drivers_dir);
 
 	kobject_unregister(&mod->mkobj.kobj);
 }
@@ -2340,6 +2333,14 @@ static char *make_driver_name(struct device_driver *drv)
 	return driver_name;
 }
 
+static void module_create_drivers_dir(struct module_kobject *mk)
+{
+	if (!mk || mk->drivers_dir)
+		return;
+
+	mk->drivers_dir = kobject_add_dir(&mk->kobj, "drivers");
+}
+
 void module_add_driver(struct module *mod, struct device_driver *drv)
 {
 	char *driver_name;
@@ -2367,6 +2368,7 @@ void module_add_driver(struct module *mod, struct device_driver *drv)
 	no_warn = sysfs_create_link(&drv->kobj, &mk->kobj, "module");
 	driver_name = make_driver_name(drv);
 	if (driver_name) {
+		module_create_drivers_dir(mk);
 		no_warn = sysfs_create_link(mk->drivers_dir, &drv->kobj,
 					    driver_name);
 		kfree(driver_name);
kernel/params.c +0 -1
index 737b7c5..cbaac85 100644
@@ -565,7 +565,6 @@ static void __init kernel_param_sysfs_setup(const char *name,
 	ret = kobject_add(&mk->kobj);
 	BUG_ON(ret < 0);
 	param_sysfs_setup(mk, kparam, num_params, name_skip);
-	mk->drivers_dir = kobject_add_dir(&mk->kobj, "drivers");
 	kobject_uevent(&mk->kobj, KOBJ_ADD);
 }
 
Signature · 1 line
1.4.4.4
#9 lore raw

PATCH 9/28 PCI: add the sysfs driver name to all modules

From: Greg Kroah-Hartman <[email protected]>

This adds the module name to all PCI drivers, if they are built into the kernel or not. It will show up in /sys/modules/MODULE_NAME/drivers/

It also fixes up the IDE core, which was calling __pci_register_driver() directly.

---
 drivers/ide/setup-pci.c  |    7 ++++---
 drivers/pci/pci-driver.c |    4 +++-
 include/linux/ide.h      |    4 ++--
 include/linux/pci.h      |    5 +++--
 4 files changed, 12 insertions(+), 8 deletions(-)
drivers/ide/setup-pci.c +4 -3
index 695e239..a52c80f 100644
@@ -783,10 +783,11 @@ static LIST_HEAD(ide_pci_drivers);
  *	Returns are the same as for pci_register_driver
  */
 
-int __ide_pci_register_driver(struct pci_driver *driver, struct module *module)
+int __ide_pci_register_driver(struct pci_driver *driver, struct module *module,
+			      const char *mod_name)
 {
 	if(!pre_init)
-		return __pci_register_driver(driver, module);
+		return __pci_register_driver(driver, module, mod_name);
 	driver->driver.owner = module;
 	list_add_tail(&driver->node, &ide_pci_drivers);
 	return 0;
@@ -862,6 +863,6 @@ void __init ide_scan_pcibus (int scan_direction)
 	{
 		list_del(l);
 		d = list_entry(l, struct pci_driver, node);
-		__pci_register_driver(d, d->driver.owner);
+		__pci_register_driver(d, d->driver.owner, d->driver.mod_name);
 	}
 }
drivers/pci/pci-driver.c +3 -1
index 92d5e8d..3587668 100644
@@ -422,7 +422,8 @@ static struct kobj_type pci_driver_kobj_type = {
  * If no error occurred, the driver remains registered even if 
  * no device was claimed during registration.
  */
-int __pci_register_driver(struct pci_driver *drv, struct module *owner)
+int __pci_register_driver(struct pci_driver *drv, struct module *owner,
+			  const char *mod_name)
 {
 	int error;
 
@@ -430,6 +431,7 @@ int __pci_register_driver(struct pci_driver *drv, struct module *owner)
 	drv->driver.name = drv->name;
 	drv->driver.bus = &pci_bus_type;
 	drv->driver.owner = owner;
+	drv->driver.mod_name = mod_name;
 	drv->driver.kobj.ktype = &pci_driver_kobj_type;
 
 	if (pci_multithread_probe)
include/linux/ide.h +2 -2
index e26a039..827688f 100644
@@ -1192,8 +1192,8 @@ void ide_init_disk(struct gendisk *, ide_drive_t *);
 extern int ideprobe_init(void);
 
 extern void ide_scan_pcibus(int scan_direction) __init;
-extern int __ide_pci_register_driver(struct pci_driver *driver, struct module *owner);
-#define ide_pci_register_driver(d) __ide_pci_register_driver(d, THIS_MODULE)
+extern int __ide_pci_register_driver(struct pci_driver *driver, struct module *owner, const char *mod_name);
+#define ide_pci_register_driver(d) __ide_pci_register_driver(d, THIS_MODULE, KBUILD_MODNAME)
 void ide_pci_setup_ports(struct pci_dev *, struct ide_pci_device_s *, int, ata_index_t *);
 extern void ide_setup_pci_noise (struct pci_dev *dev, struct ide_pci_device_s *d);
 
include/linux/pci.h +3 -2
index f3c617e..cb899eb 100644
@@ -573,10 +573,11 @@ int __must_check pci_bus_alloc_resource(struct pci_bus *bus,
 void pci_enable_bridges(struct pci_bus *bus);
 
 /* Proper probing supporting hot-pluggable devices */
-int __must_check __pci_register_driver(struct pci_driver *, struct module *);
+int __must_check __pci_register_driver(struct pci_driver *, struct module *,
+				       const char *mod_name);
 static inline int __must_check pci_register_driver(struct pci_driver *driver)
 {
-	return __pci_register_driver(driver, THIS_MODULE);
+	return __pci_register_driver(driver, THIS_MODULE, KBUILD_MODNAME);
 }
 
 void pci_unregister_driver(struct pci_driver *);
Signature · 1 line
1.4.4.4
#10 lore raw

PATCH 10/28 SERIO: add the sysfs driver name to all modules

From: Greg Kroah-Hartman <[email protected]>

This adds the module name to all SERIO drivers, if they are built into the kernel or not. It will show up in /sys/modules/MODULE_NAME/drivers/

---
 drivers/input/serio/serio.c |    6 ++++--
 include/linux/serio.h       |    5 +++++
 2 files changed, 9 insertions(+), 2 deletions(-)
drivers/input/serio/serio.c +4 -2
index f0ce822..17c8c63 100644
@@ -45,7 +45,7 @@ EXPORT_SYMBOL(serio_interrupt);
 EXPORT_SYMBOL(__serio_register_port);
 EXPORT_SYMBOL(serio_unregister_port);
 EXPORT_SYMBOL(serio_unregister_child_port);
-EXPORT_SYMBOL(serio_register_driver);
+EXPORT_SYMBOL(__serio_register_driver);
 EXPORT_SYMBOL(serio_unregister_driver);
 EXPORT_SYMBOL(serio_open);
 EXPORT_SYMBOL(serio_close);
@@ -789,12 +789,14 @@ static void serio_attach_driver(struct serio_driver *drv)
 			drv->driver.name, error);
 }
 
-int serio_register_driver(struct serio_driver *drv)
+int __serio_register_driver(struct serio_driver *drv, struct module *owner, const char *mod_name)
 {
 	int manual_bind = drv->manual_bind;
 	int error;
 
 	drv->driver.bus = &serio_bus;
+	drv->driver.owner = owner;
+	drv->driver.mod_name = mod_name;
 
 	/*
 	 * Temporarily disable automatic binding because probing
include/linux/serio.h +5 -0
index 0f478a8..ac2c70e 100644
@@ -86,6 +86,11 @@ static inline void serio_register_port(struct serio *serio)
 void serio_unregister_port(struct serio *serio);
 void serio_unregister_child_port(struct serio *serio);
 
+int __serio_register_driver(struct serio_driver *drv, struct module *owner, const char *mod_name);
+static inline int serio_register_driver(struct serio_driver *drv)
+{
+	return __serio_register_driver(drv, THIS_MODULE, KBUILD_MODNAME);
+}
 int serio_register_driver(struct serio_driver *drv);
 void serio_unregister_driver(struct serio_driver *drv);
 
Signature · 1 line
1.4.4.4
#11 lore raw

PATCH 11/28 USB: add the sysfs driver name to all modules

From: Greg Kroah-Hartman <[email protected]>

This adds the module name to all USB drivers, if they are built into the kernel or not. It will show up in /sys/modules/MODULE_NAME/drivers/

---
 drivers/usb/core/driver.c |    4 +++-
 include/linux/usb.h       |    5 +++--
 2 files changed, 6 insertions(+), 3 deletions(-)
drivers/usb/core/driver.c +3 -1
index d6eb5ce..d505926 100644
@@ -750,7 +750,8 @@ EXPORT_SYMBOL_GPL(usb_deregister_device_driver);
  * usb_register_dev() to enable that functionality.  This function no longer
  * takes care of that.
  */
-int usb_register_driver(struct usb_driver *new_driver, struct module *owner)
+int usb_register_driver(struct usb_driver *new_driver, struct module *owner,
+			const char *mod_name)
 {
 	int retval = 0;
 
@@ -763,6 +764,7 @@ int usb_register_driver(struct usb_driver *new_driver, struct module *owner)
 	new_driver->drvwrap.driver.probe = usb_probe_interface;
 	new_driver->drvwrap.driver.remove = usb_unbind_interface;
 	new_driver->drvwrap.driver.owner = owner;
+	new_driver->drvwrap.driver.mod_name = mod_name;
 	spin_lock_init(&new_driver->dynids.lock);
 	INIT_LIST_HEAD(&new_driver->dynids.list);
 
include/linux/usb.h +3 -2
index aab5b1b..733f38d 100644
@@ -868,10 +868,11 @@ struct usb_class_driver {
  * use these in module_init()/module_exit()
  * and don't forget MODULE_DEVICE_TABLE(usb, ...)
  */
-extern int usb_register_driver(struct usb_driver *, struct module *);
+extern int usb_register_driver(struct usb_driver *, struct module *,
+			       const char *);
 static inline int usb_register(struct usb_driver *driver)
 {
-	return usb_register_driver(driver, THIS_MODULE);
+	return usb_register_driver(driver, THIS_MODULE, KBUILD_MODNAME);
 }
 extern void usb_deregister(struct usb_driver *);
 
Signature · 1 line
1.4.4.4
#12 lore raw

PATCH 12/28 /sys/modules/*/holders

From: Kay Sievers <[email protected]>

/sys/module/usbcore/

  |-- drivers
  |   |-- usb:hub -> ../../../subsystem/usb/drivers/hub
  |   |-- usb:usb -> ../../../subsystem/usb/drivers/usb
  |   `-- usb:usbfs -> ../../../subsystem/usb/drivers/usbfs
  |-- holders
  |   |-- ehci_hcd -> ../../../module/ehci_hcd
  |   |-- uhci_hcd -> ../../../module/uhci_hcd
  |   |-- usb_storage -> ../../../module/usb_storage
  |   `-- usbhid -> ../../../module/usbhid
  |-- initstate
---
 include/linux/module.h |    1 +
 kernel/module.c        |   37 +++++++++++++++++++++++++++++++------
 kernel/params.c        |   17 ++++++++++++++++-
 3 files changed, 48 insertions(+), 7 deletions(-)
include/linux/module.h +1 -0
index 90dc254..419d3ef 100644
@@ -264,6 +264,7 @@ struct module
 	struct module_attribute *modinfo_attrs;
 	const char *version;
 	const char *srcversion;
+	struct kobject *holders_dir;
 
 	/* Exported symbols */
 	const struct kernel_symbol *syms;
kernel/module.c +31 -6
index 9de4209..8a94e05 100644
@@ -537,6 +537,8 @@ static int already_uses(struct module *a, struct module *b)
 static int use_module(struct module *a, struct module *b)
 {
 	struct module_use *use;
+	int no_warn;
+
 	if (b == NULL || already_uses(a, b)) return 1;
 
 	if (!strong_try_module_get(b))
@@ -552,6 +554,7 @@ static int use_module(struct module *a, struct module *b)
 
 	use->module_which_uses = a;
 	list_add(&use->list, &b->modules_which_use_me);
+	no_warn = sysfs_create_link(b->holders_dir, &a->mkobj.kobj, a->name);
 	return 1;
 }
 
@@ -569,6 +572,7 @@ static void module_unload_free(struct module *mod)
 				module_put(i);
 				list_del(&use->list);
 				kfree(use);
+				sysfs_remove_link(i->holders_dir, mod->name);
 				/* There can be at most one match. */
 				break;
 			}
@@ -1106,9 +1110,7 @@ static void module_remove_modinfo_attrs(struct module *mod)
 	kfree(mod->modinfo_attrs);
 }
 
-static int mod_sysfs_setup(struct module *mod,
-			   struct kernel_param *kparam,
-			   unsigned int num_params)
+static int mod_sysfs_init(struct module *mod)
 {
 	int err;
 
@@ -1125,15 +1127,30 @@ static int mod_sysfs_setup(struct module *mod,
 	kobj_set_kset_s(&mod->mkobj, module_subsys);
 	mod->mkobj.mod = mod;
 
-	/* delay uevent until full sysfs population */
 	kobject_init(&mod->mkobj.kobj);
+
+out:
+	return err;
+}
+
+static int mod_sysfs_setup(struct module *mod,
+			   struct kernel_param *kparam,
+			   unsigned int num_params)
+{
+	int err;
+
+	/* delay uevent until full sysfs population */
 	err = kobject_add(&mod->mkobj.kobj);
 	if (err)
 		goto out;
 
+	mod->holders_dir = kobject_add_dir(&mod->mkobj.kobj, "holders");
+	if (!mod->holders_dir)
+		goto out_unreg;
+
 	err = module_param_sysfs_setup(mod, kparam, num_params);
 	if (err)
-		goto out_unreg_drivers;
+		goto out_unreg_holders;
 
 	err = module_add_modinfo_attrs(mod);
 	if (err)
@@ -1144,7 +1161,9 @@ static int mod_sysfs_setup(struct module *mod,
 
 out_unreg_param:
 	module_param_sysfs_remove(mod);
-out_unreg_drivers:
+out_unreg_holders:
+	kobject_unregister(mod->holders_dir);
+out_unreg:
 	kobject_del(&mod->mkobj.kobj);
 	kobject_put(&mod->mkobj.kobj);
 out:
@@ -1157,6 +1176,8 @@ static void mod_kobject_remove(struct module *mod)
 	module_param_sysfs_remove(mod);
 	if (mod->mkobj.drivers_dir)
 		kobject_unregister(mod->mkobj.drivers_dir);
+	if (mod->holders_dir)
+		kobject_unregister(mod->holders_dir);
 
 	kobject_unregister(&mod->mkobj.kobj);
 }
@@ -1761,6 +1782,10 @@ static struct module *load_module(void __user *umod,
 	/* Now we've moved module, initialize linked lists, etc. */
 	module_unload_init(mod);
 
+	/* Initialize kobject, so we can reference it. */
+	if (mod_sysfs_init(mod) != 0)
+		goto cleanup;
+
 	/* Set up license info based on the info section */
 	set_license(mod, get_modinfo(sechdrs, infoindex, "license"));
 
kernel/params.c +16 -1
index cbaac85..553cf7d 100644
@@ -30,6 +30,8 @@
 #define DEBUGP(fmt, a...)
 #endif
 
+static struct kobj_type module_ktype;
+
 static inline char dash2underscore(char c)
 {
 	if (c == '-')
@@ -671,6 +673,19 @@ static struct sysfs_ops module_sysfs_ops = {
 	.store = module_attr_store,
 };
 
+static int uevent_filter(struct kset *kset, struct kobject *kobj)
+{
+	struct kobj_type *ktype = get_ktype(kobj);
+
+	if (ktype == &module_ktype)
+		return 1;
+	return 0;
+}
+
+static struct kset_uevent_ops module_uevent_ops = {
+	.filter = uevent_filter,
+};
+
 #else
 static struct sysfs_ops module_sysfs_ops = {
 	.show = NULL,
@@ -682,7 +697,7 @@ static struct kobj_type module_ktype = {
 	.sysfs_ops =	&module_sysfs_ops,
 };
 
-decl_subsys(module, &module_ktype, NULL);
+decl_subsys(module, &module_ktype, &module_uevent_ops);
 
 /*
  * param_sysfs_init - wrapper for built-in params support
Signature · 1 line
1.4.4.4
#13 lore raw

PATCH 13/28 driver core fixes: make_class_name() retval checks

From: Cornelia Huck <[email protected]>

Make make_class_name() return NULL on error and fixup callers in the driver core.

