Re: [PATCH 6/7] pm: Config option to disable handling of consoleduring suspend/resume

[Ben Collins]

On Tue, 2008-07-29 at 23:15 +0200, Rafael J. Wysocki wrote:
> On Tuesday, 29 of July 2008, Greg KH wrote:
> > On Sun, Jul 20, 2008 at 09:48:28PM -0400, Ben Collins wrote:
> > > Config option allows userspace to handle console during suspend/resume
> > > instead of the kernel.
> > > 
> > > Signed-off-by: Ben Collins <ben.collins@xxxxxxxxxxxxx>
> > > ---
> > >  kernel/power/Kconfig   |   15 +++++++++++++++
> > >  kernel/power/console.c |    4 ++++
> > >  2 files changed, 19 insertions(+), 0 deletions(-)
> > > 
> > > diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig
> > > index b45da40..6855299 100644
> > > --- a/kernel/power/Kconfig
> > > +++ b/kernel/power/Kconfig
> > > @@ -105,6 +105,21 @@ config SUSPEND_FREEZER
> > >  
> > >  	  Turning OFF this setting is NOT recommended! If in doubt, say Y.
> > >  
> > > +config PM_DISABLE_CONSOLE
> > > +	bool "Disable Power Management messing with the active console"
> > > +	depends on PM
> > > +	default n
> > > +	---help---
> > > +	  By default, PM will take over the active console (generally, this means
> > > +	  switching to the console when suspending from X). This can at times cause
> > > +	  problems, especially if userspace suspend scripts try to do things with
> > > +	  the console before or after suspending (e.g. calling vbestate).
> > > +
> > > +	  To work around this, enable this option so that PM will not handle the
> > > +	  console.
> > > +
> > > +	  If unsure, say N.
> > > +
> > >  config HIBERNATION
> > >  	bool "Hibernation (aka 'suspend to disk')"
> > >  	depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE
> > > diff --git a/kernel/power/console.c b/kernel/power/console.c
> > > index b8628be..8b9706f 100644
> > > --- a/kernel/power/console.c
> > > +++ b/kernel/power/console.c
> > > @@ -33,6 +33,7 @@ EXPORT_SYMBOL(pm_set_vt_switch);
> > >  
> > >  int pm_prepare_console(void)
> > >  {
> > > +#ifndef CONFIG_PM_DISABLE_CONSOLE
> > 
> > Can't you do this in a .h file instead?
> > 
> > And is this really needed?  Why has no other distro needed this change?
> > What is so wierd about your suspend setup to require this?
> 
> Reportedly, some systems have problems with graphics restoration from user
> space if the switch is done.
> 
> Anyway, I'd prefer to use the (already existing) pm_set_vt_switch() thing for
> that, maybe with the help of a command line parameter, to adding the .config
> option like this.

Ah. I didn't notice pm_set_vt_switch(). Then what I would like better is
the config option to set a default. For us, not doing the vt switch is a
good default (I don't think we have a case where suspend/resume requires
it).

Something like:

diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
index e7bea3e..6a042f8 100644
--- a/Documentation/kernel-parameters.txt
+++ b/Documentation/kernel-parameters.txt
@@ -2247,6 +2247,11 @@ and is between 256 and 4096 characters. It is defined in the file
 	vmpoff=		[KNL,S390] Perform z/VM CP command after power off.
 			Format: <command>
 
+	vt_switch=	[PM] Switch vt's during suspend/resume
+			Format: { 0 | 1 }
+			Default is normally '1', but depends on value of
+			the PM_DISABLE_CONSOLE config option.
+
 	waveartist=	[HW,OSS]
 			Format: <io>,<irq>,<dma>,<dma2>
 
diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig
index dcd165f..202f9b3 100644
--- a/kernel/power/Kconfig
+++ b/kernel/power/Kconfig
@@ -116,6 +116,25 @@ config SUSPEND_FREEZER
 
 	  Turning OFF this setting is NOT recommended! If in doubt, say Y.
 
+config PM_DISABLE_CONSOLE
+	bool "Disable Power Management messing with the active console"
+	depends on PM
+	default n
+	---help---
+	  By default, PM will take over the active console (generally, this means
+	  switching to the console when suspending from X). This can at times cause
+	  problems, especially if userspace suspend scripts try to do things with
+	  the console before or after suspending (e.g. calling vbestate).
+
+	  To work around this, enable this option so that PM will not handle the
+	  console.
+
+	  Note, this is the same as passing vt_switch=0 command line option, but
+	  just hardcodes the default. You can still pass vt_switch=1 with this
+	  option to get original behavior.
+
+	  If unsure, say N.
+
 config HIBERNATION
 	bool "Hibernation (aka 'suspend to disk')"
 	depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE
diff --git a/kernel/power/console.c b/kernel/power/console.c
index b8628be..db116d8 100644
--- a/kernel/power/console.c
+++ b/kernel/power/console.c
@@ -14,7 +14,11 @@
 #define SUSPEND_CONSOLE	(MAX_NR_CONSOLES-1)
 
 static int orig_fgconsole, orig_kmsg;
+#ifdef CONFIG_PM_DISABLE_CONSOLE
+static int disable_vt_switch = 1;
+#else
 static int disable_vt_switch;
+#endif
 
 /*
  * Normally during a suspend, we allocate a new console and switch to it.
@@ -80,4 +84,17 @@ void pm_restore_console(void)
 	release_console_sem();
 	kmsg_redirect = orig_kmsg;
 }
+
+static int __init set_vt_switch(char *str)
+{
+	if (str[0] == '0')
+		disable_vt_switch = 1;
+	else if (str[0] == '1')
+		disable_vt_switch = 0;
+	else
+		return -EINVAL;
+
+	return 0;
+}
+__setup("vt_switch=", set_vt_switch);
 #endif


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/