On Fri, Nov 04, 2011 at 12:34:58PM -0500, Andrew Watts wrote:
> On Fri, Nov 04, 2011 at 09:14:31AM -0700, Tejun Heo wrote:
> > Andrew, can you please kill the X server after the hang and see
> > whether that brings the system back? I think sshd should still work
> > and if not you can write a script to kill the X server after 30secs
> > after resume (and kill that script if resume succeeds).
>
> Hi Tejun & Jerome.
>
> Tejun: My script killed X after a small wait from a bad resume. I am returned
> to the funcioning calling console. I can then re-start X and all subsystems
> seem to working fine.
>
> Jerome: I posted a post-bad-resume dmesg snippet with a GPU lockup warning
> and trace here:
>
> http://article.gmane.org/gmane.linux.kernel/1210057
>
> Did you need to see more?
>
> Thanks.
>
> ~ Andy
>
Attached patch might hide the issue.
Cheers,
Jerome
>From a13af3d0bff9e093160da63f22c6b44e436aa570 Mon Sep 17 00:00:00 2001
From: Jerome Glisse <jglisse@xxxxxxxxxx>
Date: Fri, 4 Nov 2011 13:41:37 -0400
Subject: [PATCH] drm/radeon/kms: make sure GPU is idle and working at end of
resume
This might hide some GPU lockup issue deadlock on resume.
Signed-off-by: Jerome Glisse <jglisse@xxxxxxxxxx>
---
drivers/gpu/drm/radeon/radeon_device.c | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index c33bc91..1158e55 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -947,6 +947,7 @@ int radeon_resume_kms(struct drm_device *dev)
list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
drm_helper_connector_dpms(connector, DRM_MODE_DPMS_ON);
}
+ radeon_fence_wait_last(rdev);
return 0;
}
--
1.7.6.4
_______________________________________________
linux-pm mailing list
linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linuxfoundation.org/mailman/listinfo/linux-pm
