Oneshot disabling of IPS when CRC capturing is started is insufficient.
IPS may get re-enabled by any plane update, and hence tests that keep
CRC capturing on across plane updates will start to see inconsistent
results as soon as IPS kicks back in. Add a new knob into the crtc state
to make sure IPS stays disabled as long as CRC capturing is enabled.
Forcing a modeset is the easiest way to handle this since that's already
how we do the panel fitter workaround. It's a little heavy handed just
for IPS, but seeing as we might already do the panel fitter workaround
I think it's better to follow that. We migth want to optimize both cases
later if someone gets too upset by the extra delay from the modeset.
v2: Check the right thing when deciding whether to force a modeset
v3: Rebase, check HAS_IPS before forcing a modeset,
move ips_force_disable check into pipe_config_supports_ips()
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Marta Lofstedt <marta.lofstedt@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101664
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Marta Lofsted <marta.lofstedt@intel.com> #v2
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170817145509.15549-1-ville.syrjala@linux.intel.com
We use WC pages for coherent writes into the ppGTT on !llc
architectures. However, to create a WC page requires a stop_machine(),
i.e. is very slow. To compensate we currently keep a per-vm cache of
recently freed pages, but we still see the slow startup of new contexts.
We can amoritize that cost slightly by allocating WC pages in small
batches (PAGEVEC_SIZE == 14) and since creating a WC page implies a
stop_machine() there is no penalty for keeping that stash global.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170822173828.5932-1-chris@chris-wilson.co.uk
Coffee Lake CPU on Kaby Lake PCH is possible.
It does exist, and it does work.
The only missed case was this warning here noticed
by Wendy who could get one system with this configuration
and reported the issue for us:
Hardware Configuration
Board ID KBL S DDR4 UDIMM EV CRB
Processor Intel® Processor code named Coffee Lake S, (6+2), 6 cores 12 threads, GT2, A0 (Internal) (QNJ4)
[ 3.220585] WARNING: CPU: 10 PID: 206 at drivers/gpu/drm/i915/i915_drv.c:340 i915_driver_load+0x1210/0x1660 [i915]
[ 3.221312] Modules linked in: hid_generic usbhid i915 i2c_algo_bit drm_kms_helper e1000e syscopyarea sysfillrect sysimgblt nvme fb_sys_fops ptp ahci i2c_hid drm pps_core nvme_core libahci wmi hid video
[ 3.222050] CPU: 10 PID: 206 Comm: systemd-udevd Not tainted 4.13.0-rc5-intel-next+ #1
[ 3.222706] Hardware name: Intel Corporation Kabylake Client platform/KBL S DDR4 UDIMM EV CRB, BIOS KBLSE2R1.R00.X089.P00.1705051000 05/05/2017
Cc: Wendy Wang <wendy.wang@intel.com>
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170821235056.9015-1-rodrigo.vivi@intel.com
If we miss the current vblank because the gpu was busy, that may cause a
jitter as the frame rate temporarily drops. We try to limit the impact
of this by then boosting the GPU clock to deliver the frame as quickly
as possible. Originally done in commit 6ad790c0f5 ("drm/i915: Boost GPU
frequency if we detect outstanding pageflips") but was never forward
ported to atomic and finally dropped in commit fd3a40242e ("drm/i915:
Rip out legacy page_flip completion/irq handling").
One of the most typical use-cases for this is a mostly idle desktop.
Rendering one frame of the desktop's frontbuffer can easily be
accomplished by the GPU running at low frequency, but often exceeds
the time budget of the desktop compositor. The result is that animations
such as opening the menu, doing a fullscreen switch, or even just trying
to move a window around are slow and jerky. We need to respond within a
frame to give the best impression of a smooth UX, as a compromise we
instead respond if that first frame misses its goal. The result should
be a near-imperceivable initial delay and a smooth animation even
starting from idle. The cost, as ever, is that we spend more power than
is strictly necessary as we overestimate the required GPU frequency and
then try to ramp down.
