Skip to contentSkip to conversation

pm_register() and pm_send_all()

DISCUSSION Andrew Morton 6 participants 14 messages Mar 12, 2008 – Mar 21, 2008

Subsystems (5)
x86 MIPS Suspend to RAM Hibernation +1 more subsystems
lore
Share & tools
lore
#0 lore raw

This:

kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178)

is getting a bit boring. These were marked deprecated at least three years ago.

pm_register() no longer has any callers and afacit can be removed immediately.

pm_send_all() has just a handful of callers:

arch/x86/kernel/apm_32.c
arch/mips/au1000/common/power.c
arch/frv/kernel/pm.c

could we please get a bit of a push on this? How should these pm_send_all() calls be replaced?

Thanks.

(I almost have a warning-free x86_64 allmodconfig, but the one in drivers/message/i2o/i2o_config.c is going to be a problem).

#1 lore raw
Quoted 15 lines · Andrew Morton

This:

kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178)

is getting a bit boring. These were marked deprecated at least three years ago.

pm_register() no longer has any callers and afacit can be removed immediately.

Quoted 8 lines

pm_send_all() has just a handful of callers:

arch/x86/kernel/apm_32.c
arch/mips/au1000/common/power.c
arch/frv/kernel/pm.c

could we please get a bit of a push on this? How should these pm_send_all() calls be replaced?

Well, I was going to do that some time ago, but I'm not really familiar with the old code. Some assistance would be welcome.

Thanks,
Rafael

#2 lore raw
In reply to Andrew Morton
Quoted 12 lines · Andrew Morton

This:

kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178)

is getting a bit boring. These were marked deprecated at least three years ago.

Quoted 8 lines

pm_register() no longer has any callers and afacit can be removed immediately.

pm_send_all() has just a handful of callers:

arch/x86/kernel/apm_32.c
arch/mips/au1000/common/power.c
arch/frv/kernel/pm.c

AFAICT pm_send_all is a nop when noone uses pm_register...

Hmm.. can we just force CONFIG_PM_LEGACY=n, and see what happens?

Or maybe this is better idea? It may break build somewhere, but it should be easy to fix... (it builds here, i386 and x86-64).

arch/frv/kernel/pm.c +0 -8
index c57ce3f..73f3aee 100644
@@ -163,14 +163,11 @@ static int sysctl_pm_do_suspend(ctl_tabl
 	if ((mode != 1) && (mode != 5))
 		return -EINVAL;
 
-	retval = pm_send_all(PM_SUSPEND, (void *)3);
-
 	if (retval == 0) {
 		if (mode == 5)
 		    retval = pm_do_bus_sleep();
 		else
 		    retval = pm_do_suspend();
-		pm_send_all(PM_RESUME, (void *)0);
 	}
 
 	return retval;
@@ -183,9 +180,6 @@ static int try_set_cmode(int new_cmode)
 	if (!(clock_cmodes_permitted & (1<<new_cmode)))
 		return -EINVAL;
 
-	/* tell all the drivers we're suspending */
-	pm_send_all(PM_SUSPEND, (void *)3);
-
 	/* now change cmode */
 	local_irq_disable();
 	frv_dma_pause_all();
@@ -201,8 +195,6 @@ #endif
 	frv_dma_resume_all();
 	local_irq_enable();
 
-	/* tell all the drivers we're resuming */
-	pm_send_all(PM_RESUME, (void *)0);
 	return 0;
 }
 
arch/mips/au1000/common/power.c +2 -15
index 54047d6..a9f7f63 100644
@@ -258,7 +258,6 @@ #endif
 static int pm_do_sleep(ctl_table * ctl, int write, struct file *file,
 		       void __user *buffer, size_t * len, loff_t *ppos)
 {
-	int retval = 0;
 #ifdef SLEEP_TEST_TIMEOUT
 #define TMPBUFLEN2 16
 	char buf[TMPBUFLEN2], *p;
@@ -278,33 +277,21 @@ #ifdef SLEEP_TEST_TIMEOUT
 		p = buf;
 		sleep_ticks = simple_strtoul(p, &p, 0);
 #endif
-		retval = pm_send_all(PM_SUSPEND, (void *) 2);
-
-		if (retval)
-			return retval;
 
