Re: 2.6.24-rc2 STD with s2disk fails to activate suspended system after loading - now 2.6.24-rc3

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

On Tuesday, 20 of November 2007, Ingo Molnar wrote:
> 
> * Rafael J. Wysocki <rjw@sisk.pl> wrote:
> 
> > > increasing CONFIG_BLK_DEV_RAM_SIZE from 7777 to 131072 hasn't 
> > > changed the non-functioning of 2.6.24-rc3
> > > 
> > > s2disk works with 2.6.23.8 ; I tested 4 cycles in a row, 2 from 
> > > console and 2 from within X
> > 
> > I've attached a patch to the bugzilla entry, please test it.
> [...]
> > --- linux-2.6.orig/init/do_mounts_initrd.c
> > +++ linux-2.6/init/do_mounts_initrd.c
> > @@ -55,6 +55,8 @@ static void __init handle_initrd(void)
> >  	sys_mount(".", "/", NULL, MS_MOVE, NULL);
> >  	sys_chroot(".");
> >  
> > +	current->flags |= PF_NOFREEZE;
> > +
> >  	pid = kernel_thread(do_linuxrc, "/linuxrc", SIGCHLD);
> >  	if (pid > 0)
> >  		while (pid != sys_wait4(-1, NULL, 0, NULL)) {
> 
> this is not the first time (and not the last time) that a missing 
> PF_NOFREEZE is causing hard to debug suspend problems. I think we should 
> be more robust about this and at minimum include some debug mechanism 
> that determines when a PF_NOFREEZE annotation is missing. (or at least 
> detect the condition somehow and report it) This bug took 10 days to 
> track down.

In fact all of the needed pieces are in there, I only needed the right dmesg
output to figure out what's wrong.

Also, I don't think we should use PF_NOFREEZE in there, as this is an
exceptional case.  I'd prefer to use the appended fix, which also has been
tested.

Thanks,
Rafael


---
 init/do_mounts_initrd.c |   12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

Index: linux-2.6/init/do_mounts_initrd.c
===================================================================
--- linux-2.6.orig/init/do_mounts_initrd.c
+++ linux-2.6/init/do_mounts_initrd.c
@@ -55,12 +55,18 @@ static void __init handle_initrd(void)
 	sys_mount(".", "/", NULL, MS_MOVE, NULL);
 	sys_chroot(".");
 
+	/*
+	 * In case that a resume from disk is carreid out by linuxrc or one of
+	 * its children, we need to tell the freezer not to wait for us
+	 */
+	current->flags |= PF_FREEZER_SKIP;
+
 	pid = kernel_thread(do_linuxrc, "/linuxrc", SIGCHLD);
 	if (pid > 0)
-		while (pid != sys_wait4(-1, NULL, 0, NULL)) {
-			try_to_freeze();
+		while (pid != sys_wait4(-1, NULL, 0, NULL))
 			yield();
-		}
+
+	current->flags &= ~PF_FREEZER_SKIP;
 
 	/* move initrd to rootfs' /old */
 	sys_fchdir(old_fd);