Hi,

On Saturday, 9 December 2006 16:35, Rafael J. Wysocki wrote:
> On Saturday, 9 December 2006 00:36, Pavel Machek wrote:
> > Hi!
> > 
> > > > > I wonder if we should start a test suite ;-).
> > > > > 
> > > > > > This means, however, that with this patch the behavior of a process 
> > > > > > (gdb)
> > > > > > after the resume may be different to its normal behavior, which is 
> > > > > > wrong.
> > > > > 
> > > > > Yep.
> > > 
> > > Okay, I think I know what to do so that it works.  The above symptoms are 
> > > not
> > > present with the appended patch.
> > 
> > Looks good to me. Thanks for you work!
> 
> Well, there's still a race possible in there, if SIGCONT comes after we have
> forced the SIGSTOP and before we call signal_wake_up().
> 
> I think we should force the SIGSTOP and call signal_wake_up() without
> releasing the lock.  I'll try to do something along these lines later today.

I think something like the appended patch will do.

Greetings,
Rafael


Signed-off-by: Rafael J. Wysocki <[EMAIL PROTECTED]>
---
 include/linux/sched.h  |    1 +
 kernel/power/process.c |   15 ++++++++-------
 kernel/signal.c        |   16 +++++++++++++++-
 3 files changed, 24 insertions(+), 8 deletions(-)

Index: linux-2.6.19-rc6-mm2/kernel/power/process.c
===================================================================
--- linux-2.6.19-rc6-mm2.orig/kernel/power/process.c    2006-12-08 
23:20:35.000000000 +0100
+++ linux-2.6.19-rc6-mm2/kernel/power/process.c 2006-12-10 00:37:24.000000000 
+0100
@@ -28,8 +28,7 @@ static inline int freezeable(struct task
        if ((p == current) || 
            (p->flags & PF_NOFREEZE) ||
            (p->exit_state == EXIT_ZOMBIE) ||
-           (p->exit_state == EXIT_DEAD) ||
-           (p->state == TASK_STOPPED))
+           (p->exit_state == EXIT_DEAD))
                return 0;
        return 1;
 }
@@ -61,9 +60,13 @@ static inline void freeze_process(struct
        unsigned long flags;
 
        if (!freezing(p)) {
-               freeze(p);
                spin_lock_irqsave(&p->sighand->siglock, flags);
-               signal_wake_up(p, 0);
+               freeze(p);
+               if (p->state == TASK_STOPPED) {
+                       force_sigstop_and_wake_up(p);
+               } else {
+                       signal_wake_up(p, 0);
+               }
                spin_unlock_irqrestore(&p->sighand->siglock, flags);
        }
 }
@@ -103,9 +106,7 @@ static unsigned int try_to_freeze_tasks(
                        if (frozen(p))
                                continue;
 
-                       if (p->state == TASK_TRACED &&
-                           (frozen(p->parent) ||
-                            p->parent->state == TASK_STOPPED)) {
+                       if (p->state == TASK_TRACED && frozen(p->parent)) {
                                cancel_freezing(p);
                                continue;
                        }
Index: linux-2.6.19-rc6-mm2/kernel/signal.c
===================================================================
--- linux-2.6.19-rc6-mm2.orig/kernel/signal.c   2006-12-08 23:20:35.000000000 
+0100
+++ linux-2.6.19-rc6-mm2/kernel/signal.c        2006-12-10 00:43:39.000000000 
+0100
@@ -962,6 +962,18 @@ force_sig_specific(int sig, struct task_
 }
 
 /*
+ * Force a SIGSTOP and make the task wake up.
+ *
+ * This is needed for the freezing of stopped tasks, because we want them to
+ * enter the refrigerator and be stopped again immediately after leaving it.
+ */
+void force_sigstop_and_wake_up(struct task_struct *t)
+{
+       specific_send_sig_info(SIGSTOP, SEND_SIG_FORCED, t);
+       signal_wake_up(t, 1);
+}
+
+/*
  * Test if P wants to take SIG.  After we've checked all threads with this,
  * it's equivalent to finding no threads not blocking SIG.  Any threads not
  * blocking SIG were ruled out because they are not running and already
@@ -1829,7 +1841,9 @@ finish_stop(int stop_count)
                read_unlock(&tasklist_lock);
        }
 
-       schedule();
+       do {
+               schedule();
+       } while (try_to_freeze());
        /*
         * Now we don't run again until continued.
         */
Index: linux-2.6.19-rc6-mm2/include/linux/sched.h
===================================================================
--- linux-2.6.19-rc6-mm2.orig/include/linux/sched.h     2006-12-10 
00:20:07.000000000 +0100
+++ linux-2.6.19-rc6-mm2/include/linux/sched.h  2006-12-10 00:42:44.000000000 
+0100
@@ -1350,6 +1350,7 @@ extern int kill_pg_info(int, struct sigi
 extern void do_notify_parent(struct task_struct *, int);
 extern void force_sig(int, struct task_struct *);
 extern void force_sig_specific(int, struct task_struct *);
+extern void force_sigstop_and_wake_up(struct task_struct *t);
 extern int send_sig(int, struct task_struct *, int);
 extern void zap_other_threads(struct task_struct *p);
 extern int kill_pg(pid_t, int, int);

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Suspend-devel mailing list
Suspend-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/suspend-devel

Reply via email to