 		au_sleep();
-		retval = pm_send_all(PM_RESUME, (void *) 0);
 	}
-	return retval;
+	return 0;
 }
 
 static int pm_do_suspend(ctl_table * ctl, int write, struct file *file,
 			 void __user *buffer, size_t * len, loff_t *ppos)
 {
-	int retval = 0;
-
 	if (!write) {
 		*len = 0;
 	} else {
-		retval = pm_send_all(PM_SUSPEND, (void *) 2);
-		if (retval)
-			return retval;
 		suspend_mode = 1;
-
-		retval = pm_send_all(PM_RESUME, (void *) 0);
 	}
-	return retval;
+	return 0;
 }
 
 
arch/x86/kernel/apm_32.c +0 -15
index d4438ef..d7e92bf 100644
@@ -1189,19 +1189,6 @@ static int suspend(int vetoable)
 	int err;
 	struct apm_user	*as;
 
-	if (pm_send_all(PM_SUSPEND, (void *)3)) {
-		/* Vetoed */
-		if (vetoable) {
-			if (apm_info.connection_version > 0x100)
-				set_system_power_state(APM_STATE_REJECT);
-			err = -EBUSY;
-			ignore_sys_suspend = 0;
-			printk(KERN_WARNING "apm: suspend was vetoed.\n");
-			goto out;
-		}
-		printk(KERN_CRIT "apm: suspend was vetoed, but suspending anyway.\n");
-	}
-
 	device_suspend(PMSG_SUSPEND);
 	local_irq_disable();
 	device_power_down(PMSG_SUSPEND);
@@ -1224,7 +1211,6 @@ static int suspend(int vetoable)
 	device_power_up();
 	local_irq_enable();
 	device_resume();
-	pm_send_all(PM_RESUME, (void *)0);
 	queue_event(APM_NORMAL_RESUME, NULL);
  out:
 	spin_lock(&user_list_lock);
@@ -1337,7 +1323,6 @@ #endif
 			if ((event != APM_NORMAL_RESUME)
 			    || (ignore_normal_resume == 0)) {
 				device_resume();
-				pm_send_all(PM_RESUME, (void *)0);
 				queue_event(event, NULL);
 			}
 			ignore_normal_resume = 0;
kernel/power/Kconfig +0 -10
index e2d7b39..d6d8f8b 100644
@@ -19,16 +19,6 @@ config PM
 	  will issue the hlt instruction if nothing is to be done, thereby
 	  sending the processor to sleep and saving power.
 
-config PM_LEGACY
-	bool "Legacy Power Management API (DEPRECATED)"
-	depends on PM
-	default n
-	---help---
-	   Support for pm_register() and friends.  This old API is obsoleted
-	   by the driver model.
-
-	   If unsure, say N.
-
 config PM_DEBUG
 	bool "Power Management Debug Support"
 	depends on PM
kernel/power/Makefile +0 -1
index e5693d6..523abb9 100644
@@ -4,7 +4,6 @@ EXTRA_CFLAGS	+=	-DDEBUG
 endif
 
 obj-y				:= main.o
-obj-$(CONFIG_PM_LEGACY)		+= pm.o
 obj-$(CONFIG_PM_SLEEP)		+= process.o console.o sleepy.o
 obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
 
kernel/power/pm.c deleted +0 -205

205 lines not on this page · raw message

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#3 lore raw
Quoted 6 lines · Pavel Machek

AFAICT pm_send_all is a nop when noone uses pm_register...

Hmm.. can we just force CONFIG_PM_LEGACY=n, and see what happens?

Or maybe this is better idea? It may break build somewhere, but it should be easy to fix... (it builds here, i386 and x86-64).

Looks correct to me, thus ACK.

Spinning that a little further I wonder how useful the rest of the Alchemy code - which seems to only deal with devices - is. If at all. I'm going to check with some Alchemy users if that code actually works or can be removed. I think the latter. Also there is the APM emulation available which provides similar functionality.

Ralf

#4 lore raw
Quoted 7 lines · Ralf Baechle

Looks correct to me, thus ACK.

Spinning that a little further I wonder how useful the rest of the Alchemy code - which seems to only deal with devices - is. If at all. I'm going to check with some Alchemy users if that code actually works or can be removed. I think the latter. Also there is the APM emulation available which provides similar functionality.

So here a patch do cleanup a part. Probably more to come

From: Ralf Baechle <[email protected]>

[MIPS] Alchey: Crapectomy after removal of pm_send_all calls.

 arch/mips/au1000/common/power.c |   20 --------------------
 1 file changed, 20 deletions(-)
arch/mips/au1000/common/power.c +0 -20
Index: linux-mips/arch/mips/au1000/common/power.c
===================================================================
@@ -283,18 +283,6 @@ static int pm_do_sleep(ctl_table * ctl, 
 	return 0;
 }
 
-static int pm_do_suspend(ctl_table * ctl, int write, struct file *file,
-			 void __user *buffer, size_t * len, loff_t *ppos)
-{
-	if (!write) {
-		*len = 0;
-	} else {
-		suspend_mode = 1;
-	}
-	return 0;
-}
-
-
 static int pm_do_freq(ctl_table * ctl, int write, struct file *file,
 		      void __user *buffer, size_t * len, loff_t *ppos)
 {
@@ -408,14 +396,6 @@ static int pm_do_freq(ctl_table * ctl, i
 
 static struct ctl_table pm_table[] = {
 	{
-		.ctl_name 	= CTL_UNNUMBERED,
-		.procname	= "suspend",
-		.data		= NULL,
-		.maxlen		= 0,
-		.mode		= 0600,
-		.proc_handler	= &pm_do_suspend
-	},
-	{
 		.ctl_name	= CTL_UNNUMBERED,
 		.procname	= "sleep",
 		.data		= NULL,
#5 lore raw
Quoted 10 lines · Ralf Baechle
Quoted 7 lines · Ralf Baechle

Looks correct to me, thus ACK.

Spinning that a little further I wonder how useful the rest of the Alchemy code - which seems to only deal with devices - is. If at all. I'm going to check with some Alchemy users if that code actually works or can be removed. I think the latter. Also there is the APM emulation available which provides similar functionality.

Quoted 7 lines

So here a patch do cleanup a part. Probably more to come

From: Ralf Baechle <[email protected]>

[MIPS] Alchey: Crapectomy after removal of pm_send_all calls.

ACK.

Quoted 42 lines
 arch/mips/au1000/common/power.c |   20 --------------------
 1 file changed, 20 deletions(-)

Index: linux-mips/arch/mips/au1000/common/power.c
===================================================================

arch/mips/au1000/common/power.c +0 -20
@@ -283,18 +283,6 @@ static int pm_do_sleep(ctl_table * ctl, 
 	return 0;
 }
 
-static int pm_do_suspend(ctl_table * ctl, int write, struct file *file,
-			 void __user *buffer, size_t * len, loff_t *ppos)
-{
-	if (!write) {
-		*len = 0;
-	} else {
-		suspend_mode = 1;
-	}
-	return 0;
-}
-
-
 static int pm_do_freq(ctl_table * ctl, int write, struct file *file,
 		      void __user *buffer, size_t * len, loff_t *ppos)
 {
@@ -408,14 +396,6 @@ static int pm_do_freq(ctl_table * ctl, i
 
 static struct ctl_table pm_table[] = {
 	{
-		.ctl_name 	= CTL_UNNUMBERED,
-		.procname	= "suspend",
-		.data		= NULL,
-		.maxlen		= 0,
-		.mode		= 0600,
-		.proc_handler	= &pm_do_suspend
-	},
-	{
 		.ctl_name	= CTL_UNNUMBERED,
 		.procname	= "sleep",
 		.data		= NULL,
Signature · 3 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
pomozte zachranit klanovicky les:  http://www.ujezdskystrom.info/
#6 lore raw
In reply to Pavel Machek
Quoted 16 lines · Pavel Machek
Quoted 15 lines · Andrew Morton

This:

kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178)

is getting a bit boring. These were marked deprecated at least three years ago.

pm_register() no longer has any callers and afacit can be removed immediately.

Quoted 7 lines
Quoted 5 lines

pm_send_all() has just a handful of callers:

arch/x86/kernel/apm_32.c
arch/mips/au1000/common/power.c
arch/frv/kernel/pm.c

AFAICT pm_send_all is a nop when noone uses pm_register...

Apparently on 2.6.24 it's not used anywehere, at least according to LXR.

Quoted 1 line

Hmm.. can we just force CONFIG_PM_LEGACY=n, and see what happens?

Do you want me to push this patch and the following one from Ralf to Len?

Quoted 375 lines

Or maybe this is better idea? It may break build somewhere, but it should be easy to fix... (it builds here, i386 and x86-64).

arch/frv/kernel/pm.c +0 -8
index c57ce3f..73f3aee 100644
@@ -163,14 +163,11 @@ static int sysctl_pm_do_suspend(ctl_tabl
 	if ((mode != 1) && (mode != 5))
 		return -EINVAL;
 
-	retval = pm_send_all(PM_SUSPEND, (void *)3);
-
 	if (retval == 0) {
 		if (mode == 5)
 		    retval = pm_do_bus_sleep();
 		else
 		    retval = pm_do_suspend();
-		pm_send_all(PM_RESUME, (void *)0);
 	}
 
 	return retval;
@@ -183,9 +180,6 @@ static int try_set_cmode(int new_cmode)
 	if (!(clock_cmodes_permitted & (1<<new_cmode)))
 		return -EINVAL;
 
-	/* tell all the drivers we're suspending */
-	pm_send_all(PM_SUSPEND, (void *)3);
-
 	/* now change cmode */
 	local_irq_disable();
 	frv_dma_pause_all();
@@ -201,8 +195,6 @@ #endif
 	frv_dma_resume_all();
 	local_irq_enable();
 
-	/* tell all the drivers we're resuming */
-	pm_send_all(PM_RESUME, (void *)0);
 	return 0;
 }
 
arch/mips/au1000/common/power.c +2 -15
index 54047d6..a9f7f63 100644
@@ -258,7 +258,6 @@ #endif
 static int pm_do_sleep(ctl_table * ctl, int write, struct file *file,
 		       void __user *buffer, size_t * len, loff_t *ppos)
 {
-	int retval = 0;
 #ifdef SLEEP_TEST_TIMEOUT
 #define TMPBUFLEN2 16
 	char buf[TMPBUFLEN2], *p;
@@ -278,33 +277,21 @@ #ifdef SLEEP_TEST_TIMEOUT
 		p = buf;
 		sleep_ticks = simple_strtoul(p, &p, 0);
 #endif
-		retval = pm_send_all(PM_SUSPEND, (void *) 2);
-
-		if (retval)
-			return retval;
 
 		au_sleep();
-		retval = pm_send_all(PM_RESUME, (void *) 0);
 	}
-	return retval;
+	return 0;
 }
 
 static int pm_do_suspend(ctl_table * ctl, int write, struct file *file,
 			 void __user *buffer, size_t * len, loff_t *ppos)
 {
-	int retval = 0;
-
 	if (!write) {
 		*len = 0;
 	} else {
-		retval = pm_send_all(PM_SUSPEND, (void *) 2);
-		if (retval)
-			return retval;
 		suspend_mode = 1;
-
-		retval = pm_send_all(PM_RESUME, (void *) 0);
 	}
-	return retval;
+	return 0;
 }
 
 
arch/x86/kernel/apm_32.c +0 -15
index d4438ef..d7e92bf 100644
@@ -1189,19 +1189,6 @@ static int suspend(int vetoable)
 	int err;
 	struct apm_user	*as;
 
