[Intel-gfx] [regression] [git pull] drm for 4.3

[Daniel Vetter]

On Tue, Sep 29, 2015 at 09:07:22PM -0400, davej at codemonkey.org.uk wrote:
> On Thu, Sep 24, 2015 at 04:26:28PM +0300, Jani Nikula wrote:
>  > On Thu, 24 Sep 2015, "davej at codemonkey.org.uk" <davej at codemonkey.org.uk> wrote:
>  > > On Wed, Sep 23, 2015 at 11:07:56AM +0000, Lankhorst, Maarten wrote:
>  > >  > Hey,
>  > >  > 
>  > >  > Dave Jones schreef op di 22-09-2015 om 21:49 [-0400]:
>  > >  > > On Tue, Sep 22, 2015 at 09:15:58AM -0700, Matt Roper wrote:
>  > >  > >  > On Tue, Sep 22, 2015 at 05:13:55PM +0200, Daniel Vetter wrote:
>  > >  > >  > > On Tue, Sep 22, 2015 at 08:00:17AM -0700, Jesse Barnes wrote:
>  > >  > >  > > > Cc'ing Maarten and Matt; I'm guessing this may be related to one of
>  > >  > >  > > > their recent patches.
>  > >  > >  > 
>  > >  > >  > Sounds like this showed up before my recent work, but I think I might
>  > >  > >  > have seen similar problems while working on atomic watermarks; the
>  > >  > >  > issues I was seeing were because the initial hardware readout could
>  > >  > >  > leave primary->visible set to true even when the CRTC was off.  My
>  > >  > >  > series (which is still under development) contains this patch to fix
>  > >  > >  > that:
>  > >  > >  > 
>  > >  > >  >         http://patchwork.freedesktop.org/patch/59564/
>  > >  > >  > 
>  > >  > >  > Does applying that help with the problems reported here?
>  > >  > > 
>  > >  > > No difference at all for me.
>  > >  > Looks like a (reopened) dup of 91952?
>  > >  > 
>  > >  > Can you apply "[PATCH] drm/i915: Add primary plane to mask if it's
>  > >  > visible", and get me the results?
>  > >
>  > > This doesn't apply on top of Linus' current tree.
>  > > If you let me know what it's dependant on, I'll do a build with
>  > > those patches tomorrow.
>  > 
>  > It's now part of the drm-intel-fixes pull request [1], maybe it's
>  > easiest to pull that in? Just four commits on top of
>  > v4.3-rc2. Alternatively pick it up from our repo [2].
> 
> The warning on boot seems to be gone as of rc3, but I can now trigger this pretty easily..

http://patchwork.freedesktop.org/patch/60618/

Cheers, Daniel

> 
> WARNING: CPU: 2 PID: 28911 at drivers/gpu/drm/drm_atomic.c:889 drm_atomic_get_property+0x244/0x2d0()
> CPU: 2 PID: 28911 Comm: trinity-c313 Not tainted 4.3.0-rc3-think+ #14
>  0000000000000379 ffff8801a1377c88 ffffffff8e35d5ec 0000000000000000
>  ffff8801a1377cc0 ffffffff8e07a862 ffff880500b392b8 ffff880500a13008
>  ffff880500b39290 ffff8804fe3806d8 ffff88003fa45668 ffff8801a1377cd0
> Call Trace:
>  [<ffffffff8e35d5ec>] dump_stack+0x4e/0x82
>  [<ffffffff8e07a862>] warn_slowpath_common+0x82/0xc0
>  [<ffffffff8e07a95a>] warn_slowpath_null+0x1a/0x20
>  [<ffffffff8e48efc4>] drm_atomic_get_property+0x244/0x2d0
>  [<ffffffff8e47c7fc>] drm_object_property_get_value+0x6c/0x70
>  [<ffffffff8e47a41f>] dpms_show+0x2f/0x70
>  [<ffffffff8e54ce30>] dev_attr_show+0x20/0x50
>  [<ffffffff8e27ae91>] ? sysfs_file_ops+0x41/0x60
>  [<ffffffff8e27b1a7>] sysfs_kf_seq_show+0xb7/0x110
>  [<ffffffff8e2799a6>] kernfs_seq_show+0x26/0x30
>  [<ffffffff8e21bba6>] seq_read+0xe6/0x430
>  [<ffffffff8e27a3f7>] kernfs_fop_read+0x127/0x170
>  [<ffffffff8e7e987b>] ? mutex_lock_nested+0x26b/0x3f0
>  [<ffffffff8e1f3af8>] __vfs_read+0x28/0xe0
>  [<ffffffff8e7e9897>] ? mutex_lock_nested+0x287/0x3f0
>  [<ffffffff8e214c09>] ? __fdget_pos+0x49/0x50
>  [<ffffffff8e214c09>] ? __fdget_pos+0x49/0x50
>  [<ffffffff8e1f43c6>] vfs_read+0x86/0x130
>  [<ffffffff8e1f50a9>] SyS_read+0x49/0xb0
>  [<ffffffff8e7eded7>] entry_SYSCALL_64_fastpath+0x12/0x6f
> ---[ end trace e053063c697a1355 ]---
> 
>  887         case DRM_MODE_OBJECT_CONNECTOR: {
>  888                 struct drm_connector *connector = obj_to_connector(obj);
>  889                 WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
>  890                 ret = drm_atomic_connector_get_property(connector,
>  891                                 connector->state, property, val);
>  892                 break;
>  893         }
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch