We disabled rps while it appeared to be a contributing factor to system
instablity, as that is now resolved, re-enable RPS and see how we fare.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123135604.1402572-1-chris@chris-wilson.co.uk
Using a clear page for scratch means that we have relatively benign
errors in case it is accidentally used, but that can be rather too
benign for debugging. If we poison the scratch, ideally it quickly
results in an obvious error.
v2: Set each page individually just in case we are using highmem for our
scratch page.
v3: Pick a new scratch register as MI_STORE_REGISTER_MEM does not work
with GPR0 on gen7, unbelievably.
v4: Haswell still considers 3DPRIM a privileged register!
Suggested-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Reviewed-by: Matthew Auld <matthew.william.auld@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200124115133.53360-1-chris@chris-wilson.co.uk
Due to the asynchronous nature of releasing our wakerefs, we can signal
the main GT wakeref as complete before the individual engines have
cleared their own wakerefs. During shutdown we assert that the engines
are indeed parked before we release them, but for this to be always true
we need to flush their workers as well.
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/20200124143339.140988-1-chris@chris-wilson.co.uk
Perhaps in some cases the BIOS/GOP or other firmware may turn on
PHY A but may not program the MUX correctly. Therefore, re-program
PHY A if it is determined after reading the VBT that the value
programmed for the MUX bit does not match the expected value.
Cc: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Vivek Kasireddy <vivek.kasireddy@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200121235848.8457-1-vivek.kasireddy@intel.com
A recent change in BSpec allow us to change EXTLINE while transcoder
is enabled so this allow us to change it even when doing the first
fastset after taking over previous hardware state set by BIOS.
BIOS don't enable PSR, so if sink supports PSR it will be enabled on
the first fastset, so moving the EXTLINE compute and set to PSR flows
allow us to simplfy a bunch of code.
This will save a lot of time in all the IGT tests that uses CRC, as
when PSR2 is enabled CRCs are not generated, so we switch to PSR1, so
the previous code would compute dc3co_exitline=0 causing a full
modeset that would shutdown pipe, enable and train link.
v2: only programming EXTLINE when DC3CO is enabled
BSpec: 49196
Cc: Imre Deak <imre.deak@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Reviewed-by: Anshuman Gupta <anshuman.gupta@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200122182617.18597-2-jose.souza@intel.com
This will calculaet the DC3CO exit delay only once per full modeset.
Cc: Imre Deak <imre.deak@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Reviewed-by: Anshuman Gupta <anshuman.gupta@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200122182617.18597-1-jose.souza@intel.com
Remove the i2c_bus_num >= 0 check from the adapter lookup function
as this would prevent ACPI bus number override. This check was mainly
there to return early if the bus number has already been found but we
anyway return in the next line if the slave address does not match.
Fixes: 8cbf89db29 ("drm/i915/dsi: Parse the I2C element from the VBT MIPI sequence block (v3)")
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Nabendu Maiti <nabendu.bikash.maiti@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Bob Paauwe <bob.j.paauwe@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Vivek Kasireddy <vivek.kasireddy@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200118005848.20382-1-vivek.kasireddy@intel.com
Optimistically wait for the prior vma activity before taking the mutex
to minimise the mutex hold time while unbinding. We will then verify the
vma is idle with a second wait under the mutex to ensure it is safe to
unbind.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123224459.38128-2-chris@chris-wilson.co.uk
Only assert that the i915_vma is now idle if and only if no other pins
are present. If another user has the i915_vma pinned, they may submit
more work to the i915_vma skipping the vm->mutex used to serialise the
unbind. We need to wait again, if we want to continue and unbind this
vma.
However, if we own the i915_vma (we hold the vm->mutex for the unbind
and the pin_count is 0), we can assert that the vma remains idle as we
unbind.
Fixes: 2850748ef8 ("drm/i915: Pull i915_vma_pin under the vm->mutex")
Closes: https://gitlab.freedesktop.org/drm/intel/issues/530
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123224459.38128-1-chris@chris-wilson.co.uk
If the ctx->vm is freed before we can acquire a local reference to it,
we proceed to call i915_vm_put(NULL), which is invalid.
Reported-by: Colin Ian King <colin.king@canonical.com>
Fixes: 5dbd2b7be6 ("drm/i915/gem: Convert vm idr to xarray")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123152602.1432282-1-chris@chris-wilson.co.uk
Include the current RC6 residency counter in the error message, so that
if we fail to park and manually enter RC6 we can see if the counter has
a particularly suspect value (such as 0).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Andi Shyti <andi.shyti@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123145755.1420622-1-chris@chris-wilson.co.uk
In the port sync mode, for the master crtc, the master_transcoder is INVALID.
In that case since its value is -1, do not set the bit in the bitmask.
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Fixes: d0eed1545f ("drm/i915: Fix post-fastset modeset check for port sync")
Signed-off-by: Manasi Navare <manasi.d.navare@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200123002415.31478-1-manasi.d.navare@intel.com
Add convenience helpers for the most common uncore operations with
struct drm_i915_private * as context rather than struct intel_uncore *.
The goal is to replace all instances of I915_READ(),
I915_POSTING_READ(), I915_WRITE(), I915_READ_FW(), and I915_WRITE_FW()
in display/ with these, to finally be able to get rid of the implicit
dev_priv local parameter use.
The idea is that any non-u32 reads or writes are special enough that
they can use the intel_uncore_* functions directly.
v2:
- rename the file intel_de.h
- move intel_de_wait_for_* there too
- also add de fw helpers
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200121113915.9813-1-jani.nikula@intel.com
We've already pinned the vma and fence by the time we try to
deal with implicit fencing. Properly unpin the vma and fence
if the fence setup fails instead of just bailing straight out
from .prepare_fb(). As can be expected
drm_atomic_helper_prepare_planes() will not call .cleanup_fb()
for the plane whose .prepare_fb() failed so we must do the
cleanup ourself.
v2: Rebase
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200110183228.8199-6-ville.syrjala@linux.intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
intel_prepare_plane_fb() bails early if there is no fb (or rather
no obj, which is the same thing). intel_cleanup_plane_fb() does not.
This means the steps performed by intel_cleanup_plane_fb() aren't
balanced with with what was done intel_prepare_plane_fb() if there
is no fb for the plane. These hooks get called for every plane in
the state regardless of whether they have an fb or not.
Add a matching null obj check to intel_cleanup_plane_fb() to restore
the balance.
Note that intel_cleanup_plane_fb() has sufficient protections
already in place that the imbalance doesn't cause any real problems.
But having things be in balance seems nicer anyway, and might help
avoid some surprises in the future.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200110183228.8199-5-ville.syrjala@linux.intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Switch over to using explicit old/new planes states instead of
digging the old state out via plane->state. The main issue is that
plane->state will point to the uapi state which we generally don't
even want to look at.
Also it sets a bad example as using plane->state during commit_tail()
would be a bug. Here we're still holding the modeset locks so it's
actually safe, but best not give people bad ideas.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200110183228.8199-3-ville.syrjala@linux.intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Let's do the intel_plane_copy_uapi_to_hw_state() before we bail out
due to both old and new uapi.crtc being NULL. This will drop the
reference to the old hw.fb for planes that are transitioning from
being a slave plane to simply being disabled.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200110183228.8199-2-ville.syrjala@linux.intel.com
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Replace the vm_idr + vm_idr_mutex to an XArray. The XArray data
structure is now used to implement IDRs, and provides its own locking.
We can simply remove the IDR wrapper and in the process also remove our
extra mutex.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200122161531.508903-1-chris@chris-wilson.co.uk
If we encounter a hang on a virtual engine, as we process the hang the
request may already have been moved back to the virtual engine (we are
processing the hang on the physical engine). We need to reclaim the
request from the virtual engine so that the locking is consistent and
local to the real engine on which we will hold the request for error
state capturing.
v2: Pull the reclamation into execlists_hold() and assert that cannot be
called from outside of the reset (i.e. with the tasklet disabled).
v3: Added selftest
v4: Drop the reference owned by the virtual engine
Fixes: 748317386a ("drm/i915/execlists: Offline error capture")
Testcase: igt/gem_exec_balancer/hang
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200122140243.495621-2-chris@chris-wilson.co.uk
Thanks to preempt-to-busy, we leave the request on the HW as we submit
the preemption request. This means that the request may complete at any
moment as we process HW events, and in particular the request may be
retired as we are planning to capture it for a preemption timeout.
Be more careful while obtaining the request to capture after a
preemption timeout, and check to see if it completed before we were able
to put it on the on-hold list. If we do see it did complete just before
we capture the request, proclaim the preemption-timeout a false positive
and pardon the reset as we should hit an arbitration point momentarily
and so be able to process the preemption.
Note that even after we move the request to be on hold it may be retired
(as the reset to stop the HW comes after), so we do require to hold our
own reference as we work on the request for capture (and all of the
peeking at state within the request needs to be carefully protected).
Fixes: 32ff621fd7 ("drm/i915/gt: Allow temporary suspension of inflight requests")
Closes: https://gitlab.freedesktop.org/drm/intel/issues/997
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200122140243.495621-1-chris@chris-wilson.co.uk
We have two trace messages that rely on the function name for
distinction. However, if gcc inlines the function, the two traces end up
with the same function name and are indistinguishable. Add a different
message to each to clarify which one we hit, i.e. which phase of engine
parking we are processing.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200122124154.483444-1-chris@chris-wilson.co.uk
Add new struct drm_device based WARN* macros. These are modeled after
the core kernel device based WARN* macros. These would be preferred
over the regular WARN* macros, where possible.
These macros include device information in the backtrace, so we know
what device the warnings originate from.
Knowing the device specific information in the backtrace would be
helpful in development all around.
Signed-off-by: Pankaj Bharadiya <pankaj.laxminarayan.bharadiya@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Acked-by: Maxime Ripard <mripard@kernel.org>
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Acked-by: Sean Paul <sean@poorly.run>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200115034455.17658-2-pankaj.laxminarayan.bharadiya@intel.com
While we do flush writes to the vma before unbinding (to make sure they
go through the right detiling register), we may also be concurrently
poking at the GGTT_WRITE bit from set-domain, as we mark all GGTT vma
associated with an object. We know this is for another vma, as we
are currently unbinding this one -- so if this vma will be reused, it
will be refaulted and have its dirty bit set before the next write.
Closes: https://gitlab.freedesktop.org/drm/intel/issues/999
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200121222447.419489-1-chris@chris-wilson.co.uk
Despite the fact that the VBT appears to have a field for specifying
that a system is equipped with a panel that supports standard VESA
backlight controls over the DP AUX channel, so far every system we've
spotted DPCD backlight control support on doesn't actually set this
field correctly and all have it set to INTEL_BACKLIGHT_DISPLAY_DDI.
While we don't know the exact reason for this VBT misuse, talking with
some vendors indicated that there's a good number of laptop panels out
there that supposedly support both PWM backlight controls and DPCD
backlight controls as a workaround until Intel supports DPCD backlight
controls across platforms universally. This being said, the X1 Extreme
2nd Gen that I have here (note that Lenovo is not the hardware vendor
that informed us of this) PWM backlight controls are advertised, but
only DPCD controls actually function. I'm going to make an educated
guess here and say that on systems like this one, it's likely that PWM
backlight controls might have been intended to work but were never
really tested by QA.
Since we really need backlights to work without any extra module
parameters, let's take the risk here and rely on the standard DPCD caps
to tell us whether AUX backlight controls are supported or not. We still
check the VBT, just so we can print a debugging message on systems that
advertise DPCD backlight support on the panel but not in the VBT.
Changes since v3:
* Print a debugging message if we enable DPCD backlight control on a
device which doesn't report DPCD backlight controls in it's VBT,
instead of warning on custom panel backlight interfaces.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112376
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Perry Yuan <pyuan@redhat.com>
Cc: AceLan Kao <acelan.kao@canonical.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200117232155.135579-1-lyude@redhat.com
For a simulated preemption reset, we don't populate the request and so
do not fill in the guilty context name.
[ 79.991294] i915 0000:00:02.0: GPU HANG: ecode 9:1:e757fefe, in [0]
Just don't mention the empty string in the logs!
Fixes: 742379c0c4 ("drm/i915: Start chopping up the GPU error capture")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200121132107.267709-1-chris@chris-wilson.co.uk
Eliminate the inconsistencies in the hdcp code local variables:
- use dev_priv over dev
- use to_i915() instead of dev->dev_private
- initialize variables when declaring them
- a bit of declaration suffling to appease ocd
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191204180549.1267-10-ville.syrjala@linux.intel.com
Reviewed-by: Ramalingam C <ramalingam.c@intel.com>
Report port presence based on port presence in VBT alone, relaxing the
requirements on supported encoders (DP, DVI, or HDMI). The goal is to
make future changes easier, however there is a small risk of reporting
more ports present than before in case of dubious VBT.
Regarding the current callers of intel_bios_is_port_present(), the
potential issue might be caused by DVO_PORT_CRT being identified as port
E in dvo_port_to_port(). Hopefully no VBT has that on SKL+ which support
DP/DVI/HDMI on port E; the current CRT init code on HSW/BDW does not
care.
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/4338a29e4ed49e69f859dff1490fd85f6ae6177e.1579270868.git.jani.nikula@intel.com
