Ville Syrjälä [Mon, 26 Sep 2016 08:30:46 +0000 (11:30 +0300)]
drm/i915: Allow PCH DPLL sharing regardless of DPLL_SDVO_HIGH_SPEED
DPLL_SDVO_HIGH_SPEED must be set for SDVO/HDMI/DP, but nowhere is it
forbidden to set it for LVDS/CRT as well. So let's also set it on
CRT to make it possible to share the DPLL between HDMI and CRT.
What that bit apparently does is enable the x5 clock to the port,
which then pumps out the bits on both edges of the clock. The DAC
doesn't need that clock since it's not pumping out bits, but I don't
think it hurts to have the DPLL output that clock anyway.
This is fairly important on IVB since it has only two DPLLs with three
pipes. So trying to drive three or more PCH ports with three pipes
is only possible when at least one of the DPLLs gets shared between
two of the pipes.
SNB doesn't really need to do this since it has only two pipes. It could
be done to avoid enabling the second DPLL at all in certain cases, but
I'm not sure that's such a huge win. So let's not do it for SNB, at
least for now. On ILK it never makes sense as the DPLLs can't be shared.
v2: Just always enable the high speed clock to keep things simple (Daniel)
Beef up the commit message a bit (Daniel)
Cc: Nick Yamane <nick.diego@gmail.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: stable@vger.kernel.org Tested-by: Nick Yamane <nick.diego@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97204 Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1474878646-17711-1-git-send-email-ville.syrjala@linux.intel.com Reviewed-by: Ander Conselvan de Oliveira <conselvan2@gmail.com>
Imre Deak [Mon, 26 Sep 2016 14:54:31 +0000 (17:54 +0300)]
drm/i915/bxt: Fix HDMI DPLL configuration
a277ca7dc01d should've been a no-functional-change commit, but it
removed the initialization of the dpll_hw_state for HDMI outputs,
resulting in state mismatches and a failed modeset with blank
screen. Fix this by reinstating the dpll_hw_state initialization.
v2:
- Make bxt_ddi_hdmi_set_dpll_hw_state() static.
Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> Cc: Durgadoss R <durgadoss.r@intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Fixes: a277ca7dc01d ("drm/i915: Split bxt_ddi_pll_select()") Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Ander Conselvan de Oliveira <conselvan2@gmail.com> Link: http://patchwork.freedesktop.org/patch/msgid/1474901671-22719-1-git-send-email-imre.deak@intel.com
Paulo Zanoni [Mon, 26 Sep 2016 12:07:52 +0000 (15:07 +0300)]
drm/i915/skl: tell the user about pre-production hardware
We just removed the implementation for all the pre-production
workarounds, so now tell the user that we expect his machine to not
work properly. Also convert this to DRM_ERROR so we can more easily
spot these problems in bug reports and CI/QA runs.
drm/i915: don't report compression when fbc is disabled
When i915_fbc_status is read while fbc is disabled,
it reports compressing to be true, which is confusing.
Report compressing only when fbc is enabled.
Paulo Zanoni [Thu, 22 Sep 2016 21:00:34 +0000 (18:00 -0300)]
drm/i915/gen9: implement missing case for SKL watermarks calculation
This should affect linear and X tiled planes on really small htotal
cases. It doesn't seem to be a very feasible case, but let's implement
it since it's on the specification and it's better to have it and
never need than not have it and realize we needed it.
Paulo Zanoni [Thu, 22 Sep 2016 21:00:33 +0000 (18:00 -0300)]
drm/i915/gen9: fix the watermark res_blocks value
We forgot the "res_blocks += y_tile_minimum" that's described on step
V of our documentation.
Again, this should only affect the Y tiling cases.
It looks like the relevant code was introduced in 0fda65680e92, but
there's always the possibility that it matched our specification when
it was introduced, and then the specification changed while the code
stayed the same. So we can't really say this was a regression, but
let's try to add a "Fixes" tag anyway to help backporting.
v2: Try to add a "Fixes" tag (Maarten).
Fixes: 0fda65680e92 ("drm/i915/skl: Update watermarks for Y tiling") Cc: stable@vger.kernel.org Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Lyude <cpaul@redhat.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1474578035-424-8-git-send-email-paulo.r.zanoni@intel.com
Paulo Zanoni [Thu, 22 Sep 2016 21:00:32 +0000 (18:00 -0300)]
drm/i915/gen9: fix plane_blocks_per_line on watermarks calculations
The confusing thing is that plane_blocks_per_line is listed as part of
the method 2 calculation but is also used for other things. We
calculated it in two different places and different ways: one inside
skl_wm_method2() and the other inside skl_compute_plane_wm(). The
skl_wm_method2() implementation is the one that matches the
specification.
With this patch we fix the skl_compute_plane_wm() calculation and just
pass it as a parameter to skl_wm_method2(). We also take care to not
modify the value of plane_bytes_per_line since we're going to rely on
it having a correct value in later patches.
This should affect the watermarks for Linear and Y-tiled.
From my analysis, it looks like the two plane_blocks_per_line
variables got out of sync on 0fda65680e92, but we can't really say
that commit was a regression, it looks like just an incomplete fix.
There's always the possibility that 0fda65680e92 matched our
specification at that time, and then later the specification changed.
v2: Try to add a "Fixes" tag (Maarten).
Fixes: 0fda65680e92 ("drm/i915/skl: Update watermarks for Y tiling") Cc: stable@vger.kernel.org Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Lyude <cpaul@redhat.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1474578035-424-7-git-send-email-paulo.r.zanoni@intel.com
Paulo Zanoni [Thu, 22 Sep 2016 21:00:31 +0000 (18:00 -0300)]
drm/i915/gen9: minimum scanlines for Y tile is not always 4
During watermarks calculations, this value is used in 3 different
places. Only one of them was not using a hardcoded 4. Move the code up
so everybody can benefit from the actual value.
This should only help on situations with Y tiling + 90/270 rotation +
1 or 2 bpp or NV12.
Paulo Zanoni [Thu, 22 Sep 2016 21:00:30 +0000 (18:00 -0300)]
drm/i915/gen9: fix the WaWmMemoryReadLatency implementation
Bspec says:
"The mailbox response data may not account for memory read latency.
If the mailbox response data for level 0 is 0us, add 2 microseconds
to the result for each valid level."
This means we should only do the +2 in case wm[0] == 0, not always.
So split the sanitizing implementation from the WA implementation and
fix the WA implementation.
Paulo Zanoni [Thu, 22 Sep 2016 21:00:28 +0000 (18:00 -0300)]
drm/i915: introduce intel_has_sagv()
And use it to move knowledge about the SAGV-supporting platforms from
the callers to the SAGV code.
We'll add more platforms to intel_has_sagv(), so IMHO it makes more
sense to move all this to a single function instead of patching all
the callers every time we add SAGV support to a new platform.
v2: Move I915_SAGV_NOT_CONTROLLED to the new function (Lyude).
Paulo Zanoni [Thu, 22 Sep 2016 21:00:27 +0000 (18:00 -0300)]
drm/i915: SAGV is not SKL-only, so rename a few things
The plan is to introduce intel_has_sagv() and then use it to discover
which platforms actually support it.
I thought about keeping the functions with their current skl names,
but found two problems: (i) skl_has_sagv() would become a very
confusing name, and (ii) intel_atomic_commit_tail() doesn't seem to be
calling any functions whose name start with a platform name, so the
"intel_" naming scheme seems make more sense than the "firstplatorm_"
naming scheme here.
Paulo Zanoni [Fri, 19 Aug 2016 22:03:23 +0000 (19:03 -0300)]
drm/i915: don't forget to set intel_crtc->dspaddr_offset on SKL+
We never remembered to set it (so it was zero), but this was not a
problem in the past due to the way handled the hardware registers.
Unfortunately we changed how we set the hardware and forgot to set
intel_crtc->dspaddr_offset.
This started to reflect on a few kms_frontbuffer_tracking subtests
that relied on page flips with CRTCs that don't point to the x:0,y:0
coordinates of the frontbuffer. After the page flip the CRTC was
showing the x:0,y:0 coordinate of the frontbuffer instead of
x:500,y:500. This problem is present even if we don't enable FBC or
PSR.
While trying to bisect it I realized that the first bad commit
actually just gives me a black screen for the mentioned tests instead
of showing the wrong x:0,y:0 offsets. A few commits later the black
screen problem goes away and we get to the point where the code is
today, but I'll consider the black screen as the first bad commit
since it's the point where the IGT subtests start to fail.
Paulo Zanoni [Tue, 13 Sep 2016 13:38:57 +0000 (10:38 -0300)]
drm/i915/fbc: disable FBC on FIFO underruns
Ever since I started working on FBC I was already aware that FBC can
really amplify the FIFO underrun symptoms. On systems where FIFO
underruns were harmless error messages, enabling FBC would cause the
underruns to give black screens.
We recently tried to enable FBC on Haswell and got reports of a system
that would hang after some hours of uptime, and the first bad commit
was the one that enabled FBC. We also observed that this system had
FIFO underrun error messages on its dmesg. Although we don't have any
evidence that fixing the underruns would solve the bug and make FBC
work properly on this machine, IMHO it's better if we minimize the
amount of possible problems by just giving up FBC whenever we detect
an underrun.
v2: New version, different implementation and commit message.
v3: Clarify the fact that we run from an IRQ handler (Chris).
v4: Also add the underrun_detected check at can_choose() to avoid
misleading dmesg messages (DK).
v5: Fix Engrish, use READ_ONCE on the unlocked read (Chris).
Cc: Stefan Richter <stefanr@s5r6.in-berlin.de> Cc: Lyude <cpaul@redhat.com> Cc: stevenhoneyman@gmail.com <stevenhoneyman@gmail.com> Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1473773937-19758-1-git-send-email-paulo.r.zanoni@intel.com
DP MST provides the capability to send multiple video and audio streams
through a single port. This requires the API's between i915 and audio
drivers to distinguish between multiple audio capable displays that can be
connected to a port. Currently only the port identity is shared in the
APIs. This patch adds support for MST with an additional parameter
'int pipe'. The existing parameter 'port' does not change it's meaning.
pipe =
MST : display pipe that the stream originates from
Non-MST : -1
Affected APIs:
struct i915_audio_component_ops
- int (*sync_audio_rate)(struct device *, int port, int rate);
+ int (*sync_audio_rate)(struct device *, int port, int pipe,
+ int rate);
- int (*get_eld)(struct device *, int port, bool *enabled,
- unsigned char *buf, int max_bytes);
+ int (*get_eld)(struct device *, int port, int pipe,
+ bool *enabled, unsigned char *buf, int max_bytes);
struct i915_audio_component_audio_ops
- void (*pin_eld_notify)(void *audio_ptr, int port);
+ void (*pin_eld_notify)(void *audio_ptr, int port, int pipe);
This patch makes dummy changes in the audio drivers (thanks Libin) for
build to succeed. The audio side drivers will send the right 'pipe' values
for MST in patches that will follow.
v2:
Renamed the new API parameter from 'dev_id' to 'pipe'. (Jim, Ville)
Included Asoc driver API compatibility changes from Jeeja.
Added WARN_ON() for invalid pipe in get_saved_encoder(). (Takashi)
Added comment for av_enc_map[] definition. (Takashi)
v3:
Fixed logic error introduced while renaming 'dev_id' as 'pipe' (Ville)
Renamed get_saved_encoder() to get_saved_enc() to reduce line length
v4:
Rebased.
Parameter check for pipe < -1 values in get_saved_enc() (Ville)
Switched to for_each_pipe() in get_saved_enc() (Ville)
Renamed 'pipe' to 'dev_id' in audio side code (Takashi)
v5:
Included a comment for the dev_id arg. (Libin)
With DP MST, a digital_port can carry more than one audio stream. Hence,
more than one audio_connector needs to be attached to intel_digital_port in
such cases. However, each stream is associated with an unique encoder. So,
instead of creating an array of audio_connectors per port, move
audio_connector from struct intel_digital_port to struct intel_encoder.
This also simplifies access to the right audio_connector from codec
functions in intel_audio.c that receive intel_encoder.
v2: Removed locals that are not needed anymore.
v3: No code change except for minor change in context.
Storing the port enum in intel_encoder makes it convenient to know the
port attached to an encoder. Moving the port information up from
intel_digital_port to intel_encoder avoids unecessary intel_digital_port
access and handles MST encoders cleanly without requiring conditional
checks for them (thanks danvet).
v2:
Renamed the port enum member from 'attached_port' to 'port' (danvet)
Fixed missing initialization of port in intel_sdvo.c (danvet)
v3:
Fixed missing initialization of port in intel_crt.c (Ville)
Changing the return type from 'char' to 'enum port' in
intel_dvo_port_name() makes it easier to later move the port information to
intel_encoder. In addition, the port type conforms to what we have
elsewhere.
Removing the last conditional that handles invalid port because dvo_reg is
intialized to valid values for all DVO devices at definition.
Chris Wilson [Wed, 21 Sep 2016 13:51:08 +0000 (14:51 +0100)]
drm/i915/execlists: Reset RING registers upon resume
There is a disparity in the context image saved to disk and our own
bookkeeping - that is we presume the RING_HEAD and RING_TAIL match our
stored ce->ring->tail value. However, as we emit WA_TAIL_DWORDS into the
ring but may not tell the GPU about them, the GPU may be lagging behind
our bookkeeping. Upon hibernation we do not save stolen pages, presuming
that their contents are volatile. This means that although we start
writing into the ring at tail, the GPU starts executing from its HEAD
and there may be some garbage in between and so the GPU promptly hangs
upon resume.
Chris Wilson [Wed, 21 Sep 2016 13:51:07 +0000 (14:51 +0100)]
drm/i915: Only shrink the unbound objects during freeze
At the point of creating the hibernation image, the runtime power manage
core is disabled - and using the rpm functions triggers a warn.
i915_gem_shrink_all() tries to unbind objects, which requires device
access and so tries to how an rpm reference triggering a warning:
Chris Wilson [Wed, 21 Sep 2016 13:51:06 +0000 (14:51 +0100)]
drm/i915: Restore current RPS state after reset
Following commit 821ed7df6e2a ("drm/i915: Update reset path to fix
incomplete requests") we no longer mark the context as lost on reset as
we keep the requests (and contexts) alive. However, RPS remains reset
and we need to restore the current state to match the in-flight
requests.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97824 Fixes: 821ed7df6e2a ("drm/i915: Update reset path to fix incomplete requests") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Arun Siluvery <arun.siluvery@linux.intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20160921135108.29574-1-chris@chris-wilson.co.uk
Imre Deak [Tue, 20 Sep 2016 11:58:19 +0000 (14:58 +0300)]
drm/i915: Queue page flip work via a low latency, unbound workqueue
While user space has control over the scheduling priority of its page
flipping thread, the corresponding work the driver schedules for MMIO
flips always runs from the generic system workqueue which has some
scheduling overhead due it being CPU bound. This would hinder an
application that wants more stringent guarantees over flip timing (to
avoid missing a flip at the next frame count).
Fix this by scheduling the work from the unbound system workqueue
which provides for minimal scheduling latency.
v2:
- Use an unbound workqueue instead of a high-prio one. (Tvrtko, Chris)
v3:
- Use the system unbound wq instead of a dedicated one. (Maarten)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97775
Testcase: igt/kms_cursor_legacy CC: Chris Wilson <chris@chris-wilson.co.uk> CC: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> CC: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> (v1) Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1474372699-22841-1-git-send-email-imre.deak@intel.com
Ben Widawsky [Tue, 20 Sep 2016 13:54:33 +0000 (16:54 +0300)]
drm/i915: Try to print INSTDONE bits for all slice/subslice
v2: (Imre)
- Access only subslices that are known to exist.
- Reset explicitly the MCR selector to slice/sub-slice ID 0 after the
readout.
- Use the subslice INSTDONE bits for the hangcheck/subunits-stuck
detection too.
- Take the uncore lock for the MCR-select/subslice-readout sequence.
Ben Widawsky [Tue, 20 Sep 2016 13:54:32 +0000 (16:54 +0300)]
drm/i915: Cleanup instdone collection
Consolidate the instdone logic so we can get a bit fancier. This patch also
removes the duplicated print of INSTDONE[0].
v2: (Imre)
- Rebased on top of hangcheck INSTDONE changes.
- Move all INSTDONE registers into a single struct, store it within the
engine error struct during error capturing.
Imre Deak [Wed, 14 Sep 2016 10:04:13 +0000 (13:04 +0300)]
drm/i915: Unlock PPS registers after GPU reset
Reapply the PPS register unlock workaround after GPU reset on platforms
where the reset clobbers the display HW state. This at least gets rid of
the related WARN during LVDS encoder enabling on PNV.
Shawn Lee [Mon, 19 Sep 2016 10:35:26 +0000 (13:35 +0300)]
drm/i915/backlight: setup backlight pwm alternate increment on backlight enable
Backlight enable is supposed to do a full setup of the backlight. We
were missing the PWM alternate increment bit in the south chicken
registers on lpt+ pch. This potentially caused a PWM frequency change
when the chicken register value was lost e.g. on suspend.
v2 by Jani, rebase on the patch caching alt increment
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97486
References: https://bugs.freedesktop.org/show_bug.cgi?id=67454 Cc: Cooper Chiou <cooper.chiou@intel.com> Cc: Wei Shun Chen <wei.shun.chang@intel.com> Cc: Gary C Wang <gary.c.wang@intel.com> Cc: stable@vger.kernel.org # v4.4+ 32b421e79e6b drm/i915/backlight: setup and cache... Cc: stable@vger.kernel.org # v4.4+ Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Shawn Lee <shawn.c.lee@intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/8265f5935bd31c039ddfc82819d26c2ca1ae9cba.1474281249.git.jani.nikula@intel.com
Jani Nikula [Mon, 19 Sep 2016 12:02:28 +0000 (15:02 +0300)]
drm/i915/dsi: run backlight on/off sequences in panel enable/disable hooks
Based on the documentation alone, it's anyone's guess when exactly we
should be running these sequences. Add them where it feels logical. The
drm panel hooks don't currently offer us more granularity anyway.
Jani Nikula [Mon, 19 Sep 2016 12:02:27 +0000 (15:02 +0300)]
drm/i915/dsi: update reset and power sequences in panel prepare/unprepare hooks
Based on the documentation alone, it's anyone's guess when exactly we
should be running these sequences. Add power on/off sequences where they
feel logical and update assert/deassert reset. The drm panel hooks don't
currently offer us more granularity anyway.
Mika Kahola [Fri, 16 Sep 2016 10:39:15 +0000 (13:39 +0300)]
drm: Fix DisplayPort branch device ID kernel-doc
Fix missing parameter description for DisplayPort branch device ID.
This fixes warning of "No description found for parameter 'id[6]'" when
creating documentation by 'make htmldocs'.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
References: https://lists.freedesktop.org/archives/intel-gfx/2016-September/106645.html Fixes: 266d783baaf5 ("drm: Read DP branch device id") Signed-off-by: Mika Kahola <mika.kahola@intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1474022355-29990-1-git-send-email-mika.kahola@intel.com
Mika Kahola [Fri, 9 Sep 2016 11:10:57 +0000 (14:10 +0300)]
drm: Add DP branch device info on debugfs
Read DisplayPort branch device info from through debugfs
interface.
v2: use drm_dp_helper routines to collect data
v3: cleanup to match the drm_dp_helper.c patches introduced
earlier in this series
v4: move DP branch device info to function 'intel_dp_branch_device_info()'
v5: initial step to move debugging info from intel_dp. to drm_dp_helper.c (Daniel)
v6: read hw and sw revision without using specific drm_dp_helper routines
v7: indentation fixes (Jim Bride)
Mika Kahola [Fri, 9 Sep 2016 11:10:56 +0000 (14:10 +0300)]
drm/i915: Update bits per component for display info
DisplayPort branch device may define max supported bits per
component. Update display info based on this value if bpc
is defined.
v2: cleanup to match the drm_dp_helper.c patches introduced
earlier in this series
v3: Fill bpc for connector's display info in separate
drm_dp_helper function (Daniel)
v4: remove updating bpc for display info as it may be overridden
when parsing EDID. Instead, check bpc for DP branch device
during compute_config
v5: Indentation fixes (Jim Bride)
Mika Kahola [Fri, 9 Sep 2016 11:10:55 +0000 (14:10 +0300)]
drm/i915: Check pixel rate for DP to VGA dongle
Filter out a mode that exceeds the max pixel rate setting
for DP to VGA dongle. This is defined in DPCD register 0x81
if detailed cap info i.e. info field is 4 bytes long and
it is available for DP downstream port.
The register defines the pixel rate divided by 8 in MP/s.
v2: DPCD read outs and computation moved to drm (Ville, Daniel)
v3: Sink pixel rate computation moved to drm_dp_max_sink_dotclock()
function (Daniel)
v4: Use of drm_dp_helper.c routines to compute max pixel clock (Ville)
v5: Use of intel_dp->downstream_ports to read out port capabilities.
Code restructuring (Ville)
v6: Move DP branch device check to drm_dp_helper.c (Daniel)
v7: Cleanup as suggested by Ville
Mika Kahola [Fri, 9 Sep 2016 11:10:54 +0000 (14:10 +0300)]
drm/i915: Read DP branch device SW revision
SW revision is mandatory field for DisplayPort branch
devices. This is defined in DPCD register fields 0x50A
and 0x50B.
v2: move drm_dp_ds_revision structure to be part of
drm_dp_link structure (Daniel)
v3: remove dependency to drm_dp_helper but instead parse
DPCD and print SW revision info to dmesg (Ville)
v4: commit message fix (Jim Bride)
Mika Kahola [Fri, 9 Sep 2016 11:10:53 +0000 (14:10 +0300)]
drm/i915: Read DP branch device HW revision
HW revision is mandatory field for DisplayPort branch
devices. This is defined in DPCD register field 0x509.
v2: move drm_dp_ds_revision structure to be part of
drm_dp_link structure (Daniel)
v3: remove dependency to drm_dp_helper but instead parse
DPCD and print HW revision info to dmesg (Ville)
Mika Kahola [Fri, 9 Sep 2016 11:10:49 +0000 (14:10 +0300)]
drm: Helper to read max clock rate
Helper routine to read out maximum supported pixel rate
for DisplayPort legay VGA converter or TMDS clock rate
for other digital legacy converters. The helper returns
clock rate in kHz.
v2: Return early if detailed port cap info is not available.
Replace if-else ladder with switch-case (Ville)
Dave Gordon [Mon, 12 Sep 2016 20:19:35 +0000 (21:19 +0100)]
drm/i915: clarify PMINTRMSK/pm_intr_keep usage
No functional changes; just renaming a bit, tweaking a datatype,
prettifying layout, and adding comments, in particular in the
GuC setup code that touches this data.
Dave Gordon [Wed, 14 Sep 2016 12:10:33 +0000 (13:10 +0100)]
drm/i915: Only expand COND once in wait_for()
Commentary from Chris Wilson's original version:
> I was looking at some wait_for() timeouts on a slow system, with lots of
> debug enabled (KASAN, lockdep, mmio_debug). Thinking that we were
> mishandling the timeout, I tried to ensure that we loop at least once
> after first testing COND. However, the double test of COND either side
> of the timeout check makes that unlikely. But we can do an equivalent
> loop, that keeps the COND check after testing for timeout (required so
> that we are not preempted between testing COND and then testing for a
> timeout) without expanding COND twice.
>
> The advantage of only expanding COND once is a dramatic reduction in
> code size:
>
> text data bss dec hex
> 1308733 5184 1152 1315069 1410fd before
> 1305341 5184 1152 1311677 1403bd after
but it turned out that due to a missing iniitialiser, gcc had "gone
wild trimming undefined code" :( This version acheives a rather more
modest (but still worthwhile) gain of ~550 bytes.
Signed-off-by: Dave Gordon <david.s.gordon@intel.com> Original-idea-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Zanoni, Paulo R <paulo.r.zanoni@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1473855033-26980-1-git-send-email-david.s.gordon@intel.com Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Ville Syrjälä [Tue, 13 Sep 2016 09:22:19 +0000 (12:22 +0300)]
drm/i915: Ignore OpRegion panel type except on select machines
Turns out
commit a05628195a0d ("drm/i915: Get panel_type from OpRegion panel
details") has regressed quite a few machines. So it looks like we
can't use the panel type from OpRegion on all systems, and yet we
absolutely must use it on some specific systems.
Despite trying, I was unable to find any automagic way to determine
if the OpRegion panel type is respectable or not. The only glimmer
of hope I had was bit 8 in the SCIC response, but that turned out to
not work either (it was always 0 on both types of systems).
So, to fix the regressions without breaking the machine we know to need
the OpRegion panel type, let's just add a quirk for this. Only specific
machines known to require the OpRegion panel type will therefore use
it. Everyone else will fall bck to the VBT panel type.
The only known machine so far is a "Conrac GmbH IX45GM2". The PCI
subsystem ID on this machine is just a generic 8086:2a42, so of no use.
Instead we'll go with a DMI match.
I suspect we can now also revert
commit aeddda06c1a7 ("drm/i915: Ignore panel type from OpRegion on SKL")
but let's leave that to a separate patch.
v2: Do the DMI match in the opregion code directly, as dev_priv->quirks
gets populated too late
Cc: Rob Kramer <rob@solution-space.com> Cc: Martin van Es <martin@mrvanes.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Dave Airlie <airlied@linux.ie> Cc: Marco Krüger <krgsch@gmail.com> Cc: Sean Greenslade <sean@seangreenslade.com> Cc: Trudy Tective <bertslany@gmail.com> Cc: Robin Müller <rm1990@gmx.de> Cc: Alexander Kobel <a-kobel@a-kobel.de> Cc: Alexey Shumitsky <alexey.shumitsky@gmail.com> Cc: Emil Andersen Lauridsen <mine809@gmail.com> Cc: oceans112@gmail.com Cc: James Hogan <james@albanarts.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Cc: stable@vger.kernel.org
References: https://lists.freedesktop.org/archives/intel-gfx/2016-August/105545.html
References: https://lists.freedesktop.org/archives/dri-devel/2016-August/116888.html
References: https://lists.freedesktop.org/archives/intel-gfx/2016-June/098826.html
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94825
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97060
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97443
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97363 Fixes: a05628195a0d ("drm/i915: Get panel_type from OpRegion panel details") Tested-by: Marco Krüger <krgsch@gmail.com> Tested-by: Alexey Shumitsky <alexey.shumitsky@gmail.com> Tested-by: Sean Greenslade <sean@seangreenslade.com> Tested-by: Emil Andersen Lauridsen <mine809@gmail.com> Tested-by: Robin Müller <rm1990@gmx.de> Tested-by: oceans112@gmail.com Tested-by: Rob Kramer <rob@solution-space.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1473758539-21565-1-git-send-email-ville.syrjala@linux.intel.com
References: http://patchwork.freedesktop.org/patch/msgid/1473602239-15855-1-git-send-email-adrienverge@gmail.com Acked-by: Jani Nikula <jani.nikula@intel.com>
There are panels that needs 4 idle frames before entering PSR,
but VBT is unproperly set.
Also lately it was identified that idle frame count calculated at HW
can be off by 1, what makes the minimum of 2, at least.
Without the current vbt+1 we are with the risk of having HW calculating
0 idle frames and entering PSR when it shouldn't. Regardless the lack
of link training.
[Jani: there is some disagreement on the explanation, but the commit
regresses so revert it is.]
References: http://marc.info/?i=20160904191153.GA2328@light.dominikbrodowski.net Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Daniel Vetter <daniel.vetter@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Fixes: 1c80c25fb622 ("drm/i915/psr: Make idle_frames sensible again") Cc: drm-intel-fixes@lists.freedesktop.org # v4.8-rc1+ Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1473295351-8766-1-git-send-email-rodrigo.vivi@intel.com
drm/i915: Add support for Kabylake to function obtaining shared PLL
This adds support for KBL in the new function added in commit ID:
commit <f169660ed4e57a03e6f6ed07fe192dbcb7687a0d> that returns a
shared pll in case of DDI platforms.
drm/i915: Make DP link training channel equalization DP 1.2 Spec compliant
Fix the number of tries in channel euqalization link training sequence
according to DP 1.2 Spec. It returns a boolean depending on channel
equalization pass or failure.
drm/dp/i915: Make clock recovery in the link training compliant with DP Spec 1.2
This function cleans up clock recovery loop in link training compliant
tp Dp Spec 1.2. It tries the clock recovery 5 times for the same voltage
or until max voltage swing is reached and removes the additional non
compliant retries. This function now returns a boolean values based on
if clock recovery passed or failed.
v3:
* Better Debug prints in case of failures (Mika Kahola)
v2:
* Rebased on top of new revision of vswing patch (Manasi Navare)
Jim Bride [Wed, 7 Sep 2016 22:47:34 +0000 (15:47 -0700)]
drm/i915/dp: Add a standalone function to obtain shared dpll for HSW/BDW/SKL/BXT
Add the PLL selection code for HSW/BDW/BXT/SKL into a stand-alone function
in order to allow for the implementation of a platform neutral upfront
link training function.
v4:
* Removed dereferencing NULL pointer in case of failure (Dhinakaran Pandiyan)
v3:
* Add Hooks for all DDI platforms into this standalone function
v2:
* Change the macro to use dev_priv instead of dev (David Weinehall)
Reviewed-by: Durgadoss R <durgadoss.r@intel.com> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Signed-off-by: Jim Bride <jim.bride@linux.intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Chris Wilson [Fri, 9 Sep 2016 20:19:57 +0000 (21:19 +0100)]
drm/i915: Flush to GTT domain all GGTT bound objects after hibernation
Recently I have been applying an optimisation to avoid stalling and
clflushing GGTT objects based on their current binding. That is we only
set-to-gtt-domain upon first bind. However, on hibernation the objects
remain bound, but they are in the CPU domain. Currently (since commit 975f7ff42edf ("drm/i915: Lazily migrate the objects after hibernation"))
we only flush scanout objects as all other objects are expected to be
flushed prior to use. That breaks down in the face of the runtime
optimisation above - and we need to flush all GGTT pinned objects
(essentially ringbuffers).
To reduce the burden of extra clflushes, we only flush those objects we
cannot discard from the GGTT. Everything pinned to the scanout, or
current contexts or ringbuffers will be flushed and rebound. Other
objects, such as inactive contexts, will be left unbound and in the CPU
domain until first use after resuming.
Fixes: 7abc98fadfdd ("drm/i915: Only change the context object's domain...") Fixes: 57e885318119 ("drm/i915: Use VMA for ringbuffer tracking")
References: https://bugs.freedesktop.org/show_bug.cgi?id=94722 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: David Weinehall <david.weinehall@intel.com> Reviewed-by: Matthew Auld <matthew.auld@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20160909201957.2499-1-chris@chris-wilson.co.uk
Chris Wilson [Fri, 9 Sep 2016 13:12:01 +0000 (14:12 +0100)]
drm/i915: Serialise execbuf operation after a dma-buf reservation object
Now that we can wait upon fences before emitting the request, it becomes
trivial to wait upon any implicit fence provided by the dma-buf
reservation object.
To protect against failure, we force any asynchronous waits on a foreign
fence to timeout after 10s - so that a stall in another driver does not
permanently cripple ourselves. Still unpleasant though!
Chris Wilson [Fri, 9 Sep 2016 13:12:00 +0000 (14:12 +0100)]
drm/i915: Nonblocking request submission
Now that we have fences in place to drive request submission, we can
employ those to queue requests after their dependencies as opposed to
stalling in the middle of an execbuf ioctl. (However, we still choose to
spin before enabling the IRQ as that is faster - though contentious.)
v2: Do the fence ordering first, where we can still fail.
Chris Wilson [Fri, 9 Sep 2016 13:11:59 +0000 (14:11 +0100)]
drm/i915: Avoid incrementing hangcheck whilst waiting for external fence
If we are waiting upon an external fence, from the pov of hangcheck the
engine is stuck on the last submitted seqno. Currently we give a small
increment to the hangcheck score in order to catch a stuck waiter /
driver. Now that we both have an independent wait hangcheck and may be
stuck waiting on an external fence, resetting the GPU has little effect
on that external fence. As we cannot advance by resetting, skip
incrementing the hangcheck score.
Chris Wilson [Fri, 9 Sep 2016 13:11:57 +0000 (14:11 +0100)]
drm/i915/guc: Prepare for nonblocking execbuf submission
Currently the presumption is that the request construction and its
submission to the GuC are all under the same holding of struct_mutex. We
wish to relax this to separate the request construction and the later
submission to the GuC. This requires us to reserve some space in the
GuC command queue for the future submission. For flexibility to handle
out-of-order request submission we do not preallocate the next slot in
the GuC command queue during request construction, just ensuring that
there is enough space later.
Chris Wilson [Fri, 9 Sep 2016 13:11:56 +0000 (14:11 +0100)]
drm/i915: Prepare object synchronisation for asynchronicity
We are about to specialize object synchronisation to enable nonblocking
execbuf submission. First we make a copy of the current object
synchronisation for execbuffer. The general i915_gem_object_sync() will
be removed following the removal of CS flips in the near future.
Chris Wilson [Fri, 9 Sep 2016 13:11:54 +0000 (14:11 +0100)]
drm/i915: Drive request submission through fence callbacks
Drive final request submission from a callback from the fence. This way
the request is queued until all dependencies are resolved, at which
point it is handed to the backend for queueing to hardware. At this
point, no dependencies are set on the request, so the callback is
immediate.
A side-effect of imposing a heavier-irqsafe spinlock for execlist
submission is that we lose the softirq enabling after scheduling the
execlists tasklet. To compensate, we manually kickstart the softirq by
disabling and enabling the bh around the fence signaling.
Chris Wilson [Fri, 9 Sep 2016 13:11:53 +0000 (14:11 +0100)]
drm/i915: Update reset path to fix incomplete requests
Update reset path in preparation for engine reset which requires
identification of incomplete requests and associated context and fixing
their state so that engine can resume correctly after reset.
The request that caused the hang will be skipped and head is reset to the
start of breadcrumb. This allows us to resume from where we left-off.
Since this request didn't complete normally we also need to cleanup elsp
queue manually. This is vital if we employ nonblocking request
submission where we may have a web of dependencies upon the hung request
and so advancing the seqno manually is no longer trivial.
We change the way we count pending batches. Only the active context
involved in the reset is marked as either innocent or guilty, and not
mark the entire world as pending. By inspection this only affects
igt/gem_reset_stats (which assumes implementation details) and not
piglit.
ARB_robustness gives this guide on how we expect the user of this
interface to behave:
* Provide a mechanism for an OpenGL application to learn about
graphics resets that affect the context. When a graphics reset
occurs, the OpenGL context becomes unusable and the application
must create a new context to continue operation. Detecting a
graphics reset happens through an inexpensive query.
And with regards to the actual meaning of the reset values:
Certain events can result in a reset of the GL context. Such a reset
causes all context state to be lost. Recovery from such events
requires recreation of all objects in the affected context. The
current status of the graphics reset state is returned by
enum GetGraphicsResetStatusARB();
The symbolic constant returned indicates if the GL context has been
in a reset state at any point since the last call to
GetGraphicsResetStatusARB. NO_ERROR indicates that the GL context
has not been in a reset state since the last call.
GUILTY_CONTEXT_RESET_ARB indicates that a reset has been detected
that is attributable to the current GL context.
INNOCENT_CONTEXT_RESET_ARB indicates a reset has been detected that
is not attributable to the current GL context.
UNKNOWN_CONTEXT_RESET_ARB indicates a detected graphics reset whose
cause is unknown.
The language here is explicit in that we must mark up the guilty batch,
but is loose enough for us to relax the innocent (i.e. pending)
accounting as only the active batches are involved with the reset.
In the future, we are looking towards single engine resetting (with
minimal locking), where it seems inappropriate to mark the entire world
as innocent since the reset occurred on a different engine. Reducing the
information available means we only have to encounter the pain once, and
also reduces the information leaking from one context to another.
v2: Legacy ringbuffer submission required a reset following hibernation,
or else we restore stale values to the RING_HEAD and walked over
stolen garbage.
v3: GuC requires replaying the requests after a reset.
v4: Restore engine IRQ after reset (so waiters will be woken!)
Rearm hangcheck if resetting with a waiter.
Chris Wilson [Fri, 9 Sep 2016 13:11:52 +0000 (14:11 +0100)]
drm/i915: Replace wait-on-mutex with wait-on-bit in reset worker
Since we have a cooperative mode now with a direct reset, we can avoid
the contention on struct_mutex and instead try then sleep on the
I915_RESET_IN_PROGRESS bit. If the mutex is held and that bit is
cleared, all is fine. Otherwise, we sleep for a bit and try again. In
the worst case we sleep for an extra second waiting for the mutex to be
released (no one touching the GPU is allowed the struct_mutex whilst the
I915_RESET_IN_PROGRESS bit is set). But when we have a direct reset,
this allows us to clean up the reset worker faster.
v2: Remember to call wake_up_bit() after changing (for the faster wakeup
as promised)
Chris Wilson [Fri, 9 Sep 2016 13:11:51 +0000 (14:11 +0100)]
drm/i915: Perform a direct reset of the GPU from the waiter
If a waiter is holding the struct_mutex, then the reset worker cannot
reset the GPU until the waiter returns. We do not want to return -EAGAIN
form i915_wait_request as that breaks delicate operations like
i915_vma_unbind() which often cannot be restarted easily, and returning
-EIO is just as useless (and has in the past proven dangerous). The
remaining WARN_ON(i915_wait_request) serve as a valuable reminder that
handling errors from an indefinite wait are tricky.
We can keep the current semantic that knowing after a reset is complete,
so is the request, by performing the reset ourselves if we hold the
mutex.
uevent emission is still handled by the reset worker, so it may appear
slightly out of order with respect to the actual reset (and concurrent
use of the device).
Chris Wilson [Fri, 9 Sep 2016 13:11:50 +0000 (14:11 +0100)]
drm/i915: Mark up all locked waiters
In the next patch we want to handle reset directly by a locked waiter in
order to avoid issues with returning before the reset is handled. To
handle the reset, we must first know whether we hold the struct_mutex.
If we do not hold the struct_mtuex we can not perform the reset, but we do
not block the reset worker either (and so we can just continue to wait for
request completion) - otherwise we must relinquish the mutex.
Chris Wilson [Fri, 9 Sep 2016 13:11:48 +0000 (14:11 +0100)]
drm/i915: Drop local struct_mutex around intel_init_emon[ilk]
Access to intel_init_emon() is strictly ordered by gt_powersave, using
struct_mutex around it is overkill (and will conflict with the caller
holding struct_mutex themselves).
Chris Wilson [Fri, 9 Sep 2016 13:11:47 +0000 (14:11 +0100)]
drm/i915: Separate out reset flags from the reset counter
In preparation for introducing a per-engine reset, we can first separate
the mixing of the reset state from the global reset counter.
The loss of atomicity in updating the reset state poses a small problem
for handling the waiters. For requests, this is solved by advancing the
seqno so that a waiter waking up after the reset knows the request is
complete. For pending flips, we still rely on the increment of the
global reset epoch (as well as the reset-in-progress flag) to signify
when the hardware was reset.
The advantage, now that we do not inspect the reset state during reset
itself i.e. we no longer emit requests during reset, is that we can use
the atomic updates of the state flags to ensure that only one reset
worker is active.
v2: Mika spotted that I transformed the i915_gem_wait_for_error() wakeup
into a waiter wakeup.
Chris Wilson [Fri, 9 Sep 2016 13:11:46 +0000 (14:11 +0100)]
drm/i915: Simplify ELSP queue request tracking
Emulate HW to track and manage ELSP queue. A set of SW ports are defined
and requests are assigned to these ports before submitting them to HW. This
helps in cleaning up incomplete requests during reset recovery easier
especially after engine reset by decoupling elsp queue management. This
will become more clear in the next patch.
In the engine reset case we want to resume where we left-off after skipping
the incomplete batch which requires checking the elsp queue, removing
element and fixing elsp_submitted counts in some cases. Instead of directly
manipulating the elsp queue from reset path we can examine these ports, fix
up ringbuffer pointers using the incomplete request and restart submissions
again after reset.
Chris Wilson [Fri, 9 Sep 2016 13:11:44 +0000 (14:11 +0100)]
drm/i915: Compute the ELSP register location once
Similar to the issue with reading from the context status buffer,
see commit 26720ab97fea ("drm/i915: Move CSB MMIO reads out of the
execlists lock"), we frequently write to the ELSP register (4 writes per
interrupt) and know we hold the required spinlock and forcewake throughout.
We can further reduce the cost of writing these registers beyond the
I915_WRITE_FW() by precomputing the address of the ELSP register. We also
note that the subsequent read serves no purpose here, and are happy to
see it go.
v2: Address I915_WRITE mistakes in changelog
text data bss dec hex filename 1259784 4581 576 1264941 134d2d drivers/gpu/drm/i915/i915.ko 1259720 4581 576 1264877 134ced drivers/gpu/drm/i915/i915.ko
Chris Wilson [Fri, 9 Sep 2016 13:11:41 +0000 (14:11 +0100)]
drm/i915: Add a sw fence for collecting up dma fences
This is really a core kernel struct in disguise until we can finally
place it in kernel/. There is an immediate need for a fence collection
mechanism that is more flexible than fence-array, in particular being
able to easily drive request submission via events (and not just
interrupt driven). The same mechanism would be useful for handling
nonblocking and asynchronous atomic modesets, parallel execution and
more, but for the time being just create a local sw fence for execbuf.
Carlos Santa [Wed, 17 Aug 2016 19:30:57 +0000 (12:30 -0700)]
drm/i915: Move HAS_GUC definition to platform definition
Moving all GPU features to the platform definition allows for
- standard place when adding new features from new platform
- possible to see supported features when dumping struct
definitions
Signed-off-by: Carlos Santa <carlos.santa@intel.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Carlos Santa [Wed, 17 Aug 2016 19:30:56 +0000 (12:30 -0700)]
drm/i915: Make HWS_NEEDS_PHYSICAL the exception
Make the .hws_needs_physical the exception by switching the flag
on earlier platforms since they are fewer to support. Remove the flag on
later GPUs hardware since they all use GTT hws by default.
Switch the logic as well in the driver to reflect this change
Signed-off-by: Carlos Santa <carlos.santa@intel.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Carlos Santa [Wed, 17 Aug 2016 19:30:55 +0000 (12:30 -0700)]
drm/i915: Move HAS_GMCH_DISPLAY definition to platform
Moving all GPU features to the platform definition allows for
- standard place when adding new features from new platforms
- possible to see supported features when dumping struct
definitions
Signed-off-by: Carlos Santa <carlos.santa@intel.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
projects / linux-beck.git / log