---
 drivers/base/class.c |   21 ++++++++++++++-------
 drivers/base/core.c  |   17 +++++++++++------
 2 files changed, 25 insertions(+), 13 deletions(-)
drivers/base/class.c +14 -7
index 8bf2ca2..96def1d 100644
@@ -364,7 +364,7 @@ char *make_class_name(const char *name, struct kobject *kobj)
 
 	class_name = kmalloc(size, GFP_KERNEL);
 	if (!class_name)
-		return ERR_PTR(-ENOMEM);
+		return NULL;
 
 	strcpy(class_name, name);
 	strcat(class_name, ":");
@@ -411,8 +411,11 @@ static int make_deprecated_class_device_links(struct class_device *class_dev)
 		return 0;
 
 	class_name = make_class_name(class_dev->class->name, &class_dev->kobj);
-	error = sysfs_create_link(&class_dev->dev->kobj, &class_dev->kobj,
-				  class_name);
+	if (class_name)
+		error = sysfs_create_link(&class_dev->dev->kobj,
+					  &class_dev->kobj, class_name);
+	else
+		error = -ENOMEM;
 	kfree(class_name);
 	return error;
 }
@@ -425,7 +428,8 @@ static void remove_deprecated_class_device_links(struct class_device *class_dev)
 		return;
 
 	class_name = make_class_name(class_dev->class->name, &class_dev->kobj);
-	sysfs_remove_link(&class_dev->dev->kobj, class_name);
+	if (class_name)
+		sysfs_remove_link(&class_dev->dev->kobj, class_name);
 	kfree(class_name);
 }
 #else
@@ -863,9 +867,12 @@ int class_device_rename(struct class_device *class_dev, char *new_name)
 	if (class_dev->dev) {
 		new_class_name = make_class_name(class_dev->class->name,
 						 &class_dev->kobj);
-		sysfs_create_link(&class_dev->dev->kobj, &class_dev->kobj,
-				  new_class_name);
-		sysfs_remove_link(&class_dev->dev->kobj, old_class_name);
+		if (new_class_name)
+			sysfs_create_link(&class_dev->dev->kobj,
+					  &class_dev->kobj, new_class_name);
+		if (old_class_name)
+			sysfs_remove_link(&class_dev->dev->kobj,
+						old_class_name);
 	}
 #endif
 	class_device_put(class_dev);
drivers/base/core.c +11 -6
index c60114d..6fee3e6 100644
@@ -527,9 +527,13 @@ int device_add(struct device *dev)
 					  &dev->kobj, dev->bus_id);
 #ifdef CONFIG_SYSFS_DEPRECATED
 		if (parent) {
-			sysfs_create_link(&dev->kobj, &dev->parent->kobj, "device");
-			class_name = make_class_name(dev->class->name, &dev->kobj);
-			sysfs_create_link(&dev->parent->kobj, &dev->kobj, class_name);
+			sysfs_create_link(&dev->kobj, &dev->parent->kobj,
+							"device");
+			class_name = make_class_name(dev->class->name,
+							&dev->kobj);
+			if (class_name)
+				sysfs_create_link(&dev->parent->kobj,
+						  &dev->kobj, class_name);
 		}
 #endif
 	}
@@ -672,7 +676,9 @@ void device_del(struct device * dev)
 		if (parent) {
 			char *class_name = make_class_name(dev->class->name,
 							   &dev->kobj);
-			sysfs_remove_link(&dev->parent->kobj, class_name);
+			if (class_name)
+				sysfs_remove_link(&dev->parent->kobj,
+						  class_name);
 			kfree(class_name);
 			sysfs_remove_link(&dev->kobj, "device");
 		}
@@ -975,8 +981,7 @@ static int device_move_class_links(struct device *dev,
 
 	class_name = make_class_name(dev->class->name, &dev->kobj);
 	if (!class_name) {
-		error = PTR_ERR(class_name);
-		class_name = NULL;
+		error = -ENOMEM;
 		goto out;
 	}
 	if (old_parent) {
Signature · 1 line
1.4.4.4
#14 lore raw

PATCH 14/28 driver core fixes: device_register() retval check in platform.c

From: Cornelia Huck <[email protected]>

Check the return value of device_register() in platform_bus_init().

---
 drivers/base/platform.c |   11 +++++++++--
 1 files changed, 9 insertions(+), 2 deletions(-)
drivers/base/platform.c +9 -2
index f9c903b..30480f6 100644
@@ -611,8 +611,15 @@ EXPORT_SYMBOL_GPL(platform_bus_type);
 
 int __init platform_bus_init(void)
 {
-	device_register(&platform_bus);
-	return bus_register(&platform_bus_type);
+	int error;
+
+	error = device_register(&platform_bus);
+	if (error)
+		return error;
+	error =  bus_register(&platform_bus_type);
+	if (error)
+		device_unregister(&platform_bus);
+	return error;
 }
 
 #ifndef ARCH_HAS_DMA_GET_REQUIRED_MASK
Signature · 1 line
1.4.4.4
#15 lore raw

PATCH 15/28 driver core: Don't stop probing on ->probe errors.

From: Cornelia Huck <[email protected]>

Don't stop on the first ->probe error that is not -ENODEV/-ENXIO.

There might be a driver registered returning an unresonable return code, and this stops probing completely even though it may make sense to try the next possible driver. At worst, we may end up with an unbound device.

---
 drivers/base/dd.c |   13 ++++++-------
 1 files changed, 6 insertions(+), 7 deletions(-)
drivers/base/dd.c +6 -7
index 510e788..f705137 100644
@@ -136,18 +136,17 @@ probe_failed:
 	driver_sysfs_remove(dev);
 	dev->driver = NULL;
 
-	if (ret == -ENODEV || ret == -ENXIO) {
-		/* Driver matched, but didn't support device
-		 * or device not found.
-		 * Not an error; keep going.
-		 */
-		ret = 0;
-	} else {
+	if (ret != -ENODEV && ret != -ENXIO) {
 		/* driver matched but the probe failed */
 		printk(KERN_WARNING
 		       "%s: probe of %s failed with error %d\n",
 		       drv->name, dev->bus_id, ret);
 	}
+	/*
+	 * Ignore errors returned by ->probe so that the next driver can try
+	 * its luck.
+	 */
+	ret = 0;
 done:
 	kfree(data);
 	atomic_dec(&probe_count);
Signature · 1 line
1.4.4.4
#16 lore raw

PATCH 16/28 driver core: Change function call order in device_bind_driver().

From: Cornelia Huck <[email protected]>

Change function call order in device_bind_driver().

If we create symlinks (which might fail) before adding the device to the list we don't have to clean up afterwards (which we didn't).

---
 drivers/base/dd.c |    8 ++++++--
 1 files changed, 6 insertions(+), 2 deletions(-)
drivers/base/dd.c +6 -2
index f705137..b5bf243 100644
@@ -86,8 +86,12 @@ static void driver_sysfs_remove(struct device *dev)
  */
 int device_bind_driver(struct device *dev)
 {
-	driver_bound(dev);
-	return driver_sysfs_add(dev);
+	int ret;
+
+	ret = driver_sysfs_add(dev);
+	if (!ret)
+		driver_bound(dev);
+	return ret;
 }
 
 struct stupid_thread_structure {
Signature · 1 line
1.4.4.4
#17 lore raw

PATCH 17/28 Driver core: fix race in sysfs between sysfs_remove_file() and read()/write()

From: Oliver Neukum <[email protected]>

This patch prevents a race between IO and removing a file from sysfs. It introduces a list of sysfs_buffers associated with a file at the inode. Upon removal of a file the list is walked and the buffers marked orphaned. IO to orphaned buffers fails with -ENODEV. The driver can safely free associated data structures or be unloaded.

---
 fs/sysfs/bin.c     |    1 +
 fs/sysfs/dir.c     |    1 +
 fs/sysfs/file.c    |   67 ++++++++++++++++++++++++++++++++++++++-------------
 fs/sysfs/group.c   |    1 +
 fs/sysfs/inode.c   |   24 ++++++++++++++++++
 fs/sysfs/mount.c   |    9 +++++++
 fs/sysfs/symlink.c |    1 +
 fs/sysfs/sysfs.h   |   16 ++++++++++++
 8 files changed, 103 insertions(+), 17 deletions(-)
fs/sysfs/bin.c +1 -0
index e8f540d..9ec1c85 100644
@@ -16,6 +16,7 @@
 #include <linux/slab.h>
 
 #include <asm/uaccess.h>
+#include <asm/semaphore.h>
 
 #include "sysfs.h"
 
fs/sysfs/dir.c +1 -0
index 2bab1b4..9ff0449 100644
@@ -9,6 +9,7 @@
 #include <linux/module.h>
 #include <linux/kobject.h>
 #include <linux/namei.h>
+#include <asm/semaphore.h>
 #include "sysfs.h"
 
 DECLARE_RWSEM(sysfs_rename_sem);
fs/sysfs/file.c +50 -17

144 lines not on this page · raw message

fs/sysfs/group.c +1 -0
index 122145b..46a277b 100644
@@ -13,6 +13,7 @@
 #include <linux/dcache.h>
 #include <linux/namei.h>
 #include <linux/err.h>
+#include <asm/semaphore.h>
 #include "sysfs.h"
 
 
fs/sysfs/inode.c +24 -0
index e79e38d..1a81ebc 100644
@@ -13,6 +13,7 @@
 #include <linux/backing-dev.h>
 #include <linux/capability.h>
 #include <linux/errno.h>
+#include <asm/semaphore.h>
 #include "sysfs.h"
 
 extern struct super_block * sysfs_sb;
@@ -209,6 +210,22 @@ const unsigned char * sysfs_get_name(struct sysfs_dirent *sd)
 	return NULL;
 }
 
