Skip to contentSkip to conversation

PM / Hibernate: Fix memory corruption related to swap

PATCH In review Rafael J. Wysocki 4 participants 13 messages Dec 2, 2010 – Dec 3, 2010

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 PM / Hibernate: Fix memory corruption related to swap

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

There is a problem that swap pages allocated before the creation of a hibernation image can be released and used for storing the contents of different memory pages while the image is being saved. Since the kernel stored in the image doesn't know of that, it causes memory corruption to occur after resume from hibernation, especially on systems with relatively small RAM that need to swap often.

This issue can be addressed by keeping the GFP_IOFS bits clear in gfp_allowed_mask during the entire hibernation, including the saving of the image, until the system is finally turned off or the hibernation is aborted. Unfortunately, for this purpose it's necessary to rework the way in which the hibernate and suspend code manipulates gfp_allowed_mask.

This change is based on an earlier patch from Hugh Dickins.

---

Hi,

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

Thanks,
Rafael

---
 include/linux/gfp.h      |    4 ++--
 kernel/power/hibernate.c |   22 ++++++++++++----------
 kernel/power/suspend.c   |    5 ++---
 kernel/power/user.c      |    2 ++
 mm/page_alloc.c          |   19 ++++++++++++-------
 5 files changed, 30 insertions(+), 22 deletions(-)
kernel/power/hibernate.c +12 -10
Index: linux-2.6/kernel/power/hibernate.c
===================================================================
@@ -327,7 +327,6 @@ static int create_image(int platform_mod
 int hibernation_snapshot(int platform_mode)
 {
 	int error;
-	gfp_t saved_mask;
 
 	error = platform_begin(platform_mode);
 	if (error)
@@ -339,7 +338,7 @@ int hibernation_snapshot(int platform_mo
 		goto Close;
 
 	suspend_console();
-	saved_mask = clear_gfp_allowed_mask(GFP_IOFS);
+	pm_restrict_gfp_mask();
 	error = dpm_suspend_start(PMSG_FREEZE);
 	if (error)
 		goto Recover_platform;
@@ -348,7 +347,10 @@ int hibernation_snapshot(int platform_mo
 		goto Recover_platform;
 
 	error = create_image(platform_mode);
-	/* Control returns here after successful restore */
+	/*
+	 * Control returns here (1) after the image has been created or the
+	 * image creation has failed and (2) after a successful restore.
+	 */
 
  Resume_devices:
 	/* We may need to release the preallocated image pages here. */
@@ -357,7 +359,10 @@ int hibernation_snapshot(int platform_mo
 
 	dpm_resume_end(in_suspend ?
 		(error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE);
-	set_gfp_allowed_mask(saved_mask);
+
+	if (error || !in_suspend)
+		pm_restore_gfp_mask();
+
 	resume_console();
  Close:
 	platform_end(platform_mode);
@@ -452,17 +457,16 @@ static int resume_target_kernel(bool pla
 int hibernation_restore(int platform_mode)
 {
 	int error;
-	gfp_t saved_mask;
 
 	pm_prepare_console();
 	suspend_console();
-	saved_mask = clear_gfp_allowed_mask(GFP_IOFS);
+	pm_restrict_gfp_mask();
 	error = dpm_suspend_start(PMSG_QUIESCE);
 	if (!error) {
 		error = resume_target_kernel(platform_mode);
 		dpm_resume_end(PMSG_RECOVER);
 	}
-	set_gfp_allowed_mask(saved_mask);
+	pm_restore_gfp_mask();
 	resume_console();
 	pm_restore_console();
 	return error;
@@ -476,7 +480,6 @@ int hibernation_restore(int platform_mod
 int hibernation_platform_enter(void)
 {
 	int error;
-	gfp_t saved_mask;
 
 	if (!hibernation_ops)
 		return -ENOSYS;
@@ -492,7 +495,6 @@ int hibernation_platform_enter(void)
 
 	entering_platform_hibernation = true;
 	suspend_console();
-	saved_mask = clear_gfp_allowed_mask(GFP_IOFS);
 	error = dpm_suspend_start(PMSG_HIBERNATE);
 	if (error) {
 		if (hibernation_ops->recover)
@@ -536,7 +538,6 @@ int hibernation_platform_enter(void)
  Resume_devices:
 	entering_platform_hibernation = false;
 	dpm_resume_end(PMSG_RESTORE);
-	set_gfp_allowed_mask(saved_mask);
 	resume_console();
 