This of course is reactionary, too little, too late; nevertheless it is
surprisingly effective.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102199
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170817123706.6777-1-chris@chris-wilson.co.uk
Tested-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com>
The enable/disable/etc. encoder hooks aren't supposed to alter the
state(s), so pass them as const. Unfortunately C lacks any kind of deep
const thingy, so this can't catch all abuses. But at least it acts as a
hint to the reader telling them not to mess about with the state(s).
v2: Update intel_tv_mode_find() and ironlake_edp_pll_on() as well
v3: Deal with intel_sdvo_connector_state
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170818134958.15502-9-ville.syrjala@linux.intel.com
Disabling the video DIP when shutting the port down seems like a good
idea.
Bspec says:
"When disabling both the DIP port and DIP transmission,
first disable the port and then disable DIP."
and
"Restriction : GCP is only supported with HDMI when the bits per color is
not equal to 8. GCP must be enabled prior to enabling TRANS_DDI_FUNC_CTL
for HDMI with bits per color not equal to 8 and disabled after disabling
TRANS_DDI_FUNC_CTL"
So let's do it in the .post_disable() hook.
v2: Remove double "dpms off" caused by rebase fail
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Shashank Sharma <shashank.sharma@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170822140914.24413-1-ville.syrjala@linux.intel.com
The corruption in CSB mmio reads we were seeing has been tracked down to
incorrectly touching forcewake of all domains, following an engine reset.
It is still a mistery why we only catched this in Broxton, since it
could happen in any platform.
With that fix already merged, commit 4055dc75d6 ("drm/i915: Stop
touching forcewake following a gen6+ engine reset"), lets try to enable
per-engine resets in Broxton one more time.
This reverts commit f188258bde0f ("drm/i915: Disable per-engine reset for
Broxton").
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Michel Thierry <michel.thierry@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170818172342.7282-1-michel.thierry@intel.com
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This bit enables hardware that will change the approximation used for distances
calculations for AA wide lines so that they are rendered more accurately.
The default value for this bit leaves the legacy behavior. There is no good
reason to not enable the new approximation except if comparing to previous GEN
rendered images.
v2: Rebase
v3: Fix author.
Rebased by Rodrigo who also added a comment as suggested by Oscar.
Since it is surrounded by Workarounds let's just add a comment to
make clear it is not an Wa.
Cc: Oscar Mateo <oscar.mateo@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Oscar Mateo <oscar.mateo@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170815231651.975-4-rodrigo.vivi@intel.com
Let's inherit workarounds from previous platforms that
according to wa_database and BSpec are still valid for
Cannonlake.
v2: Add missed workarounds.
v3: Rebase
v4: Remove bad chunk that was added to rc6 disable. (Ander)
Also remove A0 W/a that are not needed anymore.
v5: Rebase on top of CFL.
v6: Remove empty gen9_init_perctx_bb and gen9_init_indirectctx_bb
since they don't carry any gen10 related W/a. (by Oscar).
Also Remove A0 exclusive workaround.
v7: Remove more A0 exclusive workarounds. As pointed out by Oscar
many workarounds were changed to be A0 only so let's remove
them.
Cc: Oscar Mateo <oscar.mateo@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Oscar Mateo <oscar.mateo@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170815231651.975-1-rodrigo.vivi@intel.com
The commit 213e08ad60
("drm/i915/bxt: add bxt dsi gpio element support")
enables GPIO support for Broxton based platforms.
While using that API we might get into troubles in the future, because
we can't rely on label name in the driver since vendor firmware might
provide any GPIO pin there, e.g. "reset", and even mark it in _DSD (in
which case the request will fail).
To avoid inconsistency and potential issues we have two options:
a) generate GPIO ACPI mapping table and supply it via
acpi_dev_add_driver_gpios(), or
b) just pass NULL as connection ID.
The b) approach is much simpler and would work since the driver relies
on GPIO indices only. Moreover, the _CRS fallback mechanism, when
requesting GPIO, has been made stricter, and supplying non-NULL
connection ID when neither _DSD, nor GPIO ACPI mapping is present, is
making request fail.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101921
Fixes: f10e4bf663 ("gpio: acpi: Even more tighten up ACPI GPIO lookups")
Cc: Mika Kahola <mika.kahola@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Tested-by: Mika Kahola <mika.kahola@intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170817105541.63914-1-andriy.shevchenko@linux.intel.com
In a synchronous setup, we may retire the last request before we
complete allocating the next request. As the last request is retired, we
queue a timer to mark the device as idle, and promptly have to execute
ad cancel that timer once we complete allocating the request and need to
keep the device awake. If we rearrange the mark_busy() to occur before
we retire the previous request, we can skip this ping-pong.
v2: Joonas pointed out that unreserve_seqno() was now doing more than
doing seqno handling and should be renamed to reflect its wider purpose.
That also highlighted the new asymmetry with reserve_seqno(), so fixup
that and rename both to [un]reserve_engine().
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170817144719.10968-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
This was the competing idea long ago, but it was only with the rewrite
of the idr as an radixtree and using the radixtree directly ourselves,
along with the realisation that we can store the vma directly in the
radixtree and only need a list for the reverse mapping, that made the
patch performant enough to displace using a hashtable. Though the vma ht
is fast and doesn't require any extra allocation (as we can embed the node
inside the vma), it does require a thread for resizing and serialization
and will have the occasional slow lookup. That is hairy enough to
investigate alternatives and favour them if equivalent in peak performance.
One advantage of allocating an indirection entry is that we can support a
single shared bo between many clients, something that was done on a
first-come first-serve basis for shared GGTT vma previously. To offset
the extra allocations, we create yet another kmem_cache for them.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170816085210.4199-5-chris@chris-wilson.co.uk
When userspace is doing most of the work, avoiding relocs (using
NO_RELOC) and opting out of implicit synchronisation (using ASYNC), we
still spend a lot of time processing the arrays in execbuf, even though
we now should have nothing to do most of the time. One issue that
becomes readily apparent in profiling anv is that iterating over the
large execobj[] is unfriendly to the loop prefetchers of the CPU and it
much prefers iterating over a pair of arrays rather than one big array.
v2: Clear vma[] on construction to handle errors during vma lookup
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170816085210.4199-3-chris@chris-wilson.co.uk
Forcewake is not affected by the engine reset on gen6+. Indeed the
reason why we added intel_uncore_forcewake_reset() to
gen6_reset_engines() was to keep the bookkeeping intact because the
reset did not touch the forcewake bit (yet we cancelled the forcewake
consumers)! This was done in commit 521198a2e7:
Author: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Fri Aug 23 16:52:30 2013 +0300
drm/i915: sanitize forcewake registers on reset
In reset we try to restore the forcewake state to
pre reset state, using forcewake_count. The reset
doesn't seem to clear the forcewake bits so we
get warn on forcewake ack register not clearing.
That futzing of the forcewake bookkeeping was dropped in commit
0294ae7b44 ("drm/i915: Consolidate forcewake resetting to a single
function"), but it did not make the realisation that the remaining
intel_uncore_forcewake_reset() was redundant.
The new danger with using intel_uncore_forcewake_reset() with per-engine
resets is that the driver and hw are still in an active state as we
perform the reset. We may be using the forcewake to read protected
registers elsewhere and those results may be clobbered by the concurrent
dropping of forcewake.
Reported-by: Michel Thierry <michel.thierry@intel.com>
Fixes: 142bc7d99b ("drm/i915: Modify error handler for per engine hang recovery")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170817173229.20324-1-chris@chris-wilson.co.uk
Reviewed-by: Michel Thierry <michel.thierry@intel.com>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
For a bunch of reasons[1] I've decided to step down as maintainer and
let some other folks enjoy the reputation and hang out in the
spotlight.
Jani is going to stick around with his expertise in kms and having
done the fixes flow for a long time now. Joonas will join and bring in
his knowledge on all things GEM. Rodrigo has been less visible because
he's been doing tons of work taking care of the internal branch, and
it'd be good to have more continuity between these two worlds also on
the maintainer side.
1: They all boil down to: This is going to happen sooner or later
anyway, we have a great team, with the process improvements over the
last few years things work rather well, now is as good as any time to
do this. With that change I'll have more time for other aspects of the
stack development than maintainership.
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Dave Airlie <airlied@gmail.com>
Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Acked-by: Jani Nikula <jani.nikula@linux.intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170815160101.1683-1-daniel.vetter@ffwll.ch
Sometimes it would be most enlightening to debug systems by replacing
the VBT to be used. For example, in the referenced bug the BIOS provides
different VBT depending on the boot mode (UEFI vs. legacy). It would be
interesting to try the failing boot mode with the VBT from the working
boot, and see if that makes a difference.
Add a module parameter to load the VBT using the firmware loader, not
unlike the EDID firmware mechanism.
As a starting point for experimenting, one can pick up the BIOS provided
VBT from /sys/kernel/debug/dri/0/i915_opregion/i915_vbt.
v2: clarify firmware load return value check (Bob)
v3: kfree the loaded firmware blob
References: https://bugs.freedesktop.org/show_bug.cgi?id=97822#c83
Reviewed-by: Bob Paauwe <bob.j.paauwe@intel.com>
Acked-by: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170817115209.25912-1-jani.nikula@intel.com
Ever since we've parsed VBT child devices, starting from 6acab15a7b
("drm/i915: use the HDMI DDI buffer translations from VBT"), we've
ignored the child device information if more than one child device
references the same port. The rationale for this seems lost in time.
Since commit 311a20949f ("drm/i915: don't init DP or HDMI when not
supported by DDI port") we started using this information more to skip
HDMI/DP init if the port wasn't there per VBT child devices. However, at
the same time it added port defaults without further explanation.
Thus, if the child device info was skipped due to multiple child devices
referencing the same port, the device info would be retrieved from the
somewhat arbitrary defaults.
Finally, when commit bb1d132935 ("drm/i915/vbt: split out defaults
that are set when there is no VBT") stopped initializing the defaults
whenever VBT is present, thus trusting the VBT more, we stopped
initializing ports which were referenced by more than one child device.
Apparently at least Asus UX305UA, UX305U, and UX306U laptops have VBT
child device blocks which cause this behaviour. Arguably they were
shipped with a broken VBT.
Relax the rules for multiple references to the same port, and use the
first child device info to reference a port. Retain the logic to debug
log about this, though.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101745
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=196233
Fixes: bb1d132935 ("drm/i915/vbt: split out defaults that are set when there is no VBT")
Tested-by: Oliver Weißbarth <mail@oweissbarth.de>
Reported-by: Oliver Weißbarth <mail@oweissbarth.de>
Reported-by: Didier G <didierg-divers@orange.fr>
Reported-by: Giles Anderson <agander@gmail.com>
Cc: Manasi Navare <manasi.d.navare@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: <stable@vger.kernel.org> # v4.12+
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170811113907.6716-1-jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
This commit adds support for waiting on or signaling DRM syncobjs as
part of execbuf. It does so by hijacking the currently unused cliprects
pointer to instead point to an array of i915_gem_exec_fence structs
which containe a DRM syncobj and a flags parameter which specifies
whether to wait on it or to signal it. This implementation
theoretically allows for both flags to be set in which case it waits on
the dma_fence that was in the syncobj and then immediately replaces it
with the dma_fence from the current execbuf.
v2:
- Rebase on new syncobj API
v3:
- Pull everything out into helpers
- Do all allocation in gem_execbuffer2
- Pack the flags in the bottom 2 bits of the drm_syncobj*
v4:
- Prevent a potential race on syncobj->fence
Testcase: igt/gem_exec_fence/syncobj*
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Link: https://patchwork.freedesktop.org/patch/msgid/1499289202-25441-1-git-send-email-jason.ekstrand@intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20170815145733.4562-1-chris@chris-wilson.co.uk
The wait-ioctl is optionally supplied a timeout with nanosecond
precision in a s64 field. We use nsecs_to_jiffies64() to convert that
into the jiffies consumed by the scheduler, but internally
nsecs_to_jiffies64() does not guard against overflow (as it's purpose is
for use by the scheduler and not drivers!). So we must guard against the
overflow ourselves, and in the process note that we may then return
much earlier than the timeout selected by the user, so don't report
ETIME unless we do hit the timeout. (Woe betold us though if the user
waits for a year (32bit) and the request is still not complete!)
v2: Refine overflow detection (to not include an overffow itself)
Reported-by: Jason Ekstrand <jason.ekstrand@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20170811105731.9482-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
