On Thursday, November 29, 2012 03:03:04 PM Jan-Matthias Braun wrote:
> Hi Rafael,
>
> thanks for looking into this.
>
> Am Donnerstag, 29. November 2012, 01:20:04 schrieb Rafael J. Wysocki:
> > On Tuesday, November 27, 2012 09:26:11 PM Jiri Kosina wrote:
> > > On Sun, 25 Nov 2012, Jan-Matthias Braun wrote:
> > > > I have now done a git bisection from 3.0 to 3.1 and have found commit
> > > > 1e2ef05bb8cf851a694d38e9170c89e7ff052741 PM: Limit race conditions
> > > > between runtime PM and system sleep (v2)
> > > > to be the first one to introduce the necessity of a module reload after
> > > > resume.
> >
> > So with that commit your device doesn't work after resume from system
> > suspend
> > unless you reload the driver?
>
> Yes. Additionally, with later kernel revisions even this won't help, but I
> could imagine this to be a consequence of the same problem.
>
> > > > I hope that this helps in finding a solution. As I (again) don't
> > > > immediatly
> > > > know how to go on, I kindly ask you to give me some hints for
> > > > testing/patching
> > > > or even a possible solution. :-)
> >
> > Can you apply the patch below to the current mainline kernel and see if it
> > makes any
> > difference, please?
>
> It does: Current mainline kernel (version >3.7.0-rc7 from git) is not showing
> the problem, if the patch is applied.
If you apply the patch below instead of the previous one, does it make
the problem return, or is it still good?
Rafael
---
drivers/base/power/main.c | 9 ++++-----
1 file changed, 4 insertions(+), 5 deletions(-)
Index: linux/drivers/base/power/main.c
===================================================================
--- linux.orig/drivers/base/power/main.c
+++ linux/drivers/base/power/main.c
@@ -513,6 +513,8 @@ static int device_resume_early(struct de
Out:
TRACE_RESUME(error);
+
+ pm_runtime_enable(dev);
return error;
}
@@ -589,8 +591,6 @@ static int device_resume(struct device *
if (!dev->power.is_suspended)
goto Unlock;
- pm_runtime_enable(dev);
-
if (dev->pm_domain) {
info = "power domain ";
callback = pm_op(&dev->pm_domain->ops, state);
@@ -930,6 +930,8 @@ static int device_suspend_late(struct de
pm_callback_t callback = NULL;
char *info = NULL;
+ __pm_runtime_disable(dev, false);
+
if (dev->power.syscore)
return 0;
@@ -1133,11 +1135,8 @@ static int __device_suspend(struct devic
Complete:
complete_all(&dev->power.completion);
-
if (error)
async_error = error;
- else if (dev->power.is_suspended)
- __pm_runtime_disable(dev, false);
return error;
}
--
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