  Close:
@@ -647,6 +648,7 @@ int hibernate(void)
 		if (!error)
 			power_down();
 		in_suspend = 0;
+		pm_restore_gfp_mask();
 	} else {
 		pr_debug("PM: Image restored successfully.\n");
 	}
kernel/power/user.c +2 -0
Index: linux-2.6/kernel/power/user.c
===================================================================
@@ -263,6 +263,7 @@ static long snapshot_ioctl(struct file *
 	case SNAPSHOT_UNFREEZE:
 		if (!data->frozen || data->ready)
 			break;
+		pm_restore_gfp_mask();
 		thaw_processes();
 		usermodehelper_enable();
 		data->frozen = 0;
@@ -275,6 +276,7 @@ static long snapshot_ioctl(struct file *
 			error = -EPERM;
 			break;
 		}
+		pm_restore_gfp_mask();
 		error = hibernation_snapshot(data->platform_support);
 		if (!error)
 			error = put_user(in_suspend, (int __user *)arg);
mm/page_alloc.c +12 -7
Index: linux-2.6/mm/page_alloc.c
===================================================================
@@ -104,19 +104,24 @@ gfp_t gfp_allowed_mask __read_mostly = G
  * only be modified with pm_mutex held, unless the suspend/hibernate code is
  * guaranteed not to run in parallel with that modification).
  */
-void set_gfp_allowed_mask(gfp_t mask)
+
+static gfp_t saved_gfp_mask;
+
+void pm_restore_gfp_mask(void)
 {
 	WARN_ON(!mutex_is_locked(&pm_mutex));
-	gfp_allowed_mask = mask;
+	if (saved_gfp_mask) {
+		gfp_allowed_mask = saved_gfp_mask;
+		saved_gfp_mask = 0;
+	}
 }
 
-gfp_t clear_gfp_allowed_mask(gfp_t mask)
+void pm_restrict_gfp_mask(void)
 {
-	gfp_t ret = gfp_allowed_mask;
-
 	WARN_ON(!mutex_is_locked(&pm_mutex));
-	gfp_allowed_mask &= ~mask;
-	return ret;
+	WARN_ON(saved_gfp_mask);
+	saved_gfp_mask = gfp_allowed_mask;
+	gfp_allowed_mask &= ~GFP_IOFS;
 }
 #endif /* CONFIG_PM_SLEEP */
 
include/linux/gfp.h +2 -2
Index: linux-2.6/include/linux/gfp.h
===================================================================
@@ -360,7 +360,7 @@ void drain_local_pages(void *dummy);
 
 extern gfp_t gfp_allowed_mask;
 
-extern void set_gfp_allowed_mask(gfp_t mask);
-extern gfp_t clear_gfp_allowed_mask(gfp_t mask);
+extern void pm_restrict_gfp_mask(void);
+extern void pm_restore_gfp_mask(void);
 
 #endif /* __LINUX_GFP_H */
kernel/power/suspend.c +2 -3
Index: linux-2.6/kernel/power/suspend.c
===================================================================
@@ -206,7 +206,6 @@ static int suspend_enter(suspend_state_t
 int suspend_devices_and_enter(suspend_state_t state)
 {
 	int error;
-	gfp_t saved_mask;
 
 	if (!suspend_ops)
 		return -ENOSYS;
@@ -217,7 +216,7 @@ int suspend_devices_and_enter(suspend_st
 			goto Close;
 	}
 	suspend_console();
-	saved_mask = clear_gfp_allowed_mask(GFP_IOFS);
+	pm_restrict_gfp_mask();
 	suspend_test_start();
 	error = dpm_suspend_start(PMSG_SUSPEND);
 	if (error) {
@@ -234,7 +233,7 @@ int suspend_devices_and_enter(suspend_st
 	suspend_test_start();
 	dpm_resume_end(PMSG_RESUME);
 	suspend_test_finish("resume devices");
-	set_gfp_allowed_mask(saved_mask);
+	pm_restore_gfp_mask();
 	resume_console();
  Close:
 	if (suspend_ops->end)
maintainer
#1 lore raw
Quoted 21 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

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

There is a problem that swap pages allocated before the creation of a hibernation image can be released and used for storing the contents of different memory pages while the image is being saved. Since the kernel stored in the image doesn't know of that, it causes memory corruption to occur after resume from hibernation, especially on systems with relatively small RAM that need to swap often.

This issue can be addressed by keeping the GFP_IOFS bits clear in gfp_allowed_mask during the entire hibernation, including the saving of the image, until the system is finally turned off or the hibernation is aborted. Unfortunately, for this purpose it's necessary to rework the way in which the hibernate and suspend code manipulates gfp_allowed_mask.

This change is based on an earlier patch from Hugh Dickins.

Quoted 7 lines

---

Hi,

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

It looks OK to me for 2.6.37 but for 2.6.38 please let's make everything here a 100% no-op for CONFIG_PM_SLEEP=n builds. Specifically the slight overhead in __alloc_pages_nodemask.

Because given the global nature of saved_gfp_mask and the unlocked way in which it is accessed, this facility won't be at all useful for anything other than suspend.

maintainer
#2 lore raw
In reply to Rafael J. Wysocki
Quoted 21 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

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

There is a problem that swap pages allocated before the creation of a hibernation image can be released and used for storing the contents of different memory pages while the image is being saved. Since the kernel stored in the image doesn't know of that, it causes memory corruption to occur after resume from hibernation, especially on systems with relatively small RAM that need to swap often.

This issue can be addressed by keeping the GFP_IOFS bits clear in gfp_allowed_mask during the entire hibernation, including the saving of the image, until the system is finally turned off or the hibernation is aborted. Unfortunately, for this purpose it's necessary to rework the way in which the hibernate and suspend code manipulates gfp_allowed_mask.

This change is based on an earlier patch from Hugh Dickins.

Quoted 7 lines

---

Hi,

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

It looks OK to me for 2.6.37 but for 2.6.38 please let's make everything here a 100% no-op for CONFIG_PM_SLEEP=n builds. Specifically the slight overhead in __alloc_pages_nodemask.

Because given the global nature of saved_gfp_mask and the unlocked way in which it is accessed, this facility won't be at all useful for anything other than suspend.

#3 lore raw
Quoted 28 lines · Andrew Morton
Quoted 25 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

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

There is a problem that swap pages allocated before the creation of a hibernation image can be released and used for storing the contents of different memory pages while the image is being saved. Since the kernel stored in the image doesn't know of that, it causes memory corruption to occur after resume from hibernation, especially on systems with relatively small RAM that need to swap often.

This issue can be addressed by keeping the GFP_IOFS bits clear in gfp_allowed_mask during the entire hibernation, including the saving of the image, until the system is finally turned off or the hibernation is aborted. Unfortunately, for this purpose it's necessary to rework the way in which the hibernate and suspend code manipulates gfp_allowed_mask.

This change is based on an earlier patch from Hugh Dickins.

---

Hi,

Quoted 7 lines
Quoted 3 lines

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

It looks OK to me for 2.6.37 but for 2.6.38 please let's make everything here a 100% no-op for CONFIG_PM_SLEEP=n builds. Specifically the slight overhead in __alloc_pages_nodemask.

OK, I'll take care of it.

Quoted 3 lines

Because given the global nature of saved_gfp_mask and the unlocked way in which it is accessed, this facility won't be at all useful for anything other than suspend.

Right.

