We need to make sure that userspace keeps on following the contract,
otherwise we won't be able to use the reserved fields at all.

v2: Add DRM_DEBUG (Chris)

Testcase: igt/gem_exec_params/*-dirt
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

A bit tricky since 0 is also a valid constant ...

v2: Add DRM_DEBUG (Chris)

Testcase: igt/gem_exec_params/rel-constants-*
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Currently we catch it, but silently succeed. Our userspace is
better than this.

v2: Add DRM_DEBUG (Chris)

Testcase: igt/gem_exec_params/sol-reset-*
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

If we include the expected values for the failing ring register checks,
it makes it marginally easier to see which is the culprit.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NOscar Mateo <oscar.mateo@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

During module load, if we fail to initialise the rings, we abort the
load reporting EIO. However during resume, even though we report EIO as
we fail to reinitialize the ringbuffers, the resume continues and the
device is restored - albeit in a non-functional state. As we cannot
execute any commands on the GPU, it is effectively wedged, mark it so.

As we now preserve the ringbuffers across resume, this should prevent
UXA from falling into the trap of repeatedly sending invalid
batchbuffers and dropping all further rendering into /dev/null.
Reported-and-tested-by: NJiri Kosina <jkosina@suse.cz>
References: https://bugs.freedesktop.org/show_bug.cgi?id=76554Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NOscar Mateo <oscar.mateo@intel.com>
[danvet: Drop unused error, spotted by Oscar.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Even without enabling the ringbuffers to allow command execution, we can
still control the display engines to enable modesetting. So make the
ringbuffer initialization failure soft, and mark the GPU as wedged
instead.

v2: Only treat an EIO from ring initialisation as a soft failure, and
abort module load for any other failure, such as allocation failures.

v3: Add an *ERROR* prior to declaring the GPU wedged so that it stands
out like a sore thumb in the logs
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: NOscar Mateo <oscar.mateo@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Tearing down the ring buffers across resume is overkill, risks
unnecessary failure and increases fragmentation.

After failure, since the device is still active we may end up trying to
write into the dangling iomapping and trigger an oops.

v2: stop_ringbuffers() was meant to call stop(ring) not
cleanup(ring) during resume!
Reported-by: NJae-hyeon Park <jhyeon@gmail.com>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=72351
References: https://bugs.freedesktop.org/show_bug.cgi?id=76554Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NOscar Mateo <oscar.mateo@intel.com>
[danvet: s/ring->obj == NULL/!intel_ring_initialized(ring)/ as
suggested by Oscar.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

For readibility and guess at the meaning behind the constants.

v2: Claim only the meagerest connections with reality.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NOscar Mateo <oscar.mateo@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

I don't think this is necessary; at least it doesn't appear to be on my
BYT.  Dropping it speeds up our shutdown code a little, in some cases
resulting in faster init times.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Remove occurrences of unused struct qxl_device pointer in functions
qxl_ttm_fault() and qxl_init_mem_type().

Detected by Coverity: CID 1019128, CID 1019129.
Signed-off-by: NChristian Engelmayer <cengelma@gmx.at>
Signed-off-by: NDave Airlie <airlied@redhat.com>

For QXL hw we really want the bits to be replaced as we change
the preferred mode on the fly, and the same goes for virgl when
I get to it, however the original fix for this seems to have caused
a wierd regression on Intel G33 that in a stunning display of failure
at opposition to his normal self, Daniel failed to diagnose.

So we are left doing this, ugly ugly ugly ugly, Daniel you fixed
that G33 yet?, ugly, ugly.
Tested-by: NMarc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Use %pad for dma_addr_t, because a dma_addr_t type can vary
based on build options. So, it prevents possible build warnings
in printks.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Reviewed-by: NDaniel Kurtz <djkurtz@chromium.org>
Signed-off-by: NInki Dae <inki.dae@samsung.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

devm_ioremap_resource() returns an error pointer, not NULL. Thus,
the result should be checked with IS_ERR().
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

exynos_drm_crtc_mode_set assigns primary framebuffer to plane without
taking reference. Then during framebuffer removal it is dereferenced twice,
causing oops. The patch fixes it.
Signed-off-by: NAndrzej Hajda <a.hajda@samsung.com>
Signed-off-by: NInki Dae <inki.dae@samsung.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

In commit a51435a3
Author: Naresh Kumar Kachhi <naresh.kumar.kachhi@intel.com>
Date:   Wed Mar 12 16:39:40 2014 +0530

    drm/i915: disable rings before HW status page setup

we reordered stopping the rings to do so before we set the HWS register.
However, there is an extra workaround for g45 to reset the rings twice,
and for consistency we should apply that workaround before setting the
HWS to be sure that the rings are truly stopped.

Reference: http://lkml.kernel.org/r/20140423202248.GA3621@amd.pavel.ucw.czTested-by: NPavel Machek <pavel@ucw.cz>
Cc: Naresh Kumar Kachhi <naresh.kumar.kachhi@intel.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>

The status bits are unconditionally set, the control bits only enable
the actual interrupt generation. Which means if we get some random
other interrupts we'll bogusly complain about them.

