make disable_console_suspend runtime configurable

PATCH In review Stefan Seyfried 4 participants 13 messages Jun 13, 2007 – Jun 21, 2007

Subsystems (3)

Suspend to RAM Hibernation Console

lore

Share & tools

lore

Stefan Seyfried

author

2007-06-13 20:06 UTC #0 lore raw

PATCH 1/1 make disable_console_suspend runtime configurable

I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.

Signed-off-by Stefan Seyfried <[email protected]>

Signed-off-by Frank Seidel <[email protected]>

---
Second try, no longer uses a sysctl in /proc, but a tunable in /sys/power. Thanks Frank for moving it quicker to sysfs than i would have ever been able to :-)

 Documentation/power/basic-pm-debugging.txt |   16 +++++++++-------
 drivers/serial/serial_core.c               |    8 ++------
 include/linux/console.h                    |    7 ++-----
 kernel/power/Kconfig                       |   11 -----------
 kernel/power/main.c                        |   27 +++++++++++++++++++++++++++
 kernel/printk.c                            |    7 +++++--
 6 files changed, 45 insertions(+), 31 deletions(-)

Documentation/power/basic-pm-debugging.txt +9 -7

@@ -78,13 +78,15 @@ c) Advanced debugging
 In case the STD does not work on your system even in the minimal configuration
 and compiling more drivers as modules is not practical or some modules cannot
 be unloaded, you can use one of the more advanced debugging techniques to find
-the problem.  First, if there is a serial port in your box, you can set the
-CONFIG_DISABLE_CONSOLE_SUSPEND kernel configuration option and try to log kernel
-messages using the serial console.  This may provide you with some information
-about the reasons of the suspend (resume) failure.  Alternatively, it may be
-possible to use a FireWire port for debugging with firescope
-(ftp://ftp.firstfloor.org/pub/ak/firescope/).  On i386 it is also possible to
-use the PM_TRACE mechanism documented in Documentation/s2ram.txt .
+the problem.  First, if there is a serial port in your box, you can
+
+# echo 1 > /sys/power/disable_console_suspend
+
+and try to log kernel messages using the serial console.  This may provide you
+with some information about the reasons of the suspend (resume) failure.
+Alternatively, it may be possible to use a FireWire port for debugging with
+firescope (ftp://ftp.firstfloor.org/pub/ak/firescope/).  On i386 it is also
+possible to use the PM_TRACE mechanism documented in Documentation/s2ram.txt .
 
 2. Testing suspend to RAM (STR)

drivers/serial/serial_core.c +2 -6

@@ -1934,12 +1934,10 @@ int uart_suspend_port(struct uart_driver
 
 	mutex_lock(&state->mutex);
 
-#ifdef CONFIG_DISABLE_CONSOLE_SUSPEND
-	if (uart_console(port)) {
+	if (disable_console_suspend && uart_console(port)) {
 		mutex_unlock(&state->mutex);
 		return 0;
 	}
-#endif
 
 	if (state->info && state->info->flags & UIF_INITIALIZED) {
 		const struct uart_ops *ops = port->ops;

@@ -1982,12 +1980,10 @@ int uart_resume_port(struct uart_driver 
 
 	mutex_lock(&state->mutex);
 
-#ifdef CONFIG_DISABLE_CONSOLE_SUSPEND
-	if (uart_console(port)) {
+	if (disable_console_suspend && uart_console(port)) {
 		mutex_unlock(&state->mutex);
 		return 0;
 	}
-#endif
 
 	uart_change_pm(state, 0);

include/linux/console.h +2 -5

@@ -64,6 +64,8 @@ extern const struct consw vga_con;	/* VG
 extern const struct consw newport_con;	/* SGI Newport console  */
 extern const struct consw prom_con;	/* SPARC PROM console */
 
+extern int disable_console_suspend;
+
 int con_is_bound(const struct consw *csw);
 int register_con_driver(const struct consw *csw, int first, int last);
 int unregister_con_driver(const struct consw *csw);

@@ -120,14 +122,9 @@ extern void console_stop(struct console 
 extern void console_start(struct console *);
 extern int is_console_locked(void);
 
-#ifndef CONFIG_DISABLE_CONSOLE_SUSPEND
 /* Suspend and resume console messages over PM events */
 extern void suspend_console(void);
 extern void resume_console(void);
-#else
-static inline void suspend_console(void) {}
-static inline void resume_console(void) {}
-#endif /* CONFIG_DISABLE_CONSOLE_SUSPEND */
 
 int mda_console_init(void);
 void prom_con_init(void);

kernel/power/Kconfig +0 ~~-11~~

@@ -37,17 +37,6 @@ config PM_DEBUG
 	code. This is helpful when debugging and reporting various PM bugs, 
 	like suspend support.
 
