Skip to contentSkip to conversation

Android PM fixes

PATCH In review Arve Hjønnevåg 3 participants 12 messages Jan 14, 2009 – Feb 8, 2009

Subsystems (2)
lore
Share & tools
lore
#1 lore raw

PATCH 1/2 PM: Wait for console in resume.

Avoids later waking up to a blinking cursor if the device woke up and returned to sleep before the console switch happened.

---
 kernel/power/console.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
kernel/power/console.c +6 -0
index b8628be..a3961b2 100644
@@ -78,6 +78,12 @@ void pm_restore_console(void)
 	}
 	set_console(orig_fgconsole);
 	release_console_sem();
+
+	if (vt_waitactive(orig_fgconsole)) {
+		pr_debug("Resume: Can't switch VCs.");
+		return;
+	}
+
 	kmsg_redirect = orig_kmsg;
 }
 #endif
Signature · 6 lines
1.6.1

_______________________________________________
linux-pm mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/linux-pm
#2 lore raw

PATCH 2/2 PM: Fix suspend_console/resume_console to use only one semaphore.

This fixes a race if another thread acquired the console while it was suspended and released it after it was resumed the secondary console semaphre would be left locked, and the primary semaphore would be released twice. This in turn caused the console switch on suspend or resume to hang forever.

Note that suspend_console does not actually lock the console for clients that use acquire_console_sem, it only locks it for clients that use try_acquire_console_sem. If we change suspend_console to fully lock the console, then the kernel may deadlock on suspend.

---
 kernel/printk.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
kernel/printk.c +9 -6
index 7015733..118ee77 100644
@@ -73,7 +73,6 @@ EXPORT_SYMBOL(oops_in_progress);
  * driver system.
  */
 static DECLARE_MUTEX(console_sem);
-static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
@@ -896,12 +895,14 @@ void suspend_console(void)
 	printk("Suspending console(s) (use no_console_suspend to debug)\n");
 	acquire_console_sem();
 	console_suspended = 1;
+	up(&console_sem);
 }
 
 void resume_console(void)
 {
 	if (!console_suspend_enabled)
 		return;
+	down(&console_sem);
 	console_suspended = 0;
 	release_console_sem();
 }
@@ -917,11 +918,9 @@ void resume_console(void)
 void acquire_console_sem(void)
 {
 	BUG_ON(in_interrupt());
-	if (console_suspended) {
-		down(&secondary_console_sem);
-		return;
-	}
 	down(&console_sem);
+	if (console_suspended)
+		return;
 	console_locked = 1;
 	console_may_schedule = 1;
 }
@@ -931,6 +930,10 @@ int try_acquire_console_sem(void)
 {
 	if (down_trylock(&console_sem))
 		return -1;
+	if (console_suspended) {
+		up(&console_sem);
+		return -1;
+	}
 	console_locked = 1;
 	console_may_schedule = 0;
 	return 0;
@@ -984,7 +987,7 @@ void release_console_sem(void)
 	unsigned wake_klogd = 0;
 
 	if (console_suspended) {
-		up(&secondary_console_sem);
+		up(&console_sem);
 		return;
 	}
 
Signature · 6 lines
1.6.1

_______________________________________________
linux-pm mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/linux-pm
#3 lore raw

Re: [PATCH 2/2] PM: Fix suspend_console/resume_console to use only one semaphore.

Quoted 6 lines · Arve Hj??nnev??g

This fixes a race if another thread acquired the console while it was suspended and released it after it was resumed the secondary console semaphre would be left locked, and the primary semaphore would be released twice. This in turn caused the console switch on suspend or resume to hang forever.

Quoted 7 lines

Note that suspend_console does not actually lock the console for clients that use acquire_console_sem, it only locks it for clients that use try_acquire_console_sem. If we change suspend_console to fully lock the console, then the kernel may deadlock on suspend.

This probably needs to go to lkml for proper review. Getting rid of secondary_console_sem would be nice cleanup in itself...

