GitHub/exynos8895/android_kernel_samsung_universal8895.git
9 years agodrm/i915: Dump pfit state as hex
Ville Syrjälä [Thu, 3 Sep 2015 18:50:16 +0000 (21:50 +0300)]
drm/i915: Dump pfit state as hex

The pfit state is stored as register values, so dump them as hex instead
of decimal to make some sense of the error messages.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: access the PP_ON_DELAYS/PP_OFF_DELAYS regs only pre GEN5
Imre Deak [Thu, 3 Sep 2015 13:24:36 +0000 (16:24 +0300)]
drm/i915: access the PP_ON_DELAYS/PP_OFF_DELAYS regs only pre GEN5

These registers exist only before GEN5, so currently we may access
undefined registers on VLV/CHV and BXT. Apply the workaround only pre
GEN5.

Since the workaround is relevant only when LVDS is present, for clarity
apply it only if this is the case.

This triggered an unclaimed register access warning on BXT.

v2: (Ville)
- move the workaround to the LVDS init code
- print a debug note about the workaround

Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: access the PP_CONTROL reg only pre GEN5
Imre Deak [Thu, 3 Sep 2015 13:24:35 +0000 (16:24 +0300)]
drm/i915: access the PP_CONTROL reg only pre GEN5

This register exists only pre GEN5, but atm we also access it on
VLV/BXT/CHV. Prevent accessing it on these latter platforms.

Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Refactor common ringbuffer allocation code
Chris Wilson [Thu, 3 Sep 2015 12:01:39 +0000 (13:01 +0100)]
drm/i915: Refactor common ringbuffer allocation code

A small, very small, step to sharing the duplicate code between
execlists and legacy submission engines, starting with the ringbuffer
allocation code.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Arun Siluvery <arun.siluvery@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Dave Gordon <david.s.gordon@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: use the yesno helper for logging
Jani Nikula [Thu, 3 Sep 2015 08:16:09 +0000 (11:16 +0300)]
drm/i915: use the yesno helper for logging

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: ignore link rate in TPS3 selection
Jani Nikula [Thu, 3 Sep 2015 08:16:08 +0000 (11:16 +0300)]
drm/i915: ignore link rate in TPS3 selection

TPS3 is mandatory for downstream devices that support HBR2, and Intel
platforms that support HBR2 also support TPS3. Whenever TPS3 is
supported by both the source and sink, it should be used. In other
words, whenever the source and sink are capable of 5.4 Gbps link, we
should anyway go for TPS3, regardless of the link rate being selected.

Log an error if the sink has advertized HBR2 capability without TPS3
capability.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/dp: move TPS3 logic to where it's used
Jani Nikula [Thu, 3 Sep 2015 08:16:07 +0000 (11:16 +0300)]
drm/i915/dp: move TPS3 logic to where it's used

There is no need to have a separate flag for tps3 as the information is
only used at one location. Move the logic there to make it easier to
follow.

Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Future proof panel fitter.
Rodrigo Vivi [Wed, 2 Sep 2015 22:19:26 +0000 (15:19 -0700)]
drm/i915: Future proof panel fitter.

This is another case where we can consider the default is the
newest available and not actually a missed case.

No functional change.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Future proof uncore_init.
Rodrigo Vivi [Thu, 3 Sep 2015 21:38:00 +0000 (14:38 -0700)]
drm/i915: Future proof uncore_init.

Unless future specs tells otherwise we can assume future gens
inherit some stuff from the previous so let's handle
missed cases when we know tehy should't be there and assume
default equals newest one.

No functional changes.

v2: Remove useless case as pointed out by Ville.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Future proof interrupt handler.
Rodrigo Vivi [Wed, 2 Sep 2015 22:19:24 +0000 (15:19 -0700)]
drm/i915: Future proof interrupt handler.

These functions are already being called for gen >= 9,
so let's be sure when this happens we use whatever is
there already for the latest platform.

No functional change.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/gtt: Avoid calling kcalloc in a loop when allocating temp bitmaps
Michał Winiarski [Thu, 3 Sep 2015 17:22:18 +0000 (19:22 +0200)]
drm/i915/gtt: Avoid calling kcalloc in a loop when allocating temp bitmaps

On each call to gen8_alloc_va_range_3lvl we're allocating temporary
bitmaps needed for error handling. Unfortunately, when we increase
address space size (48b ppgtt) we do additional (512 - 4) calls to
kcalloc, increasing latency between exec and actual start of execution
on the GPU. Let's just do a single kcalloc, we can also drop the size
from free_gen8_temp_bitmaps since it's no longer used.

v2: Use GFP_TEMPORARY to make the allocations reclaimable.
v3: Drop the 2D array, just allocate a single block.
v4: Rebase to handle gen8_preallocate_top_level_pdps.
v5: Align misaligned bracket.

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Correct kcalloc arguments as suggested by Chris.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/skl+: Add YUV pixel format in Capability list
Kumar, Mahesh [Thu, 3 Sep 2015 10:47:09 +0000 (16:17 +0530)]
drm/i915/skl+: Add YUV pixel format in Capability list

GEN >= 9 supports YUV format for all planes, but it's not exported in
Capability list of primary plane. Add YUV formats in skl_primary_formats
list.

Testcase: igt/kms_universal_plane.c

Signed-off-by: Kumar, Mahesh <mahesh1.kumar@intel.com>
Cc: Konduru, Chandra <chandra.konduru@intel.com>
Reviewed-by: Sonika Jindal <sonika.jindal@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/skl: Avoid using un-initialized bits_per_pixel
Kumar, Mahesh [Thu, 3 Sep 2015 10:47:08 +0000 (16:17 +0530)]
drm/i915/skl: Avoid using un-initialized bits_per_pixel

Don't rely on fb->bits_per_pixel as intel_framebuffer_init is not
filling bits_per_pixel field of fb-struct for YUV pixel format.
This leads to divide by zero error during watermark calculation.

Signed-off-by: Kumar, Mahesh <mahesh1.kumar@intel.com>
Cc: Konduru, Chandra <chandra.konduru@intel.com>
Reviewed-by: Sonika Jindal <sonika.jindal@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Don't call intel_get_hpd_pins() when there's no hotplug interrupt
Ville Syrjälä [Fri, 28 Aug 2015 19:59:08 +0000 (22:59 +0300)]
drm/i915: Don't call intel_get_hpd_pins() when there's no hotplug interrupt

On GMCH plaforms we are now getting the following spew on aux
interrupts:
[drm:intel_get_hpd_pins] hotplug event received, stat 0x00000000, dig 0x00000000, pins 0x00000000
[drm:intel_get_hpd_pins] hotplug event received, stat 0x00000000, dig 0x00000000, pins 0x00000000
[drm:intel_get_hpd_pins] hotplug event received, stat 0x00000000, dig 0x00000000, pins 0x00000000
[drm:intel_get_hpd_pins] hotplug event received, stat 0x00000000, dig 0x00000000, pins 0x00000000
[drm:intel_get_hpd_pins] hotplug event received, stat 0x00000000, dig 0x00000000, pins 0x00000000
[drm:intel_dp_aux_ch] dp_aux_ch timeout status 0x71450064

Prevent it by not calling intel_get_hpd_pins() unless one of the HPD
interrupt bits are actually set.

I already fixed similar annoyance once with
4bca26d0a6518d51a9abe64fbde4b12f04c74053 drm/i915: Use HOTPLUG_INT_STATUS_G4X on VLV/CHV

but another source for it got added in
fd63e2a972c670887e5e8a08440111d3812c0996 drm/i915: combine i9xx_get_hpd_pins and pch_get_hpd_pins

due to pch_get_hpd_pins() being chosen over i9xx_get_hpd_pins() to
serve as the new unified piece of code. pch_get_hpd_pins() had the debug
print, and i9xx_get_hpd_pins() didn't.

Cc: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Rewrite BXT HPD code to conform to pre-existing style
Ville Syrjälä [Thu, 27 Aug 2015 20:56:11 +0000 (23:56 +0300)]
drm/i915: Rewrite BXT HPD code to conform to pre-existing style

Rewrite the BXT hpd setup to match the way we do it on other platforms:
- Throw out BXT_HOTPLUG_CTL since it's the same as PCH_PORT_HOTPLUG
- Enable the HPD bits in the DE port IER in gen8_de_irq_postinstall()
- Update DE port IMR using bdw_update_port_irq()

Also throw out port D from bxt_port_hotplug_long_detect() since BXT only
goes up to C.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Refactor the hpd irq handling functions
Ville Syrjälä [Thu, 27 Aug 2015 20:56:10 +0000 (23:56 +0300)]
drm/i915: Refactor the hpd irq handling functions

A lot of the hpd irq handling is duplicated code, so refactor it a bit
by observing that in several places the only difference is the hpd[]
array. So pull the code to a few functions and pass in the hpd[] array
from the caller. Another option would be to determine the correct array
to use within the functions themselves, but somehow passing it in felt
nicer.

Further code reduction could be achieved by passing in the hotplug
register offset, and the long pulse detection function pointer. But that
didn't feel as good for some reason, so I left it at the middle ground.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Rewrite bxt_hpd_handler() to look like everyone else
Ville Syrjälä [Thu, 27 Aug 2015 20:56:09 +0000 (23:56 +0300)]
drm/i915: Rewrite bxt_hpd_handler() to look like everyone else

bxt_hpd_handler() looks different to everyone else for no good reason.
Rewrite it to use the standard variable namees etc.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Reinitialize HPD after runtime D3
Ville Syrjälä [Thu, 27 Aug 2015 20:56:08 +0000 (23:56 +0300)]
drm/i915: Reinitialize HPD after runtime D3

Runtime suspends disabled all interrupts, so in order to get them back
fully we need to also do the HPD irq setup on runtime resume. Except
on VLV/CHV where the display interrupt initialization is part of the
display power well powerup.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add port A HPD support for SPT
Ville Syrjälä [Thu, 27 Aug 2015 20:56:07 +0000 (23:56 +0300)]
drm/i915: Add port A HPD support for SPT

On SKL the port A HPD has moved to the PCH. Hook it up.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add port A HPD support for BDW
Ville Syrjälä [Thu, 27 Aug 2015 20:56:06 +0000 (23:56 +0300)]
drm/i915: Add port A HPD support for BDW

Wire up the port A HPD for BDW. Compared to earlier platforms the
interrupt setup is a bit different, but basically everything else
looks the same.

v2: 0 initialize pin_mask/long_mask due to intel_get_hpd_pins() changes
    Check for BDW before processing the HPD to not break BXT
    Set found=true when processing port A HPD
    Sort out the mess I made of the irq setup in v1
    Warn about bad irq mask vs. enable bits in bdw_update_port_irq()  (Paulo)

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: LPT:LP needs port A HPD enabled in both north and south
Ville Syrjälä [Thu, 27 Aug 2015 20:56:05 +0000 (23:56 +0300)]
drm/i915: LPT:LP needs port A HPD enabled in both north and south

If the CPU and PCH are on the same package we must enabled the port A
HPD also in the south hotplug register. To identify the package type
we simply look at the PCH type: LPT-H means separate package, and
LPT-LP means multi chip package (MCP).

v2: Add comment and pimp commit message

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add port A HPD support for IVB/HSW
Ville Syrjälä [Thu, 27 Aug 2015 20:56:04 +0000 (23:56 +0300)]
drm/i915: Add port A HPD support for IVB/HSW

As with ILK/SNB wire up the port A HPD on IVB/HSW.

This might be more important on HSW with PSR. BSpec tells us that if the
automagic link training performed by the hardware fails for some reason,
we're going to get a short HPD and are supposed to re-train the link
manyally.

v2: 0 initialize pin_mask/long_mask due to intel_get_hpd_pins() changes
    Add a comment about the pulse duration bits being reserved on HSW+
    like we have for LPT+ in ibx_hpd_irq_setup()

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add port A HPD support for ILK/SNB
Ville Syrjälä [Thu, 27 Aug 2015 20:56:03 +0000 (23:56 +0300)]
drm/i915: Add port A HPD support for ILK/SNB

ILK/SNB support port A HPD. While HPD is optional on eDP let's at least
try to wite it up so that we might notice if the link has issues.

The eDP spec suggests that if HPD is not wired up, one should poll the
link status instead. We don't even do that currently.

v2: 0 initialize pin_mask/long_mask due to intel_get_hpd_pins() changes

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Introduce spt_irq_handler()
Ville Syrjälä [Thu, 27 Aug 2015 20:56:02 +0000 (23:56 +0300)]
drm/i915: Introduce spt_irq_handler()

Starting from SPT the only interrupts living in the south are GMBUS and
HPD. What's worse some of the SPT specific new bits conflict with some
other bits on earlier PCH generations. So better not use the
cpt_irq_handler() for SPT+ anymore.

Also kill the hand rolled port E handling with something more
standardish. This also avoids accidentally confusing port B and port E
long pulses since the bits occupy the same positions, just in different
registers.

Also add a comment noting that the short pulse duration bits are
reserved on LPT+. The 2ms value we program is 0, so no issue wrt. the
MBZ in the spec.

v2: Call intel_hpd_irq_handler() only once (Paulo)

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Move {pin, long}_mask initialization to caller from intel_get_hpd_pins()
Ville Syrjälä [Fri, 28 Aug 2015 18:26:27 +0000 (21:26 +0300)]
drm/i915: Move {pin, long}_mask initialization to caller from intel_get_hpd_pins()

Move the 0 initialization of pin_mask and long_mask from
intel_get_hpd_pins() into each caller. This we we can call
intel_get_hpd_pins() multiple times to accumulate more pins from several
sources.

v2: Add a comment explaining the dangers of intel_get_hpd_pins() (Paulo)

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Rename BXT PORTA HPD defines
Ville Syrjälä [Thu, 27 Aug 2015 20:56:00 +0000 (23:56 +0300)]
drm/i915: Rename BXT PORTA HPD defines

The PORTA HPD defines are not BXT specific. They also exist on SPT,
and partially already on LPT:LP.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add HAS_PCH_LPT_LP() macro
Ville Syrjälä [Thu, 27 Aug 2015 20:55:59 +0000 (23:55 +0300)]
drm/i915: Add HAS_PCH_LPT_LP() macro

Make LPT:LP checks look neater by wrapping the details in a
new HAS_PCH_LPT_LP() macro.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Factor out ilk_update_display_irq()
Ville Syrjälä [Thu, 27 Aug 2015 20:55:58 +0000 (23:55 +0300)]
drm/i915: Factor out ilk_update_display_irq()

Extract the core of ironlake_{enable,disable}_display_irq() into a new
function. We'll have further use for it later.

v2: Warn about invalid mask vs. enable bits (Paulo)

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Extract intel_hpd_enabled_irqs()
Ville Syrjälä [Thu, 27 Aug 2015 20:55:57 +0000 (23:55 +0300)]
drm/i915: Extract intel_hpd_enabled_irqs()

Eliminate a bunch of duplicated code that calculates the currently
enabled HPD interrupt bits.

v2: s/;/:/ in patch subject (Paulo)

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Clean up various HPD defines
Ville Syrjälä [Thu, 27 Aug 2015 20:55:56 +0000 (23:55 +0300)]
drm/i915: Clean up various HPD defines

Indent the PORTx_HOTPLUG_... defines appropriately, and fix some space
vs. tab issues.

v2: Document pre-HSW/LPT bits, and order another tab (Paulo)

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Update comments around base bpp
Daniel Vetter [Wed, 26 Aug 2015 16:57:26 +0000 (18:57 +0200)]
drm/i915: Update comments around base bpp

Forgot to do that in

commit d328c9d78d64ca11e744fe227096990430a88477
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Fri Apr 10 16:22:37 2015 +0200

    drm/i915: Select starting pipe bpp irrespective or the primary plane

and it's confusing. Fix it.

Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
9 years agoMerge tag 'drm-intel-next-fixes-2015-09-02' into drm-intel-next-queued
Daniel Vetter [Wed, 2 Sep 2015 12:33:42 +0000 (14:33 +0200)]
Merge tag 'drm-intel-next-fixes-2015-09-02' into drm-intel-next-queued

Backmerge -fixes since there's more DDI-E related cleanups on top of
the pile of -fixes for skl that just landed for 4.3.

Conflicts:
drivers/gpu/drm/i915/intel_display.c
drivers/gpu/drm/i914/intel_dp.c
drivers/gpu/drm/i915/intel_lrc.c

Conflicts are all fairly harmless adjacent line stuff.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
9 years agodrm/i915: Also record time difference if vblank evasion fails, v2.
Maarten Lankhorst [Tue, 1 Sep 2015 10:15:33 +0000 (12:15 +0200)]
drm/i915: Also record time difference if vblank evasion fails, v2.

This makes the error message slightly more useful.

Changes since v1:
- Use ktime_get() while irqs are still disabled. (vsyrjala)

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Remove start frame argument to pipe_update_begin/end.
Maarten Lankhorst [Mon, 31 Aug 2015 11:04:25 +0000 (13:04 +0200)]
drm/i915: Remove start frame argument to pipe_update_begin/end.

There's already a per crtc member that can be used for it.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: guest i915 notification for Intel GVT-g
Zhiyuan Lv [Fri, 28 Aug 2015 07:41:18 +0000 (15:41 +0800)]
drm/i915: guest i915 notification for Intel GVT-g

When i915 drivers run inside a VM with Intel GVT-g, some explicit
notifications are needed from guest to host device model through PV
INFO page write. The notifications include:

PPGTT create
PPGTT destroy

They are used for the shadow implementation of PPGTT. Intel GVT-g
needs to write-protect the guest pages of PPGTT, and clear the write
protection when they end their life cycle.

v2:
- Use lower_32_bits()/upper_32_bits() for qword operations;
- Remove the notification of guest context creation/destroy;

Signed-off-by: Zhiyuan Lv <zhiyuan.lv@intel.com>
Signed-off-by: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Update PV INFO page definition for Intel GVT-g
Zhiyuan Lv [Fri, 28 Aug 2015 07:41:17 +0000 (15:41 +0800)]
drm/i915: Update PV INFO page definition for Intel GVT-g

Some more definitions in the PV info page are added. They are mainly
for the guest notification to Intel GVT-g device model. They are used
for Broadwell enabling.

The notification of PPGTT page table creation/destroy is to notify
GVT-g device model the life cycle of guest page tables. Then device
model will implement shadow page table for guests.

The notification of context create/destroy is optional. If it is used,
the device model will create/destroy shadow context corresponding to
the context's life cycle. Guest driver needs to make sure that the
context's LRCA and backing storage address unchanged. If it is not
used, the device model will perform the context shadow work in the
context scheduling time.

Signed-off-by: Zhiyuan Lv <zhiyuan.lv@intel.com>
Signed-off-by: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Always enable execlists on BDW for vgpu
Zhiyuan Lv [Fri, 28 Aug 2015 07:41:16 +0000 (15:41 +0800)]
drm/i915: Always enable execlists on BDW for vgpu

Broadwell hardware supports both ring buffer mode and execlist mode.
When i915 runs inside a VM with Intel GVT-g, we allow execlist mode
only.

The main reason of EXECLIST only is that GVT-g does not support the
dynamic mode switch between ring buffer mode and execlist mode when
running multiple virtual machines.

v2:
- Adjust the position of vgpu check in sanitize function (Joonas)
- Add vgpu error check in context initialization. (Joonas, Daniel)

Signed-off-by: Zhiyuan Lv <zhiyuan.lv@intel.com>
Signed-off-by: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: preallocate pdps for 32 bit vgpu
Zhiyuan Lv [Fri, 28 Aug 2015 07:41:14 +0000 (15:41 +0800)]
drm/i915: preallocate pdps for 32 bit vgpu

This is based on Mika Kuoppala's patch below:
http://article.gmane.org/gmane.comp.freedesktop.xorg.drivers.intel/61104/match=workaround+hw+preload

The patch will preallocate the page directories for 32-bit PPGTT when
i915 runs inside a virtual machine with Intel GVT-g. With this change,
the root pointers in EXECLIST context will always keep the same.

The change is needed for vGPU because Intel GVT-g will do page table
shadowing, and needs to track all the page table changes from guest
i915 driver. However, if guest PPGTT is modified through GPU commands
like LRI, it is not possible to trap the operations in the right time,
so it will be hard to make shadow PPGTT to work correctly.

Shadow PPGTT could be much simpler with this change. Meanwhile
hypervisor could simply prohibit any attempt of PPGTT modification
through GPU command for security.

The function gen8_preallocate_top_level_pdps() in the patch is from
Mika, with only one change to set "used_pdpes" to avoid duplicated
allocation later.

Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Dave Gordon <david.s.gordon@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Zhiyuan Lv <zhiyuan.lv@intel.com>
Signed-off-by: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: add yesno utility function
Jani Nikula [Thu, 27 Aug 2015 13:23:30 +0000 (16:23 +0300)]
drm/i915: add yesno utility function

Add a common function to return "yes" or "no" string based on the
argument, and drop the local versions of it.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: move intel_hrawclk() to intel_display.c
Jani Nikula [Wed, 26 Aug 2015 07:58:20 +0000 (10:58 +0300)]
drm/i915: move intel_hrawclk() to intel_display.c

Make it available outside of intel_dp.c.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Clint Taylor <Clinton.A.Taylor@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Notify GuC rc6 state
Alex Dai [Tue, 18 Aug 2015 21:34:47 +0000 (14:34 -0700)]
drm/i915: Notify GuC rc6 state

If rc6 is enabled, notify GuC so it can do proper forcewake before
command submission.

Signed-off-by: Alex Dai <yu.dai@intel.com>
Reviewed-by: Tom O'Rourke <Tom.O'Rourke@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/guc: Support GuC version 4.3
Alex Dai [Tue, 18 Aug 2015 21:32:35 +0000 (14:32 -0700)]
drm/i915/guc: Support GuC version 4.3

The firmware layout changes that now it only has css header +
uCode + RSA signature. Plus, other trivial changes to support
GuC V4.3.

Signed-off-by: Alex Dai <yu.dai@intel.com>
Reviewed-by: Dave Gordon <david.s.gordon@intel.com>
Reviewed-by: Dave Gordon <david.s.gordon@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Fix module initialisation, v2.
Maarten Lankhorst [Thu, 27 Aug 2015 13:15:15 +0000 (15:15 +0200)]
drm/i915: Fix module initialisation, v2.

The driver doesn't support UMS any more, so set DRIVER_MODESET by default,
remove the legacy s/r callbacks, and rename the s/r functions to make it more clear
they're only in use by switcheroo now.

Also remove an obsolete comment about atomic. Normal updates are supported only
async updates aren't yet.

v2: Don't unconditionally set DRIVER_ATOMIC, we're not yet there.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Factor out intel_crtc_has_encoders()
Ville Syrjälä [Wed, 26 Aug 2015 16:39:19 +0000 (19:39 +0300)]
drm/i915: Factor out intel_crtc_has_encoders()

Make the code mode readable by pulling the "does this crtc have any
encoders?" deduction into a separate function.

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Fix clock readout when pipes are enabled w/o ports
Ville Syrjälä [Wed, 26 Aug 2015 16:39:18 +0000 (19:39 +0300)]
drm/i915: Fix clock readout when pipes are enabled w/o ports

The BIOS sometimes likes to enable pipes w/o any ports, at least on
older machines. Currently we fail to assign anything sensible to
crtc->hwmode.crtc_clock which leads to complaints from the vblank code.
Deal with active pipes w/o ports and assign something sensible to
crtc_clock in i9xx_get_pipe_config(). The encoder .get_config() will
override this if the port is enabled.

Gets rid of rest of these on my gen4:
[drm:drm_calc_timestamping_constants [drm]] *ERROR* crtc 24: Can't calculate constants, dotclock = 0!
[drm:i915_get_vblank_timestamp] crtc 1 is disabled

v2: Fill out crtc_clock already in i9xx_get_pipe_config() (Maarten)

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add CHV PHY LDO power sanity checks
Ville Syrjälä [Wed, 8 Jul 2015 20:46:01 +0000 (23:46 +0300)]
drm/i915: Add CHV PHY LDO power sanity checks

At various points when changing the DPIO lane/phy power states,
construct an expected value of the DISPLAY_PHY_STATUS register
and compare it with the real thing.

To construct the expected value we look at our shadow PHY_CONTROL
register value (which should match what we've just written to the
hardware), and we also need to look at the actual state of the cmn
power wells as a disabled power well causes the relevant LDO status
to be reported as 'on' in DISPLAY_PHY_STATUS.

When initially powering up the PHY it performs various internal
calibrations for which it fully powers up. That means that if we check
for the expetected power state immediately upon releasing cmnreset we
would get the occasional false positive. But we can of course
poll until the expected value appears. It shouldn't be too long so
this shouldn't make modesets substantially longer.

One extra complication is introduced when we cross the streams, ie.
drive port B with pipe B. In this case we trick CL2 (where the DPLL lives)
into life by temporaily powering up the lanes in the second channel,
and once the pipe is up and runnign we release the lane power override.
At that point the power state of CL2 has somehow gotten entangled with
the power state of the first channel. That means that constructing the
expected DISPLAY_PHY_STATUS value is a bit tricky since based on the
lane power states in the second channel, CL2 should also be powered
down. But we can use the DPLL enable bit to determine when CL2 should
be alive even if the lanes are powered down. However the power state
of CL2 isn't actually tied in with the DPLL state, but to the state
of the lanes in first channel, so we have to avoid checking the
expected state between shutting down the DPLL and powering down
the lanes in the first channel. So no calling assert_chv_phy_status()
before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(),
but after the write is a safe time to check.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add some CHV DPIO lane power state asserts
Ville Syrjälä [Wed, 8 Jul 2015 20:46:00 +0000 (23:46 +0300)]
drm/i915: Add some CHV DPIO lane power state asserts

Add some checks that the state of the DPIO lanes is more or less what we
expect based on the overrides.

The hardware only provides two bits per channel indicating whether all
or some of the lanes are powered down, so we can't do an exact check.

Additionally, CL2 powering down before we can check it adds another
twist. To work around this we simply check for the 0 value of the
CL2 register (which is what we get when it's powered down) and
adjust our expectations.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Clean up CHV lane soft reset programming
Ville Syrjälä [Thu, 9 Jul 2015 17:14:11 +0000 (20:14 +0300)]
drm/i915: Clean up CHV lane soft reset programming

Currently we release the lane soft reset before lane stagger settings
have been programmed. I believe that means we don't actually do lane
staggering. So move the soft reset deassert to happen after lane
staggering has been programmed.

The one confusing thing in this is that when we remove the power down
override from the lanes, they power up with defaul register values,
which do not have the soft reset overrides enabled. And according to
some docs by default the data lane resets are tied to cmnreset. So that
would mean that lanes would come out of reset without staggering as
soon as the power down overrides are removed. But since we can't access
either the lane stagger register nor the soft reset override registers
until the lanes are powered on, we can't really do anything about it.
So let's just set the soft reset overrides as soon as the lane is
powered on and hope for the best.

v2: Fix typos in commit message (Daniel)

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agoi915: Set ddi_pll_sel in DP MST path
Ander Conselvan de Oliveira [Mon, 31 Aug 2015 08:23:28 +0000 (11:23 +0300)]
i915: Set ddi_pll_sel in DP MST path

The DP MST encoder config function never sets ddi_pll_sel, even though
its value is programmed in its ->pre_enable() hook. That used to work
because a new pipe_config was kzalloc'ed at every modeset, and the value
of zero selects the highest clock for the PLL. Starting with the commit
below, the value of ddi_pll_sel is preserved through modesets, and since
the correct value wasn't properly setup by the MST code, it could lead
to warnings and blank screens.

commit 8504c74c7ae48b4b8ed1f1c0acf67482a7f45c93
Author: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Date:   Fri May 15 11:51:50 2015 +0300

    drm/i915: Preserve ddi_pll_sel when allocating new pipe_config

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91628
Cc: stable@vger.kernel.org # 7e6313a2516d drm/i915: Don't use link_bw for PLL setup
Cc: stable@vger.kernel.org
Cc: Timo Aaltonen <tjaalton@ubuntu.com>
Cc: Luciano Coelho <luciano.coelho@intel.com>
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: Bump command parser version number.
Francisco Jerez [Mon, 15 Jun 2015 11:03:29 +0000 (14:03 +0300)]
drm/i915: Bump command parser version number.

This was forgotten in

commit d351f6d94893f3ba98b1b20c5ef44c35fc1da124
Author: Francisco Jerez <currojerez@riseup.net>
Date:   Fri May 29 16:44:15 2015 +0300

    drm/i915: Add SCRATCH1 and ROW_CHICKEN3 to the register whitelist.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/dp: use the drm dp helper for determining sink tps3 support
Jani Nikula [Thu, 27 Aug 2015 10:25:37 +0000 (13:25 +0300)]
drm/i915/dp: use the drm dp helper for determining sink tps3 support

No functional changes.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
[danvet: s/intel_dp_tps/drm_dp_tps/.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Don't use link_bw for PLL setup
Ville Syrjälä [Tue, 11 Aug 2015 17:21:46 +0000 (20:21 +0300)]
drm/i915: Don't use link_bw for PLL setup

Use port_clock instead of link_bw when picking the PLL parameters for
DP. link_bw may be zero with an eDP 1.4 sink that supports
DP_LINK_RATE_SET so we shouldn't use it for anything other than feed it
to the sink appropriately.

v2: Fix typo in commit message (Sivakumar)

Reviewed-by: Sivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
[Jani: cherry-picked from future.]
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/dp: add drm_dp_tps3_supported helper
Jani Nikula [Wed, 26 Aug 2015 11:33:31 +0000 (14:33 +0300)]
drm/dp: add drm_dp_tps3_supported helper

Cc: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Preserve SSC earlier
Lukas Wunner [Wed, 15 Jul 2015 11:57:35 +0000 (13:57 +0200)]
drm/i915: Preserve SSC earlier

Commit 92122789b2d6 ("drm/i915: preserve SSC if previously set v3")
added code to intel_modeset_gem_init to override the SSC status read
from VBT with the SSC status set by BIOS.

However, intel_modeset_gem_init is invoked *after* intel_modeset_init,
which calls intel_setup_outputs, which *modifies* SSC status by way of
intel_init_pch_refclk. So unlike advertised, intel_modeset_gem_init
doesn't preserve the SSC status set by BIOS but whatever
intel_init_pch_refclk decided on.

This is a problem on dual gpu laptops such as the MacBook Pro which
require either a handler to switch DDC lines, or the discrete gpu
to proxy DDC/AUX communication: Both the handler and the discrete
gpu may initialize after the i915 driver, and consequently, an LVDS
connector may initially seem disconnected and the SSC therefore
is disabled by intel_init_pch_refclk, but on reprobe the connector
may turn out to be connected and the SSC must then be enabled.

Due to 92122789b2d6 however, the SSC is not enabled on reprobe since
it is assumed BIOS disabled it while in fact it was disabled by
intel_init_pch_refclk.

Also, because the SSC status is preserved so late, the preserved value
only ever gets used on resume but not on panel initialization:
intel_modeset_init calls intel_init_display which indirectly calls
intel_panel_use_ssc via multiple subroutines, *before* the BIOS value
overrides the VBT value in intel_modeset_gem_init (intel_panel_use_ssc
is the sole user of dev_priv->vbt.lvds_use_ssc).

Fix this by moving the code introduced by 92122789b2d6 from
intel_modeset_gem_init to intel_modeset_init before the invocation
of intel_setup_outputs and intel_init_display.

Add a DRM_DEBUG_KMS as suggested way back by Jani:
http://lists.freedesktop.org/archives/intel-gfx/2014-June/046666.html

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=88861
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=61115
Tested-by: Paul Hordiienko <pvt.gord@gmail.com>
    [MBP  6,2 2010  intel ILK + nvidia GT216  pre-retina]
Tested-by: William Brown <william@blackhats.net.au>
    [MBP  8,2 2011  intel SNB + amd turks     pre-retina]
Tested-by: Lukas Wunner <lukas@wunner.de>
    [MBP  9,1 2012  intel IVB + nvidia GK107  pre-retina]
Tested-by: Bruno Bierbaumer <bruno@bierbaumer.net>
    [MBP 11,3 2013  intel HSW + nvidia GK107  retina -- work in progress]
Fixes: 92122789b2d6 ("drm/i915: preserve SSC if previously set v3")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915/skl: Adding DDI_E power well domain
Xiong Zhang [Thu, 13 Aug 2015 10:00:12 +0000 (18:00 +0800)]
drm/i915/skl: Adding DDI_E power well domain

From B spec, DDI_E port belong to PowerWell 2, but
DDI_E share the powerwell_req/staus register bit with
DDI_A which belong to DDI_A_E_POWER_WELL.

In order to communicate with the connector on DDI-E, both
DDI_A_E_POWER_WELL and POWER_WELL_2 must be enabled.

Currently intel_dp_power_get(DDI_E) only enable
DDI_A_E_POWER_WELL, this patch will not only enable
DDI_a_E_POWER_WELL but also enable POWER_WELL_2.

This patch also fix the DDI-E hotplug function.

Signed-off-by: Xiong Zhang <xiong.y.zhang@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: eDP can be present on DDI-E
Rodrigo Vivi [Thu, 6 Aug 2015 07:51:39 +0000 (15:51 +0800)]
drm/i915: eDP can be present on DDI-E

Enable eDP on DDI-E.

Also let's remove duplicated definitions to avoid later confusion.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Xiong Zhang <xiong.y.zhang@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915/skl: Enable DDI-E
Rodrigo Vivi [Sat, 8 Aug 2015 00:35:21 +0000 (17:35 -0700)]
drm/i915/skl: Enable DDI-E

There are OEMs using DDI-E out there,
so let's enable it.

Unfortunately there is no detection bit for DDI-E
So we need to rely on VBT for that.

I also need to give credits to Xiong since before seing
his approach to check info->support_* I was creating an ugly
vbt->ddie_sfuse_strap in order to propagate the ddi presence info

v2: Rebased as last patch in the series. since all other patches
in this series are needed for anything working propperly on DDI-E.

Credits-to: "Zhang, Xiong Y" <xiong.y.zhang@intel.com>
Cc: "Zhang, Xiong Y" <xiong.y.zhang@intel.com>
Reviewed-by: Xiong Zhang <xiong.y.zhang@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: Enable HDMI on DDI-E
Xiong Zhang [Mon, 17 Aug 2015 08:04:04 +0000 (16:04 +0800)]
drm/i915: Enable HDMI on DDI-E

DDI-E doesn't have the correspondent GMBUS pin.

We rely on VBT to tell us which one it being used instead.

The DVI/HDMI on shared port couldn't exist.

This patch isn't tested without hardware wchich has HDMI
on DDI-E.

v2: fix trailing whitespace
v3: MISSING_CASE take place of BUG()

Signed-off-by: Xiong Zhang <xiong.y.zhang@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: apply the PCI_D0/D3 hibernation workaround everywhere on pre GEN6
Imre Deak [Tue, 30 Jun 2015 14:06:47 +0000 (17:06 +0300)]
drm/i915: apply the PCI_D0/D3 hibernation workaround everywhere on pre GEN6

commit da2bc1b9db3351addd293e5b82757efe1f77ed1d
Author: Imre Deak <imre.deak@intel.com>
Date:   Thu Oct 23 19:23:26 2014 +0300

    drm/i915: add poweroff_late handler

introduced a regression on old platforms during hibernation. A workaround was
added in

commit ab3be73fa7b43f4c3648ce29b5fd649ea54d3adb
Author: Imre Deak <imre.deak@intel.com>
Date:   Mon Mar 2 13:04:41 2015 +0200

    drm/i915: gen4: work around hang during hibernation

using an explicit blacklist for the GENs/BIOS vendors where the issue was
reported. Later there we had reports of the same failure on platforms not on
this list.

To my best knowledge the correct thing to do is still to put the device to PCI
D3 state during hibernation, see [1] and [2] for the reasons. This also aligns
with our future plans to unify more the runtime and system suspend/resume
paths. Since an exact blacklist seems to be impractical (multiple GENs and
BIOS vendors are affected) apply the workaround on everything pre GEN6.

[1] http://lists.freedesktop.org/archives/intel-gfx/2015-February/060710.html
[2] https://lkml.org/lkml/2015/6/22/274

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=95061
Reported-by: Ilya Tumaykin <itumaykin@gmail.com>
Reported-by: Dirk Griesbach <spamthis@freenet.de>
Reported-by: Pavel Machek <pavel@ucw.cz>
Reported-by: Mikko Rapeli <mikko.rapeli@iki.fi>
Tested-by: Mikko Rapeli <mikko.rapeli@iki.fi>
Reported-by: Paul Bolle <pebolle@tiscali.nl>
CC: stable@vger.kernel.org
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: Check DP link status on long hpd too
Ville Syrjälä [Thu, 20 Aug 2015 16:37:29 +0000 (19:37 +0300)]
drm/i915: Check DP link status on long hpd too

We are no longer checkling the DP link status on long hpd. We used to do
that from the .hot_plug() handler, but it was removed when MST got
introduced.

If there's no userspace we now fail to retrain the link if the sink
power is toggled (or cable yanked and replugged), meaning the user is
left staring at a blank screen. With the retraining put back that should
be fixed.

Also remove the leftover comment that referred to the old retraining
from .hot_plug().

Fixes a regression introduced in:
commit 0e32b39ceed665bfa4a77a4bc307b6652b991632
Author: Dave Airlie <airlied@redhat.com>
Date:   Fri May 2 14:02:48 2014 +1000

    drm/i915: add DP 1.2 MST support (v0.7)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89453
Tested-by: Palmer Dabbelt <palmer@dabbelt.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91407
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89461
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89594
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85641
Cc: Dave Airlie <airlied@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: set CDCLK if DPLL0 enabled during resuming from S3
Gary Wang [Fri, 28 Aug 2015 08:40:34 +0000 (16:40 +0800)]
drm/i915: set CDCLK if DPLL0 enabled during resuming from S3

Since BIOS RC 1.4 it would enable CDCLK PLL during BIOS S3 resume, then
driver needs to set CDCLK to avoid display corruption if DPLL0 enabled.

References: https://bugs.freedesktop.org/show_bug.cgi?id=91697
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Cooper Chiou <cooper.chiou@intel.com>
Reviewed-by: Wei Shun Chang <wei.shun.chang@intel.com>
Tested-by: Gary Wang <gary.c.wang@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Gavin Hindman <gavin.hindman@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Xiong Y Zhang <xiong.y.zhang@intel.com>
Signed-off-by: Gary Wang <gary.c.wang@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: Update DRIVER_DATE to 20150828
Daniel Vetter [Fri, 28 Aug 2015 16:05:26 +0000 (18:05 +0200)]
drm/i915: Update DRIVER_DATE to 20150828

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agoPartially revert "drm/i915: Use full atomic modeset."
Maarten Lankhorst [Wed, 26 Aug 2015 07:29:56 +0000 (09:29 +0200)]
Partially revert "drm/i915: Use full atomic modeset."

This partially reverts commit 74c090b1bdc57b1c9f1361908cca5a3d8a80fb08.

The DRIVER_ATOMIC cap cannot yet be exported because i915 lacks async
support.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: gen 9 can check for unclaimed registers too
Paulo Zanoni [Tue, 25 Aug 2015 22:03:41 +0000 (19:03 -0300)]
drm/i915: gen 9 can check for unclaimed registers too

Dear git bisect user,

Even though this is the patch that introduced the WARN() you're
bisecting, please notice that it's very likely that the problem you're
facing was already present before this commit. In other words: this
commit adds code to detect errors and give WARN()s about them, but the
errors were already there.

In order to continue your debug, please use the i915.mmio_debug
option, check the backtraces and try to discover which read or write
operation is causing the error message. Then check if this is
happening because the register does not exist or because its power
well is down when the operation is being done.

On my SKL machine, if I use i915.mmio_debug=999, this patch triggers
42 WARNs just by booting. I didn't investigate them yet. Normal users
are only going to get a single WARN due to the default i915.mmio_debug
setting.

Thank you for your comprehension,
Paulo

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Force CL2 off in CHV x1 PHY
Ville Syrjälä [Wed, 8 Jul 2015 20:45:58 +0000 (23:45 +0300)]
drm/i915: Force CL2 off in CHV x1 PHY

We can choose to leave the display PHY CL2 powerdown up to some hardware
signals, or we can force it. The BXT code forces the nonexistent CL2 in
the x1 PHY to power down. Follow suit on CHV. Maybe it can still save
some extra power by disabling some extra logic in CL1, or something.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Enable DPIO SUS clock gating on CHV
Ville Syrjälä [Wed, 8 Jul 2015 20:45:57 +0000 (23:45 +0300)]
drm/i915: Enable DPIO SUS clock gating on CHV

CHV has supports some form of automagic clock gating for the
DPIO SUS clock. We can simply enable the magic bits and the
hardware should take care of the rest.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Force common lane on for the PPS kick on CHV
Ville Syrjälä [Fri, 10 Jul 2015 07:56:24 +0000 (10:56 +0300)]
drm/i915: Force common lane on for the PPS kick on CHV

With DPIO powergating active the DPLL can't be accessed unless
something else is keeping the common lane in the channel on.
That means the PPS kick procedure could fail to enable the PLL.

Power up some data lanes to force the common lane to power up
so that the PLL can be enabled temporarily.

v2: Avoid gcc uninitilized variable warning

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Trick CL2 into life on CHV when using pipe B with port B
Ville Syrjälä [Wed, 8 Jul 2015 20:45:55 +0000 (23:45 +0300)]
drm/i915: Trick CL2 into life on CHV when using pipe B with port B

Normmally the common lane in a PHY channel gets powered up when some
of the data lanes get powered up. But when we're driving port B with
pipe B we don't want to enabled any of the data lanes, and just want
the DPLL in the common lane to be active.

To make that happens we have to temporarily enable some data lanes
after which we can access the DPLL registers in the common lane. Once
the pipe is up and running we can drop the power override on the data
lanes allowing them to shut down. From this point forward the common
lane will in fact stay powered on until the data lanes in the other
channel get powered down.

Ville's extended explanation from the review thread:

On Wed, Aug 19, 2015 at 07:47:41AM +0530, Deepak wrote:
> One Q, why only for port B? Port C is also in same common lane right?

Port B is in the first PHY channel which also houses CL1. CL1 always
powers up whenever any lanes in either PHY channel are powered up.
CL2 only powers up if lanes in the second channel (ie. the one with
port C) powers up.

So in this scenario (pipe B->port B) we want the DPLL from CL2, but
ideally we only want to power up the lanes for port B. Powering up
port B lanes will only power up CL1, but as we need CL2 instead we
need to, temporarily, power up some lanes in port C as well.

Crossing the streams the other way (pipe A->port C) is not a problem
since CL1 powers up whenever anything else powers up. So powering up
some port C lanes is enough on its own to make the CL1 DPLL
operational, even though CL1 and the lanes live in separate channels.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
[danvet: Amend commit message with extended explanation.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Implement PHY lane power gating for CHV
Ville Syrjälä [Wed, 8 Jul 2015 20:45:54 +0000 (23:45 +0300)]
drm/i915: Implement PHY lane power gating for CHV

Powergate the PHY lanes when they're not needed. For HDMI all four lanes
are needed always, but for DP we can enable only the needed lanes. To
power down the unused lanes we use some power down override bits in the
DISPLAY_PHY_CONTROL register. Without the overrides it appears that the
hardware always powers on all the lanes. When the port is disabled the
power down override is not needed and the lanes will shut off on their
own. That also means the override is critical to actually be able to
access the DPIO registers before the port is actually enabled.

Additionally the common lanes will power down when not needed. CL1
remains on as long as anything else is on, CL2 will shut down when
all the lanes in the same channel will shut down. There is one exception
for CL2 that will be dealt in a separate patch for clarity.

With potentially some lanes powered down, the DP code now has to check
the number of active lanes before accessing PCS/TX registers. All
registers in powered down blocks will reads as 0xffffffff, and soe we
would drown in warnings from vlv_dpio_read() if we allowed the code
to access all those registers.

Another important detail in the DP code is the "TX latency optimal"
setting. Normally the second TX lane acts as some kind of reset master,
with the other lanes as slaves. But when only a single lane is enabled,
that single lane obviously has to be the master.

A bit of extra care is needed to reconstruct the initial state of the
DISPLAY_PHY_CONTROL register since it can't be read safely. So instead
read the actual lane status from the DPLL/PHY_STATUS registers and
use that to determine which lanes ought to be powergated initially.

We also need to switch the PHY power modes to "deep PSR" to avoid
a hard system hang when powering down the single channel PHY.

Also sprinkle a few debug prints around so that we can monitor the
DISPLAY_PHY_STATUS changes without having to read it and risk
corrupting it.

v2: Add locking to chv_powergate_phy_lanes()
v3: Actually enable dynamic powerdown in the PHY and deal with the
    fallout

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Move DPLL ref/cri/VGA mode frobbing to the disp2d well enable
Ville Syrjälä [Mon, 29 Jun 2015 12:25:53 +0000 (15:25 +0300)]
drm/i915: Move DPLL ref/cri/VGA mode frobbing to the disp2d well enable

Bunch of stuff needs the DPLL ref/cri clocks on both VLV and CHV,
and having VGA mode enabled causes some problems for CHV. So let's just
pull the code to configure those bits into the disp2d well enable hook.
With the DPLL disable code also fixed to leave those bits alone we
should now have a consistent DPLL state all the time even if the DPLL
is disabled.

This also neatly removes some duplicated code between the VLV and
CHV codepaths.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Sivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Make some string arrays const
Ville Syrjälä [Fri, 21 Aug 2015 17:45:29 +0000 (20:45 +0300)]
drm/i915: Make some string arrays const

Most of our char* arrays are markes as const already, but a few slipped
through the cracks. Fix it.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Use ARRAY_SIZE() instead of hand rolling it
Ville Syrjälä [Fri, 21 Aug 2015 17:45:28 +0000 (20:45 +0300)]
drm/i915: Use ARRAY_SIZE() instead of hand rolling it

A couple of hand rolled ARRAY_SIZE()s caught my eye. Get rid of them.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Fix some gcc warnings
Ville Syrjälä [Fri, 21 Aug 2015 17:45:27 +0000 (20:45 +0300)]
drm/i915: Fix some gcc warnings

Simple one:
drivers/gpu/drm/i915/i915_debugfs.c:2449:57: warning: Using plain integer as NULL pointer

And something a bit more peculiar:
drivers/gpu/drm/i915/i915_debugfs.c:4953:18: warning: Variable length array is used.
drivers/gpu/drm/i915/i915_debugfs.c:4953:32: warning: Variable length array is used.

We pass a 'const int' as the array size which results in the warning,
dropping the const gets rid of the warning. Weird, but I think getting
rid of the warnings is better than holding on to the const.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/bxt: Use correct live status register for BXT platform
Jani Nikula [Thu, 20 Aug 2015 07:47:42 +0000 (10:47 +0300)]
drm/i915/bxt: Use correct live status register for BXT platform

BXT platform uses live status bits from 0x44440 register to obtain DP
status on hotplug. The existing g4x_digital_port_connected() uses a
different register and hence misses DP hotplug events on BXT
platform. This patch fixes it by using the appropriate register(0x44440)
and live status bits(3:5).

Based on a patch by Durgadoss R <durgadoss.r@intel.com>, from whom the
commit message is shamelessly copy pasted.

Reported-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: split g4x_digital_port_connected to g4x and vlv variants
Jani Nikula [Thu, 20 Aug 2015 07:47:41 +0000 (10:47 +0300)]
drm/i915: split g4x_digital_port_connected to g4x and vlv variants

Choose the right function at the intel_digital_port_connected level.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: split ibx_digital_port_connected to ibx and cpt variants
Jani Nikula [Thu, 20 Aug 2015 07:47:40 +0000 (10:47 +0300)]
drm/i915: split ibx_digital_port_connected to ibx and cpt variants

Choose the right function at the intel_digital_port_connected level.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: add common intel_digital_port_connected function
Jani Nikula [Thu, 20 Aug 2015 07:47:39 +0000 (10:47 +0300)]
drm/i915: add common intel_digital_port_connected function

Add a common intel_digital_port_connected() that splits out to functions
for different platforms. No functional changes.

v2: make the function return a boolean

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: add MISSING_CASE annotation to ibx_digital_port_connected
Jani Nikula [Thu, 20 Aug 2015 07:47:38 +0000 (10:47 +0300)]
drm/i915: add MISSING_CASE annotation to ibx_digital_port_connected

With the case added for eDP on port A (always connected from this
function's point of view), we should not be hitting any of the default
cases in ibx_digital_port_connected, so add MISSING_CASE annotation.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: make g4x_digital_port_connected return boolean status
Jani Nikula [Thu, 20 Aug 2015 07:47:37 +0000 (10:47 +0300)]
drm/i915: make g4x_digital_port_connected return boolean status

We should not be hitting any of the default cases in
g4x_digital_port_connected, so add MISSING_CASE annotation and return
boolean status. The current behaviour is just cargo culting from the
days of yonder when the display port support was added to i915.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: move ibx_digital_port_connected to intel_dp.c
Jani Nikula [Thu, 20 Aug 2015 07:47:36 +0000 (10:47 +0300)]
drm/i915: move ibx_digital_port_connected to intel_dp.c

The function can be made static there. No functional changes.

Reviewed-by: Durgadoss R <durgadoss.r@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: DVO pixel clock check
Mika Kahola [Tue, 18 Aug 2015 11:37:02 +0000 (14:37 +0300)]
drm/i915: DVO pixel clock check

It is possible the we request to have a mode that has
higher pixel clock than our HW can support. This patch
checks if requested pixel clock is lower than the one
supported by the HW. The requested mode is discarded
if we cannot support the requested pixel clock.

This patch applies to DVO.

V2:
- removed computation for max pixel clock

V3:
- cleanup by removing unnecessary lines

V4:
- clock check against max dotclock moved inside 'if (fixed_mode)'

V5:
- dot clock check against fixed_mode clock when available

Signed-off-by: Mika Kahola <mika.kahola@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: DSI pixel clock check
Mika Kahola [Tue, 18 Aug 2015 11:37:01 +0000 (14:37 +0300)]
drm/i915: DSI pixel clock check

It is possible the we request to have a mode that has
higher pixel clock than our HW can support. This patch
checks if requested pixel clock is lower than the one
supported by the HW. The requested mode is discarded
if we cannot support the requested pixel clock.

This patch applies to DSI.

V2:
- removed computation for max pixel clock

V3:
- cleanup by removing unnecessary lines

V4:
- max_pixclk variable renamed as max_dotclk
- moved dot clock checking inside 'if (fixed_mode)'

V5:
- dot clock checked against fixed_mode clock

Signed-off-by: Mika Kahola <mika.kahola@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: LVDS pixel clock check
Mika Kahola [Tue, 18 Aug 2015 11:37:00 +0000 (14:37 +0300)]
drm/i915: LVDS pixel clock check

It is possible the we request to have a mode that has
higher pixel clock than our HW can support. This patch
checks if requested pixel clock is lower than the one
supported by the HW. The requested mode is discarded
if we cannot support the requested pixel clock.

This patch applies to LVDS.

V2:
- removed computation for max pixel clock

V3:
- cleanup by removing unnecessary lines

V4:
- moved supported dotclock check from mode_valid() to intel_lvds_init()

V5:
- dotclock check moved back to mode_valid() function
- dotclock check for fixed mode

Signed-off-by: Mika Kahola <mika.kahola@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Store max dotclock
Mika Kahola [Tue, 18 Aug 2015 11:36:59 +0000 (14:36 +0300)]
drm/i915: Store max dotclock

Store max dotclock into dev_priv structure so we are able
to filter out the modes that are not supported by our
platforms.

V2:
- limit the max dot clock frequency to max CD clock frequency
  for the gen9 and above
- limit the max dot clock frequency to 90% of the max CD clock
  frequency for the older gens
- for Cherryview the max dot clock frequency is limited to 95%
  of the max CD clock frequency
- for gen2 and gen3 the max dot clock limit is set to 90% of the
  2X max CD clock frequency

V3:
- max_dotclk variable renamed as max_dotclk_freq in i915_drv.h
- in intel_compute_max_dotclk() the rounding method changed from
  round up to round down when computing max dotclock

V4:
- Haswell and Broadwell supports now dot clocks up to max CD clock
  frequency

Signed-off-by: Mika Kahola <mika.kahola@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add vlv_dport_to_phy()
Ville Syrjälä [Wed, 8 Jul 2015 20:45:53 +0000 (23:45 +0300)]
drm/i915: Add vlv_dport_to_phy()

Add vlv_dport_to_phy() and fix up the return values of
vlv_dport_to_channel() and vlv_pipe_to_channel() to use
the appropriate enums.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Move VLV/CHV prepare_pll later
Ville Syrjälä [Wed, 8 Jul 2015 20:45:52 +0000 (23:45 +0300)]
drm/i915: Move VLV/CHV prepare_pll later

With DPIO powergating active on CHV, we can't even access the DPIO PLL
registers until the lane power state overrides have been enabled. That
will happen from the encoder .pre_pll_enable() hook, so move
chv_prepare_pll() to happen after that point, which puts it just before
chv_enable_pll() actually.

Do the same for VLV to avoid accumulating weird differences between the
platforms. Both platforms seem happy with the new arrangement.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add locking around chv_phy_control_init()
Ville Syrjälä [Wed, 8 Jul 2015 20:45:51 +0000 (23:45 +0300)]
drm/i915: Add locking around chv_phy_control_init()

dev_priv->chv_phy_control is protected by the power_domains->lock
elsewhere, so also grab it when initializing chv_phy_control.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Move DPIO port init earlier
Ville Syrjälä [Wed, 8 Jul 2015 20:45:50 +0000 (23:45 +0300)]
drm/i915: Move DPIO port init earlier

To implement DPIO lane power gating on CHV we're going to need to access
DPIO registers from the cmn power well enable hook. That gets called
rather early, so we need to move the DPIO port IOSF sideband port
assignment earlier as well.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Add encoder->post_pll_disable() hooks and move CHV clock buffer disables...
Ville Syrjälä [Wed, 8 Jul 2015 20:45:49 +0000 (23:45 +0300)]
drm/i915: Add encoder->post_pll_disable() hooks and move CHV clock buffer disables there

Move the CHV clock buffer disable from chv_disable_pll() to the new
encoder .post_pll_disable() hook. This is more symmetric since the
clock buffer enable happens from the .pre_pll_enable() hook.

We'll have more use for the new hook soon.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Always program unique transition scale for CHV
Ville Syrjälä [Wed, 8 Jul 2015 20:45:48 +0000 (23:45 +0300)]
drm/i915: Always program unique transition scale for CHV

The docs give you the impression that the unique transition scale
value shouldn't matter when unique transition scale is enabled. But
as Imre found on BXT (and I verfied also on BSW) the value does
matter. So from now on just program the same value 0x9a always.

Cc: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Always program m2 fractional value on CHV
Ville Syrjälä [Wed, 8 Jul 2015 20:45:47 +0000 (23:45 +0300)]
drm/i915: Always program m2 fractional value on CHV

When fractional m2 divider isn't used on CHV the fractional part
is ignore by the hardware. Despite that, program the fractional
value (0 in this case) to the hardware register just to keep
things a bit more consistent. Might at least make register dumps
a bit less confusing when there isn't some stale fractional part
hanging around.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: fix driver's versions of WARN_ON & WARN_ON_ONCE
Dave Gordon [Mon, 17 Aug 2015 16:30:52 +0000 (17:30 +0100)]
drm/i915: fix driver's versions of WARN_ON & WARN_ON_ONCE

The current versions of these two macros don't work correctly if the
argument expression happens to contain a modulo operator (%) -- when
stringified, it gets interpreted as a printf formatting character!
With a specifically crafted parameter, this could probably cause a
kernel OOPS; consider WARN_ON(p%s) or WARN_ON(f %*pEp).

Instead, we should use an explicit "%s" format, with the stringified
expression as the coresponding literal-string argument.

Signed-off-by: Dave Gordon <david.s.gordon@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Put back lane_count into intel_dp and add link_rate too
Ville Syrjälä [Mon, 17 Aug 2015 15:05:12 +0000 (18:05 +0300)]
drm/i915: Put back lane_count into intel_dp and add link_rate too

With MST there won't be a crtc assigned to the main link encoder, so
trying to dig up the pipe_config from there is a recipe for an oops.

Instead store the parameters (lane_count and link_rate) in the encoder,
and use those values during link training etc. Since those parameters
are now assigned only when the link is actually enabled,
.compute_config() won't clobber them as it did before.

Hardware state readout is still bonkers though as we don't transfer the
link parameters from pipe_config intel_dp. We should do that during
encoder sanitation. But since we don't even do a proper job of reading
out the main link encoder state for MST there's littel point in
worrying about this now.

Fixes a regression with MST caused by:
 commit 90a6b7b052b1aa17fbb98b049e9c8b7f729c35a7
 Author: Ville Syrjälä <ville.syrjala@linux.intel.com>
 Date:   Mon Jul 6 16:39:15 2015 +0300

    drm/i915: Move intel_dp->lane_count into pipe_config

v2: Different apporoach that should keep intel_dp_check_mst_status()
    somewhat less oopsy

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reported-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Tested-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/bxt: don't allow cached GEM mappings on A stepping
Imre Deak [Fri, 14 Aug 2015 15:43:30 +0000 (18:43 +0300)]
drm/i915/bxt: don't allow cached GEM mappings on A stepping

Due to a coherency issue on BXT A steppings we can't guarantee a
coherent view of cached (CPU snooped) GPU mappings, so fail such
requests. User space is supposed to fall back to uncached mappings in
this case.

v2:
- limit the WA to A steppings, on later stepping this HW issue is fixed
v3:
- return error instead of trying to work around the issue in kernel,
  since that could confuse user space (Chris)

Testcast: igt/gem_store_dword_batches_loop/cached-mapping
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/bxt: work around HW coherency issue when accessing GPU seqno
Imre Deak [Fri, 14 Aug 2015 15:35:27 +0000 (18:35 +0300)]
drm/i915/bxt: work around HW coherency issue when accessing GPU seqno

By running igt/store_dword_loop_render on BXT we can hit a coherency
problem where the seqno written at GPU command completion time is not
seen by the CPU. This results in __i915_wait_request seeing the stale
seqno and not completing the request (not considering the lost
interrupt/GPU reset mechanism). I also verified that this isn't a case
of a lost interrupt, or that the command didn't complete somehow: when
the coherency issue occured I read the seqno via an uncached GTT mapping
too. While the cached version of the seqno still showed the stale value
the one read via the uncached mapping was the correct one.

Work around this issue by clflushing the corresponding CPU cacheline
following any store of the seqno and preceding any reading of it. When
reading it do this only when the caller expects a coherent view.

v2:
- fix using the proper logical && instead of a bitwise & (Jani, Mika)
- limit the workaround to A stepping, on later steppings this HW issue
  is fixed
v3:
- use a separate get_seqno/set_seqno vfunc (Chris)

Testcase: igt/store_dword_loop_render
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915/skl: enable DDI-E hotplug
Xiong Zhang [Mon, 17 Aug 2015 07:55:50 +0000 (15:55 +0800)]
drm/i915/skl: enable DDI-E hotplug

v2: fix one error found by checkpath.pl
v3: Add one ignored break for switch-case. DDI-E hotplug
    function doesn't work after updating drm-intel tree,
    I checked the code and found this missing which isn't
    the root cause for broke DDI-E hp.  The broken
    DDI-E hp function is fixed by "Adding DDI_E power
    well domain".

Signed-off-by: Xiong Zhang <xiong.y.zhang@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Tested-by: Timo Aaltonen <timo.aaltonen@canonical.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: Fix build warning on 32-bit
Thierry Reding [Fri, 14 Aug 2015 10:35:23 +0000 (12:35 +0200)]
drm/i915: Fix build warning on 32-bit

The gtt.stolen_size field is of type size_t, and so should be printed
using %zu to avoid build warnings on either 32-bit and 64-bit builds.

Signed-off-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915/skl: Update DDI buffer translation programming.
Rodrigo Vivi [Mon, 24 Aug 2015 23:48:44 +0000 (16:48 -0700)]
drm/i915/skl: Update DDI buffer translation programming.

SKL-Y can now use the same programming for all VccIO values after an
adjustment to I_boost.  SKL-U DP table adjustments.

1. Remove SKL Y 0.95V from "SKL H and S" columns in all tables. The
   other SKL Y column removes the "0.85V VccIO" so it now applies to all
   voltages.

2. DP table changes SKL U 400mV+0db dword 0 value from 2016h to 201Bh.

3. DP table changes SKL U 600mv+0db dword 0 value from 2016h to 201Bh.

4. DP table increases I_boost to level 3 for SKL Y 400mv+9.5db.

v2: Fix compilation warnings as pointed by Paulo.

Reference: Graphics Spec Change r97962
Cc: Arthur Runyan <arthur.j.runyan@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
[Jani: reformatted commit message for shorter lines.]
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
9 years agodrm/i915: Also call frontbuffer flip when disabling planes.
Rodrigo Vivi [Mon, 24 Aug 2015 23:38:23 +0000 (16:38 -0700)]
drm/i915: Also call frontbuffer flip when disabling planes.

We also need to call the frontbuffer flip to trigger proper
invalidations when disabling planes. Otherwise we will miss
screen updates when disabling sprites or cursor.

On core platforms where HW tracking also works, this issue
is totally masked because HW tracking triggers PSR exit
however on VLV/CHV that has only SW tracking we miss screen
updates when disabling planes.

It was caught with kms_psr_sink_crc sprite_plane_onoff
and cursor_plane_onoff subtests running on VLV/CHV.

This is probably a regression since I can also get this
with the manual test case, but with so many changes on atomic
modeset I couldn't track exactly when this was introduced.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
9 years agodrm/i915: Change SRM, LRM instructions to use correct length
Arun Siluvery [Tue, 4 Aug 2015 15:22:20 +0000 (16:22 +0100)]
drm/i915: Change SRM, LRM instructions to use correct length

MI_STORE_REGISTER_MEM, MI_LOAD_REGISTER_MEM instructions are not really
variable length instructions unlike MI_LOAD_REGISTER_IMM where it expects
(reg, addr) pairs so use fixed length for these instructions.

v2: rebase

Cc: Dave Gordon <david.s.gordon@intel.com>
Signed-off-by: Arun Siluvery <arun.siluvery@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
[danvet: Appease checkpatch as Mika spotted in i915_reg.h - it seems
terminally unhappy about i915_cmd_parser.c so that would be a separate
patch.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>