-config DISABLE_CONSOLE_SUSPEND
-	bool "Keep console(s) enabled during suspend/resume (DANGEROUS)"
-	depends on PM && PM_DEBUG
-	default n
-	---help---
-	This option turns off the console suspend mechanism that prevents
-	debug messages from reaching the console during the suspend/resume
-	operations.  This may be helpful when debugging device drivers'
-	suspend/resume routines, but may itself lead to problems, for example
-	if netconsole is used.
-
 config PM_TRACE
 	bool "Suspend/resume event tracing"
 	depends on PM && PM_DEBUG && X86_32 && EXPERIMENTAL

kernel/printk.c +5 -2

@@ -76,6 +76,7 @@ struct console *console_drivers;
  * locked without the console sempahore held
  */
 static int console_locked, console_suspended;
+int disable_console_suspend;
 
 /*
  * logbuf_lock protects log_buf, log_start, log_end, con_start and logged_chars

@@ -726,7 +727,6 @@ int __init add_preferred_console(char *n
 	return 0;
 }
 
-#ifndef CONFIG_DISABLE_CONSOLE_SUSPEND
 /**
  * suspend_console - suspend the console subsystem
  *

@@ -734,6 +734,8 @@ int __init add_preferred_console(char *n
  */
 void suspend_console(void)
 {
+	if (disable_console_suspend)
+		return;
 	printk("Suspending console(s)\n");
 	acquire_console_sem();
 	console_suspended = 1;

@@ -741,10 +743,11 @@ void suspend_console(void)
 
 void resume_console(void)
 {
+	if (disable_console_suspend)
+		return;
 	console_suspended = 0;
 	release_console_sem();
 }
-#endif /* CONFIG_DISABLE_CONSOLE_SUSPEND */
 
 /**
  * acquire_console_sem - lock the console system for exclusive use.

kernel/power/main.c +27 -0

@@ -308,6 +308,31 @@ static ssize_t state_store(struct kset *
 
 power_attr(state);
 
+/**
+ * disable_console_suspend - control console suspend
+ *
+ * If set the consoles are not suspended so it is
+ * still possible to see and collect debug info.
+ */
+static ssize_t disable_console_suspend_show(struct kset *kset, char *buf)
+{
+	return sprintf(buf, "%i\n", disable_console_suspend);
+}
+
+static ssize_t disable_console_suspend_store(struct kset *kset, const char *buf, size_t n)
+{
+	int disable_console;
+
+	if (sscanf(buf, "%i", &disable_console) == 1) {
+		disable_console_suspend = disable_console;
+		return n;
+	}
+
+	return -EINVAL;
+}
+
+power_attr(disable_console_suspend);
+
 #ifdef CONFIG_PM_TRACE
 int pm_trace_enabled;

@@ -332,12 +357,14 @@ power_attr(pm_trace);
 
 static struct attribute * g[] = {
 	&state_attr.attr,
+	&disable_console_suspend_attr.attr,
 	&pm_trace_attr.attr,
 	NULL,
 };
 #else
 static struct attribute * g[] = {
 	&state_attr.attr,
+	&disable_console_suspend_attr.attr,
 	NULL,
 };
 #endif /* CONFIG_PM_TRACE */

Signature · 6 lines

Stefan Seyfried
QA / R&D Team Mobile Devices        |              "Any ideas, John?"
SUSE LINUX Products GmbH, Nürnberg  | "Well, surrounding them's out." 

This footer brought to you by insane German lawmakers:
SUSE Linux Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)

Rafael J. Wysocki

2007-06-13 21:31 UTC #1 lore raw

Quoted 10 lines · Stefan Seyfried

I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.
Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>
---
Second try, no longer uses a sysctl in /proc, but a tunable in /sys/power. Thanks Frank for moving it quicker to sysfs than i would have ever been able to :-)

ACK

Greetings,
Rafael

Signature · 1 line

"Premature optimization is the root of all evil." - Donald Knuth

Pavel Machek

2007-06-13 22:08 UTC #2 lore raw

In reply to Stefan Seyfried

Hi!

Quoted 6 lines

I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.
Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>

I wonder if there's a better name?

Or maybe this should not be /sys configurable, but just have value for each console "this console can work while suspended"?

(serial can, vesafb can, netconsole can't)?

Exporting "crash-me" option to user does not seem that cool to me.
Pavel

Signature · 2 lines

(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Stefan Seyfried

author

2007-06-13 22:23 UTC #3 lore raw

Quoted 10 lines · Pavel Machek

Hi!
Quoted 6 lines
I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.
Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>
I wonder if there's a better name?

Suggest one.

Quoted 4 lines

Or maybe this should not be /sys configurable, but just have value for each console "this console can work while suspended"?
(serial can, vesafb can, netconsole can't)?

Go ahead, submit a patch. It won't be that trivial. And i wonder if it is actually worth the hassle. This is a debugging facility.

Quoted 1 line

Exporting "crash-me" option to user does not seem that cool to me.

We have "echo c > /proc/sysrq-trigger" also.
This is a debugging option, and forcing users to recompile the kernel just to debug suspend problems (not resume problems, the "it does not even go to sleep" stuff is where this matters most) is IMO a bad idea.

We can also make this a boot parameter, i don't care, but i want to disable console suspend without recompiling the kernel.

Signature · 6 lines

Stefan Seyfried
QA / R&D Team Mobile Devices        |              "Any ideas, John?"
SUSE LINUX Products GmbH, Nürnberg  | "Well, surrounding them's out." 

This footer brought to you by insane German lawmakers:
SUSE Linux Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)

Rafael J. Wysocki

2007-06-13 23:02 UTC #4 lore raw

Quoted 24 lines · Stefan Seyfried

Quoted 10 lines · Pavel Machek
Hi!
Quoted 6 lines
I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.
Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>
I wonder if there's a better name?
Suggest one.
Quoted 4 lines
Or maybe this should not be /sys configurable, but just have value for each console "this console can work while suspended"?
(serial can, vesafb can, netconsole can't)?
Go ahead, submit a patch. It won't be that trivial. And i wonder if it is actually worth the hassle. This is a debugging facility.
Quoted 1 line
Exporting "crash-me" option to user does not seem that cool to me.

Quoted 7 lines

We have "echo c > /proc/sysrq-trigger" also.
This is a debugging option, and forcing users to recompile the kernel just to debug suspend problems (not resume problems, the "it does not even go to sleep" stuff is where this matters most) is IMO a bad idea.
We can also make this a boot parameter, i don't care, but i want to disable console suspend without recompiling the kernel.

It's a bit similar to the pm_trace thing.

As long as there is a sane default (ie. the consoles are disabled), I don't see a big problem with that.

Moreover, distributions will probably want to have a switch so that they can tell the user "please do 'echo 1 > ...' and retest" and if they start to add such things by themselves that wouldn't be nice.

Greetings,
Rafael

Signature · 1 line

"Premature optimization is the root of all evil." - Donald Knuth

Rafael J. Wysocki

2007-06-14 12:08 UTC #5 lore raw

In reply to Stefan Seyfried

Quoted 7 lines · Stefan Seyfried

Quoted 6 lines · Pavel Machek
Hi!
Quoted 3 lines
I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.

Quoted 6 lines

Quoted 4 lines
Quoted 2 lines
Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>
I wonder if there's a better name?
Suggest one.

Hmm, you could call it 'console_suspend' with
'echo disabled > /sys/power/console_suspend' for switching that off (and analogously with 'enabled') and with 'cat /sys/power/console_suspend' printing the current status.

Greetings,
Rafael

Signature · 1 line

"Premature optimization is the root of all evil." - Donald Knuth

Frank Seidel

2007-06-14 14:00 UTC #6 lore raw

From: Stefan Seyfried <[email protected]>

I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.

Signed-off-by Stefan Seyfried <[email protected]>

Signed-off-by Frank Seidel <[email protected]>

---
Third try, renamed sysfs interface to console_suspend reporting and expecting either "enabled" or "disabled"

 Documentation/power/basic-pm-debugging.txt |   16 +++++++++-------
 drivers/serial/serial_core.c               |    8 ++------
 include/linux/console.h                    |    7 ++-----
 kernel/power/Kconfig                       |   11 -----------
 kernel/power/main.c                        |   28 ++++++++++++++++++++++++++++
 kernel/printk.c                            |    7 +++++--
 6 files changed, 46 insertions(+), 31 deletions(-)

Documentation/power/basic-pm-debugging.txt +9 -7

@@ -78,13 +78,15 @@ c) Advanced debugging
 In case the STD does not work on your system even in the minimal configuration
 and compiling more drivers as modules is not practical or some modules cannot
 be unloaded, you can use one of the more advanced debugging techniques to find
-the problem.  First, if there is a serial port in your box, you can set the
-CONFIG_DISABLE_CONSOLE_SUSPEND kernel configuration option and try to log kernel
-messages using the serial console.  This may provide you with some information
-about the reasons of the suspend (resume) failure.  Alternatively, it may be
-possible to use a FireWire port for debugging with firescope
-(ftp://ftp.firstfloor.org/pub/ak/firescope/).  On i386 it is also possible to
-use the PM_TRACE mechanism documented in Documentation/s2ram.txt .
+the problem.  First, if there is a serial port in your box, you can
+
+# echo disabled > /sys/power/console_suspend
+
+and try to log kernel messages using the serial console.  This may provide you
+with some information about the reasons of the suspend (resume) failure.
+Alternatively, it may be possible to use a FireWire port for debugging with
+firescope (ftp://ftp.firstfloor.org/pub/ak/firescope/).  On i386 it is also
+possible to use the PM_TRACE mechanism documented in Documentation/s2ram.txt .
 
 2. Testing suspend to RAM (STR)

drivers/serial/serial_core.c +2 -6

@@ -1934,12 +1934,10 @@ int uart_suspend_port(struct uart_driver
 
 	mutex_lock(&state->mutex);
 
-#ifdef CONFIG_DISABLE_CONSOLE_SUSPEND
-	if (uart_console(port)) {
+	if (disable_console_suspend && uart_console(port)) {
 		mutex_unlock(&state->mutex);
 		return 0;
 	}
-#endif
 
 	if (state->info && state->info->flags & UIF_INITIALIZED) {
 		const struct uart_ops *ops = port->ops;

@@ -1982,12 +1980,10 @@ int uart_resume_port(struct uart_driver 
 
 	mutex_lock(&state->mutex);
 
-#ifdef CONFIG_DISABLE_CONSOLE_SUSPEND
-	if (uart_console(port)) {
+	if (disable_console_suspend && uart_console(port)) {
 		mutex_unlock(&state->mutex);
 		return 0;
 	}
-#endif
 
 	uart_change_pm(state, 0);

include/linux/console.h +2 -5

@@ -64,6 +64,8 @@ extern const struct consw vga_con;	/* VG
 extern const struct consw newport_con;	/* SGI Newport console  */
 extern const struct consw prom_con;	/* SPARC PROM console */
 
+extern int disable_console_suspend;
+
 int con_is_bound(const struct consw *csw);
 int register_con_driver(const struct consw *csw, int first, int last);
 int unregister_con_driver(const struct consw *csw);

@@ -120,14 +122,9 @@ extern void console_stop(struct console 
 extern void console_start(struct console *);
 extern int is_console_locked(void);
 
-#ifndef CONFIG_DISABLE_CONSOLE_SUSPEND
 /* Suspend and resume console messages over PM events */
 extern void suspend_console(void);
 extern void resume_console(void);
-#else
-static inline void suspend_console(void) {}
-static inline void resume_console(void) {}
-#endif /* CONFIG_DISABLE_CONSOLE_SUSPEND */
 
 int mda_console_init(void);
 void prom_con_init(void);

kernel/power/Kconfig +0 ~~-11~~

@@ -37,17 +37,6 @@ config PM_DEBUG
 	code. This is helpful when debugging and reporting various PM bugs, 
 	like suspend support.
 
-config DISABLE_CONSOLE_SUSPEND
-	bool "Keep console(s) enabled during suspend/resume (DANGEROUS)"
-	depends on PM && PM_DEBUG
-	default n
-	---help---
-	This option turns off the console suspend mechanism that prevents
-	debug messages from reaching the console during the suspend/resume
-	operations.  This may be helpful when debugging device drivers'
-	suspend/resume routines, but may itself lead to problems, for example
-	if netconsole is used.
-
 config PM_TRACE
 	bool "Suspend/resume event tracing"
 	depends on PM && PM_DEBUG && X86_32 && EXPERIMENTAL

kernel/printk.c +5 -2

@@ -76,6 +76,7 @@ struct console *console_drivers;
  * locked without the console sempahore held
  */
 static int console_locked, console_suspended;
+int disable_console_suspend;
 
 /*
  * logbuf_lock protects log_buf, log_start, log_end, con_start and logged_chars

@@ -726,7 +727,6 @@ int __init add_preferred_console(char *n
 	return 0;
 }
 
-#ifndef CONFIG_DISABLE_CONSOLE_SUSPEND
 /**
  * suspend_console - suspend the console subsystem
  *

@@ -734,6 +734,8 @@ int __init add_preferred_console(char *n
  */
 void suspend_console(void)
 {
+	if (disable_console_suspend)
+		return;
 	printk("Suspending console(s)\n");
 	acquire_console_sem();
 	console_suspended = 1;

@@ -741,10 +743,11 @@ void suspend_console(void)
 
 void resume_console(void)
 {
+	if (disable_console_suspend)
+		return;
 	console_suspended = 0;
 	release_console_sem();
 }
-#endif /* CONFIG_DISABLE_CONSOLE_SUSPEND */
 
 /**
  * acquire_console_sem - lock the console system for exclusive use.

kernel/power/main.c +28 -0

@@ -308,6 +308,32 @@ static ssize_t state_store(struct kset *
 
 power_attr(state);
 
+/**
+ * disable_console_suspend - control console suspend
+ *
+ * If set the consoles are not suspended so it is
+ * still possible to see and collect debug info.
+ */
+static ssize_t console_suspend_show(struct kset *kset, char *buf)
+{
+	return sprintf(buf, disable_console_suspend ? "disabled\n" : "enabled\n");
+}
+
+static ssize_t console_suspend_store(struct kset *kset, const char *buf, size_t n)
+{
+	if (strncmp(buf, "disabled", 8) == 0) {
+		disable_console_suspend = 1;
+		return n;
+	} else if (strncmp(buf, "enabled", 7) == 0) {
+		disable_console_suspend = 0;
+		return n;
+	}
+
+	return -EINVAL;
+}
+
+power_attr(console_suspend);
+
 #ifdef CONFIG_PM_TRACE
 int pm_trace_enabled;

@@ -332,12 +358,14 @@ power_attr(pm_trace);
 
 static struct attribute * g[] = {
 	&state_attr.attr,
+	&console_suspend_attr.attr,
 	&pm_trace_attr.attr,
 	NULL,
 };
 #else
 static struct attribute * g[] = {
 	&state_attr.attr,
+	&console_suspend_attr.attr,
 	NULL,
 };
 #endif /* CONFIG_PM_TRACE */

Rafael J. Wysocki

2007-06-14 22:16 UTC #7 lore raw

Quoted 7 lines · Frank Seidel

From: Stefan Seyfried <[email protected]>
I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.

Quoted 5 lines

Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>
---
Third try, renamed sysfs interface to console_suspend reporting and expecting either "enabled" or "disabled"

Thanks a lot for redoing it.

I have no objections. Pavel?

Greetings,
Rafael

Signature · 1 line

"Premature optimization is the root of all evil." - Donald Knuth

Pavel Machek

2007-06-17 21:36 UTC #8 lore raw

Quoted 14 lines · Rafael J. Wysocki

Quoted 7 lines · Frank Seidel
From: Stefan Seyfried <[email protected]>
I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.
Quoted 5 lines
Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>
---
Third try, renamed sysfs interface to console_suspend reporting and expecting either "enabled" or "disabled"

Quoted 3 lines

Thanks a lot for redoing it.
I have no objections. Pavel?

I still think that patch is bad. I should have screamed when CONFIG_DISABLE_CONSOLE_SUSPEND went into kernel. That beast should _not_ be configurable, it should just do the right thing.

But I realized that too late. And this only makes it works, making that mistake part of user-kernel interface.

Sorry for not screaming when CONFIG_DISABLE_CONSOLE_SUSPEND went in, but please lets solve this correctly....

Like every console says if it needs to be disabled during suspend, and if at least one active console says so, we disable. Nothing to configure, we just do the right thing. And it should not be that much work, either.
Pavel

Signature · 2 lines

(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Pavel Machek

2007-06-17 21:51 UTC #9 lore raw

Hi!

Quoted 19 lines

Quoted 11 lines
Quoted 10 lines
I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.
Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>
---
Third try, renamed sysfs interface to console_suspend reporting and expecting either "enabled" or "disabled"
Quoted 3 lines
Thanks a lot for redoing it.
I have no objections. Pavel?
I still think that patch is bad. I should have screamed when CONFIG_DISABLE_CONSOLE_SUSPEND went into kernel. That beast should _not_ be configurable, it should just do the right thing.