#4 lore raw
In reply to Andrew Morton
Quoted 28 lines · Andrew Morton
Quoted 25 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

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

There is a problem that swap pages allocated before the creation of a hibernation image can be released and used for storing the contents of different memory pages while the image is being saved. Since the kernel stored in the image doesn't know of that, it causes memory corruption to occur after resume from hibernation, especially on systems with relatively small RAM that need to swap often.

This issue can be addressed by keeping the GFP_IOFS bits clear in gfp_allowed_mask during the entire hibernation, including the saving of the image, until the system is finally turned off or the hibernation is aborted. Unfortunately, for this purpose it's necessary to rework the way in which the hibernate and suspend code manipulates gfp_allowed_mask.

This change is based on an earlier patch from Hugh Dickins.

---

Hi,

Quoted 7 lines
Quoted 3 lines

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

It looks OK to me for 2.6.37 but for 2.6.38 please let's make everything here a 100% no-op for CONFIG_PM_SLEEP=n builds. Specifically the slight overhead in __alloc_pages_nodemask.

OK, I'll take care of it.

Quoted 3 lines

Because given the global nature of saved_gfp_mask and the unlocked way in which it is accessed, this facility won't be at all useful for anything other than suspend.

Right.

#5 lore raw
In reply to Andrew Morton
Quoted 5 lines · Andrew Morton
Quoted 3 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

(It will need backport to 2.6.35-stable and 2.6.36-stable:
IIRC it doesn't quite apply cleanly to those, so we'll need to send a separate version.)

Quoted 4 lines

It looks OK to me for 2.6.37 but for 2.6.38 please let's make everything here a 100% no-op for CONFIG_PM_SLEEP=n builds. Specifically the slight overhead in __alloc_pages_nodemask.

I expect you're right that the CONFIG_PM would better be CONFIG__PM_SLEEP; but I think you misunderstand gfp_allowed_mask in __alloc_pages_nodemask:
it came from slab & slub, to fix some early bootup issues, and predates Rafael's recent use of it in suspend and hibernation.

Quoted 4 lines

Because given the global nature of saved_gfp_mask and the unlocked way in which it is accessed, this facility won't be at all useful for anything other than suspend.

... and bootup.

Hugh

#6 lore raw
In reply to Andrew Morton
Quoted 5 lines · Andrew Morton
Quoted 3 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

(It will need backport to 2.6.35-stable and 2.6.36-stable:
IIRC it doesn't quite apply cleanly to those, so we'll need to send a separate version.)

Quoted 4 lines

It looks OK to me for 2.6.37 but for 2.6.38 please let's make everything here a 100% no-op for CONFIG_PM_SLEEP=n builds. Specifically the slight overhead in __alloc_pages_nodemask.

I expect you're right that the CONFIG_PM would better be CONFIG__PM_SLEEP; but I think you misunderstand gfp_allowed_mask in __alloc_pages_nodemask:
it came from slab & slub, to fix some early bootup issues, and predates Rafael's recent use of it in suspend and hibernation.

Quoted 4 lines

Because given the global nature of saved_gfp_mask and the unlocked way in which it is accessed, this facility won't be at all useful for anything other than suspend.

... and bootup.

Hugh

#7 lore raw
Quoted 10 lines · Hugh Dickins
Quoted 5 lines · Andrew Morton
Quoted 3 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

(It will need backport to 2.6.35-stable and 2.6.36-stable:
IIRC it doesn't quite apply cleanly to those, so we'll need to send a separate version.)

That's correct. I'll backport it when it hits the Linus' tree.

Rafael

#8 lore raw
In reply to Hugh Dickins
Quoted 10 lines · Hugh Dickins
Quoted 5 lines · Andrew Morton
Quoted 3 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

(It will need backport to 2.6.35-stable and 2.6.36-stable:
IIRC it doesn't quite apply cleanly to those, so we'll need to send a separate version.)

That's correct. I'll backport it when it hits the Linus' tree.

Rafael

maintainer
#9 lore raw
In reply to Hugh Dickins
Quoted 22 lines · On Thu, 2 Dec 2010 16:37:21 -0800 (PST) Hugh Dickins <[email protected]> wrote:
Quoted 5 lines · Andrew Morton
Quoted 3 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

(It will need backport to 2.6.35-stable and 2.6.36-stable:
IIRC it doesn't quite apply cleanly to those, so we'll need to send a separate version.)