So restrict the WARN to platforms with a sane hotplug interrupt
handling scheme. And even more important also don't attempt to process
the hpd bit since we've detected a storm already. Instead just clear
the bit silently.

This WARN has been introduced in

commit b8f102e8
Author: Egbert Eich <eich@suse.de>
Date:   Fri Jul 26 14:14:24 2013 +0200

    drm/i915: Add messages useful for HPD storm detection debugging (v2)

before that we silently handled the hpd event and so partially
defeated the storm detection.

v2: Pimp commit message (Jani)

Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Egbert Eich <eich@suse.de>
Cc: bitlord <bitlord0xff@gmail.com>
Reported-by: Nbitlord <bitlord0xff@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>

The hw cursor is relatively adept at triggering underflows, which
manifest as a "blue flash" (since blue is configured as the underflow
color).  Juggle a few things around to tighten up the timing for setting
cursor registers in DONE irq.

And most importantly, don't ever disable the hw cursor.  Instead flip it
to a blank/empty cursor.  This seems far more reliable, as even simply
clearing the cursor-enable bit (with no other updates in previous/
following frames) can in some cases cause underflow.

v1: original
v2: add missing locking spotted by Micah

Cc: Micah Richert <richert@braincorporation.com>
Signed-off-by: NRob Clark <robdclark@gmail.com>

Since X11 is going to create an XR24 fb, if the pixel formats do not
match then crtc helpers will think it is a full modeset even if mode is
the same, which prevents smooth/flickerless handover from fbcon/plymouth
to X11.
Signed-off-by: NRob Clark <robdclark@gmail.com>

Signed-off-by: NMicah Richert <richert@braincorporation.com>
Signed-off-by: NRob Clark <robdclark@gmail.com>

In Matt Ropers primary plane series a set of prep patches like

commit af2b653b
Author: Matt Roper <matthew.d.roper@intel.com>
Date:   Tue Apr 1 15:22:32 2014 -0700

    drm/i915: Restrict plane loops to only operate on overlay planes (v2)

ensured that all exisiting users of the mode_config->plane_list
wouldn't change behaviour. Unfortunately tegra seems to have fallen
through the cracks. Fix it.

This regression was introduced in

commit e13161af
Author: Matt Roper <matthew.d.roper@intel.com>
Date:   Tue Apr 1 15:22:38 2014 -0700

    drm: Add drm_crtc_init_with_planes() (v2)

The result was that we've unref'ed the fb for the primary plane twice,
leading to a use-after free bug. This is because the drm core will
already set crtc->primary->fb to NULL and do the unref for us, and the
crtc disable hook is called by the drm crtc helpers for exactly this
case.

Aside: Now that the fbdev helpers clean up planes there's no longer a
need to do this in drivers. So this could probably be nuked entirely
in linux-next.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: NMatt Roper <matthew.d.roper@intel.com>
Tested-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>

When PPGTT was disabled by default, the patch also prevented the user
from overriding this behavior via module parameter. Being able to test
this on arbitrary kernels is extremely beneficial to track down the
remaining bugs. The patch that prevented this was:

commit 93a25a9e
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Thu Mar 6 09:40:43 2014 +0100

    drm/i915: Disable full ppgtt by default

By default PPGTT is set to -1. 0 means off, 1 means aliasing only, 2
means full, all other values are reserved.
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>

If the inherited BIOS framebuffer is smaller than the mode selected for
fbdev, then if we continue to use it then we cause display corruption as
we do not setup the panel fitter to upscale.

Regression from commit d978ef14
Author: Jesse Barnes <jbarnes@virtuousgeek.org>
Date:   Fri Mar 7 08:57:51 2014 -0800

    drm/i915: Wrap the preallocated BIOS framebuffer and preserve for KMS fbcon v12

v2: Add a debug message to track the discard of the BIOS fb.
v3: Ville pointed out the difference between ref/unref
Reported-by: NKnut Petersen <Knut_Petersen@t-online.de>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=77767Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>

We already check that the buffer object we're accessing is registered with
the file. Now also make sure that we can't DMA across buffer object boundaries.

v2: Code commenting update.

Cc: stable@vger.kernel.org
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: NJakob Bornecrantz <jakob@vmware.com>

Especially not on modesetting drivers - this is used to size
the driver private structure for legacy drm buffers.
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Reviewed-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

With the last patch to ditch the ->get_name callbacks the last
user is now gone.
Reviewed-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The only user is the info debugfs file, so we only need something
human readable. Now for both pci and platform devices we've used the
name of the underlying device driver, which matches the name of the
drm driver in all cases. So we can just use that instead.

The exception is usb, which used a generic "USB". Not to harmful with
just one usb driver, but better to use "udl", too.

With that converted we can rip out all the ->get_name implementations.
Reviewed-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This was only ever used to pretty-print the irq driver name. And on
kms systems due to set_version bonghits we never set up the prettier
name, ever. Which make this a bit pointless.

Also, we can always dig out the driver-instance/irq relationship
through other means, so this isn't that useful. So just rip it out to
simplify the set_version/set_busid insanity a bit.

Also delete the temporary busname from drm_pci_set_busid, it's now
unused.

v2: Rebase on top of the new host1x drm_bus for tegra.
Reviewed-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This is only used for drm versions 1.0, and kms drivers have never
been there. So we can appropriately restrict this to legacy and hence
pci devices and inline everything.

v2: Make the dummy function actually return something, caught by Wu
Fengguang's 0-day tester.

v3: Fix spelling in comment (Thierry)
Reviewed-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Now that they're all unused we can get rid of them, including the
dummy version in drm_usb.c.
Reviewed-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Unfortunately this requires a drm-wide change, and I didn't see a sane
way around that. Luckily it's fairly simple, we just need to inline
the respective get_irq implementation from either drm_pci.c or
drm_platform.c.

With that we can now also remove drm_dev_to_irq from drm_irq.c.
Reviewed-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

It's only ever called for legacy drivers, which are all pci.
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

To get rid of the dev->bus->get_irq callback we need to pass in the
desired irq explicitly into drm_irq_install. To avoid having to do the
same for drm_irq_unistall just track it internally. That leaves
drivers with less room to botch things up.

v2: Add the hunk lost in an earlier patch to this one (Thierry).

v3: Fix up the totally fumbled logic in drm_irq_install and use the
local variable consistently. Spotted by both Thierry and Laurent.
Shame on me for failing to properly test the rebase version of this
patch ...

Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Since really that's all it protects - legacy horror stories in
drm_bufs.c. Since I don't want to waste any more time on this I didn't
bother to actually look at what it protects in there, but it's at
least contained now.

v2: Move the spurious hunk to the right patch (Thierry).

Cc: Thierry Reding <thierry.reding@gmail.com>
Reviewed-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

So I just wanted to add a new field to struct drm_device and
accidentally stumbled over something. According to comments
dev->open_count is protected by dev->count_lock, but that's totally
not the case. It's protected by drm_global_mutex.

Unfortunately the vga switcheroo callbacks took this comment at face
value. The problem is that we can't just take the drm_global_mutex
because:
- It would lead to a locking inversion with the driver load/unload
  paths.
- It wouldn't actually protect anything, for that we'd need to wrap
  the entire vga switcheroo code in the drm_global_mutex. And I'm not
  sure whether that would actually solve anything.

What we probably want is a try_to_grab_switcheroo reference kind of
thing which is used in the driver's ->open callback. Then we could
move all that ->can_switch madness into the vga switcheroo core where
it really belongs.

But since that would amount to real work take the easy way out and
just add a comment. It's definitely not going to make anything worse
since doing switcheroo state changes while restarting X just isn't
recommended. Even though the delayed switching code does exactly that.

v2:
- Simplify the ->can_switch implementations more (Thierry)
- Fix comment about the dev->open_count locking (Thierry)

Cc: Thierry Reding <treding@nvidia.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> (v1)
Reviewed-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

If I unplug the eDP monitor, the BIOS of my machine will enable the
VDD bit, then when the driver loads it will think VDD is enabled. It
will detect that the eDP is not enabled and return false from
intel_edp_init_connector. This will trigger a call to
edp_panel_vdd_off_sync(), which trigger a WARN saying that the
refcount of the power domain is less than zero.

The problem happens because the driver gets a refcount whenever it
enables the VDD bit, and puts the refcount whenever it disables the
VDD bit. But on this case, the BIOS enabled VDD, so all we do is to
call put() without calling get() first, so the code added is there to
make sure we always have the get() in case the BIOS enabled the bit.

This regression was introduced in
commit e9cb81a2
Author: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date:   Thu Nov 21 13:47:23 2013 -0200

    drm/i915: get a runtime PM reference when the panel VDD is on

v2: - Rebase

Tested-by: Chris Wilson <chris@chris-wilson.co.uk> (v1)
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Cc: stable@vger.kernel.org (v3.13+)
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>

... our current modeset code isn't good enough yet to handle this. The
scenario is:

1. BIOS sets up a cloned config with lvds+external screen on the same
pipe, e.g. pipe B.

2. We read out that state for pipe B and assign the gmch_pfit state to
it.

3. The initial modeset switches the lvds to pipe A but due to lack of
atomic modeset we don't recompute the config of pipe B.

-> both pipes now claim (in the sw pipe config structure) to use the
gmch_pfit, which just won't work.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=74081Tested-by: Nmax <manikulin@gmail.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>

vgaswitcheroo and the ATPX ACPI methods are required to
power down the dGPU.

bug:
https://bugzilla.kernel.org/show_bug.cgi?id=73901Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Some newer PX laptops have the pci device class
set to DISPLAY_OTHER rather than DISPLAY_VGA.  This
properly detects ATPX on those laptops.

Based on a patch from: Pali Rohár <pali.rohar@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Cc: airlied@gmail.com

Avoids a crash in certain cases when thermal irqs are generated
before the display structures have been initialized.

v2: fix the vblank and vrefresh helpers as well

bug:
https://bugzilla.kernel.org/show_bug.cgi?id=73931Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Need to properly unregister the hwmon device on driver
unload.

v2: minor clean up

bug:
https://bugzilla.kernel.org/show_bug.cgi?id=73931Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org