+static inline void orphan_all_buffers(struct inode *node)
+{
+	struct sysfs_buffer_collection *set = node->i_private;
+	struct sysfs_buffer *buf;
+
+	mutex_lock(&node->i_mutex);
+	if (node->i_private) {
+		list_for_each_entry(buf, &set->associates, associates) {
+			down(&buf->sem);
+			buf->orphaned = 1;
+			up(&buf->sem);
+		}
+	}
+	mutex_unlock(&node->i_mutex);
+}
+
 
 /*
  * Unhashes the dentry corresponding to given sysfs_dirent
@@ -217,16 +234,23 @@ const unsigned char * sysfs_get_name(struct sysfs_dirent *sd)
 void sysfs_drop_dentry(struct sysfs_dirent * sd, struct dentry * parent)
 {
 	struct dentry * dentry = sd->s_dentry;
+	struct inode *inode;
 
 	if (dentry) {
 		spin_lock(&dcache_lock);
 		spin_lock(&dentry->d_lock);
 		if (!(d_unhashed(dentry) && dentry->d_inode)) {
+			inode = dentry->d_inode;
+			spin_lock(&inode->i_lock);
+			__iget(inode);
+			spin_unlock(&inode->i_lock);
 			dget_locked(dentry);
 			__d_drop(dentry);
 			spin_unlock(&dentry->d_lock);
 			spin_unlock(&dcache_lock);
 			simple_unlink(parent->d_inode, dentry);
+			orphan_all_buffers(inode);
+			iput(inode);
 		} else {
 			spin_unlock(&dentry->d_lock);
 			spin_unlock(&dcache_lock);
fs/sysfs/mount.c +9 -0
index e503f85..a1a58b9 100644
@@ -8,6 +8,7 @@
 #include <linux/mount.h>
 #include <linux/pagemap.h>
 #include <linux/init.h>
+#include <asm/semaphore.h>
 
 #include "sysfs.h"
 
@@ -18,9 +19,12 @@ struct vfsmount *sysfs_mount;
 struct super_block * sysfs_sb = NULL;
 struct kmem_cache *sysfs_dir_cachep;
 
+static void sysfs_clear_inode(struct inode *inode);
+
 static struct super_operations sysfs_ops = {
 	.statfs		= simple_statfs,
 	.drop_inode	= generic_delete_inode,
+	.clear_inode	= sysfs_clear_inode,
 };
 
 static struct sysfs_dirent sysfs_root = {
@@ -31,6 +35,11 @@ static struct sysfs_dirent sysfs_root = {
 	.s_iattr	= NULL,
 };
 
+static void sysfs_clear_inode(struct inode *inode)
+{
+	kfree(inode->i_private);
+}
+
 static int sysfs_fill_super(struct super_block *sb, void *data, int silent)
 {
 	struct inode *inode;
fs/sysfs/symlink.c +1 -0
index f50e3cc..4869f61 100644
@@ -7,6 +7,7 @@
 #include <linux/module.h>
 #include <linux/kobject.h>
 #include <linux/namei.h>
+#include <asm/semaphore.h>
 
 #include "sysfs.h"
 
fs/sysfs/sysfs.h +16 -0
index bd7cec2..39c623f 100644
@@ -33,6 +33,22 @@ struct sysfs_symlink {
 	struct kobject * target_kobj;
 };
 
+struct sysfs_buffer {
+	struct list_head		associates;
+	size_t				count;
+	loff_t				pos;
+	char				* page;
+	struct sysfs_ops		* ops;
+	struct semaphore		sem;
+	int				orphaned;
+	int				needs_read_fill;
+	int				event;
+};
+
+struct sysfs_buffer_collection {
+	struct list_head	associates;
+};
+
 static inline struct kobject * to_kobj(struct dentry * dentry)
 {
 	struct sysfs_dirent * sd = dentry->d_fsdata;
Signature · 1 line
1.4.4.4
#18 lore raw

PATCH 18/28 sysfs: suppress lockdep warnings

From: Frederik Deweerdt <[email protected]>

Lockdep issues the following warning:

[    9.064000] =============================================
[    9.064000] [ INFO: possible recursive locking detected ]
[    9.064000] 2.6.20-rc3-mm1 #3
[    9.064000] ---------------------------------------------
[    9.064000] init/1 is trying to acquire lock:
[    9.064000]  (&sysfs_inode_imutex_key){--..}, at: [<c03e6afc>] mutex_lock+0x1c/0x1f
[    9.064000]
[    9.064000] but task is already holding lock:
[    9.064000]  (&sysfs_inode_imutex_key){--..}, at: [<c03e6afc>] mutex_lock+0x1c/0x1f
[    9.065000]
[    9.065000] other info that might help us debug this:
[    9.065000] 2 locks held by init/1:
[    9.065000]  #0:  (tty_mutex){--..}, at: [<c03e6afc>] mutex_lock+0x1c/0x1f
[    9.065000]  #1:  (&sysfs_inode_imutex_key){--..}, at: [<c03e6afc>] mutex_lock+0x1c/0x1f
[    9.065000]
[    9.065000] stack backtrace:
[    9.065000]  [<c010390d>] show_trace_log_lvl+0x1a/0x30
[    9.066000]  [<c0103935>] show_trace+0x12/0x14
[    9.066000]  [<c0103a2f>] dump_stack+0x16/0x18
[    9.066000]  [<c0138cb8>] print_deadlock_bug+0xb9/0xc3
[    9.066000]  [<c0138d17>] check_deadlock+0x55/0x5a
[    9.066000]  [<c013a953>] __lock_acquire+0x371/0xbf0
[    9.066000]  [<c013b7a9>] lock_acquire+0x69/0x83
[    9.066000]  [<c03e6b7e>] __mutex_lock_slowpath+0x75/0x2d1
[    9.066000]  [<c03e6afc>] mutex_lock+0x1c/0x1f
[    9.066000]  [<c01b249c>] sysfs_drop_dentry+0xb1/0x133
[    9.066000]  [<c01b25d1>] sysfs_hash_and_remove+0xb3/0x142
[    9.066000]  [<c01b30ed>] sysfs_remove_file+0xd/0x10
[    9.067000]  [<c02849e0>] device_remove_file+0x23/0x2e
[    9.067000]  [<c02850b2>] device_del+0x188/0x1e6
[    9.067000]  [<c028511b>] device_unregister+0xb/0x15
[    9.067000]  [<c0285318>] device_destroy+0x9c/0xa9
[    9.067000]  [<c0261431>] vcs_remove_sysfs+0x1c/0x3b
[    9.067000]  [<c0267a08>] con_close+0x5e/0x6b
[    9.067000]  [<c02598f2>] release_dev+0x4c4/0x6e5
[    9.067000]  [<c0259faa>] tty_release+0x12/0x1c
[    9.067000]  [<c0174872>] __fput+0x177/0x1a0
[    9.067000]  [<c01746f5>] fput+0x3b/0x41
[    9.068000]  [<c0172ee1>] filp_close+0x36/0x65
[    9.068000]  [<c0172f73>] sys_close+0x63/0xa4
[    9.068000]  [<c0102a96>] sysenter_past_esp+0x5f/0x99
[    9.068000]  =======================

This is due to sysfs_hash_and_remove() holding dir->d_inode->i_mutex before calling sysfs_drop_dentry() which calls orphan_all_buffers() which in turn takes node->i_mutex.

---
 fs/sysfs/inode.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
fs/sysfs/inode.c +2 -2
index 1a81ebc..dbd820f 100644
@@ -215,7 +215,7 @@ static inline void orphan_all_buffers(struct inode *node)
 	struct sysfs_buffer_collection *set = node->i_private;
 	struct sysfs_buffer *buf;
 
-	mutex_lock(&node->i_mutex);
+	mutex_lock_nested(&node->i_mutex, I_MUTEX_CHILD);
 	if (node->i_private) {
 		list_for_each_entry(buf, &set->associates, associates) {
 			down(&buf->sem);
@@ -272,7 +272,7 @@ int sysfs_hash_and_remove(struct dentry * dir, const char * name)
 		return -ENOENT;
 
 	parent_sd = dir->d_fsdata;
-	mutex_lock(&dir->d_inode->i_mutex);
+	mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT);
 	list_for_each_entry(sd, &parent_sd->s_children, s_sibling) {
 		if (!sd->s_element)
 			continue;
Signature · 1 line
1.4.4.4
#19 lore raw

PATCH 19/28 sysfs: kobject_put cleanup

From: Mariusz Kozlowski <[email protected]>

This patch removes redundant argument checks for kobject_put().

---
 fs/sysfs/bin.c  |    5 ++---
 fs/sysfs/file.c |    5 ++---
 2 files changed, 4 insertions(+), 6 deletions(-)
fs/sysfs/bin.c +2 -3
index 9ec1c85..d3b9f5f 100644
@@ -147,7 +147,7 @@ static int open(struct inode * inode, struct file * file)
  Error:
 	module_put(attr->attr.owner);
  Done:
-	if (error && kobj)
+	if (error)
 		kobject_put(kobj);
 	return error;
 }
@@ -158,8 +158,7 @@ static int release(struct inode * inode, struct file * file)
 	struct bin_attribute * attr = to_bin_attr(file->f_path.dentry);
 	u8 * buffer = file->private_data;
 
-	if (kobj) 
-		kobject_put(kobj);
+	kobject_put(kobj);
 	module_put(attr->attr.owner);
 	kfree(buffer);
 	return 0;
fs/sysfs/file.c +2 -3
index cba4c1c..46618f8 100644
@@ -361,7 +361,7 @@ static int sysfs_open_file(struct inode *inode, struct file *file)
 	error = -EACCES;
 	module_put(attr->owner);
  Done:
-	if (error && kobj)
+	if (error)
 		kobject_put(kobj);
 	return error;
 }
@@ -375,8 +375,7 @@ static int sysfs_release(struct inode * inode, struct file * filp)
 
 	if (buffer)
 		remove_from_collection(buffer, inode);
-	if (kobj) 
-		kobject_put(kobj);
+	kobject_put(kobj);
 	/* After this point, attr should not be accessed. */
 	module_put(owner);
 
Signature · 1 line
1.4.4.4
#20 lore raw

PATCH 20/28 kobject: kobject_put cleanup

From: Mariusz Kozlowski <[email protected]>

This patch removes redundant argument checks for kobject_put().

---
 lib/kobject.c |    6 ++----
 1 files changed, 2 insertions(+), 4 deletions(-)
lib/kobject.c +2 -4
index c033dc8..74b8dbc 100644
@@ -195,8 +195,7 @@ int kobject_add(struct kobject * kobj)
 	if (error) {
 		/* unlink does the kobject_put() for us */
 		unlink(kobj);
-		if (parent)
-			kobject_put(parent);
+		kobject_put(parent);
 
 		/* be noisy on error issues */
 		if (error == -EEXIST)
@@ -420,8 +419,7 @@ void kobject_cleanup(struct kobject * kobj)
 		t->release(kobj);
 	if (s)
 		kset_put(s);
-	if (parent) 
-		kobject_put(parent);
+	kobject_put(parent);
 }
 
 static void kobject_release(struct kref *kref)
Signature · 1 line
1.4.4.4
#21 lore raw

PATCH 21/28 sysfs: error handling in sysfs, fill_read_buffer()

From: Oliver Neukum <[email protected]>

if a driver returns an error in fill_read_buffer(), the buffer will be marked as filled. Subsequent reads will return eof. But there is no data because of an error, not because it has been read. Not marking the buffer filled is the obvious fix.

---
 fs/sysfs/file.c |   10 ++++++----
 1 files changed, 6 insertions(+), 4 deletions(-)
fs/sysfs/file.c +6 -4
index 46618f8..c0e1176 100644
@@ -83,7 +83,8 @@ remove_from_collection(struct sysfs_buffer *buffer, struct inode *node)
  *	Allocate @buffer->page, if it hasn't been already, then call the
  *	kobject's show() method to fill the buffer with this attribute's 
  *	data. 
- *	This is called only once, on the file's first read. 
+ *	This is called only once, on the file's first read unless an error
+ *	is returned.
  */
 static int fill_read_buffer(struct dentry * dentry, struct sysfs_buffer * buffer)
 {
@@ -101,12 +102,13 @@ static int fill_read_buffer(struct dentry * dentry, struct sysfs_buffer * buffer
 
 	buffer->event = atomic_read(&sd->s_event);
 	count = ops->show(kobj,attr,buffer->page);
-	buffer->needs_read_fill = 0;
 	BUG_ON(count > (ssize_t)PAGE_SIZE);
-	if (count >= 0)
+	if (count >= 0) {
+		buffer->needs_read_fill = 0;
 		buffer->count = count;
-	else
+	} else {
 		ret = count;
+	}
 	return ret;
 }
 
Signature · 1 line
1.4.4.4
#22 lore raw

PATCH 22/28 HOWTO: Add a reference to Harbison and Steele

From: Robert P. J. Day <[email protected]>

Add a reference to Harbison and Steele's C book.

---
 Documentation/HOWTO |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
Documentation/HOWTO +1 -0
index 8d51c14..48123db 100644
@@ -30,6 +30,7 @@ are not a good substitute for a solid C education and/or years of
 experience, the following books are good for, if anything, reference:
  - "The C Programming Language" by Kernighan and Ritchie [Prentice Hall]
  - "Practical C Programming" by Steve Oualline [O'Reilly]
+ - "C:  A Reference Manual" by Harbison and Steele [Prentice Hall]
 
 The kernel is written using GNU C and the GNU toolchain.  While it
 adheres to the ISO C89 standard, it uses a number of extensions that are
Signature · 1 line
1.4.4.4
#23 lore raw

PATCH 23/28 SYSFS: Fix missing include of list.h in sysfs.h

From: Frank Haverkamp <[email protected]>

Sysfs.h uses definitions (e.g. struct list_head s_sibling) from list.h but does not include it.

---
 include/linux/sysfs.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
include/linux/sysfs.h +1 -0
index 2129d1b..eee4859 100644
@@ -11,6 +11,7 @@
 #define _SYSFS_H_
 
 #include <linux/compiler.h>
+#include <linux/list.h>
 #include <asm/atomic.h>
 
 struct kobject;
Signature · 1 line
1.4.4.4
#24 lore raw

PATCH 24/28 Driver core: add uevent vars for devices of a class

From: Kay Sievers <[email protected]>

Devices converted from class_device to device should have the same uevent keys as the original class_device had. We search up the parents until we find the first bus device and add the (already deprecated) PHYDEV* values.

---
 drivers/base/core.c |   46 ++++++++++++++++++++++++++++++++++------------
 1 files changed, 34 insertions(+), 12 deletions(-)
drivers/base/core.c +34 -12
index 6fee3e6..7a5336f 100644
@@ -154,25 +154,47 @@ static int dev_uevent(struct kset *kset, struct kobject *kobj, char **envp,
 			       "MINOR=%u", MINOR(dev->devt));
 	}
 
-#ifdef CONFIG_SYSFS_DEPRECATED
-	/* add bus name (same as SUBSYSTEM, deprecated) */
-	if (dev->bus)
-		add_uevent_var(envp, num_envp, &i,
-			       buffer, buffer_size, &length,
-			       "PHYSDEVBUS=%s", dev->bus->name);
-#endif
-
-	/* add driver name (PHYSDEV* values are deprecated)*/
-	if (dev->driver) {
+	if (dev->driver)
 		add_uevent_var(envp, num_envp, &i,
 			       buffer, buffer_size, &length,
 			       "DRIVER=%s", dev->driver->name);
+
 #ifdef CONFIG_SYSFS_DEPRECATED
+	if (dev->class) {
+		struct device *parent = dev->parent;
+
+		/* find first bus device in parent chain */
+		while (parent && !parent->bus)
+			parent = parent->parent;
+		if (parent && parent->bus) {
+			const char *path;
+
+			path = kobject_get_path(&parent->kobj, GFP_KERNEL);
+			add_uevent_var(envp, num_envp, &i,
+				       buffer, buffer_size, &length,
+				       "PHYSDEVPATH=%s", path);
+			kfree(path);
+
+			add_uevent_var(envp, num_envp, &i,
+				       buffer, buffer_size, &length,
+				       "PHYSDEVBUS=%s", parent->bus->name);
+
+			if (parent->driver)
+				add_uevent_var(envp, num_envp, &i,
+					       buffer, buffer_size, &length,
+					       "PHYSDEVDRIVER=%s", parent->driver->name);
+		}
+	} else if (dev->bus) {
 		add_uevent_var(envp, num_envp, &i,
 			       buffer, buffer_size, &length,
-			       "PHYSDEVDRIVER=%s", dev->driver->name);
-#endif
+			       "PHYSDEVBUS=%s", dev->bus->name);
+
+		if (dev->driver)
+			add_uevent_var(envp, num_envp, &i,
+				       buffer, buffer_size, &length,
+				       "PHYSDEVDRIVER=%s", dev->driver->name);
 	}
+#endif
 
 	/* terminate, set to next free slot, shrink available space */
 	envp[i] = NULL;
Signature · 1 line
1.4.4.4
#25 lore raw

PATCH 25/28 Driver core: add device_type to struct device

From: Kay Sievers <[email protected]>

This allows us to add type specific attributes, uevent vars and release funtions.

A subsystem can carry different types of devices like the "block" subsys has disks and partitions. Both types create a different set of attributes, but belong to the same subsystem.

