Dave Airlie [Mon, 22 Apr 2013 08:48:45 +0000 (18:48 +1000)]
Merge branch 'drm-intel-fixes' of git://people.freedesktop.org/~danvet/drm-intel into drm-next
Daniel writes:
As promised a stash of (mostly) fixes. Two pieces of non-fixes included:
- A notch more gtt refactoring from Ben, beating to death with igt in our
nightly testing.
- Support for display display-less server chips (again from Ben). New hw
support which is only likely to break itself ;-)
Otherwise just tons of fixes:
- hpd irq storm mitigation from Egbert Eich. Your -next tree already has
the infrastructure, this here just supplies the logic.
- sdvo hw state check fix from Egbert Eich
- fb cb tune settings for the pch pll clocks on cpt/ppt
- "Bring a bigger gun" coherence workaround for multi-threade, mulit-core
& thrashing tiled gtt cpu access from Chris.
- Update haswell mPHY code.
- l3$ caching for context objects on ivb/hsw (Chris).
- dp aux refclock fix for haswell (Jani)
- moar overclocking fixes for snb/ivb (Ben)
- ecobits ppgtt pte caching control fixes from Ville
- fence stride check fixes and limit improvements (Ville)
- fix up crtc force restoring, potentially resulting in tons of hw state
check WARNs
- OOPS fix for NULL derefencing of fb pointers when force-restoring a crtc
when other crtcs are disabled and the force-restored crtc is _not_ the
first one.
- Fix pfit disabling on gen2/3.
- Haswell ring freq scaling fixes (Chris).
- backlight init/teardown fix (failed eDP init killed the lvds backlight)
from Jani
- cpt/ppt fdi polarity fixes from Paulo (should help a lot of the FDI link
train failures).
- And a bunch of smaller things all over.
* 'drm-intel-fixes' of git://people.freedesktop.org/~danvet/drm-intel: (56 commits)
drm/i915: fix bpc vs. bpp confusion in intel_crtc_compute_config
drm/i915: move cpu_transcoder to the pipe configuration
drm/i915: preserve the PBC bits of TRANS_CHICKEN2
drm/i915: set CPT FDI RX polarity bits based on VBT
drm/i915: Add Reenable Timer to turn Hotplug Detection back on (v4)
drm/i915: Disable HPD interrupt on pin when irq storm is detected (v3)
drm/i915: Mask out the HPD irq bits before setting them individually.
drm/i915: (re)init HPD interrupt storm statistics
drm/i915: Add HPD IRQ storm detection (v5)
drm/i915: WARN when LPT-LP is not paired with ULT CPU
drm/i915: don't intel_crt_init on any ULT machines
drm/i915: remove comment about IVB link training from intel_pm.c
drm/i915: VLV doesn't have LLC
drm/i915: Scale ring, rather than ia, frequency on Haswell
drm/i915: shorten debugfs output simple attributes
drm/i915: Fixup pfit disabling for gen2/3
drm/i915: Fixup Oops in the pipe config computation
drm/i915: ensure single initialization and cleanup of backlight device
drm/i915: don't touch the PF regs if the power well is down
drm/i915: add intel_using_power_well
...
Daniel Vetter [Wed, 17 Apr 2013 18:15:07 +0000 (20:15 +0200)]
drm/i915: move cpu_transcoder to the pipe configuration
For a bunch of reason we need to more accurately track this:
- hw pipe state readout for Haswell needs the cpu transcoder.
- We need to know the right cpu transcoder in a bunch of places in
->disable and other modeset callbacks.
In the future we need to add hw state readout&check support, too. But
to avoid ugly merge conflicts do the rote sed job now without any
functional changes.
v2: Preserve the cpu_transcoder value when overwriting crtc->config.
Reported by Paulo.
Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> (v1)
[danvet: Removed rough whitespace that Chris spotted.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo Zanoni [Mon, 8 Apr 2013 18:48:08 +0000 (15:48 -0300)]
drm/i915: preserve the PBC bits of TRANS_CHICKEN2
Bits 30 and 24:0 are PBC, so don't zero them. Some of the other bits
are being zeroed, but I couldn't find a reason for this, so leave them
as they are for now to avoid regressions.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com>
[danvet: Delete the redudant #define that Imre spotted in his review.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drm/i915: Add Reenable Timer to turn Hotplug Detection back on (v4)
We disable hoptplug detection when we encounter a hotplug event
storm. Still hotplug detection is required on some outputs (like
Display Port). The interrupt storm may be only temporary (on certain
Dell Laptops for instance it happens at certain charging states of
the system). Thus we enable it after a certain grace period (2 minutes).
Should the interrupt storm persist it will be detected immediately
and it will be disabled again.
v2: Reordered drm_i915_private: moved hotplug_reenable_timer to hpd state tracker.
v3: Clarified loop start value,
Removed superfluous test for Ivybridge and Haswell,
Restructured loop to avoid deep nesting (all suggested by Ville Syrjälä)
v4: Fixed two bugs pointed out by Jani Nikula.
Signed-off-by: Egbert Eich <eich@suse.de> Reviewed-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drm/i915: Disable HPD interrupt on pin when irq storm is detected (v3)
This patch disables hotplug interrupts if an 'interrupt storm'
has been detected.
Noise on the interrupt line renders the hotplug interrupt useless:
each hotplug event causes the devices to be rescanned which will
will only increase the system load.
Thus disable the hotplug interrupts and fall back to periodic
device polling.
v2: Fixed cleanup typo.
v3: Fixed format issues, clarified a variable name,
changed pr_warn() to DRM_INFO() as suggested by
Jani Nikula <jani.nikula@linux.intel.com>.
Signed-off-by: Egbert Eich <eich@suse.de> Reviewed-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
When an encoder is shared on several connectors there is only
one hotplug line, thus this line needs to be shared among these
connectors.
If HPD detect only works reliably on a subset of those connectors,
we want to poll the others. Thus we need to make sure that storm
detection doesn't mess up the settings for those connectors.
Therefore we store the settings in the intel_connector struct and
restore them from there.
If nothing is set but the encoder has a hpd_pin set we assume this
connector is hotplug capable.
On init/reset we make sure the polled state of the connectors
is (re)set to the default value, the HPD interrupts are marked
enabled.
Signed-off-by: Egbert Eich <eich@suse.de> Reviewed-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Add a hotplug IRQ storm detection (triggered when a hotplug interrupt
fires more than 5 times / sec).
Rationale:
Despite of the many attempts to fix the problem with noisy hotplug
interrupt lines we are still seeing systems which have issues:
Once cause of noise seems to be bad routing of the hotplug line
on the board: cross talk from other signals seems to cause erronous
hotplug interrupts. This has been documented as an erratum for the
the i945GM chipset and thus hotplug support was disabled for this
chipset model but others seem to have this problem, too.
We have seen this issue on a G35 motherboard for example:
Even different motherboards of the same model seem to behave
differently: while some only see only around 10-100 interrupts/s
others seem to see 5k or more.
We've also observed a dependency on the selected video mode.
Also on certain laptops interrupt noise seems to occur duing
battery charging when the battery is at a certain charge levels.
Thus we add a simple algorithm here that detects an 'interrupt storm'
condition.
v2: Fixed comment.
v3: Reordered drm_i915_private: moved hpd state tracking to hotplug work stuff.
v4: Followed by Jesse Barnes to use a time_..() macro.
v5: Fixed coding style as suggested by Jani Nikula.
Signed-off-by: Egbert Eich <eich@suse.de> Acked-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo Zanoni [Fri, 12 Apr 2013 21:16:53 +0000 (18:16 -0300)]
drm/i915: don't intel_crt_init on any ULT machines
We may have DDI_BUF_CTL(PORT_A) configured with 2 lanes and still not
have CRT, so just check for !IS_ULT. This problem happened on a real
machine and resulted in a very ugly dmesg.
Cc: stable@vger.kernel.org Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
I don't have a VLV to test it with, Jesse, Ken, can one of you test?
Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Cc: Kenneth Graunke <kenneth@whitecape.org> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Chris Wilson [Fri, 12 Apr 2013 18:10:13 +0000 (19:10 +0100)]
drm/i915: Scale ring, rather than ia, frequency on Haswell
Haswell introduces a separate frequency domain for the ring (uncore). So
where we used to increase the CPU (IA) clock with GPU busyness, we now
need to scale the ring frequency directly instead. As the ring limits
our memory bandwidth, it is vital for performance that when the GPU is
busy, we increase the frequency of the ring to increase the available
memory bandwidth.
v2: Fix the algorithm to actually use the scaled gpu frequency for the ring.
v3: s/max_ring_freq/min_ring_freq/ as that is what it is
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[danvet: Add space checkpatch complained about.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Fix this up by properly computing the pipe the pfit is on. Also
extract the logic into its own function, add a debug assert to check
that the pipe is off (mostly just documentation) and add some debug
output.
If pipe A was disabled after pipe B was set up, the panel fitter will
be disabled. Now most userspace doesn't do modesets in this order,
which is why I couldn't ever reproduce this and why it took me so long
to figure out.
We really need hw state readout and check support for the pannel
fitter ...
Reported-by: Hans de Bruin <jmdebruin@xmsnet.nl> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: Hans de Bruin <jmdebruin@xmsnet.nl>
References: http://permalink.gmane.org/gmane.comp.freedesktop.xorg.drivers.intel/19049 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Fri, 12 Apr 2013 16:48:43 +0000 (18:48 +0200)]
drm/i915: Fixup Oops in the pipe config computation
Yet again our current confusion between doing the modeset globally,
but only having the new parameters for one crtc at a time.
So that intel_set_mode essentially already does a global modeset:
intel_modeset_affected_pipes compares the current state with where we
want to go to (which is carefully set up by intel_crtc_set_config) and
then goes through the modeset sequence for any crtc which needs
updating.
Now the issue is that the actual interface with the remaining code
still only works on one crtc, and so we only pass in one fb and one
mode. In intel_set_mode we also only compute one intel_crtc_config
(which should be the one for the crtc we're doing a modeset on).
The reason for that mismatch is twofold:
- We want to eventually do all modeset as global state changes, so
it's just infrastructure prep.
- But even the old semantics can change more than one crtc when you
e.g. move a connector from crtc A to crtc B, then both crtc A and B
need to be updated. Usually that means one pipe is disabled and the
other enabled. This is also the reason why the hack doesn't touch the
disable_pipes mask.
Now hilarity ensued in our kms config restore paths when we actually
try to do a modeset on all crtcs: If the first crtc should be off and
the second should be on, then the call on the first crtc will notice
that the 2nd one should be switched on and so tries to compute the
pipe_config. But due to a lack of passed-in fb (crtc 1 should be off
after all) it only results in tears.
This case is ridiculously easy to hit on gen2/3 where the lvds output
is restricted to pipe B. Note that before the pipe_config bpp rework
gen2/3 didn't care really about the fb->depth, so this is a regression
brought to light with
drm/i915: precompute pipe bpp before touching the hw
But apparently Ajax also managed to blow up pch platforms, probably
with some randomized configs, and pch platforms trip up over the lack
of an fb even in the old code. So this actually goes back to the first
introduction of the new modeset restore code in
Fix this mess by now by justing shunting all the cool new global
modeset logic in intel_modeset_affected_pipes.
v2: Improve commit message and clean up all the comments in
intel_modeset_affected_pipes - since the introduction of the modeset
restore code they've been a bit outdated.
Bugzill: https://bugzilla.redhat.com/show_bug.cgi?id=917725 Cc: stable@vger.kernel.org
References: http://www.mail-archive.com/stable@vger.kernel.org/msg38084.html Tested-by: Richard Cochran <richardcochran@gmail.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Jani Nikula [Fri, 12 Apr 2013 12:18:38 +0000 (15:18 +0300)]
drm/i915: ensure single initialization and cleanup of backlight device
Backlight cleanup in the eDP connector destroy callback caused the
backlight device to be removed on some systems that first initialized LVDS
and then attempted to initialize eDP. Prevent multiple backlight
initializations, and ensure backlight cleanup is only done once by moving
it to modeset cleanup.
A small wrinkle is the introduced asymmetry in backlight
setup/cleanup. This could be solved by adding refcounting, but it seems
overkill considering that there should only ever be one backlight device.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=55701 Signed-off-by: Jani Nikula <jani.nikula@intel.com> Tested-by: Peter Verthez <peter.verthez@skynet.be> Cc: stable@vger.kernel.org Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Thu, 11 Apr 2013 18:22:50 +0000 (20:22 +0200)]
drm/i915: don't check inconsistent modeset state when force-restoring
It will be only consistent once we've restored all the crtcs. Since a
bunch of other callers also want to just restore a single crtc, add a
boolean to disable checking only where it doesn't make sense.
Note that intel_modeset_setup_hw_state already has a call to
intel_modeset_check_state at the end, so we don't reduce the amount of
checking.
v2: Try harder not to create a big patch (Chris).
v3: Even smaller (still Chris). Also fix a trailing space.
References: https://lkml.org/lkml/2013/3/16/60 Cc: Tomas Melin <tomas.melin@iki.fi> Cc: Richard Cochran <richardcochran@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: stable@vger.kernel.org Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Tested-by: Tomas Melin <tomas.melin@iki.fi> Tested-by: Richard Cochran <richardcochran@gmail.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ville Syrjälä [Tue, 9 Apr 2013 10:02:47 +0000 (13:02 +0300)]
drm/i915: IVB/HSW have 32 fence register
Increase the number of fence registers to 32 on IVB/HSW. VLV however
only has 16 fence registers according to the docs.
Increasing the number of fences was attempted before [1], but there was
some uncertainty about the maximum CPU fence number for FBC. Since then
BSpec has been updated to state that there are in fact 32 fence registers,
and the CPU fence number field in the SNB_DPFC_CTL_SA register is 5 bits,
and the CPU fence number field in the ILK_DPFC_CONTROL register must be
zero. So now it all makes sense.
changed the sysfs read semantics for 'gt_max_freq_mhz'. By
always returning overclock max instead of stored value.
Fix this by returning the stored value. Separate sysfs entry
should be considered for overclocking max freq.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=63415 Cc: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ville Syrjälä [Thu, 4 Apr 2013 12:13:42 +0000 (15:13 +0300)]
drm/i915: Configure GAM_ECOCHK appropriatly for Gen7
IVB and HSW use different encodings for the PPGTT cacheability bits in
the GAM_ECOCHK register.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ville Syrjälä [Thu, 4 Apr 2013 12:13:41 +0000 (15:13 +0300)]
drm/i915: Set GAC_ECO_BITS register on Gen7+
According to BSpec GAC_ECO_BITS register exists on Gen7 platforms as
well. Configure it accordingly.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ville Syrjälä [Thu, 4 Apr 2013 12:13:40 +0000 (15:13 +0300)]
drm/i915: Add ECOBITS_SNB_BIT
GAC_ECO_BITS has a bit similar to GAM_ECOCHK's ECOCHK_SNB_BIT. Add
the define, and enable it on SNB.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Fri, 5 Apr 2013 21:29:23 +0000 (14:29 -0700)]
drm/i915: Don't default to overclock max
Requested-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Fri, 5 Apr 2013 21:29:22 +0000 (14:29 -0700)]
drm/i915: Better overclock support
Most importantly this will allow users to set overclock frequencies in
sysfs. Previously the max was limited by the RP0 max as opposed to the
overclock max. This is useful if one wants to either limit the max
overclock frequency, or set the minimum frequency to be in the overclock
range. It also fixes an issue where if one sets the max frequency to be
below the overclock max, they wouldn't be able to set back the proper
overclock max.
In addition I've added a couple of other bits:
Show the overclock freq. as max in sysfs
Print the overclock max in debugfs.
Print a warning if the user sets the min frequency to be in the
overclock range.
In this patch I've decided to store the hw_max when we read it from the
pcode at init. The reason I do this is the pcode reads can fail, and are
slow.
v2: Report when user requested overclocked max (Daniel)
Remove when user sets min to overclock range (Daniel)
Reported-by: freezer from #intel-gfx on irc Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
[danvet: Fixup the s/100MHz/50MHz/ confusion in an unrelated comment
that Mika spotted.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Jani Nikula [Tue, 9 Apr 2013 05:11:00 +0000 (08:11 +0300)]
drm/i915: use lower aux clock divider on non-ULT HSW
Workaround to avoid intermittent aux channel failures, per spec change.
v2: Don't mess with cpu dp aux divider (Paulo Zanoni)
Signed-off-by: Jani Nikula <jani.nikula@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
[danvet: Kill spurious tab spotted by Paulo.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 9 Apr 2013 01:43:56 +0000 (18:43 -0700)]
drm/i915: Allow PPGTT enable to fail
I'm really not happy that we have to support this, but this will be the
simplest way to handle cases where PPGTT init can fail, which I promise
will be coming in the future.
v2: Resolve conflicts due to patch series reordering.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v1) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 9 Apr 2013 01:43:55 +0000 (18:43 -0700)]
drm/i915: NULL aliasing_ppgtt on cleanup
This will allow us to carry on if we've cleaned up the PPGTT. The usage
for this is coming up - it simplifies handling a failed PPGTT init.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: Spill the secrets about failing ppgtt init.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 9 Apr 2013 01:43:51 +0000 (18:43 -0700)]
drm/i915/ppgtt: Set scratch page "globally"
The PPGTT scratch page is used for all gens, and doing it in the global
part of our PPGTT setup makes the code a bit nicer.
This was in a patch submitted earlier as part of the PPGTT cleanups.
Grumpy maintainer must have missed it, and I didn't yell when
appropriate. Apologies for everyone :-)
v2: Update commit message
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 9 Apr 2013 01:43:49 +0000 (18:43 -0700)]
drm/i915: Map registers before GTT init
This will allow us to read/write registers in GTT init.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: Fix up error handling. We really should look into devres for
this stuff ...] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 9 Apr 2013 01:43:48 +0000 (18:43 -0700)]
drm/i915: Call out GEN6 PTE specificity
We can assume that the PTE layout, and size changes for future
generations. To avoid confusion with the existing GEN6 PTE typedef, give
it a GEN6_ prefix.
v2: Fixup checkpatch warning and bikeshed commit message slightly.
v3: Rebase on top of Imre's for_each_sg_pages rework.
v4: Fixup conflicts in patch series reordering.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v1) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Ben Widawsky [Tue, 9 Apr 2013 01:43:47 +0000 (18:43 -0700)]
drm/i915: generalize pte vs. register BAR allocation
All gen6+ parts so far have 1 BAR which holds both the register space
and the GTT PTEs. Up until now, that was a 4MB BAR with half allocated
to each.
I have a strong hunch (wink, nod, wink) that future gens will also keep
a similar 50-50 split though the sizes may change. To help this along
change the code to obey the rule of half the total size instead of a
hard-coded 2MB.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Chris Wilson [Mon, 8 Apr 2013 13:28:40 +0000 (14:28 +0100)]
drm/i915: Use MLC (l3$) for context objects
Enabling context support increases SwapBuffers latency by about 20%
(measured on an i7-3720qm). We can offset that loss slightly by enabling
faster caching for the contexts. As they are not backed by any
particular cache (such as the sampler or render caches) our only option
is to select the generic mid-level cache. This reduces the latency of
the swap by about 5%.
Oddly this effect can be observed running smokin-guns on IVB at
1280x1024:
Using BLT copies for swaps: 151.67 fps
Using Render copies for swaps (unpatched): 141.70 fps
With contexts disabled: 150.23 fps
With contexts in L3$: 150.77 fps
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Kenneth Graunke <kenneth@whitecape.org> Reviewed-by: Kenneth Graunke <kenneth@whitecape.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Chris Wilson [Thu, 4 Apr 2013 20:31:03 +0000 (21:31 +0100)]
drm/i915: Workaround incoherence between fences and LLC across multiple CPUs
In order to fully serialize access to the fenced region and the update
to the fence register we need to take extreme measures on SNB+, and
manually flush writes to memory prior to writing the fence register in
conjunction with the memory barriers placed around the register write.
Fixes i-g-t/gem_fence_thrash
v2: Bring a bigger gun
v3: Switch the bigger gun for heavier bullets (Arjan van de Ven)
v4: Remove changes for working generations.
v5: Reduce to a per-cpu wbinvd() call prior to updating the fences.
v6: Rewrite comments to ellide forgotten history.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62191 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Tested-by: Jon Bloomfield <jon.bloomfield@intel.com> (v2) Cc: stable@vger.kernel.org Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
v2: Fix the spelling in the added comment (Chris).
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=63246 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Thu, 4 Apr 2013 20:20:34 +0000 (22:20 +0200)]
drm/i915: set CB tuning also for the reduce clock
Since the ratio is different, we also need to pass in the parameters
for the reduced clock. Might or might not reduce flicker for the
auto-downclocking on lvds/eDP.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Daniel Vetter [Thu, 4 Apr 2013 20:20:33 +0000 (22:20 +0200)]
drm/i915: fix FP CB tuning limits for lvds
Only on IBX should we set the limiting factor to 25 unconditionally
for dual-channel mode, on CPT/PPT 25 only applies when the lvds
refclock is 100MHz.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drm/i915: Fix SDVO connector and encoder get_hw_state functions
The connector associated with the encoder is considered active when the
output associtated with this connector is active on the encoder. The
encoder itself is considered active when either there is an active
output on it or the respective SDVO channel is active.
Having active outputs when the SDVO channel is inactive seems to be
inconsistent: such states can be found when intel_modeset_setup_hw_state()
collects the hardware state set by the BIOS.
This inconsistency will be fixed in intel_sanitize_crtc()
(when intel_crtc_update_dpms() is called), this however only happens
when the encoder is associated with a crtc.
Dave Airlie [Wed, 17 Apr 2013 05:18:32 +0000 (15:18 +1000)]
Merge branch 'gma500-next' of git://github.com/patjak/drm-gma500 into drm-next
Patrik writes:
I haven't had much review or testing on other platforms than Poulsbo but
at least the following Cedarview bug has been squashed and no
regressions reported: https://bugs.freedesktop.org/show_bug.cgi?id=58527
* 'gma500-next' of git://github.com/patjak/drm-gma500:
drm/gma500: Add debugging info to psb_gtt_restore()
drm/gma500: Check connector status before restoring sdvo
gma500:fix build failure for 3.9-rc5
drm/gma500: Fix hibernation problems on sdvo encoders
drm/gma500: Add hooks for hibernation
drm/gma500: Activate the gtt rebuild on suspend/resume
drm/gma500: Add support for rebuilding the gtt
drm/gma500: Change fb name so pm-utils doesn't apply quirks
gma500: Make VGA and HDMI connector hotpluggable
drm/gma500: Clean up various defines
drm/gma500: Remove unnecessary function exposure
drm/gma500: Type clock limits directly into array and remove defines
drm/gma500: Calculate clock in one function instead of three identical
drm/gma500: Remove unused i8xx clock limits
gma500: medfield: Fix possible NULL pointer dereference
drivers: gpu: drm: gma500: Replaced calls kzalloc & memcpy with kmemdup
gma500: remove unused drm_psb_no_fb
Dave Airlie [Tue, 16 Apr 2013 22:48:23 +0000 (08:48 +1000)]
Merge branch 'drm-next-3.10' of git://people.freedesktop.org/~agd5f/linux into drm-next
Alex writes:
This is the initial 3.10 pull request for radeon. The big changes here
are UVD support and proper tiling support for SI. The rest is
bug fixes. I hope to have another pull request later in the week with
some new things we've been working on internally.
* 'drm-next-3.10' of git://people.freedesktop.org/~agd5f/linux: (28 commits)
drm/radeon: Always flush the VM
drm/radeon: re-enable PTE/PDE packet for set_page on cayman/TN
drm/radeon: cleanup properly if mmio mapping fails
drm/radeon/evergreen+: don't enable HPD interrupts on eDP/LVDS
drm/radeon: add si tile mode array query v3
drm/radeon: add ring working query
drm/radeon: handle broken disabled rb mask gracefully
drm/radeon: add pcie set/get lanes callbacks for newer asics
drm/radeon: update r600 set/get pcie lane config
drm/radeon/kms: replace *REG32_PCIE_P with *REG32_PCIE_PORT
drm/radeon: remove unused blit remnants from si.c
drm/radeon: add UVD tiling addr config v2
drm/radeon: init UVD clocks to sane defaults
drm/radeon: add set_uvd_clocks callback for r7xx v3
drm/radeon: add set_uvd_clocks callback for SI
drm/radeon: add set_uvd_clocks callback for evergreen
drm/radeon: add set_uvd_clocks callback for ON/LN/TN (v4)
drm/radeon: add radeon_atom_get_clock_dividers helper
drm/radeon: add pm callback for setting uvd clocks
drm/radeon: UVD bringup v8
...
This is slightly cleaned up version of Jerome's patch.
There seems to be an issue tracking the last flush of
the VM which results in hangs in certain cases when
VM is used. For now just flush the VM for every IB.
Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org
Alex Deucher [Tue, 16 Apr 2013 14:42:15 +0000 (10:42 -0400)]
drm/radeon: re-enable PTE/PDE packet for set_page on cayman/TN
PTE/PDE doesn't support a single update (count = 1). We had
previously disabled it since it we were hitting that case which
let to hangs. The PTE/PDE packet is much more efficient for VM
updates where it can be used.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Tue, 16 Apr 2013 03:24:25 +0000 (13:24 +1000)]
drm/qxl: make lots of things static.
/usr/lib/gcc/x86_64-linux-gnu/4.7/include/stddef.h:414:9: sparse: preprocessor token offsetof redefined
include/linux/stddef.h:17:9: this was the original definition
>> drivers/gpu/drm/qxl/qxl_drv.c:49:5: sparse: symbol 'qxl_modeset' was not declared. Should it be static?
Reported-by: kbuild test robot. Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Tue, 16 Apr 2013 03:20:03 +0000 (13:20 +1000)]
Merge tag 'omapdss-for-3.10' of git://gitorious.org/linux-omap-dss2/linux into drm-next
Omapdss patches for 3.10 merge window
The biggest changes are:
* DSI video mode: automatic clock and timing calculation
* Lots of platform data related panel driver cleanups, to prepare for DT
* tag 'omapdss-for-3.10' of git://gitorious.org/linux-omap-dss2/linux: (69 commits)
drm/omap: add statics to a few structs
drm/omap: Fix and improve crtc and overlay manager correlation
drm/omap: Take a fb reference in omap_plane_update()
drm/omap: Make fixed resolution panels work
drm/omap: fix modeset_init if a panel doesn't satisfy omapdrm requirements
OMAPDSS: DPI: widen the pck search when using dss fck
OMAPDSS: fix dss_fck clock rate rounding
omapdss: use devm_clk_get()
OMAPDSS: nec-nl8048 panel: Use dev_pm_ops
OMAPDSS: DISPC: Revert to older DISPC Smart Standby mechanism for OMAP5
OMAPDSS: DISPC: Configure doublestride for NV12 when using 2D Tiler buffers
omapdss: Features: Fix some parameter ranges
omapdss: DISPC: add max pixel clock limits for LCD and TV managers
OMAPDSS: DSI: Use devm_clk_get()
drivers: video: omap2: dss: Use PTR_RET function
OMAPDSS: VENC: remove platform_enable/disable calls
OMAPDSS: n8x0 panel: remove use of platform_enable/disable
OMAPDSS: n8x0 panel: handle gpio data in panel driver
OMAPDSS: picodlp panel: remove platform_enable/disable callbacks
OMAPDSS: picodlp panel: handle gpio data in panel driver
...
Chris Wilson [Fri, 12 Apr 2013 09:39:59 +0000 (10:39 +0100)]
drm: Perform ioctl command validation on the stored kernel values
Userspace is free to pass in any command bits it feels like through the
ioctl cmd, and for example trinity likes to fuzz those bits to create
conflicting commands. So instead of relying upon userspace to pass along
the correct IN/OUT flags for the ioctl, use the flags as expected by the
kernel.
This does have a side-effect that NULL pointers can not be substituted
by userspace in place of a struct. This feature was not being used by
any driver, but instead exposed all of the command handlers to a user
triggerable OOPS.
Paul Sokolovsky [Fri, 12 Apr 2013 23:28:26 +0000 (18:28 -0500)]
drm.h: Fix DRM compilation with bare-metal toolchain.
An ifdef in drm.h expects to be compiled with full-fledged Linux
toolchain, but it's common to compile kernel with just bare-metal
toolchain which doesn't define __linux__. So, also add __KERNEL__
check.
[nm@ti.com: port forward to 3.9-rc6 and post to dri devel for feedback as RFC] Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org> Signed-off-by: Dave Airlie <airlied@redhat.com>
drm: Destroy property blobs at mode config cleanup time
Property blob objects need to be destroyed when cleaning up to avoid
memory leaks. Go through the list of all blobs in the
drm_mode_config_cleanup() function and destroy them.
The drm_mode_config_cleanup() function needs to be moved after the
drm_property_destroy_blob() declaration. Move drm_mode_config_init() as
well to keep the functions together.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Tue, 16 Apr 2013 03:03:47 +0000 (13:03 +1000)]
Merge tag 'drm-intel-next-2013-04-06' of git://people.freedesktop.org/~danvet/drm-intel into drm-next
Daniel writes:
Since I expect Linus to open the merge window in about a week I guess this
is the last i915 feature pull for 3.10. Highlights:
Updated testing tree for -next. Highlights:
- Corner case fixes discovered with static analyzers (Damien).
- More fixes to combat unclaimed register errors on Haswell (Paulo).
- Some small fixes to the gpu turbo code (Rodrigo+Ben), Ben has more
fixes for overclocking support pending.
- More prep work for fastboot from Chris.
- VT-switchless suspend/resume from Jesse.
- The prep work of Egbert Eich's hpd irq storm handling. Hopefully we can
squeeze in the actual storm handling code for 3.10 ...
- More convenience helpers for Imre's sg iterator. Core parts acked by
Andrew Morton.
- A bit of backlight code cleanup from Jani.
- Fixed ilk gpu reset (Jesse).
- Reduced color range handling fixes for VLV (Ville).
The big item here is though the introduction of pipe_config to properly
pre-compute the desired modeset state before touching the hw. Together
with some very basic support to read out the current config from the hw
and compare the state with the sw tracking. This is all prep work for more
reliable fastboot, atomic modesets and other cool features. Stuff
converted to the new world includes:
- Most simple pipe attributes (reduce color range, pixel multiplier).
- Pipe bpp/dither handling.
- Some convenience flags like ->has_pch_encoder to simplify the code flow.
- (Almost) DP clock handling, had to be reverted since part of a prep
patch was lost in rebasing ...
Expect a lot of patches for this throughout 3.11, there's tons of work
till we have all state properly tracked for fastbooting to woExpect a lot
of patches for this throughout 3.11, there's tons of work till we have all
state properly tracked for fastbooting to work.
For 3.10 I have a bunch of fixes queued up and I plan to send them all out
at the end of this week. I need to shuffle patches in my -next queue a bit
so that we don't but feature-y stuff in there, too. The main thing I'd
like to sneak in is Egbert's hpd irq storm handling, which should be
pretty low-risk since all the infrastructure work has landed already. I
also have the oops fix pending, but that only mustered review before the
w/e and giving how hairy that part of our modeset code is, I want to give
it some more testing before forwarding.
Note: annarchy.fd.o seems to run out of disk space, so couldn't push the
usual for-airlied branch. Tag should work though.
Note 2: I've had to do a backmerge since conflicts grew too ugly, but the
upstream -rc I've backmerged is already in your drm-next.
* tag 'drm-intel-next-2013-04-06' of git://people.freedesktop.org/~danvet/drm-intel: (75 commits)
drm/i915: info level for simulated gpu hang dmesg notice
drm/i915: revert eDP bpp clamping code changes
Revert "drm/i915: fix DP get_hw_state return value"
drm/i915: Don't use the HDMI port color range bit on Valleyview
drm/i915: Set PIPECONF color range bit on Valleyview
drm/i915: extract i9xx_set_pipeconf
drm/i915: Add no-lvds quirk for Fujitsu Esprimo Q900
drm/i915: create pipe_config->dpll for clock state
drm/i915: hw readout support for ->has_pch_encoders
drm/i915: add hw state readout/checking for pipe_config
drm/i915: rip out superflous is_dp&is_cpu_edp tracking
drm/i915: remove leaky eDP functions
drm/i915: track dp target_clock in pipe_config
drm/i915: move dp_m_n computation to dp_encoder->compute_config
drm/i915: clear up the fdi/dp set_m_n confusion
drm/i915: Fix sdvo connector get_hw_state function
drm/i915: drop DPFLIPSTAT enables on VLV v3
drm/i915: add Punit read/write routines for VLV v2
drm/i915: panel power sequencing for VLV eDP v2
drm/i915/dp: fix up VLV DP handling v2
...
drm/gma500: Check connector status before restoring sdvo
This patch fixes a bug introduced by:
commit 749387dc8d8270b279f27a0a794cdf4f4a4aa774
Author: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Date: Sun Apr 7 16:35:50 2013 +0200
drm/gma500: Fix hibernation problems on sdvo encoders
The bug is triggered when we do a mode set on a sdvo encoder with all
connectors in the disconnected state. A crtc is considered enabled by
drm even though all of its connectors are disconnected. Work around
this by adding a check in our sdvo restore function.
Also remove the unneeded dpms on. Prepare and Commit will take care of
that.
drm: cleanup: use drm_framebuffer_reference instead of a kref_get
drm_framebuffer_lookup() does a kref_get() for the framebuffer if it finds one
corresponding to the fb id passed to it. Use drm_framebuffer_reference() instead
for clarity since it's the function used in other places to take a reference.
Signed-off-by: Archit Taneja <archit@ti.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
drm: Add 1600x1200 (UXGA) screen resolution to the built-in EDIDs
The 1600x1200 (UXGA) screen resolution was lacking in the set of
built-in selectable EDID screen resolutions that can be used to
repair misbehaving monitor firmware.
This patch adds the related data set and expands the documentation.
Signed-off-by: Carsten Emde <C.Emde@osadl.org> Acked-by: Paul Menzel <paulepanter@users.sourceforge.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
In order to make it easier to port the code to other operating systems
(like the BSDs), relicense the HDMI helpers under the more permissive
MIT license.
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de> Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Mon, 25 Feb 2013 04:47:55 +0000 (14:47 +1000)]
drm: add new QXL driver. (v1.4)
QXL is a paravirtual graphics device used by the Spice virtual desktop
interface.
The drivers uses GEM and TTM to manage memory, the qxl hw fencing however
is quite different than normal TTM expects, we have to keep track of a number
of non-linear fence ids per bo that we need to have released by the hardware.
The releases are freed from a workqueue that wakes up and processes the
release ring.
releases are suballocated from a BO, there are 3 release categories, drawables,
surfaces and cursor cmds. The hw also has 3 rings for commands, cursor and release handling.
The hardware also have a surface id tracking mechnaism and the driver encapsulates it completely inside the kernel, userspace never sees the actual hw surface
ids.
This requires a newer version of the QXL userspace driver, so shouldn't be
enabled until that has been placed into your distro of choice.
Authors: Dave Airlie, Alon Levy
v1.1: fixup some issues in the ioctl interface with padding
v1.2: add module device table
v1.3: fix nomodeset, fbcon leak, dumb bo create, release ring irq,
don't try flush release ring (broken hw), fix -modesetting.
v1.4: fbcon cpu usage reduction + suitable accel flags.
Signed-off-by: Alon Levy <alevy@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Christian König [Tue, 9 Apr 2013 14:35:42 +0000 (10:35 -0400)]
drm/radeon: add ring working query
Add new ioctl option and bumb minor version number.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Jerome Glisse <jglisse@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If the disabled rb mask register is not properly initialized
program a sane default based on the number of RBs for the
asic. This avoids a potential divide by 0 when calculating
the backend mask.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Archit Taneja [Tue, 26 Mar 2013 13:45:19 +0000 (19:15 +0530)]
drm/omap: Fix and improve crtc and overlay manager correlation
The omapdrm driver currently takes a config/module arg to figure out the number
of crtcs it needs to create. We could create as many crtcs as there are overlay
managers in the DSS hardware, but we don't do that because each crtc eats up
one DSS overlay, and that reduces the number of planes we can attach to a single
crtc.
Since the number of crtcs may be lesser than the number of hardware overlay
managers, we need to figure out which overlay managers to use for our crtcs. The
current approach is to use pipe2chan(), which returns a higher numbered manager
for the crtc.
The problem with this approach is that it assumes that the overlay managers we
choose will connect to the encoders the platform's panels are going to use,
this isn't true, an overlay manager connects only to a few outputs/encoders, and
choosing any overlay manager for our crtc might lead to a situation where the
encoder cannot connect to any of the crtcs we have chosen. For example, an
omap5-panda board has just one hdmi output. If num_crtc is set to 1, with the
current approach, pipe2chan will pick up the LCD2 overlay manager, which cannot
connect to the hdmi encoder at all. The only manager that could have connected
to hdmi was the TV overlay manager.
Therefore, there is a need to choose our overlay managers keeping in mind the
panels we have on that platform. The new approach iterates through all the
available panels, creates encoders and connectors for them, and then tries to
get a suitable overlay manager to create a crtc which can connect to the
encoders.
We use the dispc_channel field in omap_dss_output to retrieve the desired
overlay manager's channel number, we then check whether the manager had already
been assigned to a crtc or not. If it was already assigned to a crtc, we assume
that out of all the encoders which intend use this crtc, only one will run at a
time. If the overlay manager wan't assigned to a crtc till then, we create a
new crtc and link it with the overlay manager.
This approach just looks for the best dispc_channel for each encoder. On DSS HW,
some encoders can connect to multiple overlay managers. Since we don't try
looking for alternate overlay managers, there is a greater possibility that 2
or more encoders end up asking for the same crtc, causing only one encoder to
run at a time.
Also, this approach isn't the most optimal one, it can do either good or bad
depending on the sequence in which the panels/outputs are parsed. The optimal
way would be some sort of back tracking approach, where we improve the set of
managers we use as we iterate through the list of panels/encoders. That's
something left for later.
Signed-off-by: Archit Taneja <archit@ti.com> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
drm/omap: Take a fb reference in omap_plane_update()
When userspace calls SET_PLANE ioctl, drm core takes a reference of the fb and
passes control to the update_plane op defined by the drm driver.
In omapdrm, we have a worker thread which queues framebuffers objects received
from update_plane and displays them at the appropriate time.
It is possible that the framebuffer is destoryed by userspace between the time
of calling the ioctl and apply-worker being scheduled. If this happens, the
apply-worker holds a pointer to a framebuffer which is already destroyed.
Take an extra refernece/unreference of the fb in omap_plane_update() to prevent
this from happening. A reference is taken of the fb passed to update_plane(),
the previous framebuffer (held by plane->fb) is unreferenced. This will prevent
drm from destroying the framebuffer till the time it's unreferenced by the
apply-worker.
This is in addition to the exisitng reference/unreference in update_pin(),
which is taken for the scanout of the plane's current framebuffer, and an
unreference the previous framebuffer.
Signed-off-by: Archit Taneja <archit@ti.com> Reviewed-by: Rob Clark <robdclark@gmail.com> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Archit Taneja [Tue, 26 Mar 2013 13:45:20 +0000 (19:15 +0530)]
drm/omap: Make fixed resolution panels work
The omapdrm driver requires omapdss panel drivers to expose ops like detect,
set_timings and check_timings. These can be NULL for fixed panel DPI, DBI, DSI
and SDI drivers. At some places, there are no checks to see if the panel driver
has these ops or not, and that leads to a crash.
The following things are done to make fixed panels work:
- The omap_connector's detect function is modified such that it considers panel
types which are generally fixed panels as always connected(provided the panel
driver doesn't have a detect op). Hence, the connector corresponding to these
panels is always in a 'connected' state.
- If a panel driver doesn't have a check_timings op, assume that it supports the
mode passed to omap_connector_mode_valid(the 'mode_valid' drm helper function)
- The function omap_encoder_update shouldn't really do anything for fixed
resolution panels, make sure that it calls set_timings only if the panel
driver has one.
Signed-off-by: Archit Taneja <archit@ti.com> Reviewed-by: Rob Clark <robdclark@gmail.com> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Archit Taneja [Tue, 26 Mar 2013 13:45:18 +0000 (19:15 +0530)]
drm/omap: fix modeset_init if a panel doesn't satisfy omapdrm requirements
modeset_init iterates through all the registered omapdss devices and has some
initial checks to see if the panel has a driver and the required driver ops for
it to be usable by omapdrm.
The function bails out from modeset_init if a panel doesn't meet the
requirements, and stops the registration of the future panels and encoders which
come after it, that isn't the correct thing to do, we should go through the rest
of the panels. Replace the 'return's with 'continue's.
Signed-off-by: Archit Taneja <archit@ti.com> Reviewed-by: Rob Clark <robdclark@gmail.com> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Tomi Valkeinen [Wed, 10 Apr 2013 11:54:54 +0000 (14:54 +0300)]
OMAPDSS: DPI: widen the pck search when using dss fck
When not using DSI PLL to generate the pixel clock, but DSS FCK, the
possible pixel clock rates are rather limited. DSS FCK is currently used
on OMAP2 and OMAP3.
When using Beagleboard with a monitor that supports high resolutions,
the clock rates do not match (at least for me) for the monitor's pixel
clocks within the current threshold in the code, which is +/- 1MHz.
This patch widens the search up to +/- 15MHz. The search is done in
steps, i.e. it first tries to find a rather exact clock, than a bit less
exact, etc. so this should not change the cases where a clock was
already found.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Tomi Valkeinen [Wed, 10 Apr 2013 11:47:38 +0000 (14:47 +0300)]
OMAPDSS: fix dss_fck clock rate rounding
DSS func clock is calculated with prate / div * m. However, the current
omapdss code calculates it with prate * m / div, which yields a slightly
different result when there's a remainder. For example, 432000000 / 14 *
2 = 61714284, but 432000000 * 2 / 14 = 61714285.
In addition to that, the clock framework wants the clock rate given with
clk_set_rate to be higher than the actual (truncated) end result. So, if
prate is 432000000, and div is 14, the real result is 30857142.8571...
We need to call clk_set_rate with 30857143, which gives us a clock of 30857142. That's why we need to use DIV_ROUND_UP() when calling
clk_set_rate.
This patch fixes the clock calculation.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Last version of this patch is not clear enough and X86 duplicated.
This patch fixes build failure of v3.9-rc5 and rc6.
When config ACPI_VIDEO as m, DRM_GMA500 as y, here comes the failure.
GMA5/600 needs acpi_video just like nouveau.
And some tab type fix by the way.
Failure message:
drivers/built-in.o: In function `psb_driver_load':
kernel-3.9-rc5/drivers/gpu/drm/gma500/psb_drv.c:340: \
undefined reference to `acpi_video_register'
make: *** [vmlinux] Error 1
Archit Taneja [Tue, 26 Mar 2013 13:45:25 +0000 (19:15 +0530)]
OMAPDSS: DISPC: Revert to older DISPC Smart Standby mechanism for OMAP5
DISPC on OMAP5 has a more optimised mechanism of asserting Mstandby to achieve
more power savings when DISPC is configured in Smart Standby mode. This
mechanism leads to underflows when multiple DISPC pipes are enabled.
There is a register field which can let us revert to the older mechanism of
asserting Mstandby. Configure this field to prevent underflows.
Signed-off-by: Archit Taneja <archit@ti.com> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Archit Taneja [Tue, 26 Mar 2013 13:45:24 +0000 (19:15 +0530)]
OMAPDSS: DISPC: Configure doublestride for NV12 when using 2D Tiler buffers
When using a DISPC video pipeline to a fetch a NV12 buffer in a 2D container, we
need to set set a doublestride bit in the video pipe's ATTRIBUTES register. This
is needed because the stride for the UV plane(using a 16 bit Tiler container) is
double the stride for the Y plane(using a 8 bit Tiler container) for the 0 or
180 degree views. The ROW_INC register is meant for the Y plane, and the HW will
calculate the row increment needed for the UV plane by using double the stride
value based on whether this bit is set or not.
Set the bit when we are using a 2D Tiler buffer and when rotation is 0 or 180
degrees. The stride value is the same for 90 and 270 degree Tiler views, hence
the bit shouldn't be set.
Signed-off-by: Archit Taneja <archit@ti.com> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Archit Taneja [Tue, 26 Mar 2013 13:45:23 +0000 (19:15 +0530)]
omapdss: Features: Fix some parameter ranges
Increase the DSS_FCLK and DSI_FCLK max supported frequencies, these come because
some frequencies were increased from OMAP5 ES1 to OMAP5 ES2. We support only
OMAP5 ES2 in the kernel, so replace the ES1 values with ES2 values. Increase the
DSI PLL Fint range, this was previously just copied from the OMAP4 param range
struct.
Fix the maximum DSS_FCLK on OMAP2, it's 133 Mhz according to the TRM.
Signed-off-by: Archit Taneja <archit@ti.com> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Archit Taneja [Tue, 26 Mar 2013 13:45:22 +0000 (19:15 +0530)]
omapdss: DISPC: add max pixel clock limits for LCD and TV managers
Each version of OMAP has a limitation on the maximum pixel clock frequency
supported by an overlay manager. This limit isn't checked by omapdss. Add
dispc feats for lcd and tv managers and check whether the target timings can
be supported or not.
The pixel clock limitations are actually more complex. They depend on which OPP
OMAP is in, and they also depend on which encoder is the manager connected to.
The OPP dependence is ignored as DSS forces the PM framework to be on OPP100
when DSS is enabled, and the encoder dependencies are ignored by DISPC for now.
These limits should come from the encoder driver.
The OMAP2 TRM doesn't mention the maximum pixel clock limit. This value is left
as half of DSS_FCLK, as OMAP2 requires the PCD to be atleast 2.
Signed-off-by: Archit Taneja <archit@ti.com> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
projects / linux-beck.git / log