Quoted 65 lines
---
 kernel/printk.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
kernel/printk.c +9 -6
index 7015733..118ee77 100644
@@ -73,7 +73,6 @@ EXPORT_SYMBOL(oops_in_progress);
  * driver system.
  */
 static DECLARE_MUTEX(console_sem);
-static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
@@ -896,12 +895,14 @@ void suspend_console(void)
 	printk("Suspending console(s) (use no_console_suspend to debug)\n");
 	acquire_console_sem();
 	console_suspended = 1;
+	up(&console_sem);
 }
 
 void resume_console(void)
 {
 	if (!console_suspend_enabled)
 		return;
+	down(&console_sem);
 	console_suspended = 0;
 	release_console_sem();
 }
@@ -917,11 +918,9 @@ void resume_console(void)
 void acquire_console_sem(void)
 {
 	BUG_ON(in_interrupt());
-	if (console_suspended) {
-		down(&secondary_console_sem);
-		return;
-	}
 	down(&console_sem);
+	if (console_suspended)
+		return;
 	console_locked = 1;
 	console_may_schedule = 1;
 }
@@ -931,6 +930,10 @@ int try_acquire_console_sem(void)
 {
 	if (down_trylock(&console_sem))
 		return -1;
+	if (console_suspended) {
+		up(&console_sem);
+		return -1;
+	}
 	console_locked = 1;
 	console_may_schedule = 0;
 	return 0;
@@ -984,7 +987,7 @@ void release_console_sem(void)
 	unsigned wake_klogd = 0;
 
 	if (console_suspended) {
-		up(&secondary_console_sem);
+		up(&console_sem);
 		return;
 	}
 
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#4 lore raw

PATCH 1/2 PM: Wait for console in resume.

Avoids later waking up to a blinking cursor if the device woke up and returned to sleep before the console switch happened.

---
 kernel/power/console.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
kernel/power/console.c +6 -0
index b8628be..a3961b2 100644
@@ -78,6 +78,12 @@ void pm_restore_console(void)
 	}
 	set_console(orig_fgconsole);
 	release_console_sem();
+
+	if (vt_waitactive(orig_fgconsole)) {
+		pr_debug("Resume: Can't switch VCs.");
+		return;
+	}
+
 	kmsg_redirect = orig_kmsg;
 }
 #endif
Signature · 6 lines
1.6.1

_______________________________________________
linux-pm mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/linux-pm
#5 lore raw

PATCH 1/2 PM: Wait for console in resume.

In reply to Pavel Machek

Avoids later waking up to a blinking cursor if the device woke up and returned to sleep before the console switch happened.

---
 kernel/power/console.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
kernel/power/console.c +6 -0
index b8628be..a3961b2 100644
@@ -78,6 +78,12 @@ void pm_restore_console(void)
 	}
 	set_console(orig_fgconsole);
 	release_console_sem();
+
+	if (vt_waitactive(orig_fgconsole)) {
+		pr_debug("Resume: Can't switch VCs.");
+		return;
+	}
+
 	kmsg_redirect = orig_kmsg;
 }
 #endif
Signature · 1 line
1.6.1
#6 lore raw

PATCH 2/2 PM: Fix suspend_console/resume_console to use only one semaphore.

This fixes a race if another thread acquired the console while it was suspended and released it after it was resumed the secondary console semaphre would be left locked, and the primary semaphore would be released twice. This in turn caused the console switch on suspend or resume to hang forever.

Note that suspend_console does not actually lock the console for clients that use acquire_console_sem, it only locks it for clients that use try_acquire_console_sem. If we change suspend_console to fully lock the console, then the kernel may deadlock on suspend.

---
 kernel/printk.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
kernel/printk.c +9 -6
index 7015733..118ee77 100644
@@ -73,7 +73,6 @@ EXPORT_SYMBOL(oops_in_progress);
  * driver system.
  */
 static DECLARE_MUTEX(console_sem);
-static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
@@ -896,12 +895,14 @@ void suspend_console(void)
 	printk("Suspending console(s) (use no_console_suspend to debug)\n");
 	acquire_console_sem();
 	console_suspended = 1;
+	up(&console_sem);
 }
 
 void resume_console(void)
 {
 	if (!console_suspend_enabled)
 		return;
+	down(&console_sem);
 	console_suspended = 0;
 	release_console_sem();
 }
@@ -917,11 +918,9 @@ void resume_console(void)
 void acquire_console_sem(void)
 {
 	BUG_ON(in_interrupt());
-	if (console_suspended) {
-		down(&secondary_console_sem);
-		return;
-	}
 	down(&console_sem);
+	if (console_suspended)
+		return;
 	console_locked = 1;
 	console_may_schedule = 1;
 }
@@ -931,6 +930,10 @@ int try_acquire_console_sem(void)
 {
 	if (down_trylock(&console_sem))
 		return -1;
+	if (console_suspended) {
+		up(&console_sem);
+		return -1;
+	}
 	console_locked = 1;
 	console_may_schedule = 0;
 	return 0;
@@ -984,7 +987,7 @@ void release_console_sem(void)
 	unsigned wake_klogd = 0;
 
 	if (console_suspended) {
-		up(&secondary_console_sem);
+		up(&console_sem);
 		return;
 	}
 
Signature · 6 lines
1.6.1

_______________________________________________
linux-pm mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/linux-pm
#7 lore raw

PATCH 2/2 PM: Fix suspend_console/resume_console to use only one semaphore.

In reply to Arve Hjønnevåg

This fixes a race if another thread acquired the console while it was suspended and released it after it was resumed the secondary console semaphre would be left locked, and the primary semaphore would be released twice. This in turn caused the console switch on suspend or resume to hang forever.

Note that suspend_console does not actually lock the console for clients that use acquire_console_sem, it only locks it for clients that use try_acquire_console_sem. If we change suspend_console to fully lock the console, then the kernel may deadlock on suspend.

---
 kernel/printk.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
kernel/printk.c +9 -6
index 7015733..118ee77 100644
@@ -73,7 +73,6 @@ EXPORT_SYMBOL(oops_in_progress);
  * driver system.
  */
 static DECLARE_MUTEX(console_sem);
-static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
@@ -896,12 +895,14 @@ void suspend_console(void)
 	printk("Suspending console(s) (use no_console_suspend to debug)\n");
 	acquire_console_sem();
 	console_suspended = 1;
+	up(&console_sem);
 }
 
 void resume_console(void)
 {
 	if (!console_suspend_enabled)
 		return;
+	down(&console_sem);
 	console_suspended = 0;
 	release_console_sem();
 }
@@ -917,11 +918,9 @@ void resume_console(void)
 void acquire_console_sem(void)
 {
 	BUG_ON(in_interrupt());
-	if (console_suspended) {
-		down(&secondary_console_sem);
-		return;
-	}
 	down(&console_sem);
+	if (console_suspended)
+		return;
 	console_locked = 1;
 	console_may_schedule = 1;
 }
@@ -931,6 +930,10 @@ int try_acquire_console_sem(void)
 {
 	if (down_trylock(&console_sem))
 		return -1;
+	if (console_suspended) {
+		up(&console_sem);
+		return -1;
+	}
 	console_locked = 1;
 	console_may_schedule = 0;
 	return 0;
@@ -984,7 +987,7 @@ void release_console_sem(void)
 	unsigned wake_klogd = 0;
 
 	if (console_suspended) {
-		up(&secondary_console_sem);
+		up(&console_sem);
 		return;
 	}
 
Signature · 1 line
1.6.1
#8 lore raw

PATCH 1/1 PM: Fix suspend_console/resume_console to use only one semaphore.

This fixes a race where a thread acquires the console while the console is suspended, and the console is resumed before this thread releases it. In this case, the secondary console semaphore would be left locked, and the primary semaphore would be released twice. This in turn would cause the console switch on suspend or resume to hang forever.

Note that suspend_console does not actually lock the console for clients that use acquire_console_sem, it only locks it for clients that use try_acquire_console_sem. If we change suspend_console to fully lock the console, then the kernel may deadlock on suspend. One client of try_acquire_console_sem is acquire_console_semaphore_for_printk, which uses it to prevent printk from using the console while it is suspended.

---
 kernel/printk.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
kernel/printk.c +9 -6
index 69188f2..e3602d0 100644
@@ -73,7 +73,6 @@ EXPORT_SYMBOL(oops_in_progress);
  * driver system.
  */
 static DECLARE_MUTEX(console_sem);
-static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
@@ -891,12 +890,14 @@ void suspend_console(void)
 	printk("Suspending console(s) (use no_console_suspend to debug)\n");
 	acquire_console_sem();
 	console_suspended = 1;
+	up(&console_sem);
 }
 
 void resume_console(void)
 {
 	if (!console_suspend_enabled)
 		return;
+	down(&console_sem);
 	console_suspended = 0;
 	release_console_sem();
 }
@@ -912,11 +913,9 @@ void resume_console(void)
 void acquire_console_sem(void)
 {
 	BUG_ON(in_interrupt());
-	if (console_suspended) {
-		down(&secondary_console_sem);
-		return;
-	}
 	down(&console_sem);
+	if (console_suspended)
+		return;
 	console_locked = 1;
 	console_may_schedule = 1;
 }
@@ -926,6 +925,10 @@ int try_acquire_console_sem(void)
 {
 	if (down_trylock(&console_sem))
 		return -1;
+	if (console_suspended) {
+		up(&console_sem);
+		return -1;
+	}
 	console_locked = 1;
 	console_may_schedule = 0;
 	return 0;
@@ -979,7 +982,7 @@ void release_console_sem(void)
 	unsigned wake_klogd = 0;
 
 	if (console_suspended) {
-		up(&secondary_console_sem);
+		up(&console_sem);
 		return;
 	}
 
Signature · 6 lines
1.6.1

_______________________________________________
linux-pm mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/linux-pm
#9 lore raw

PATCH 1/1 PM: Fix suspend_console/resume_console to use only one semaphore.

