Right now, drm_sysfs_create() returns the newly allocated "struct class"
to the caller (which is drm_core_init()), which then has to set the
global variable 'drm_class'. During cleanup, though, we call
drm_sysfs_destroy() which implicitly uses the global 'drm_class'. This is
confusing, as ownership of the global 'drm_class' is non-obvious.

This patch changes drm_sysfs_create() to drm_sysfs_init() and makes it
initialize the 'drm_class' object directly, rather than returning it.
This way, both drm_sysfs_init() and drm_sysfs_destroy() work in a similar
fashion and manage the global drm class.
Signed-off-by: NDavid Herrmann <dh.herrmann@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Simplify `foo == NULL || IS_ERR(foo)` via IS_ERR_OR_NULL(). This is
pretty commonly used all over the kernel, especially for debugfs/sysfs
cleanup paths.
Signed-off-by: NDavid Herrmann <dh.herrmann@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

drm_av_sync_delay() doesn't change the passed in mode, so make it const.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

drm_select_eld() doesn't look at the passed in mode, so don't pass it
in.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

valid_inferred_mode() don't change the modes over which it iterates,
so make the iterator const.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

When an i2c WRITE gets an i2c defer or short i2c ack reply, we are
supposed to switch the request from I2C_WRITE to I2C_WRITE_STATUS_UPDATE
when we continue to poll for the completion of the request.

v2: Don't assume DP_AUX_I2C_WRITE is 0 even though it is, to make the
    code more obvious to the casual reader (Jani)
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
[danvet: Resolve conflict due to changed context.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

A address-only I2C_WRITE can't be replied with a short i2c ack, but I
suppose it could be replied with an i2c defer. So the code should be
prepared for an address-only I2C_WRITE_STATUS_UPDATE.

Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: "Terje Bergström" <tbergstrom@nvidia.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

When we get an i2c defer or short ack for i2c-over-aux write we need
to switch to WRITE_STATUS_UPDATE to poll for the completion of the
original request.

Looks like radeon doesn't do anything special with the request type,
so hopefully just treating it the same as a i2c write is enough.

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

When we get an i2c defer or short ack for i2c-over-aux write we need
to switch to WRITE_STATUS_UPDATE to poll for the completion of the
original request.

i915 doesn't try to interpret wht request type apart from separating
reads from writes, and so we should be able to treat this the same as
a normal i2c write.
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

Rename the I2C_STATUS request to I2C_WRITE_STATUS_UPDATE to match the
spec.
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

This function can be used to duplicate an atomic state object. This is
useful for example to implement suspend/resume, where the state before
suspend can be saved and restored upon resume.

v2: move locking to caller, be more explicit about prerequisites
v3: explicitly pass lock acquisition context, improve kerneldoc
Signed-off-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Requested by Laurent.

Note that this uses the new markdown support which will only land in
kernel 4.4 (for the code snippet).

v2: A few spelling fixes I spotted myself.

v3: Big reword for commit_planes() kerneldoc based on a text from
Laurent.

Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Thierry Reding <treding@nvidia.com> (v1 on irc)
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

With drivers supporting runtime pm it's generally not a good idea to
touch the hardware when it's off. Add an option to the commit_planes
helper to support this case.

Note that the helpers already add all planes on a crtc when a modeset
happens, hence plane updates will not be lost if drivers set this to
true.

v2: Check for NULL state->crtc before chasing the pointer. Also check
both old and new crtc if there's a switch. Finally just outright
disallow switching crtcs for a plane if the plane is in active use, on
most hardware that doesn't make sense.

v3: Since commit_planes(active_only = true) is for enabling things
only after all the crtc are on we should only look at the new crtc to
decide whether to call the plane hooks - if the current CRTC isn't on
then skip. If the old crtc (when moving a plane) went down then the
plane should have been disabled as part of the pipe shutdown work
already. For which there's currently no helper really unfortunately.
Also move the check for wether a plane gets a new CRTC assigned while
still in active use out of this patch.

v4: Rebase over exynos changes.

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Thierry Reding <treding@nvidia.com>
Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: NThierry Reding <treding@nvidia.com>
Tested-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

This removes the need to separately track fb changes i915.
That will be done as a separate commit, however.

Changes since v1:
- Add dri-devel to cc.
- Fix a check in intel's prepare and cleanup fb to take rotation
  into account.
Changes since v2:
- Split out i915 changes to a separate commit.

Cc: dri-devel@lists.freedesktop.org
Signed-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: NDaniel Stone <daniels@collabora.com>
[danvet: Squash in msm fixup from Maarten.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This will make sure we get a lockdep spat in all cases
even if the context is a complete garbage pointer.
Signed-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

 drivers/gpu/drm/i2c/ch7006.ko:
-.text                        5913
+.text                        5897
-.rodata                       664
+.rodata                      7256
-.data                        6992
+.data                         416
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Make the mode names passed to drm_mode_create_tv_properties() const.

 drivers/gpu/drm/i2c/ch7006.ko:
-.rodata                       596
+.rodata                       664
-.data                        7064
+.data                        6992

 drivers/gpu/drm/nouveau/nouveau.ko:
-.rodata                      146808
+.rodata                      146904
-.data                        178624
+.data                        178528
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Make generic_edid_names[] const since it's supposed to be immutable.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Allow comma separated filenames in the edid_firmware parameter.