-	if (pm_send_all(PM_SUSPEND, (void *)3)) {
-		/* Vetoed */
-		if (vetoable) {
-			if (apm_info.connection_version > 0x100)
-				set_system_power_state(APM_STATE_REJECT);
-			err = -EBUSY;
-			ignore_sys_suspend = 0;
-			printk(KERN_WARNING "apm: suspend was vetoed.\n");
-			goto out;
-		}
-		printk(KERN_CRIT "apm: suspend was vetoed, but suspending anyway.\n");
-	}
-
 	device_suspend(PMSG_SUSPEND);
 	local_irq_disable();
 	device_power_down(PMSG_SUSPEND);
@@ -1224,7 +1211,6 @@ static int suspend(int vetoable)
 	device_power_up();
 	local_irq_enable();
 	device_resume();
-	pm_send_all(PM_RESUME, (void *)0);
 	queue_event(APM_NORMAL_RESUME, NULL);
  out:
 	spin_lock(&user_list_lock);
@@ -1337,7 +1323,6 @@ #endif
 			if ((event != APM_NORMAL_RESUME)
 			    || (ignore_normal_resume == 0)) {
 				device_resume();
-				pm_send_all(PM_RESUME, (void *)0);
 				queue_event(event, NULL);
 			}
 			ignore_normal_resume = 0;
kernel/power/Kconfig +0 -10
index e2d7b39..d6d8f8b 100644
@@ -19,16 +19,6 @@ config PM
 	  will issue the hlt instruction if nothing is to be done, thereby
 	  sending the processor to sleep and saving power.
 
-config PM_LEGACY
-	bool "Legacy Power Management API (DEPRECATED)"
-	depends on PM
-	default n
-	---help---
-	   Support for pm_register() and friends.  This old API is obsoleted
-	   by the driver model.
-
-	   If unsure, say N.
-
 config PM_DEBUG
 	bool "Power Management Debug Support"
 	depends on PM
kernel/power/Makefile +0 -1
index e5693d6..523abb9 100644
@@ -4,7 +4,6 @@ EXTRA_CFLAGS	+=	-DDEBUG
 endif
 
 obj-y				:= main.o
-obj-$(CONFIG_PM_LEGACY)		+= pm.o
 obj-$(CONFIG_PM_SLEEP)		+= process.o console.o sleepy.o
 obj-$(CONFIG_HIBERNATION)	+= swsusp.o disk.o snapshot.o swap.o user.o
 
kernel/power/pm.c deleted +0 -205
deleted file mode 100644
index 60c73fa..0000000
@@ -1,205 +0,0 @@
-/*
- *  pm.c - Power management interface
- *
- *  Copyright (C) 2000 Andrew Henroid
- *
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to the Free Software
- *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- */
-#include <linux/init.h>
-#include <linux/module.h>
-#include <linux/spinlock.h>
-#include <linux/mm.h>
-#include <linux/slab.h>
-#include <linux/pm.h>
-#include <linux/pm_legacy.h>
-#include <linux/interrupt.h>
-#include <linux/mutex.h>
-
-/*
- *	Locking notes:
- *		pm_devs_lock can be a semaphore providing pm ops are not called
- *	from an interrupt handler (already a bad idea so no change here). Each
- *	change must be protected so that an unlink of an entry doesn't clash
- *	with a pm send - which is permitted to sleep in the current architecture
- *
- *	Module unloads clashing with pm events now work out safely, the module 
- *	unload path will block until the event has been sent. It may well block
- *	until a resume but that will be fine.
- */
- 
-static DEFINE_MUTEX(pm_devs_lock);
-static LIST_HEAD(pm_devs);
-
-/**
- *	pm_register - register a device with power management
- *	@type: device type 
- *	@id: device ID
- *	@callback: callback function
- *
- *	Add a device to the list of devices that wish to be notified about
- *	power management events. A &pm_dev structure is returned on success,
- *	on failure the return is %NULL.
- *
- *      The callback function will be called in process context and
- *      it may sleep.
- */
- 
-struct pm_dev *pm_register(pm_dev_t type,
-			   unsigned long id,
-			   pm_callback callback)
-{
-	struct pm_dev *dev = kzalloc(sizeof(struct pm_dev), GFP_KERNEL);
-	if (dev) {
-		dev->type = type;
-		dev->id = id;
-		dev->callback = callback;
-
-		mutex_lock(&pm_devs_lock);
-		list_add(&dev->entry, &pm_devs);
-		mutex_unlock(&pm_devs_lock);
-	}
-	return dev;
-}
-
-/**
- *	pm_send - send request to a single device
- *	@dev: device to send to
- *	@rqst: power management request
- *	@data: data for the callback
- *
- *	Issue a power management request to a given device. The 
- *	%PM_SUSPEND and %PM_RESUME events are handled specially. The
- *	data field must hold the intended next state. No call is made
- *	if the state matches.
- *
- *	BUGS: what stops two power management requests occurring in parallel
- *	and conflicting.
- *
- *	WARNING: Calling pm_send directly is not generally recommended, in
- *	particular there is no locking against the pm_dev going away. The
- *	caller must maintain all needed locking or have 'inside knowledge'
- *	on the safety. Also remember that this function is not locked against
- *	pm_unregister. This means that you must handle SMP races on callback
- *	execution and unload yourself.
- */
- 
-static int pm_send(struct pm_dev *dev, pm_request_t rqst, void *data)
-{
-	int status = 0;
-	unsigned long prev_state, next_state;
-
-	if (in_interrupt())
-		BUG();
-
-	switch (rqst) {
-	case PM_SUSPEND:
-	case PM_RESUME:
-		prev_state = dev->state;
-		next_state = (unsigned long) data;
-		if (prev_state != next_state) {
-			if (dev->callback)
-				status = (*dev->callback)(dev, rqst, data);
-			if (!status) {
-				dev->state = next_state;
-				dev->prev_state = prev_state;
-			}
-		}
-		else {
-			dev->prev_state = prev_state;
-		}
-		break;
-	default:
-		if (dev->callback)
-			status = (*dev->callback)(dev, rqst, data);
-		break;
-	}
-	return status;
-}
-
-/*
- * Undo incomplete request
- */
-static void pm_undo_all(struct pm_dev *last)
-{
-	struct list_head *entry = last->entry.prev;
-	while (entry != &pm_devs) {
-		struct pm_dev *dev = list_entry(entry, struct pm_dev, entry);
-		if (dev->state != dev->prev_state) {
-			/* previous state was zero (running) resume or
-			 * previous state was non-zero (suspended) suspend
-			 */
-			pm_request_t undo = (dev->prev_state
-					     ? PM_SUSPEND:PM_RESUME);
-			pm_send(dev, undo, (void*) dev->prev_state);
-		}
-		entry = entry->prev;
-	}
-}
-
-/**
- *	pm_send_all - send request to all managed devices
- *	@rqst: power management request
- *	@data: data for the callback
- *
- *	Issue a power management request to a all devices. The 
- *	%PM_SUSPEND events are handled specially. Any device is 
- *	permitted to fail a suspend by returning a non zero (error)
- *	value from its callback function. If any device vetoes a 
- *	suspend request then all other devices that have suspended 
- *	during the processing of this request are restored to their
- *	previous state.
- *
- *	WARNING:  This function takes the pm_devs_lock. The lock is not dropped until
- *	the callbacks have completed. This prevents races against pm locking
- *	functions, races against module unload pm_unregister code. It does
- *	mean however that you must not issue pm_ functions within the callback
- *	or you will deadlock and users will hate you.
- *
- *	Zero is returned on success. If a suspend fails then the status
- *	from the device that vetoes the suspend is returned.
- *
- *	BUGS: what stops two power management requests occurring in parallel
- *	and conflicting.
- */
- 
-int pm_send_all(pm_request_t rqst, void *data)
-{
-	struct list_head *entry;
-	
-	mutex_lock(&pm_devs_lock);
-	entry = pm_devs.next;
-	while (entry != &pm_devs) {
-		struct pm_dev *dev = list_entry(entry, struct pm_dev, entry);
-		if (dev->callback) {
-			int status = pm_send(dev, rqst, data);
-			if (status) {
-				/* return devices to previous state on
-				 * failed suspend request
-				 */
-				if (rqst == PM_SUSPEND)
-					pm_undo_all(dev);
-				mutex_unlock(&pm_devs_lock);
-				return status;
-			}
-		}
-		entry = entry->next;
-	}
-	mutex_unlock(&pm_devs_lock);
-	return 0;
-}
-
-EXPORT_SYMBOL(pm_register);
-EXPORT_SYMBOL(pm_send_all);
-
#7 lore raw
Quoted 23 lines · Rafael J. Wysocki
Quoted 22 lines · Pavel Machek
Quoted 20 lines · Andrew Morton