This corresponds to the low level objects:

  kobject   -> device       (object/device data)
  kobj_type -> device_type  (type of object/device we are embedded in)
  kset      -> class/bus    (list of objects/devices of a subsystem)
---
 drivers/base/core.c          |   57 ++++++++++++++++++++++++++++-------------
 drivers/usb/input/hid-lgff.c |    4 +-
 include/linux/device.h       |    8 ++++++
 3 files changed, 49 insertions(+), 20 deletions(-)
drivers/base/core.c +39 -18
index 7a5336f..34ac187 100644
@@ -95,6 +95,8 @@ static void device_release(struct kobject * kobj)
 
 	if (dev->release)
 		dev->release(dev);
+	else if (dev->type && dev->type->release)
+		dev->type->release(dev);
 	else if (dev->class && dev->class->dev_release)
 		dev->class->dev_release(dev);
 	else {
@@ -206,19 +208,25 @@ static int dev_uevent(struct kset *kset, struct kobject *kobj, char **envp,
 	if (dev->bus && dev->bus->uevent) {
 		/* have the bus specific function add its stuff */
 		retval = dev->bus->uevent(dev, envp, num_envp, buffer, buffer_size);
-			if (retval) {
-			pr_debug ("%s - uevent() returned %d\n",
+		if (retval)
+			pr_debug ("%s: bus uevent() returned %d\n",
 				  __FUNCTION__, retval);
-		}
 	}
 
 	if (dev->class && dev->class->dev_uevent) {
 		/* have the class specific function add its stuff */
 		retval = dev->class->dev_uevent(dev, envp, num_envp, buffer, buffer_size);
-			if (retval) {
-				pr_debug("%s - dev_uevent() returned %d\n",
-					 __FUNCTION__, retval);
-		}
+		if (retval)
+			pr_debug("%s: class uevent() returned %d\n",
+				 __FUNCTION__, retval);
+	}
+
+	if (dev->type && dev->type->uevent) {
+		/* have the device type specific fuction add its stuff */
+		retval = dev->type->uevent(dev, envp, num_envp, buffer, buffer_size);
+		if (retval)
+			pr_debug("%s: dev_type uevent() returned %d\n",
+				 __FUNCTION__, retval);
 	}
 
 	return retval;
@@ -269,37 +277,50 @@ static void device_remove_groups(struct device *dev)
 static int device_add_attrs(struct device *dev)
 {
 	struct class *class = dev->class;
+	struct device_type *type = dev->type;
 	int error = 0;
 	int i;
 
-	if (!class)
-		return 0;
-
-	if (class->dev_attrs) {
+	if (class && class->dev_attrs) {
 		for (i = 0; attr_name(class->dev_attrs[i]); i++) {
 			error = device_create_file(dev, &class->dev_attrs[i]);
 			if (error)
 				break;
 		}
+		if (error)
+			while (--i >= 0)
+				device_remove_file(dev, &class->dev_attrs[i]);
 	}
-	if (error)
-		while (--i >= 0)
-			device_remove_file(dev, &class->dev_attrs[i]);
+
+	if (type && type->attrs) {
+		for (i = 0; attr_name(type->attrs[i]); i++) {
+			error = device_create_file(dev, &type->attrs[i]);
+			if (error)
+				break;
+		}
+		if (error)
+			while (--i >= 0)
+				device_remove_file(dev, &type->attrs[i]);
+	}
+
 	return error;
 }
 
 static void device_remove_attrs(struct device *dev)
 {
 	struct class *class = dev->class;
+	struct device_type *type = dev->type;
 	int i;
 
-	if (!class)
-		return;
-
-	if (class->dev_attrs) {
+	if (class && class->dev_attrs) {
 		for (i = 0; attr_name(class->dev_attrs[i]); i++)
 			device_remove_file(dev, &class->dev_attrs[i]);
 	}
+
+	if (type && type->attrs) {
+		for (i = 0; attr_name(type->attrs[i]); i++)
+			device_remove_file(dev, &type->attrs[i]);
+	}
 }
 
 
drivers/usb/input/hid-lgff.c +2 -2
index e474662..4f4fc3b 100644
@@ -32,7 +32,7 @@
 #include <linux/hid.h>
 #include "usbhid.h"
 
-struct device_type {
+struct dev_type {
 	u16 idVendor;
 	u16 idProduct;
 	const signed short *ff;
@@ -48,7 +48,7 @@ static const signed short ff_joystick[] = {
 	-1
 };
 
-static const struct device_type devices[] = {
+static const struct dev_type devices[] = {
 	{ 0x046d, 0xc211, ff_rumble },
 	{ 0x046d, 0xc219, ff_rumble },
 	{ 0x046d, 0xc283, ff_joystick },
include/linux/device.h +8 -0
index da72219..e1e164f 100644
@@ -328,6 +328,13 @@ extern struct class_device *class_device_create(struct class *cls,
 					__attribute__((format(printf,5,6)));
 extern void class_device_destroy(struct class *cls, dev_t devt);
 
+struct device_type {
+	struct device_attribute *attrs;
+	int (*uevent)(struct device *dev, char **envp, int num_envp,
+		      char *buffer, int buffer_size);
+	void (*release)(struct device *dev);
+};
+
 /* interface for exporting device attributes */
 struct device_attribute {
 	struct attribute	attr;
@@ -356,6 +363,7 @@ struct device {
 
 	struct kobject kobj;
 	char	bus_id[BUS_ID_SIZE];	/* position on parent bus */
+	struct device_type	*type;
 	unsigned		is_registered:1;
 	struct device_attribute uevent_attr;
 	struct device_attribute *devt_attr;
Signature · 1 line
1.4.4.4
#26 lore raw

PATCH 26/28 Driver core: allow to delay the uevent at device creation time

From: Kay Sievers <[email protected]>

For the block subsystem, we want to delay all uevents until the disk has been scanned and allpartitons are already created before the first event is sent out.

---
 drivers/base/core.c    |    3 ++-
 include/linux/device.h |    3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)
drivers/base/core.c +2 -1
index 34ac187..e136142 100644
@@ -589,7 +589,8 @@ int device_add(struct device *dev)
 		goto PMError;
 	if ((error = bus_add_device(dev)))
 		goto BusError;
-	kobject_uevent(&dev->kobj, KOBJ_ADD);
+	if (!dev->uevent_suppress)
+		kobject_uevent(&dev->kobj, KOBJ_ADD);
 	if ((error = bus_attach_device(dev)))
 		goto AttachError;
 	if (parent)
include/linux/device.h +2 -1
index e1e164f..5ca1cdb 100644
@@ -399,9 +399,10 @@ struct device {
 
 	/* class_device migration path */
 	struct list_head	node;
-	struct class		*class;		/* optional*/
+	struct class		*class;
 	dev_t			devt;		/* dev_t, creates the sysfs "dev" */
 	struct attribute_group	**groups;	/* optional groups */
+	int			uevent_suppress;
 
 	void	(*release)(struct device * dev);
 };
Signature · 1 line
1.4.4.4
#27 lore raw

PATCH 27/28 Driver Core: Increase the default timeout value of the firmware subsystem

From: Dave Jones <[email protected]>

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=174589

The ipw driver sometimes takes a long time to load its firmware. Whilst the ipw driver should be using the async interface of the firmware loader to make this a non-issue, this is a minimal fix.

---
 drivers/base/firmware_class.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/base/firmware_class.c +1 -1
index 64558f4..c0a979a 100644
@@ -35,7 +35,7 @@ enum {
 	FW_STATUS_READY_NOHOTPLUG,
 };
 
-static int loading_timeout = 10;	/* In seconds */
+static int loading_timeout = 60;	/* In seconds */
 
 /* fw_lock could be moved to 'struct firmware_priv' but since it is just
  * guarding for corner cases a global lock should be OK */
Signature · 1 line
1.4.4.4
#28 lore raw

PATCH 28/28 sysfs: Shadow directory support

From: Eric W. Biederman <[email protected]>

The problem. When implementing a network namespace I need to be able to have multiple network devices with the same name. Currently this is a problem for /sys/class/net/*.

What I want is a separate /sys/class/net directory in sysfs for each network namespace, and I want to name each of them /sys/class/net.

I looked and the VFS actually allows that. All that is needed is for /sys/class/net to implement a follow link method to redirect lookups to the real directory you want.

Implementing a follow link method that is sensitive to the current network namespace turns out to be 3 lines of code so it looks like a clean approach. Modifying sysfs so it doesn't get in my was is a bit trickier.

I am calling the concept of multiple directories all at the same path in the filesystem shadow directories. With the directory entry really at that location the shadow master.

The following patch modifies sysfs so it can handle a directory structure slightly different from the kobject tree so I can implement the shadow directories for handling /sys/class/net/.

---
 fs/sysfs/dir.c          |  207 +++++++++++++++++++++++++++++++++++++++--------
 fs/sysfs/group.c        |    1 +
 fs/sysfs/inode.c        |   10 +++
 fs/sysfs/mount.c        |    2 +-
 fs/sysfs/sysfs.h        |    5 +
 include/linux/kobject.h |    4 +
 include/linux/sysfs.h   |   23 +++++-
 lib/kobject.c           |   42 +++++++++-
 8 files changed, 249 insertions(+), 45 deletions(-)
fs/sysfs/dir.c +172 -35
index 9ff0449..9dcdf55 100644
@@ -33,8 +33,7 @@ static struct dentry_operations sysfs_dentry_ops = {
 /*
  * Allocates a new sysfs_dirent and links it to the parent sysfs_dirent
  */
-static struct sysfs_dirent * sysfs_new_dirent(struct sysfs_dirent * parent_sd,
-						void * element)
+static struct sysfs_dirent * __sysfs_new_dirent(void * element)
 {
 	struct sysfs_dirent * sd;
 
@@ -46,12 +45,28 @@ static struct sysfs_dirent * sysfs_new_dirent(struct sysfs_dirent * parent_sd,
 	atomic_set(&sd->s_count, 1);
 	atomic_set(&sd->s_event, 1);
 	INIT_LIST_HEAD(&sd->s_children);
-	list_add(&sd->s_sibling, &parent_sd->s_children);
+	INIT_LIST_HEAD(&sd->s_sibling);
 	sd->s_element = element;
 
 	return sd;
 }
 
+static void __sysfs_list_dirent(struct sysfs_dirent *parent_sd,
+			      struct sysfs_dirent *sd)
+{
+	if (sd)
+		list_add(&sd->s_sibling, &parent_sd->s_children);
+}
+
+static struct sysfs_dirent * sysfs_new_dirent(struct sysfs_dirent *parent_sd,
+						void * element)
+{
+	struct sysfs_dirent *sd;
+	sd = __sysfs_new_dirent(element);
+	__sysfs_list_dirent(parent_sd, sd);
+	return sd;
+}
+
 /*
  *
  * Return -EEXIST if there is already a sysfs element with the same name for
@@ -78,14 +93,14 @@ int sysfs_dirent_exist(struct sysfs_dirent *parent_sd,
 }
 
 
-int sysfs_make_dirent(struct sysfs_dirent * parent_sd, struct dentry * dentry,
-			void * element, umode_t mode, int type)
+static struct sysfs_dirent *
+__sysfs_make_dirent(struct dentry *dentry, void *element, mode_t mode, int type)
 {
 	struct sysfs_dirent * sd;
 
-	sd = sysfs_new_dirent(parent_sd, element);
+	sd = __sysfs_new_dirent(element);
 	if (!sd)
-		return -ENOMEM;
+		goto out;
 
 	sd->s_mode = mode;
 	sd->s_type = type;
@@ -95,7 +110,19 @@ int sysfs_make_dirent(struct sysfs_dirent * parent_sd, struct dentry * dentry,
 		dentry->d_op = &sysfs_dentry_ops;
 	}
 
-	return 0;
+out:
+	return sd;
+}
+
+int sysfs_make_dirent(struct sysfs_dirent * parent_sd, struct dentry * dentry,
+			void * element, umode_t mode, int type)
+{
+	struct sysfs_dirent *sd;
+
+	sd = __sysfs_make_dirent(dentry, element, mode, type);
+	__sysfs_list_dirent(parent_sd, sd);
+
+	return sd ? 0 : -ENOMEM;
 }
 
 static int init_dir(struct inode * inode)
@@ -166,11 +193,11 @@ int sysfs_create_subdir(struct kobject * k, const char * n, struct dentry ** d)
 
 /**
  *	sysfs_create_dir - create a directory for an object.
- *	@parent:	parent parent object.
  *	@kobj:		object we're creating directory for. 
+ *	@shadow_parent:	parent parent object.
  */
 
-int sysfs_create_dir(struct kobject * kobj)
+int sysfs_create_dir(struct kobject * kobj, struct dentry *shadow_parent)
 {
 	struct dentry * dentry = NULL;
 	struct dentry * parent;
@@ -178,7 +205,9 @@ int sysfs_create_dir(struct kobject * kobj)
 
 	BUG_ON(!kobj);
 
-	if (kobj->parent)
+	if (shadow_parent)
+		parent = shadow_parent;
+	else if (kobj->parent)
 		parent = kobj->parent->dentry;
 	else if (sysfs_mount && sysfs_mount->mnt_sb)
 		parent = sysfs_mount->mnt_sb->s_root;
@@ -299,21 +328,12 @@ void sysfs_remove_subdir(struct dentry * d)
 }
 
 
-/**
- *	sysfs_remove_dir - remove an object's directory.
- *	@kobj:	object. 
- *
- *	The only thing special about this is that we remove any files in 
- *	the directory before we remove the directory, and we've inlined
- *	what used to be sysfs_rmdir() below, instead of calling separately.
- */
-
-void sysfs_remove_dir(struct kobject * kobj)
+static void __sysfs_remove_dir(struct dentry *dentry)
 {
-	struct dentry * dentry = dget(kobj->dentry);
 	struct sysfs_dirent * parent_sd;
 	struct sysfs_dirent * sd, * tmp;
 
+	dget(dentry);
 	if (!dentry)
 		return;
 
@@ -334,32 +354,60 @@ void sysfs_remove_dir(struct kobject * kobj)
 	 * Drop reference from dget() on entrance.
 	 */
 	dput(dentry);
+}
+
+/**
+ *	sysfs_remove_dir - remove an object's directory.
+ *	@kobj:	object.
+ *
+ *	The only thing special about this is that we remove any files in
+ *	the directory before we remove the directory, and we've inlined
+ *	what used to be sysfs_rmdir() below, instead of calling separately.
+ */
+
+void sysfs_remove_dir(struct kobject * kobj)
+{
+	__sysfs_remove_dir(kobj->dentry);
 	kobj->dentry = NULL;
 }
 
-int sysfs_rename_dir(struct kobject * kobj, const char *new_name)
+int sysfs_rename_dir(struct kobject * kobj, struct dentry *new_parent,
+		     const char *new_name)
 {
 	int error = 0;
-	struct dentry * new_dentry, * parent;
+	struct dentry * new_dentry;
 
-	if (!strcmp(kobject_name(kobj), new_name))
-		return -EINVAL;
-
-	if (!kobj->parent)
-		return -EINVAL;
+	if (!new_parent)
+		return -EFAULT;
 
 	down_write(&sysfs_rename_sem);
-	parent = kobj->parent->dentry;
+	mutex_lock(&new_parent->d_inode->i_mutex);
 
-	mutex_lock(&parent->d_inode->i_mutex);
-
-	new_dentry = lookup_one_len(new_name, parent, strlen(new_name));
+	new_dentry = lookup_one_len(new_name, new_parent, strlen(new_name));
 	if (!IS_ERR(new_dentry)) {
-  		if (!new_dentry->d_inode) {
+		/* By allowing two different directories with the
+		 * same d_parent we allow this routine to move
+		 * between different shadows of the same directory
+		 */
+		if (kobj->dentry->d_parent->d_inode != new_parent->d_inode)
+			return -EINVAL;
+		else if (new_dentry->d_parent->d_inode != new_parent->d_inode)
+			error = -EINVAL;
+		else if (new_dentry == kobj->dentry)
+			error = -EINVAL;
+		else if (!new_dentry->d_inode) {
 			error = kobject_set_name(kobj, "%s", new_name);
 			if (!error) {
+				struct sysfs_dirent *sd, *parent_sd;
+
 				d_add(new_dentry, NULL);
 				d_move(kobj->dentry, new_dentry);
+
+				sd = kobj->dentry->d_fsdata;
+				parent_sd = new_parent->d_fsdata;
+
+				list_del_init(&sd->s_sibling);
+				list_add(&sd->s_sibling, &parent_sd->s_children);
 			}
 			else
 				d_drop(new_dentry);
@@ -367,7 +415,7 @@ int sysfs_rename_dir(struct kobject * kobj, const char *new_name)
 			error = -EEXIST;
 		dput(new_dentry);
 	}
-	mutex_unlock(&parent->d_inode->i_mutex);
+	mutex_unlock(&new_parent->d_inode->i_mutex);
 	up_write(&sysfs_rename_sem);
 
 	return error;
@@ -546,6 +594,95 @@ static loff_t sysfs_dir_lseek(struct file * file, loff_t offset, int origin)
 	return offset;
 }
 
