Fallout from an intermediate patch revision that I deemed worth saving.
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Currently we do a full re-init of all interrupts after a gpu hang.
Which is pretty bad since we don't restore the interrupts we've
enabled at runtime correctly. Even with that addressed it's rather
horribly race.
But on g4x and later we only reset the gt and not the entire gpu.
Which means we only need to reset the GT interrupt bits. Which has the
nice benefit that vblank waits, pipe CRC interrupts and everything
else display related just keeps on working.
The downside is that gt interrupt handling (i.e. ring->get/put_irq) is
still racy. But as long as the gpu hang reliably wakes all waters and
we have a short time where the refcount drops to 0 we'll recover. So
not that bad really.
v2: Ville noticed that GTIMR and PMIMR don't get cleared, only the
subordinate per-ring registers. So let's rip out all the interrupt dancing.
The FIXME comment is still required though since the ring irq handling
happens at the per-ring interrupt mask registers, too.
Testcase: igt/kms_flip/vblank-vs-hang
Testcase: igt/kms_pipe_crc_basic/hang-*
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ville figured out that it needs a full display reset since apparently
a lot more goes down than just the GT. Until that's address it's
better to just diable gpu reset.
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
So apparently this is tricky.
We need to consider:
- We start out with all the hw enabling bits disabled, both the
individual fifo underrun interrupts and the shared display error
interrupts masked. Otherwise if the bios config is broken we'll blow
up with a NULL deref in our interrupt handler since the crtc
structures aren't set up yet at driver load time.
- On gmch we need to mask fifo underruns on the sw side, so always
need to set that in sanitize_crtc for those platforms.
- On other platforms we try to set the sw tracking so that it reflects
the real state. But since a few platforms have shared bits we must
_not_ disable fifo underrun reporting. Otherwise we'll never enable
the shared error interrupt.
This is the state before out patch, but unfortunately this is not good
enough. But after a suspend resume operation this is broken:
1. We don't enable the hw interrupts since the same code runs on
resume as on driver load.
2. The fifo underrun state adjustments we do in sanitize_crtc doesn't
fire on resume since (except for hilarious firmware) all pipes are off
at that point. But they also don't hurt since the subsequent crtc
enabling due to force_restore will enable fifo underruns.
Which means when we enable fifo underrun reporting we notice that the
per-crtc state is already correct and short-circuit everthing out. And
the interrupt doesn't get enabled.
A similar problem would happen if the bios doesn't light up anything
when the driver loads. Which is exactly what happens when we reload
the driver since our unload functions disables all outputs.
Now we can't just rip out the short-circuit logic and unconditionally
update the fifo underrun reporting interrupt masking: We have some
checks for shared error interrupts to catch issues that happened when
the shared error interrupt was disabled.
The right fix is to push down this logic so that we can always update
the hardware state, but only check for missed fifo underruns on a real
enabled->disabled transition and ignore them when we're already
disabled.
On platforms with shared error interrupt the pipe CRC interrupts are
grouped together with the fifo underrun reporting this fixes pipe CRC
support after suspend and driver reloads.
Testcase: igt/kms_pipe_crc_basic/suspend-*
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
On platforms with shared interrupt enable bits (which are shared even
with the pipe CRC logic) there's some tricky corner cases. Add
information to make debugging those easier.
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
All drm_fb_helper_restore_fbdev_mode() call sites, save one, do the same
locking. Simplify this into drm_fb_helper_restore_fbdev_mode_unlocked().
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
For atomic, it will be quite necessary to not need to care so much
about locking order. And 'state' gives us a convenient place to stash a
ww_ctx for any sort of update that needs to grab multiple crtc locks.
Because we will want to eventually make locking even more fine grained
(giving locks to planes, connectors, etc), split out drm_modeset_lock
and drm_modeset_acquire_ctx to track acquired locks.
Atomic will use this to keep track of which locks have been acquired
in a transaction.
v1: original
v2: remove a few things not needed until atomic, for now
v3: update for v3 of connection_mutex patch..
v4: squash in docbook
v5: doc tweaks/fixes
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This should avoid races between connector probing and HPD
irqs in the future, currently mode_config.mutex blocks this
possibility.
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Just flushing out my pile of random drm patches for the merge window,
nothing big. And it all hung around in drm-intel trees for a while (only
just rebased now).
* tag 'topic/core-stuff-2014-06-02' of git://anongit.freedesktop.org/drm-intel:
imx-drm: imx-tve: remove unused variable
drm: Missed clflushopt in drm_clflush_virt_range
drm/plane: Fix a couple of checkpatch warnings
drm/plane: Fix sparse warnings
drm/exynos: Fix double locks at PM resume
drm/ast: Fix double lock at PM resume
drm/dp-helper: Deprecate old i2c-over-dp_aux heleprs
Pretty small pull this time around for msm. Adds some useful debugfs
I'd been carrying around on a branch for a while, plus few fixes. And
Kconfig update for the great ARCH_MSM -> ARCH_QCOM split.
* 'msm-next' of git://people.freedesktop.org/~robclark/linux:
drm/msm: use correct gfp flag for vram allocation
drm/msm/mdp5: fix error return value
drm/msm: remove redundant private plane cleanup
drm/msm: add perf logging debugfs
drm/msm: add rd logging debugfs
drm/msm: update for ARCH_MSM -> ARCH_QCOM
drm/msm/hdmi: use gpio and HPD polling
drm/msm/mdp5: fix crash in error/unload paths
The drm core shouldn't depend upon any helpers, and we make sure this
doesn't accidentally happen by moving them into the helper-only
drm_kms_helper.ko module.
v2: Don't break the build for vmwgfx, spotted by Matt.
v3: Unbreak the depency loop around CONFIG_FB (not actually a loop
since it involves select). Reported by Chris.
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Thomas Hellstrom <thellstrom@vmware.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
After the split-out of crtc locks from the big mode_config.mutex
there's still two major areas it protects:
- Various connector probe states, like connector->status, EDID
properties, probed mode lists and similar information.
- The links from connector->encoder and encoder->crtc and other
modeset-relevant connector state (e.g. properties which control the
panel fitter).
The later is used by modeset operations. But they don't really care
about the former since it's allowed to e.g. enable a disconnected VGA
output or with a mode not in the probed list.
Thus far this hasn't been a problem, but for the atomic modeset
conversion Rob Clark needs to convert all modeset relevant locks into
w/w locks. This is required because the order of acquisition is
determined by how userspace supplies the atomic modeset data. This has
run into troubles in the detect path since the i915 load detect code
needs _both_ protections offered by the mode_config.mutex: It updates
probe state and it needs to change the modeset configuration to enable
the temporary load detect pipe.
The big deal here is that for the probe/detect users of this lock a
plain mutex fits best, but for atomic modesets we really want a w/w
mutex. To fix this lets split out a new connection_mutex lock for the
modeset relevant parts.
For simplicity I've decided to only add one additional lock for all
connector/encoder links and modeset configuration states. We have
piles of different modeset objects in addition to those (like bridges
or panels), so adding per-object locks would be much more effort.
Also, we're guaranteed (at least for now) to do a full modeset if we
need to acquire this lock. Which means that fine-grained locking is
fairly irrelevant compared to the amount of time the full modeset will
take.
I've done a full audit, and there's just a few things that justify
special focus:
- Locking in drm_sysfs.c is almost completely absent. We should
sprinkle mode_config.connection_mutex over this file a bit, but
since it already lacks mode_config.mutex this patch wont make the
situation any worse. This is material for a follow-up patch.
- omap has a omap_framebuffer_flush function which walks the
connector->encoder->crtc links and is called from many contexts.
Some look like they don't acquire mode_config.mutex, so this is
already racy. Again fixing this is material for a separate patch.
- The radeon hot_plug function to retrain DP links looks at
connector->dpms. Currently this happens without any locking, so is
already racy. I think radeon_hotplug_work_func should gain
mutex_lock/unlock calls for the mode_config.connection_mutex.
- Same applies to i915's intel_dp_hot_plug. But again, this is already
racy.
- i915 load_detect code needs to acquire this lock. Which means the
w/w dance due to Rob's work will be nicely contained to _just_ this
function.
I've added fixme comments everywhere where it looks suspicious but in
the sysfs code. After a quick irc discussion with Dave Airlie it
sounds like the lack of locking in there is due to sysfs cleanup fun
at module unload.
v1: original (only compile tested)
v2: missing mutex_init(), etc (from Rob Clark)
v3: i915 needs more care in the conversion:
- Protect the edp pp logic with the connection_mutex.
- Use connection_mutex in the backlight code due to
get_pipe_from_connector.
- Use drm_modeset_lock_all in suspend/resume paths.
- Update lock checks in the overlay code.
Cc: Alex Deucher <alexdeucher@gmail.com>
Cc: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Rob Clark <robdclark@gmail.com>
I find myself making this change locally whenever debugging FB reference
counting. Which seems a bit silly.
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
An object property is an id (idr) for a drm mode object. This
will allow a property to be used set/get a framebuffer, CRTC, etc.
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
If we continue to use bitmask for type, we will quickly run out of room
to add new types. Split this up so existing part of bitmask range
continues to function as before, but reserve a chunk of the remaining
space for an integer type-id. Wrap this all up in some type-check
helpers to keep the backwards-compat uglyness contained.
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
No longer used or needed as the structs have a name field.
Acked-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
spice-server and downstream code expect that the primary surface
will always have surface_id = 0, while in reality, once allocated, the
surface_id in qxl.ko is NEVER 0. In a dual head environment, all
monitors render portions of the primary surface.
However, when the monitor config events are generated and sent,
the primary surface is only mapped to the correct identifier
(i.e. 0) for the primary head (where crtc index is 0).
The fix is to look at the "primary" flag in the bo and always
use id 0, irrespective of which head is being configured.
[airlied: qxl hw really needs to be fixed to scanout surfaces]
Signed-off-by: Dave Airlie <airlied@redhat.com>
Instead of trying to flip inside the vblank period when
the buffer is idle, offload blocking for idle to a kernel
thread and program the flip directly into the hardware.
v2: add error handling, fix EBUSY handling
v3: add proper exclusive_lock handling
v4: update crtc->primary->fb when the flip actually happens
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Highlights:
- GPUVM opimtizations
- HDMI audio cleanups
- Deep color HDMI support
- more bug fixes, cleanups
* 'drm-next-3.16' of git://people.freedesktop.org/~agd5f/linux: (29 commits)
drm/edid: Add quirk for Sony PVM-2541A to get 12 bpc hdmi deep color.
drm/edid: Parse and handle HDMI deep color modes.
drm/radeon: Limit hdmi deep color bit depth to 12 bpc.
drm/radeon: Setup HDMI_CONTROL for hdmi deep color gcp's (v2)
drm/radeon: fix pll setup for hdmi deep color (v7)
drm/radeon: use hw cts/n values for deep color
drm/radeon: only apply hdmi bpc pll flags when encoder mode is hdmi
drm/radeon/atom: fix dithering on certain panels
drm/radeon: optimize CIK VM handling v2
drm/radeon: optimize SI VM handling
drm/radeon: add define for flags used in R600+ GTT
drm/radeon: rework page flip handling v3
drm/radeon: separate vblank and pflip crtc handling
drm/radeon: split page flip and pending callback
drm/radeon: remove drm_vblank_get|put from pflip handling
drm/radeon: remove (pre|post)_page_flip callbacks
drm/radeon/dp: fix lane/clock setup for dp 1.2 capable devices
drm/radeon: fix typo in radeon_connector_is_dp12_capable()
radeon: Remove useless quirk for zx1/FireGL X1 combo introduced with fdo #7770
vgaswitcheroo: switch the mux to the igp on power down when runpm is enabled
...
The Sony PVM-2541A OLED high precision color display supports
both 10 bpc and 12 bpc hdmi deep color input, but its edid
does not signal any deep color support.
Add a quirk to force it being treated as a 12 bpc panel.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Check the HDMI cea block for deep color mode bits. If available,
assign the highest supported bpc for a hdmi display, corresponding
to the given deep color modes.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
DCE-4/5/6 can't support more than 12 bpc deep color over hdmi,
so clamp to 12 bpc when a hdmi deep color capable display is
connected. This even makes sense on DCE-8+, which could do up
to 16 bpc, as driving with more than 12 bpc would only waste
video bandwidth as long as we don't support framebuffers with
more than 12 bpc depth.
On pre-DCE4 we clamp hdmi bit depth to 8 bpc, as those asics
don't support hdmi deep color.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Program HDMI_CONTROL to send general control packets
for hdmi deep color mode signalling at every video
frame if bpc > 8.
This is only supported on evergreen / DCE-4 and later.
v2: rebase
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Need to adjust the pll up for deep color modes.
Additionally, the atom bpc defines were wrong in certain
cases.
v2: set the adjusted clock to the pll clock for hdmi deep
color. This fixes display and audio issues with deep color
as reported by Andy Furniss <adf.lists@gmail.com>
v3: set crtc_clock as well
v4: setcrtcinfo on the adjusted mode
v5: just use the adjusted clock for setting the pll
v6: only use the adjusted clock for hdmi
v7: only DCE5 and DCE6 and bpc > 8
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
I'm not really sure how these should be calculated
for deep color. The hw generated values seem to work.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fill VM page tables from the GART page table if applicable.
v2: fix copy&paste error
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fill VM page tables from the GART page table if applicable.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Instead of trying to flip inside the vblank period when
the buffer is idle, offload blocking for idle to a kernel
thread and program the flip directly into the hardware.
v2: add error handling, fix EBUSY handling
v3: add proper exclusive_lock handling
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We activate the VBLANK irq manually anyway, so this is unnecessary.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
They are doing the same on all generations anyway.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We were checking the ext clock rather than the display clock.
Noticed by ArtForz on IRC.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Removes useless quirk a7f465f73363fce409870f62173d518b1bc02ae6 introduced with
fdo #7770 as a failed attempt to minimize stability issues with hp zx1 chipset/
ATI FireGL X1 graphics adapter configuration
(see http://marc.info/?l=linux-ia64&m=140077543819871&w=2 for details/reason)
Signed-off-by: Émeric MASCHINO <emeric.maschino@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Daniel Vetter
Rob Clark
Dave Airlie
Jani Nikula
David Mansfield
Christian König
Mario Kleiner
Alex Deucher
Émeric MASCHINO