Re: [GIT PULL pm-next] freezer: fix various bugs and simplify implementation

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Sunday, August 21, 2011, Tejun Heo wrote:
> Hello, Rafafel.
> 
> On Sat, Aug 20, 2011 at 06:33:38PM +0200, Rafael J. Wysocki wrote:
> > >   ssh://master.kernel.org/pub/scm/linux/kernel/git/tj/misc.git freezer
> > 
> > Pulled and stored in the pm-freezer branch in my tree, and merged into
> > the linux-next branch.
> 
> Cool.
> 
> > > FYI, this patchset will cause a conflict with s390 TIF flag fix patch.
> > > The conflict is trivial and Stephen should be able to handle it
> > > without any problem.  Also, I'm planning on doing some further work on
> > > cgroup freezer and then will try to bridge it with job control.  If
> > > that plan fans out, I might ask Oleg to pull from the pm tree.
> > 
> > I'm not sure if Linus likes it.  He generally doesn't want the trees
> > that he pulls from to be entangled this way.
> 
> The job control portion has to go through Linus anyway, so let's see
> how that flies.
> 
> > > This shouldn't matter too much either way but it *might* be a good idea to
> > > keep this line of patches in a separate branch.
> > 
> > I'm going to keep it in the pm-freezer branch anyway (there may be patches
> > on top of it, though)
> 
> Yeah, I'm pretty sure it will need some fix too.

Speaking of which, the addition of might_sleep() to try_to_freeze()
causes a badly looking backtrace to appear during reboot on ARM,
so I'd prefer it to go into __refrigerator().

Please tell me what you think of the patch below.

Rafael

---
From: Rafael J. Wysocki <rjw@sisk.pl>
Subject: PM / Freezer: Move might_sleep() from try_to_freeze()

There are some code paths that call try_to_freeze() from interrupt
context, but doing so they know that the current process cannot
possible be freezing (e.g. during reboot on ARM).  However, the
recently added might_sleep() annotation in try_to_freeze()
triggers in those cases, making it look like there were bugs in
those places, which really isn't the case.

Therefore move might_sleep() from try_to_freeze() to
__refrigerator() so that it doesn't produce false positives.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
---
 include/linux/freezer.h |    1 -
 kernel/freezer.c        |    2 ++
 2 files changed, 2 insertions(+), 1 deletion(-)

Index: linux/include/linux/freezer.h
===================================================================
--- linux.orig/include/linux/freezer.h
+++ linux/include/linux/freezer.h
@@ -41,7 +41,6 @@ extern void thaw_processes(void);
 
 static inline bool try_to_freeze(void)
 {
-	might_sleep();
 	if (likely(!freezing(current)))
 		return false;
 	return __refrigerator(false);
Index: linux/kernel/freezer.c
===================================================================
--- linux.orig/kernel/freezer.c
+++ linux/kernel/freezer.c
@@ -54,6 +54,8 @@ bool __refrigerator(bool check_kthr_stop
 	bool was_frozen = false;
 	long save;
 
+	might_sleep();
+
 	/*
 	 * No point in checking freezing() again - the caller already did.
 	 * Proceed to enter FROZEN.