Let's make things a bit easier to debug when things go bad (potentially
under console_lock).
Signed-off-by: NRob Clark <robdclark@gmail.com>
Reviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Andrew Morton wrote:
> On Wed, 12 Nov 2014 13:08:55 +0900 Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> wrote:
>
> > Andrew Morton wrote:
> > > Poor ttm guys - this is a bit of a trap we set for them.
> >
> > Commit a91576d7 ("drm/ttm: Pass GFP flags in order to avoid deadlock.")
> > changed to use sc->gfp_mask rather than GFP_KERNEL.
> >
> > -       pages_to_free = kmalloc(npages_to_free * sizeof(struct page *),
> > -                       GFP_KERNEL);
> > +       pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), gfp);
> >
> > But this bug is caused by sc->gfp_mask containing some flags which are not
> > in GFP_KERNEL, right? Then, I think
> >
> > -       pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), gfp);
> > +       pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), gfp & GFP_KERNEL);
> >
> > would hide this bug.
> >
> > But I think we should use GFP_ATOMIC (or drop __GFP_WAIT flag)
>
> Well no - ttm_page_pool_free() should stop calling kmalloc altogether.
> Just do
>
> 	struct page *pages_to_free[16];
>
> and rework the code to free 16 pages at a time.  Easy.

Well, ttm code wants to process 512 pages at a time for performance.
Memory footprint increased by 512 * sizeof(struct page *) buffer is
only 4096 bytes. What about using static buffer like below?
----------
>From d3cb5393c9c8099d6b37e769f78c31af1541fe8c Mon Sep 17 00:00:00 2001
From: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Date: Thu, 13 Nov 2014 22:21:54 +0900
Subject: [PATCH] drm/ttm: Avoid memory allocation from shrinker functions.

Commit a91576d7 ("drm/ttm: Pass GFP flags in order to avoid
deadlock.") caused BUG_ON() due to sc->gfp_mask containing flags
which are not in GFP_KERNEL.

  https://bugzilla.kernel.org/show_bug.cgi?id=87891

Changing from sc->gfp_mask to (sc->gfp_mask & GFP_KERNEL) would
avoid the BUG_ON(), but avoiding memory allocation from shrinker
function is better and reliable fix.

Shrinker function is already serialized by global lock, and
clean up function is called after shrinker function is unregistered.
Thus, we can use static buffer when called from shrinker function
and clean up function.
Signed-off-by: NTetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: stable <stable@kernel.org> [2.6.35+]
Signed-off-by: NDave Airlie <airlied@redhat.com>

sizeof(u8) is always 1.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Commit f9b9faf6d94dd29eab8c128905c7d091f955481d "drm: flip-work: change
drm_flip_work_init prototype" changed the drm_flip_work_init prototype
to a void function, which makes 'ret' an unused variable.
Signed-off-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Implement crtc page_flip callback for bochsdrm.  The qemu stdvga has no
vblank signaling, so we have to fake it.  We do so by instantly calling
drm_send_vblank_event.  Tested with kmscon.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Remove the mapping offset from the bo backing the fbdev framebuffer.
Wire up fbdev mmap function to map the backing bo using ttm_fbdev_mmap.
With that patch in place mmap(/dev/fb0) works as expected.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Recently (qemu 2.2+) the qemu stdvga got a register to switch the vga
framebuffer endianness.  This patch adds code to explicitly set the
endianness of the framebuffer.  In most cases this has no effect as
the default is guest architecture endianness.  It is needed though in
case a architecture supports both big and little endian, i.e. for
ppc64le.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Drop some leftover, commented code.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Added a4xx GPU support.
Signed-off-by: NAravind Ganesan <aravindg@codeaurora.org>
Signed-off-by: NRob Clark <robdclark@gmail.com>