This:

kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:203: warning: 'pm_register' is deprecated (declared at kernel/power/pm.c:62) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178) kernel/power/pm.c:204: warning: 'pm_send_all' is deprecated (declared at kernel/power/pm.c:178)

is getting a bit boring. These were marked deprecated at least three years ago.

pm_register() no longer has any callers and afacit can be removed immediately.

pm_send_all() has just a handful of callers:

arch/x86/kernel/apm_32.c
arch/mips/au1000/common/power.c
arch/frv/kernel/pm.c

Quoted 8 lines
Quoted 1 line

AFAICT pm_send_all is a nop when noone uses pm_register...

Apparently on 2.6.24 it's not used anywehere, at least according to LXR.

Quoted 1 line

Hmm.. can we just force CONFIG_PM_LEGACY=n, and see what happens?

Do you want me to push this patch and the following one from Ralf to Len?

Yes... I think we can kill pm_send_all etc...

Signature · 3 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
pomozte zachranit klanovicky les:  http://www.ujezdskystrom.info/
#10 lore raw
Quoted 6 lines · On Thu, 13 Mar 2008 12:31:37 +0000 David Howells <[email protected]> wrote:
Quoted 1 line · Andrew Morton

pm_send_all() has just a handful of callers:

What should be done instead of calling pm_send_all()?

I don't know, and neither does Rafael. Hopefully Pavel does?

#11 lore raw
In reply to David Howells
Quoted 5 lines · David Howells
Quoted 1 line · Andrew Morton

pm_send_all() has just a handful of callers:

What should be done instead of calling pm_send_all()?

pm_send_all is currently nop.

device_suspend/resume are closest equivalents.

Signature · 3 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
pomozte zachranit klanovicky les:  http://www.ujezdskystrom.info/
#12 lore raw
Quoted 8 lines · On Thu, 13 Mar 2008 19:05:58 +0100 Pavel Machek <[email protected]> wrote:
Quoted 5 lines · David Howells
Quoted 1 line · Andrew Morton

pm_send_all() has just a handful of callers:

What should be done instead of calling pm_send_all()?

pm_send_all is currently nop.

So we could have deleted pm_send_all() and pm_register() ages ago? And we've all been needlessly staring at those warnings all this time? Grumble.

Could we please have a volunteer to take care of making these things go away asap?

#13 lore raw
Quoted 8 lines · Andrew Morton
Quoted 5 lines · On Thu, 13 Mar 2008 19:05:58 +0100 Pavel Machek <[email protected]> wrote:
Quoted 4 lines · David Howells
Quoted 1 line · Andrew Morton

pm_send_all() has just a handful of callers:

Quoted 7 lines
Quoted 3 lines
Quoted 1 line

What should be done instead of calling pm_send_all()?

pm_send_all is currently nop.

So we could have deleted pm_send_all() and pm_register() ages ago? And we've all been needlessly staring at those warnings all this time? Grumble.

we could have deleted pm_send_all when users of pm_register were gone, yes.... just nobody noticed :-(.
Pavel

Signature · 3 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
pomozte zachranit klanovicky les:  http://www.ujezdskystrom.info/