Quoted 5 lines

But I realized that too late. And this only makes it works, making that mistake part of user-kernel interface.
Sorry for not screaming when CONFIG_DISABLE_CONSOLE_SUSPEND went in, but please lets solve this correctly....

Ouch and sorry for not screaming at "try 1" time. But it still does not make the patch right, and I believe that even patch authors agree that "no-config-needed" is superior solution.
Pavel

Signature · 2 lines

(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Stefan Seyfried

author

2007-06-18 06:43 UTC #10 lore raw

Quoted 24 lines · Pavel Machek

Hi!
Quoted 22 lines
Quoted 11 lines
Quoted 10 lines
I hate having to recompile the kernel, just to be able to debug suspend. Remove CONFIG_DISABLE_CONSOLE_SUSPEND, replace it by a tunable in /sys/power/disable_console_suspend.
Signed-off-by Stefan Seyfried <[email protected]>
Signed-off-by Frank Seidel <[email protected]>
---
Third try, renamed sysfs interface to console_suspend reporting and expecting either "enabled" or "disabled"
Quoted 3 lines
Thanks a lot for redoing it.
I have no objections. Pavel?
I still think that patch is bad. I should have screamed when CONFIG_DISABLE_CONSOLE_SUSPEND went into kernel. That beast should _not_ be configurable, it should just do the right thing.
But I realized that too late. And this only makes it works, making that mistake part of user-kernel interface.

Quoted 6 lines

Quoted 2 lines
Sorry for not screaming when CONFIG_DISABLE_CONSOLE_SUSPEND went in, but please lets solve this correctly....
Ouch and sorry for not screaming at "try 1" time. But it still does not make the patch right, and I believe that even patch authors agree that "no-config-needed" is superior solution.

No, i don't agree at all.

In this case, "no config needed" == "not possible to debug suspend problems".

IMO this is the same as issue as with "sysrq-C". You can crash the machine by other means, but it sometimes is just handy to have a mechanism to do it.

I do not understand what's the problem with this option. If you want to avoid that people use it for something else than debugging, i can add a patch that crashes the machine ten seconds after resume if this option is set.

Signature · 6 lines

Stefan Seyfried
QA / R&D Team Mobile Devices        |              "Any ideas, John?"
SUSE LINUX Products GmbH, Nürnberg  | "Well, surrounding them's out." 

This footer brought to you by insane German lawmakers:
SUSE Linux Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)

Pavel Machek

2007-06-21 13:20 UTC #11 lore raw

Hi!

Quoted 11 lines

Quoted 6 lines
Quoted 2 lines
Sorry for not screaming when CONFIG_DISABLE_CONSOLE_SUSPEND went in, but please lets solve this correctly....
Ouch and sorry for not screaming at "try 1" time. But it still does not make the patch right, and I believe that even patch authors agree that "no-config-needed" is superior solution.
No, i don't agree at all.
In this case, "no config needed" == "not possible to debug suspend problems".

No, sorry.

My proposed solution is "figure out which console drivers can survive being on while machines go down, and keep them on".

So, "no config needed" == "kernel always does the right thing, keeping console during suspend when possible" == "possible to debug suspend problems without having to change CONFIG_ or /sys/*".

Pavel

Signature · 2 lines

(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Stefan Seyfried

author

2007-06-21 21:14 UTC #12 lore raw

Quoted 7 lines · Pavel Machek

Hi!
Quoted 4 lines
No, i don't agree at all.
In this case, "no config needed" == "not possible to debug suspend problems".

Quoted 8 lines

No, sorry.
My proposed solution is "figure out which console drivers can survive being on while machines go down, and keep them on".
So, "no config needed" == "kernel always does the right thing, keeping console during suspend when possible" == "possible to debug suspend problems without having to change CONFIG_ or /sys/*".

Ok. Deal. Once you fixed all the console drivers, i'll gladly send a patch that reverts the patch we are discussing now.

Note that this patch actually helps fixing those drivers, since you can test much easier if a given driver survives suspend ;-)

Signature · 6 lines

Stefan Seyfried
QA / R&D Team Mobile Devices        |              "Any ideas, John?"
SUSE LINUX Products GmbH, Nürnberg  | "Well, surrounding them's out." 

This footer brought to you by insane German lawmakers:
SUSE Linux Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)