[PATCH] drm/fsl-dcu: use flat regmap cache

[Daniel Vetter]

On Sat, Jun 4, 2016 at 3:51 AM, Stefan Agner <stefan at agner.ch> wrote:
> On 2016-06-03 16:00, Daniel Vetter wrote:
>> On Fri, Jun 03, 2016 at 03:43:19PM -0700, Stefan Agner wrote:
>>> Using flat regmap cache instead of RB-tree to avoid the following
>>> lockdep warning on driver load:
>>> WARNING: CPU: 0 PID: 1 at kernel/locking/lockdep.c:2755 lockdep_trace_alloc+0x15c/0x160()
>>> DEBUG_LOCKS_WARN_ON(irqs_disabled_flags(flags))
>>>
>>> The RB-tree regmap cache needs to allocate new space on first
>>> writes. However, allocations in an atomic context (e.g. when a
>>> spinlock is held) are not allowed. The function regmap_write
>>> calls map->lock, which acquires a spinlock in the fast_io case.
>>> Since the FSL DCU driver uses MMIO, the regmap bus of type
>>> regmap_mmio is being used which has fast_io set to true.
>>>
>>> Use flat regmap cache and specify max register to be large
>>> enouth to cover all registers available in LS1021a and Vybrids
>>> register space.
>>>
>>> Signed-off-by: Stefan Agner <stefan at agner.ch>
>>> Cc: Mark Brown <broonie at kernel.org>
>>> Cc: stable at vger.kernel.org
>>> ---
>>> While regmap cache is used for suspend/resume only (which is
>>> broken in its current state) Mark noted that using the RB regmap
>>> cache can also cause issues during initialization of the driver.
>>> This patch migrates to flat regmap cache (which we can also use
>>> to fix the issue in stable kernels), and yet another patchset
>>> moves to the atomic suspend/resume helpers (which will not go
>>> into stable...)
>>
>> While talking suspend/resume, I highly recommend
>> drm_atomic_helper_suspend and drm_atomic_helper_resume. In my experience
>> dumb safe/restore of registers for restoring modeset state leads to tears.
>> -Daniel
>
> Since you commented to the patchset implementing exactly that I assume
> you have seen it.

Yup, that was pretty awesome coincidence ;-)

> I still would like to have this patch applied to fix the regmap cache
> warning also for older kernels.

Sure, this was just a quick comment on your remark that there's still
trouble left with suspend/resume.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch