On Thu, Dec 18, 2014 at 06:04:45PM -0800, Fengguang Wu wrote:
> 
> Hi Thierry,
> 
> FYI, here is another bisect result.

Thanks. Looking at the cirrus driver more closely it already employs the
big hammer upon failure and ends up calling drm_fb_helper_fini() via the
cirrus_driver_unload() call in cirrus_driver_load()'s cleanup path.

> commit 7bf2844ed14de049a4ccec26eb3d8b4c3d5c1971
> Author:     Thierry Reding <treding@nvidia.com>
> AuthorDate: Wed Dec 17 16:39:59 2014 +0100
> Commit:     Daniel Vetter <daniel.vetter@ffwll.ch>
> CommitDate: Wed Dec 17 22:05:08 2014 +0100
> 
>     drm/fb-helper: Propagate errors from initial config failure
>     
>     Make drm_fb_helper_initial_config() return an int rather than a bool so
>     that the error can be properly propagated. While at it, update drivers
>     to propagate errors further rather than just ignore them.
>     
>     Cc: David Airlie <airlied@linux.ie>
>     Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
>     Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
>     Cc: Rob Clark <robdclark@gmail.com>
>     Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
>     Cc: Alex Deucher <alexander.deucher@amd.com>
>     Cc: Christian König <christian.koenig@amd.com>
>     Cc: Ben Skeggs <bskeggs@redhat.com>
>     Signed-off-by: Thierry Reding <treding@nvidia.com>
>     Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
>     Reviewed-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
>     Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Using the attached patch on top of this I can no longer reproduce either
this or the earlier result. I'll send a revised patch with this squashed
in.

Thanks,
Thierry