In reply to Arve Hjønnevåg

This fixes a race where a thread acquires the console while the console is suspended, and the console is resumed before this thread releases it. In this case, the secondary console semaphore would be left locked, and the primary semaphore would be released twice. This in turn would cause the console switch on suspend or resume to hang forever.

Note that suspend_console does not actually lock the console for clients that use acquire_console_sem, it only locks it for clients that use try_acquire_console_sem. If we change suspend_console to fully lock the console, then the kernel may deadlock on suspend. One client of try_acquire_console_sem is acquire_console_semaphore_for_printk, which uses it to prevent printk from using the console while it is suspended.

---
 kernel/printk.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
kernel/printk.c +9 -6
index 69188f2..e3602d0 100644
@@ -73,7 +73,6 @@ EXPORT_SYMBOL(oops_in_progress);
  * driver system.
  */
 static DECLARE_MUTEX(console_sem);
-static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
@@ -891,12 +890,14 @@ void suspend_console(void)
 	printk("Suspending console(s) (use no_console_suspend to debug)\n");
 	acquire_console_sem();
 	console_suspended = 1;
+	up(&console_sem);
 }
 
 void resume_console(void)
 {
 	if (!console_suspend_enabled)
 		return;
+	down(&console_sem);
 	console_suspended = 0;
 	release_console_sem();
 }
@@ -912,11 +913,9 @@ void resume_console(void)
 void acquire_console_sem(void)
 {
 	BUG_ON(in_interrupt());
-	if (console_suspended) {
-		down(&secondary_console_sem);
-		return;
-	}
 	down(&console_sem);
+	if (console_suspended)
+		return;
 	console_locked = 1;
 	console_may_schedule = 1;
 }
@@ -926,6 +925,10 @@ int try_acquire_console_sem(void)
 {
 	if (down_trylock(&console_sem))
 		return -1;
+	if (console_suspended) {
+		up(&console_sem);
+		return -1;
+	}
 	console_locked = 1;
 	console_may_schedule = 0;
 	return 0;
@@ -979,7 +982,7 @@ void release_console_sem(void)
 	unsigned wake_klogd = 0;
 
 	if (console_suspended) {
-		up(&secondary_console_sem);
+		up(&console_sem);
 		return;
 	}
 
Signature · 1 line
1.6.1
#10 lore raw

Re: [PATCH] PM: Fix suspend_console/resume_console to use only one semaphore.

Quoted 7 lines · Arve Hjønnevåg

This fixes a race where a thread acquires the console while the console is suspended, and the console is resumed before this thread releases it. In this case, the secondary console semaphore would be left locked, and the primary semaphore would be released twice. This in turn would cause the console switch on suspend or resume to hang forever.

Quoted 7 lines

Note that suspend_console does not actually lock the console for clients that use acquire_console_sem, it only locks it for clients that use try_acquire_console_sem. If we change suspend_console to fully lock the console, then the kernel may deadlock on suspend. One client of try_acquire_console_sem is acquire_console_semaphore_for_printk, which uses it to prevent printk from using the console while it is suspended.

Right, we shouldn't hold the semaphore over the whole time. In fact, I even have some doubts about the need for suspending the console at all...

The way I did it back then for powerbooks was to implement a "suspended" state in the fbdev/fbcon core which the drivers call effectively causing the core to stop touching the framebuffer. Works fine for kernel consoles though of course userspace or X needs something different, but that's orthogonal.

Of course that doesn't help with serial, vgacon etc... so each drivers needs to suspend itself and be made to ignore incoming requests but that's something that has to be done anyway at some stage...

I suspect the suspend console trick is mostly helpful with vgacon..

Ben.

Quoted 73 lines
---
 kernel/printk.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
kernel/printk.c +9 -6
index 69188f2..e3602d0 100644
@@ -73,7 +73,6 @@ EXPORT_SYMBOL(oops_in_progress);
  * driver system.
  */
 static DECLARE_MUTEX(console_sem);
-static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
@@ -891,12 +890,14 @@ void suspend_console(void)
 	printk("Suspending console(s) (use no_console_suspend to debug)\n");
 	acquire_console_sem();
 	console_suspended = 1;
+	up(&console_sem);
 }
 
 void resume_console(void)
 {
 	if (!console_suspend_enabled)
 		return;
+	down(&console_sem);
 	console_suspended = 0;
 	release_console_sem();
 }
@@ -912,11 +913,9 @@ void resume_console(void)
 void acquire_console_sem(void)
 {
 	BUG_ON(in_interrupt());
-	if (console_suspended) {
-		down(&secondary_console_sem);
-		return;
-	}
 	down(&console_sem);
+	if (console_suspended)
+		return;
 	console_locked = 1;
 	console_may_schedule = 1;
 }
@@ -926,6 +925,10 @@ int try_acquire_console_sem(void)
 {
 	if (down_trylock(&console_sem))
 		return -1;
+	if (console_suspended) {
+		up(&console_sem);
+		return -1;
+	}
 	console_locked = 1;
 	console_may_schedule = 0;
 	return 0;
@@ -979,7 +982,7 @@ void release_console_sem(void)
 	unsigned wake_klogd = 0;
 
 	if (console_suspended) {
-		up(&secondary_console_sem);
+		up(&console_sem);
 		return;
 	}
 
Signature · 6 lines
1.6.1

_______________________________________________
linux-pm mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/linux-pm

_______________________________________________
linux-pm mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/linux-pm

#11 lore raw

Re: [linux-pm] [PATCH] PM: Fix suspend_console/resume_console to use only one semaphore.

In reply to Arve Hjønnevåg
Quoted 7 lines · Arve Hjønnevåg

This fixes a race where a thread acquires the console while the console is suspended, and the console is resumed before this thread releases it. In this case, the secondary console semaphore would be left locked, and the primary semaphore would be released twice. This in turn would cause the console switch on suspend or resume to hang forever.

Quoted 7 lines

Note that suspend_console does not actually lock the console for clients that use acquire_console_sem, it only locks it for clients that use try_acquire_console_sem. If we change suspend_console to fully lock the console, then the kernel may deadlock on suspend. One client of try_acquire_console_sem is acquire_console_semaphore_for_printk, which uses it to prevent printk from using the console while it is suspended.

Right, we shouldn't hold the semaphore over the whole time. In fact, I even have some doubts about the need for suspending the console at all...

The way I did it back then for powerbooks was to implement a "suspended" state in the fbdev/fbcon core which the drivers call effectively causing the core to stop touching the framebuffer. Works fine for kernel consoles though of course userspace or X needs something different, but that's orthogonal.

Of course that doesn't help with serial, vgacon etc... so each drivers needs to suspend itself and be made to ignore incoming requests but that's something that has to be done anyway at some stage...

I suspect the suspend console trick is mostly helpful with vgacon..

Ben.

Quoted 73 lines
---
 kernel/printk.c |   15 +++++++++------
 1 files changed, 9 insertions(+), 6 deletions(-)
kernel/printk.c +9 -6
index 69188f2..e3602d0 100644
@@ -73,7 +73,6 @@ EXPORT_SYMBOL(oops_in_progress);
  * driver system.
  */
 static DECLARE_MUTEX(console_sem);
-static DECLARE_MUTEX(secondary_console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
@@ -891,12 +890,14 @@ void suspend_console(void)
 	printk("Suspending console(s) (use no_console_suspend to debug)\n");
 	acquire_console_sem();
 	console_suspended = 1;
+	up(&console_sem);
 }
 
 void resume_console(void)
 {
 	if (!console_suspend_enabled)
 		return;
+	down(&console_sem);
 	console_suspended = 0;
 	release_console_sem();
 }
@@ -912,11 +913,9 @@ void resume_console(void)
 void acquire_console_sem(void)
 {
 	BUG_ON(in_interrupt());
-	if (console_suspended) {
-		down(&secondary_console_sem);
-		return;
-	}
 	down(&console_sem);
+	if (console_suspended)
+		return;
 	console_locked = 1;
 	console_may_schedule = 1;
 }
@@ -926,6 +925,10 @@ int try_acquire_console_sem(void)
 {
 	if (down_trylock(&console_sem))
 		return -1;
+	if (console_suspended) {
+		up(&console_sem);
+		return -1;
+	}
 	console_locked = 1;
 	console_may_schedule = 0;
 	return 0;
@@ -979,7 +982,7 @@ void release_console_sem(void)
 	unsigned wake_klogd = 0;
 
 	if (console_suspended) {
-		up(&secondary_console_sem);
+		up(&console_sem);
 		return;
 	}
 
Signature · 6 lines
1.6.1

_______________________________________________
linux-pm mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/linux-pm