Quoted 4 lines

It looks OK to me for 2.6.37 but for 2.6.38 please let's make everything here a 100% no-op for CONFIG_PM_SLEEP=n builds. Specifically the slight overhead in __alloc_pages_nodemask.

I expect you're right that the CONFIG_PM would better be CONFIG__PM_SLEEP; but I think you misunderstand gfp_allowed_mask in __alloc_pages_nodemask:
it came from slab & slub, to fix some early bootup issues, and predates Rafael's recent use of it in suspend and hibernation.

Quoted 1 line
Quoted 6 lines
Quoted 3 lines

Because given the global nature of saved_gfp_mask and the unlocked way in which it is accessed, this facility won't be at all useful for anything other than suspend.

... and bootup.

Oh. Who writes this crap :(

maintainer
#10 lore raw
In reply to Hugh Dickins
Quoted 22 lines · On Thu, 2 Dec 2010 16:37:21 -0800 (PST) Hugh Dickins <[email protected]> wrote:
Quoted 5 lines · Andrew Morton
Quoted 3 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

This patch is a fix for a regression and nasty memory corruption, so I'd like to push it to Linus for 2.6.37 if there are no objections.

(It will need backport to 2.6.35-stable and 2.6.36-stable:
IIRC it doesn't quite apply cleanly to those, so we'll need to send a separate version.)

Quoted 4 lines

It looks OK to me for 2.6.37 but for 2.6.38 please let's make everything here a 100% no-op for CONFIG_PM_SLEEP=n builds. Specifically the slight overhead in __alloc_pages_nodemask.

I expect you're right that the CONFIG_PM would better be CONFIG__PM_SLEEP; but I think you misunderstand gfp_allowed_mask in __alloc_pages_nodemask:
it came from slab & slub, to fix some early bootup issues, and predates Rafael's recent use of it in suspend and hibernation.

Quoted 1 line
Quoted 6 lines
Quoted 3 lines

Because given the global nature of saved_gfp_mask and the unlocked way in which it is accessed, this facility won't be at all useful for anything other than suspend.

... and bootup.

Oh. Who writes this crap :(

Reviewed-by on v1 1/1
#11 lore raw
In reply to Rafael J. Wysocki
Quoted 16 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

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

There is a problem that swap pages allocated before the creation of a hibernation image can be released and used for storing the contents of different memory pages while the image is being saved. Since the kernel stored in the image doesn't know of that, it causes memory corruption to occur after resume from hibernation, especially on systems with relatively small RAM that need to swap often.

This issue can be addressed by keeping the GFP_IOFS bits clear in gfp_allowed_mask during the entire hibernation, including the saving of the image, until the system is finally turned off or the hibernation is aborted. Unfortunately, for this purpose it's necessary to rework the way in which the hibernate and suspend code manipulates gfp_allowed_mask.

Quoted 5 lines

This change is based on an earlier patch from Hugh Dickins.

Thanks,
-Kame

Reviewed-by on v1 1/1
#12 lore raw
In reply to Rafael J. Wysocki
Quoted 16 lines · On Fri, 3 Dec 2010 00:40:36 +0100 "Rafael J. Wysocki" <[email protected]> wrote:

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

There is a problem that swap pages allocated before the creation of a hibernation image can be released and used for storing the contents of different memory pages while the image is being saved. Since the kernel stored in the image doesn't know of that, it causes memory corruption to occur after resume from hibernation, especially on systems with relatively small RAM that need to swap often.

This issue can be addressed by keeping the GFP_IOFS bits clear in gfp_allowed_mask during the entire hibernation, including the saving of the image, until the system is finally turned off or the hibernation is aborted. Unfortunately, for this purpose it's necessary to rework the way in which the hibernate and suspend code manipulates gfp_allowed_mask.

Quoted 5 lines

This change is based on an earlier patch from Hugh Dickins.

Thanks,
-Kame