+
+/**
+ *	sysfs_make_shadowed_dir - Setup so a directory can be shadowed
+ *	@kobj:	object we're creating shadow of.
+ */
+
+int sysfs_make_shadowed_dir(struct kobject *kobj,
+	void * (*follow_link)(struct dentry *, struct nameidata *))
+{
+	struct inode *inode;
+	struct inode_operations *i_op;
+
+	inode = kobj->dentry->d_inode;
+	if (inode->i_op != &sysfs_dir_inode_operations)
+		return -EINVAL;
+
+	i_op = kmalloc(sizeof(*i_op), GFP_KERNEL);
+	if (!i_op)
+		return -ENOMEM;
+
+	memcpy(i_op, &sysfs_dir_inode_operations, sizeof(*i_op));
+	i_op->follow_link = follow_link;
+
+	/* Locking of inode->i_op?
+	 * Since setting i_op is a single word write and they
+	 * are atomic we should be ok here.
+	 */
+	inode->i_op = i_op;
+	return 0;
+}
+
+/**
+ *	sysfs_create_shadow_dir - create a shadow directory for an object.
+ *	@kobj:	object we're creating directory for.
+ *
+ *	sysfs_make_shadowed_dir must already have been called on this
+ *	directory.
+ */
+
+struct dentry *sysfs_create_shadow_dir(struct kobject *kobj)
+{
+	struct sysfs_dirent *sd;
+	struct dentry *parent, *dir, *shadow;
+	struct inode *inode;
+
+	dir = kobj->dentry;
+	inode = dir->d_inode;
+	parent = dir->d_parent;
+	shadow = ERR_PTR(-EINVAL);
+	if (!sysfs_is_shadowed_inode(inode))
+		goto out;
+
+	shadow = d_alloc(parent, &dir->d_name);
+	if (!shadow)
+		goto nomem;
+
+	sd = __sysfs_make_dirent(shadow, kobj, inode->i_mode, SYSFS_DIR);
+	if (!sd)
+		goto nomem;
+
+	d_instantiate(shadow, igrab(inode));
+	inc_nlink(inode);
+	inc_nlink(parent->d_inode);
+	shadow->d_op = &sysfs_dentry_ops;
+
+	dget(shadow);		/* Extra count - pin the dentry in core */
+
+out:
+	return shadow;
+nomem:
+	dput(shadow);
+	shadow = ERR_PTR(-ENOMEM);
+	goto out;
+}
+
+/**
+ *	sysfs_remove_shadow_dir - remove an object's directory.
+ *	@shadow: dentry of shadow directory
+ *
+ *	The only thing special about this is that we remove any files in
+ *	the directory before we remove the directory, and we've inlined
+ *	what used to be sysfs_rmdir() below, instead of calling separately.
+ */
+
+void sysfs_remove_shadow_dir(struct dentry *shadow)
+{
+	__sysfs_remove_dir(shadow);
+}
+
 const struct file_operations sysfs_dir_operations = {
 	.open		= sysfs_dir_open,
 	.release	= sysfs_dir_close,
fs/sysfs/group.c +1 -0
index 46a277b..b20951c 100644
@@ -13,6 +13,7 @@
 #include <linux/dcache.h>
 #include <linux/namei.h>
 #include <linux/err.h>
+#include <linux/fs.h>
 #include <asm/semaphore.h>
 #include "sysfs.h"
 
fs/sysfs/inode.c +10 -0
index dbd820f..542d2bc 100644
@@ -33,6 +33,16 @@ static struct inode_operations sysfs_inode_operations ={
 	.setattr	= sysfs_setattr,
 };
 
+void sysfs_delete_inode(struct inode *inode)
+{
+	/* Free the shadowed directory inode operations */
+	if (sysfs_is_shadowed_inode(inode)) {
+		kfree(inode->i_op);
+		inode->i_op = NULL;
+	}
+	return generic_delete_inode(inode);
+}
+
 int sysfs_setattr(struct dentry * dentry, struct iattr * iattr)
 {
 	struct inode * inode = dentry->d_inode;
fs/sysfs/mount.c +1 -1
index a1a58b9..f6a87a8 100644
@@ -23,7 +23,7 @@ static void sysfs_clear_inode(struct inode *inode);
 
 static struct super_operations sysfs_ops = {
 	.statfs		= simple_statfs,
-	.drop_inode	= generic_delete_inode,
+	.drop_inode	= sysfs_delete_inode,
 	.clear_inode	= sysfs_clear_inode,
 };
 
fs/sysfs/sysfs.h +5 -0
index 39c623f..fe1cbfd 100644
@@ -2,6 +2,7 @@
 extern struct vfsmount * sysfs_mount;
 extern struct kmem_cache *sysfs_dir_cachep;
 
+extern void sysfs_delete_inode(struct inode *inode);
 extern struct inode * sysfs_new_inode(mode_t mode, struct sysfs_dirent *);
 extern int sysfs_create(struct dentry *, int mode, int (*init)(struct inode *));
 
@@ -112,3 +113,7 @@ static inline void sysfs_put(struct sysfs_dirent * sd)
 		release_sysfs_dirent(sd);
 }
 
+static inline int sysfs_is_shadowed_inode(struct inode *inode)
+{
+	return S_ISDIR(inode->i_mode) && inode->i_op->follow_link;
+}
include/linux/kobject.h +4 -0
index 76538fc..b850e03 100644
@@ -74,9 +74,13 @@ extern void kobject_init(struct kobject *);
 extern void kobject_cleanup(struct kobject *);
 
 extern int __must_check kobject_add(struct kobject *);
+extern int __must_check kobject_shadow_add(struct kobject *, struct dentry *);
 extern void kobject_del(struct kobject *);
 
 extern int __must_check kobject_rename(struct kobject *, const char *new_name);
+extern int __must_check kobject_shadow_rename(struct kobject *kobj,
+						struct dentry *new_parent,
+						const char *new_name);
 extern int __must_check kobject_move(struct kobject *, struct kobject *);
 
 extern int __must_check kobject_register(struct kobject *);
include/linux/sysfs.h +19 -4
index eee4859..192de3a 100644
@@ -16,6 +16,7 @@
 
 struct kobject;
 struct module;
+struct nameidata;
 
 struct attribute {
 	const char		* name;
@@ -89,13 +90,13 @@ struct sysfs_dirent {
 #ifdef CONFIG_SYSFS
 
 extern int __must_check
-sysfs_create_dir(struct kobject *);
+sysfs_create_dir(struct kobject *, struct dentry *);
 
 extern void
 sysfs_remove_dir(struct kobject *);
 
 extern int __must_check
-sysfs_rename_dir(struct kobject *, const char *new_name);
+sysfs_rename_dir(struct kobject *, struct dentry *, const char *new_name);
 
 extern int __must_check
 sysfs_move_dir(struct kobject *, struct kobject *);
@@ -127,11 +128,17 @@ int __must_check sysfs_create_group(struct kobject *,
 void sysfs_remove_group(struct kobject *, const struct attribute_group *);
 void sysfs_notify(struct kobject * k, char *dir, char *attr);
 
+
+extern int sysfs_make_shadowed_dir(struct kobject *kobj,
+	void * (*follow_link)(struct dentry *, struct nameidata *));
+extern struct dentry *sysfs_create_shadow_dir(struct kobject *kobj);
+extern void sysfs_remove_shadow_dir(struct dentry *dir);
+
 extern int __must_check sysfs_init(void);
 
 #else /* CONFIG_SYSFS */
 
-static inline int sysfs_create_dir(struct kobject * k)
+static inline int sysfs_create_dir(struct kobject * k, struct dentry *shadow)
 {
 	return 0;
 }
@@ -141,7 +148,9 @@ static inline void sysfs_remove_dir(struct kobject * k)
 	;
 }
 
-static inline int sysfs_rename_dir(struct kobject * k, const char *new_name)
+static inline int sysfs_rename_dir(struct kobject * k,
+					struct dentry *new_parent,
+					const char *new_name)
 {
 	return 0;
 }
@@ -205,6 +214,12 @@ static inline void sysfs_notify(struct kobject * k, char *dir, char *attr)
 {
 }
 
+static inline int sysfs_make_shadowed_dir(struct kobject *kobj,
+	void * (*follow_link)(struct dentry *, struct nameidata *))
+{
+	return 0;
+}
+
 static inline int __must_check sysfs_init(void)
 {
 	return 0;
lib/kobject.c +37 -5

77 lines not on this page · raw message

Signature · 1 line
1.4.4.4
maintainer
#29 lore raw

Re: [PATCH 10/28] SERIO: add the sysfs driver name to all modules

In reply to Greg KH

Hi Greg,

Quoted 5 lines · Greg KH

From: Greg Kroah-Hartman <[email protected]>

This adds the module name to all SERIO drivers, if they are built into the kernel or not. It will show up in /sys/modules/MODULE_NAME/drivers/

What is this needed for?

Signature · 1 line
Dmitry
#30 lore raw

Re: [PATCH 10/28] SERIO: add the sysfs driver name to all modules

Quoted 10 lines · Dmitry Torokhov

Hi Greg,

Quoted 5 lines · Greg KH

From: Greg Kroah-Hartman <[email protected]>

This adds the module name to all SERIO drivers, if they are built into the kernel or not. It will show up in /sys/modules/MODULE_NAME/drivers/

What is this needed for?

All modules show up today in /sys/module/ no matter if they are built into the kernel or not.

But, if they are built as a module, the /sys/module/drivers/ link was present. This is not consistant, as we always want to see that link, no matter if the driver is built in or not. So with this change, we now always see the links.

I also fixed the PCI and USB drivers to do this as well.

Hope this helps explain it.

thanks,

greg k-h