Register offsets have changed between a3xx and a4xx GPUs.
To be able access these registers in common code, we create
a lookup table, and set of read-write APIs to access the
register through the lookup table.
Signed-off-by: NAravind Ganesan <aravindg@codeaurora.org>
[robclark: remove REG_ADRENO_UNDEFINED, just use zero, and minor
tweaks for latest generated headers]
Signed-off-by: NRob Clark <robdclark@gmail.com>

Use pre-computed iova when unmapping, to reduce the places we assume iova
and mmap offset are (at the moment) the same.  And get rid of an extra
drm_gem_free_mmap_offset() call (since it is already called from
drm_gem_object_release())
Signed-off-by: NRob Clark <robdclark@gmail.com>

Convert mdp4 display controller backend to atomic helpers.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Signed-off-by: NRob Clark <robdclark@gmail.com>

The core parts for async commit.
Signed-off-by: NRob Clark <robdclark@gmail.com>

It's a problem that can't happen yet, since we don't support any
multi-planar formats yet.  But let's avoid nasty surprises when the
time comes.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Atomic wants to split the prepare/pin from where we actually program the
scanout address (so that any part that can fail is done synchronously).
Add some fb/gem apis to make this easier to use from the kms parts.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Signed-off-by: NRob Clark <robdclark@gmail.com>

Give ourselves a way to wait for certain fence #..  makes it easier to
wait on a set of bo's, which we'll need for atomic.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Simplify things a bit for atomic, gets rid of some bookkeeping, and
makes the code cleaner.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Simplify things a bit for atomic, gets rid of some bookkeeping, and
makes the code cleaner.

TODO move iterator macro somewhere common.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Since we are configuring things via MDP4_PIPE regs in the plane, it seems
like setting the dimensions of the primary plane on the OVLP/DMA regs in
crtc is unnecessary.  This will make life easier when we want to do a
nofb modeset.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Since primary-plane support in core, we can just use crtc->primary.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Since primary-plane support in core, we can just use crtc->primary.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Currently forcing the video mode from the kernel command line (for example
video=HDMI-A-1:1280x720-16@60) does not correctly set the number of bits
per pixel. This is due to a rather aggressive override in
msm_fbdev_create(). This is a particular problem for Android bring up
because the software EGL fallbacks don't support 32bpp.

Since the overrides are actually the default values anyway then this
problem can be trivially fixed by removing the overrides completely.

Change was tested by dd'ing a test image to /dev/fb0 with no video=
(still 32bpp), video=1920x1080-32@60, video=1920x1080-24@60 and
video=1920x1080-16@60 .
Signed-off-by: NDaniel Thompson <daniel.thompson@linaro.org>
Cc: David Airlie <airlied@linux.ie>
Cc: Rob Clark <robdclark@gmail.com>
Signed-off-by: NRob Clark <robdclark@gmail.com>

Currently msm does not implement gem_prime_mmap. Without this it is not
possible to draw onto a dma-buf from userspace (making its very hard to
implement the Android rendering model).

Fixing this is just a matter of adding a little boilerplate.
Signed-off-by: NDaniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: NRob Clark <robdclark@gmail.com>

Split up hdmi_init() into hdmi_init() (done at hdmi sub-device
bind/probe time) and hdmi_modeset_init() done from master driver's
modeset_init().

Anything that can fail due to dependencies on other drivers which
may be missing or not probed yet should go in hdmi_init(), so that
devm error/cleanup paths work properly.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Signed-off-by: NRob Clark <robdclark@gmail.com>

Move anything that can fail after call to base class msm_gpu_init().
This way, if we fail, active_list has already been initialized so we
don't trip 'WARN_ON(!list_empty(&gpu->active_list))' in
msm_gpu_cleanup().
Signed-off-by: NRob Clark <robdclark@gmail.com>

Fixes a potential error, spotted by Felipe with randconfig:

-----
drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c: In function ‘mdp4_kms_init’:
drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c:384:2: error: implicit declaration \
        of function ‘devm_regulator_get_exclusive’ [-Werror=implicit-function-declaration]
  mdp4_kms->vdd = devm_regulator_get_exclusive(&pdev->dev, "vdd");
  ^
drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c:384:16: error: assignment makes \
        pointer from integer without a cast [-Werror]
  mdp4_kms->vdd = devm_regulator_get_exclusive(&pdev->dev, "vdd");
                ^
-----

Also add a brief comment explaining the use of _get_exclusive()
Reported-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NRob Clark <robdclark@gmail.com>

This was pointless, forgot to remove the code.
Signed-off-by: NDave Airlie <airlied@redhat.com>

This passes the guest preferences for a where to place the
outputs through to userspace. Userspace would need to be updated
to take note of this information, X server and GNOME.
Signed-off-by: NDave Airlie <airlied@redhat.com>

Virtual GPUs would like to give the guest some indication where on the screen
the outputs are layed out. So far we only provide modes, these
properties could be exposed to userspace so the desktop environment
could use them as hints to set the correct offsets.

v2: rename properties to be more consistent.
Signed-off-by: NDave Airlie <airlied@redhat.com>

While looking through drm_crtc.c to double-check make locking changes
I've noticed that there's a few other places that would now benefit
from simplified return value handling.

So let's flatten the control flow and replace and always 0 ret with 0
where possible.
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>
Reviewed-by: NSean Paul <seanpaul@chromium.org>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Don't BUG out if the link reports an invalid (or plain unknown)
bandwidth value, but report the failure and fail gracefully.

Fixes a trivial compiler warning in case the BUG is ever compiled away.

Link: http://lkml.kernel.org/p/1415785566-12758-1-git-send-email-geert@linux-m68k.orgSigned-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

For async commit, it is *intentional* that those locks are not held.
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NRob Clark <robdclark@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Now that we're using lists instead of kfifo to store drm flip-work tasks
we do not need the size parameter passed to drm_flip_work_init function
anymore.
Moreover this function cannot fail anymore, we can thus remove the return
code.

Modify drm_flip_work_init users to take account of these changes.

[airlied: fixed two unused variable warnings]
Signed-off-by: NBoris BREZILLON <boris.brezillon@free-electrons.com>
Reviewed-by: NRob Clark <robdclark@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Make use of lists instead of kfifo in order to dynamically allocate
task entry when someone require some delayed work, and thus preventing
drm_flip_work_queue from directly calling func instead of queuing this
call.
This allow drm_flip_work_queue to be safely called even within irq
handlers.

Add new helper functions to allocate a flip work task and queue it when
needed. This prevents allocating data within irq context (which might
impact the time spent in the irq handler).
Signed-off-by: NBoris BREZILLON <boris.brezillon@free-electrons.com>
Reviewed-by: NRob Clark <robdclark@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Use the new pipe config values to calculate the updated pll dividers.

This regression was introduced in

commit 0dbdf89f27b17ae1eceed6782c2917f74cbb5d59
Author: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Date:   Wed Oct 29 11:32:33 2014 +0200

    drm/i915: Add infrastructure for choosing DPLLs before disabling crtcs

	and

	commit 00d958817dd3daaa452c221387ddaf23d1e4c06f
	Author: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
	Date:   Wed Oct 29 11:32:36 2014 +0200

	    drm/i915: Covert remaining platforms to choose DPLLS before disabling CRTCs

v2: Use intel_pipe_will_have_type() to look at new configuration - Ander
Signed-off-by: NBob Paauwe <bob.j.paauwe@intel.com>
CC: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Tested-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

dma_free_writecombine() must not be called on a buffer that couldn't be
allocated. Check for a valid virtual address before attempting to free
the memory to avoid a crash.
Reported-by: NSean Paul <seanpaul@chromium.org>
Reviewed-by: NSean Paul <seanpaul@chromium.org>
Signed-off-by: NThierry Reding <treding@nvidia.com>

This helps in determining what errors happened at specifics points in
the initialization sequence.
Signed-off-by: NThierry Reding <treding@nvidia.com>