For example:

edid_firmware=eDP-1:edid/1280x480.bin,DP-2:edid/1920x1080.bin

v2: Use strsep() to simplify parsing of comma seperated string. (Matt)
    Move initial bail before strdup. (Matt)
v3: Changed conditionals after while loop to make more readable (Jani)
    Updated kernel-parameters.txt to reflect changes (Jani)
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Reviewed-by: NMatt Roper <matthew.d.roper@intel.com>
Signed-off-by: NBob Paauwe <bob.j.paauwe@intel.com>
[danvet: Flatten else control flow and appease checkpatch.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Very strictly speaking this is possible if you have special hw and
genlocked CRTCs. In general switching a plane between two active CRTC
just won't work so well and is probably not tested at all. Just forbid
it.

I've put this into the core since right now no helper or driver copes
with it, no userspace has code for it and no one asks for it. Yes
there's piles of corner-cases where this would be possible to do this
like:
- switch from inactive crtc to active crtc
- switch from active crtc to inactive crtc
- genlocked display
- invisible plane (to do whatever)
- idle plane hw due to dsi cmd mode/psr
- whatever
but looking at details it's not that easy to implement this correctly.
Hence just put it into the core and add a comment, since the only
userspace we have right now for atomic (weston) doesn't want to use
direct plane switching either.

v2: don't bother with complexity and just outright disallow plane
switching without the intermediate OFF state. Simplifies drivers, we
don't have any hw that could do it anyway and current atomic userspace
(weston) works like this already anyway.

v3: Bikeshed function name (Ville) and add comment (Rob).

v4: Also bikeshed commit message (Rob).

v5: Fix compile warnings reported by 0-day.

Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Daniel Stone <daniels@collabora.com>
Acked-by: NDaniel Stone <daniels@collabora.com>
Reviewed-by: NRob Clark <robdclark@gmail.com>
Acked-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

Faster than recompiling.

Note that restore_fbdev_mode_unlocked is a bit special and the only
one which returns an error code when fbdev isn't there - i915 needs
that one to not fall over with some additional fbcon related restore
code. Everyone else just ignores the return value or only prints a
DRM_DEBUG level message.
Reviewed-by: NArchit Taneja <architt@codeaurora.org>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

These functions are used by drivers to release fbdev emulation
buffers. We need to make them resilient to NULL pointers to
make the fbdev compile/runtime knobs not cause Oopses on module
unload.

Cc: Archit Taneja <architt@codeaurora.org>
Reviewed-by: NRob Clark <robdclark@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Using bool and returning true upon error is very uncommon. Also an int
return value is actually what all the callers which did check it seem
to have expected.

v2: Restore hunk misplaced in a rebase, spotted by Rob.

Cc: Rob Clark <robdclark@gmail.com>
Reviewed-by: NRob Clark <robdclark@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Without the () the markup and more important hyperlinking wont happen.

v2: Also fix nearby type Laurent spotted.

v3: Actually git add. Argh!
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

The adj_start calculation for DRM_MM_CREATE_TOP should happen after
mm->color_adjust. There was an inconsistency between
drm_mm_insert_helper_range
and drm_mm_insert_helper, as the later was already updating after
color_adjust.

Didn't spot it before, as color_adjust is only done in systems without
LLC. But I'm not aware of anybody using this test case yet.
Signed-off-by: NMichel Thierry <michel.thierry@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Legacy s/r hooks are only used for shadow-attaching drivers, warn
when a KMS driver tries to use them.
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

It should be gfx_v8_0_init_compute_vmid since it's
part of the gfx block.
Acked-by: NLeo Liu <leo.liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Need to properly handle the instances for the idle
checks and soft reset.
Acked-by: NLeo Liu <leo.liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

it seems the VCE ring 1 ib test not reliable, remove it for now.
Signed-off-by: NLeo Liu <leo.liu@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Tested-and-Reviewed-by: NSonny Jiang <sonny.jiang@amd.com>

This fixes not printing VM faults.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>

This should never happen so warn when the count does
not equal the expected size.
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Some buffers (UVD/VM page tables) must be placed in VRAM,
but the byte restriction for moving buffers didn't took this
into account.

Port of radeon commit 4b095566.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NJammy Zhou <Jammy.Zhou@amd.com>

The logic was reversed.  This feature is not enabled
at the moment, but fix it now for the future.
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Using the wrong watermwark value for the secondary
watermark.  Copy paste typo.  Noticed by Mykola.
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Using the wrong watermwark value for the secondary
watermark.  Copy paste typo.  Noticed by Mykola.
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Should help avoid fragmentation of vram due to CPU access
requirements.
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

For kernel driver BOs, be explicit about whether we need
vram access up front.  This avoids unecessary migrations and
avoids using visible vram for buffers were it's not needed.

v2: line wrap fixes
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: NJammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>

Customize the insert_nop func for SDMA rings, and use burst NOP for
ring/IB submissions in other places as well
Signed-off-by: NJammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>

The insert_nop function is added to amdgpu_ring_funcs structure as
well as the default implementation
Signed-off-by: NJammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>