openwrt/staging/blogic.git
5 years agodrm/i915: Split engine setup/init into two phases
Chris Wilson [Fri, 26 Apr 2019 16:33:33 +0000 (17:33 +0100)]
drm/i915: Split engine setup/init into two phases

In the next patch, we require the engine vfuncs setup prior to
initialising the pinned kernel contexts, so split the vfunc setup from
the engine initialisation and call it earlier.

v2: s/setup_xcs/setup_common/ for intel_ring_submission_setup()

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/20190426163336.15906-6-chris@chris-wilson.co.uk
5 years agodrm/i915: Pass intel_context to intel_context_pin_lock()
Chris Wilson [Fri, 26 Apr 2019 16:33:32 +0000 (17:33 +0100)]
drm/i915: Pass intel_context to intel_context_pin_lock()

Move the intel_context_instance() to the caller so that we can decouple
ourselves from one context instance per engine.

v2: Rename pin_lock() to lock_pinned(), hopefully that is clearer.

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/20190426163336.15906-5-chris@chris-wilson.co.uk
5 years agodrm/i915/selftests: Pass around intel_context for sseu
Chris Wilson [Fri, 26 Apr 2019 16:33:31 +0000 (17:33 +0100)]
drm/i915/selftests: Pass around intel_context for sseu

Combine the (i915_gem_context, intel_engine) into a single parameter,
the intel_context for convenience and later simplification.

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/20190426163336.15906-4-chris@chris-wilson.co.uk
5 years agodrm/i915/selftests: Use the real kernel context for sseu isolation tests
Chris Wilson [Fri, 26 Apr 2019 16:33:30 +0000 (17:33 +0100)]
drm/i915/selftests: Use the real kernel context for sseu isolation tests

Simply the setup slightly for the sseu selftests to use the actual
kernel_context.

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/20190426163336.15906-3-chris@chris-wilson.co.uk
5 years agodrm/i915: Export intel_context_instance()
Chris Wilson [Fri, 26 Apr 2019 16:33:29 +0000 (17:33 +0100)]
drm/i915: Export intel_context_instance()

We want to pass in a intel_context into intel_context_pin() and that
requires us to first be able to lookup the intel_context!

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/20190426163336.15906-2-chris@chris-wilson.co.uk
5 years agodrm/i915/gvt: Pin the per-engine GVT shadow contexts
Chris Wilson [Fri, 26 Apr 2019 16:33:28 +0000 (17:33 +0100)]
drm/i915/gvt: Pin the per-engine GVT shadow contexts

Our eventual goal is to rid request construction of struct_mutex, with
the short term step of lifting the struct_mutex requirements into the
higher levels (i.e. the caller must ensure that the context is already
pinned into the GTT). In this patch, we pin GVT's shadow context upon
allocation and so keep them pinned into the GGTT for as long as the
virtual machine is alive, and so we can use the simpler request
construction path safe in the knowledge that the hard work is already
done.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426163336.15906-1-chris@chris-wilson.co.uk
5 years agoMerge drm/drm-next into drm-intel-next-queued
Jani Nikula [Fri, 26 Apr 2019 15:22:43 +0000 (18:22 +0300)]
Merge drm/drm-next into drm-intel-next-queued

Get gvt-fixes back to dinq.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
5 years agodrm/i915: Clean up cherryview_load_luts()
Ville Syrjälä [Mon, 8 Apr 2019 12:18:15 +0000 (15:18 +0300)]
drm/i915: Clean up cherryview_load_luts()

I like my functions simple, so split up the low level bits from
cherryview_load_luts() into separate functions. Also rename the
whole thing to chv_load_luts() to match the new world order.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190408121815.30142-1-ville.syrjala@linux.intel.com
Reviewed-by: Swati Sharma <swati2.sharma@intel.com>
5 years agodrm/i915: Fix ICL output CSC programming
Ville Syrjälä [Thu, 25 Apr 2019 19:24:19 +0000 (22:24 +0300)]
drm/i915: Fix ICL output CSC programming

When I refactored the code into its own function I accidentally
misplaced the <<16 shifts for some of the registers causing us
to lose the blue channel entirely.

We should really find a way to test this...

Cc: Uma Shankar <uma.shankar@intel.com>
Fixes: d2c19b06d6ea ("drm/i915: Clean up ilk/icl pipe/output CSC programming")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190425192419.24931-1-ville.syrjala@linux.intel.com
Reviewed-by: Swati Sharma <swati2.sharma@intel.com>
5 years agodrm/i915: Enable render context support for gen4 (Broadwater to Cantiga)
Chris Wilson [Fri, 19 Apr 2019 17:27:20 +0000 (18:27 +0100)]
drm/i915: Enable render context support for gen4 (Broadwater to Cantiga)

Broadwater and the rest of gen4  do support being able to saving and
reloading context specific registers between contexts, providing isolation
of the basic GPU state (as programmable by userspace). This allows
userspace to assume that the GPU retains their state from one batch to the
next, minimising the amount of state it needs to reload and manually save
across batches.

v2: CONSTANT_BUFFER woes

Running through piglit turned up an interesting issue, a GPU hang inside
the context load. The context image includes the CONSTANT_BUFFER command
that loads an address into a on-gpu buffer, and the context load was
executing that immediately. However, since it was reading from the GTT
there is no guarantee that the GTT retains the same configuration as
when the context was saved, resulting in stray reads and a GPU hang.

Having tried issuing a CONSTANT_BUFFER (to disable the command) from the
ring before saving the context to no avail, we resort to patching out
the instruction inside the context image before loading.

This does impose that gen4 always reissues CONSTANT_BUFFER commands on
each batch, but due to the use of a shared GTT that was and will remain
a requirement.

v3: ECOSKPD to the rescue

Ville found the magic bit in the ECOSKPD to disable saving and restoring
the CONSTANT_BUFFER from the context image, thereby completely avoiding
the GPU hangs from chasing invalid pointers. This appears to be the
default behaviour for gen5, and so we just need to tweak gen4 to match.

v4: Fix spelling of ECOSKPD and discover it already exists

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20190419172720.5462-1-chris@chris-wilson.co.uk
5 years agodrm/i915: Enable render context support for Ironlake (gen5)
Chris Wilson [Fri, 19 Apr 2019 11:17:48 +0000 (12:17 +0100)]
drm/i915: Enable render context support for Ironlake (gen5)

Ironlake does support being able to saving and reloading context specific
registers between contexts, providing isolation of the basic GPU state
(as programmable by userspace). This allows userspace to assume that the
GPU retains their state from one batch to the next, minimising the
amount of state it needs to reload, or manually save and restore.

v2: Fix off-by-one in reading CXT_SIZE, and add a comment that the
CXT_SIZE and context-layout do not match in bspec, but the difference is
irrelevant as we overallocate the full page anyway (Ville).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190419111749.3910-2-chris@chris-wilson.co.uk
5 years agodrm/i915/ringbuffer: EMIT_INVALIDATE *before* switch context
Chris Wilson [Fri, 19 Apr 2019 11:17:47 +0000 (12:17 +0100)]
drm/i915/ringbuffer: EMIT_INVALIDATE *before* switch context

Despite what I think the prm recommends, commit f2253bd9859b
("drm/i915/ringbuffer: EMIT_INVALIDATE after switch context") turned out
to be a huge mistake when enabling Ironlake contexts as the GPU would
hang on either a MI_FLUSH or PIPE_CONTROL immediately following the
MI_SET_CONTEXT of an active mesa context (more vanilla contexts, e.g.
simple rendercopies with igt, do not suffer).

Ville found the following clue,

  "[DevCTG+]: For the invalidate operation of the pipe control, the
   following pointers are affected. The
   invalidate operation affects the restore of these packets. If the pipe
   control invalidate operation is completed
   before the context save, the indirect pointers will not be restored from
   memory.
   1. Pipeline State Pointer
   2. Media State Pointer
   3. Constant Buffer Packet"

which suggests by us emitting the INVALIDATE prior to the MI_SET_CONTEXT,
we prevent the context-restore from chasing the dangling pointers within
the image, and explains why this likely prevents the GPU hang.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190419111749.3910-1-chris@chris-wilson.co.uk
5 years agodrm/i915: Move sandybride pcode access to intel_sideband.c
Chris Wilson [Fri, 26 Apr 2019 08:17:25 +0000 (09:17 +0100)]
drm/i915: Move sandybride pcode access to intel_sideband.c

sandybride_pcode is another sideband, so move it to their new home.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426081725.31217-8-chris@chris-wilson.co.uk
5 years agodrm/i915: Merge sandybridge_pcode_(read|write)
Chris Wilson [Fri, 26 Apr 2019 08:17:24 +0000 (09:17 +0100)]
drm/i915: Merge sandybridge_pcode_(read|write)

These routines are identical except in the nature of the value parameter.
For writes it is a pure in-param, but for a read, we need an out-param.
Since they differ in a single line, merge the two routines into one.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426081725.31217-7-chris@chris-wilson.co.uk
5 years agodrm/i915: Merge sbi read/write into a single accessor
Chris Wilson [Fri, 26 Apr 2019 08:17:23 +0000 (09:17 +0100)]
drm/i915: Merge sbi read/write into a single accessor

Since intel_sideband_read and intel_sideband_write differ by only a
couple of lines (depending on whether we feed the value in or out),
merge the two into a single common accessor.

v2: Restore vlv_flisdsi_read() lost during rebasing.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426081725.31217-6-chris@chris-wilson.co.uk
5 years agodrm/i915: Separate sideband declarations to intel_sideband.h
Chris Wilson [Fri, 26 Apr 2019 08:17:22 +0000 (09:17 +0100)]
drm/i915: Separate sideband declarations to intel_sideband.h

Split the sideback declarations out of the ginormous i915_drv.h

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426081725.31217-5-chris@chris-wilson.co.uk
5 years agodrm/i915: Replace pcu_lock with sb_lock
Chris Wilson [Fri, 26 Apr 2019 08:17:21 +0000 (09:17 +0100)]
drm/i915: Replace pcu_lock with sb_lock

We now have two locks for sideband access. The general one covering
sideband access across all generation, sb_lock, and a specific one
covering sideband access via the punit on vlv/chv. After lifting the
sb_lock around the punit into the callers, the pcu_lock is now redudant
and can be separated from its other use to regulate RPS (essentially
giving RPS a lock all of its own).

v2: Extract a couple of minor bug fixes.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426081725.31217-4-chris@chris-wilson.co.uk
5 years agodrm/i915: Lift sideband locking for vlv_punit_(read|write)
Chris Wilson [Fri, 26 Apr 2019 08:17:20 +0000 (09:17 +0100)]
drm/i915: Lift sideband locking for vlv_punit_(read|write)

Lift the sideband acquisition for vlv_punit_read and vlv_punit_write
into their callers, so that we can lock the sideband once for a sequence
of operations, rather than perform the heavyweight acquisition on each
request.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426081725.31217-3-chris@chris-wilson.co.uk
5 years agodrm/i915: Lift acquiring the vlv punit magic to a common sb-get
Chris Wilson [Fri, 26 Apr 2019 08:17:19 +0000 (09:17 +0100)]
drm/i915: Lift acquiring the vlv punit magic to a common sb-get

As we now employ a very heavy pm_qos around the punit access, we want to
minimise the number of synchronous requests by performing one for the
whole punit sequence rather than around individual accesses. The
sideband lock is used for this, so push the pm_qos into the sideband
lock acquisition and release, moving it from the lowlevel punit rw
routine to the callers. In the first step, we move the punit magic into
the common sideband lock so that we can acquire a bunch of ports
simultaneously, and if need be extend the workaround protection later.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426081725.31217-2-chris@chris-wilson.co.uk
5 years agodrm/i915: Disable preemption and sleeping while using the punit sideband
Chris Wilson [Fri, 26 Apr 2019 08:17:18 +0000 (09:17 +0100)]
drm/i915: Disable preemption and sleeping while using the punit sideband

While we talk to the punit over its sideband, we need to prevent the cpu
from sleeping in order to prevent a potential machine hang.

Note that by itself, it appears that pm_qos_update_request (via
intel_idle) doesn't provide a sufficient barrier to ensure that all core
are indeed awake (out of Cstate) and that the package is awake. To do so,
we need to supplement the pm_qos with a manual ping on_each_cpu.

v2: Restrict the heavy-weight wakeup to just the ISOF_PORT_PUNIT, there
is insufficient evidence to implicate a wider problem atm. Similarly,
restrict the w/a to Valleyview, as Cherryview doesn't have an angry cadre
of users.

The working theory, courtesy of Ville and Hans, is the issue lies within
the power delivery and so is likely to be unit and board specific and
occurs when both the unit/fw require extra power at the same time as the
cpu package is changing its own power state.

References: https://bugzilla.kernel.org/show_bug.cgi?id=109051
References: https://bugs.freedesktop.org/show_bug.cgi?id=102657
References: https://bugzilla.kernel.org/show_bug.cgi?id=195255
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190426081725.31217-1-chris@chris-wilson.co.uk
5 years agoMerge tag 'drm-intel-next-fixes-2019-04-25' of git://anongit.freedesktop.org/drm...
Dave Airlie [Fri, 26 Apr 2019 01:35:58 +0000 (11:35 +1000)]
Merge tag 'drm-intel-next-fixes-2019-04-25' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

- Use after free fix during GEM_CREATE when reporting back object size
- Icelake DP register programming order fix

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190425061312.GA2919@jlahtine-desk.ger.corp.intel.com
5 years agoMerge tag 'drm-misc-next-fixes-2019-04-24' of git://anongit.freedesktop.org/drm/drm...
Dave Airlie [Fri, 26 Apr 2019 01:35:40 +0000 (11:35 +1000)]
Merge tag 'drm-misc-next-fixes-2019-04-24' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

- fb_helper: Fix NULL deref in legacy drivers (Noralf)
- leases: Ensure lessees can't connect to objects outside their perview (Daniel)
- leases: Enforce that lessees hold the lease for implicitly set planes (Daniel)
- leases: A few non-functional cleanups (Daniel)

Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Sean Paul <sean@poorly.run>
Link: https://patchwork.freedesktop.org/patch/msgid/20190424210604.GA32581@art_vandelay
5 years agodrm/i915: Allow multiple user handles to the same VM
Chris Wilson [Thu, 25 Apr 2019 05:43:33 +0000 (06:43 +0100)]
drm/i915: Allow multiple user handles to the same VM

It was noted that we made the same mistake for VM_ID as for object
handles, whereby we ensured that we only allocated a single handle for
one ppgtt. This has the unfortunate consequence for userspace that they
need to reference count the handles to avoid destroying an active ID. If
we allow multiple handles to the same ppgtt, userspace can freely
unreference any handle they own without fear of destroying the same
handle in use elsewhere.

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/20190425054333.27299-1-chris@chris-wilson.co.uk
5 years agodrm/i915: Explicitly pin the logical context for execbuf
Chris Wilson [Thu, 25 Apr 2019 05:01:43 +0000 (06:01 +0100)]
drm/i915: Explicitly pin the logical context for execbuf

In order to separate the reservation phase of building a request from
its emission phase, we need to pull some of the request alloc activities
from deep inside i915_request to the surface, GEM_EXECBUFFER.

v2: Be frivolous, use a local drm_i915_private.

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/20190425050143.811-1-chris@chris-wilson.co.uk
5 years agodrm/i915: Invert the GEM wakeref hierarchy
Chris Wilson [Wed, 24 Apr 2019 20:07:17 +0000 (21:07 +0100)]
drm/i915: Invert the GEM wakeref hierarchy

In the current scheme, on submitting a request we take a single global
GEM wakeref, which trickles down to wake up all GT power domains. This
is undesirable as we would like to be able to localise our power
management to the available power domains and to remove the global GEM
operations from the heart of the driver. (The intent there is to push
global GEM decisions to the boundary as used by the GEM user interface.)

Now during request construction, each request is responsible via its
logical context to acquire a wakeref on each power domain it intends to
utilize. Currently, each request takes a wakeref on the engine(s) and
the engines themselves take a chipset wakeref. This gives us a
transition on each engine which we can extend if we want to insert more
powermangement control (such as soft rc6). The global GEM operations
that currently require a struct_mutex are reduced to listening to pm
events from the chipset GT wakeref. As we reduce the struct_mutex
requirement, these listeners should evaporate.

Perhaps the biggest immediate change is that this removes the
struct_mutex requirement around GT power management, allowing us greater
flexibility in request construction. Another important knock-on effect,
is that by tracking engine usage, we can insert a switch back to the
kernel context on that engine immediately, avoiding any extra delay or
inserting global synchronisation barriers. This makes tracking when an
engine and its associated contexts are idle much easier -- important for
when we forgo our assumed execution ordering and need idle barriers to
unpin used contexts. In the process, it means we remove a large chunk of
code whose only purpose was to switch back to the kernel context.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190424200717.1686-5-chris@chris-wilson.co.uk
5 years agodrm/i915: Pass intel_context to i915_request_create()
Chris Wilson [Wed, 24 Apr 2019 20:07:16 +0000 (21:07 +0100)]
drm/i915: Pass intel_context to i915_request_create()

Start acquiring the logical intel_context and using that as our primary
means for request allocation. This is the initial step to allow us to
avoid requiring struct_mutex for request allocation along the
perma-pinned kernel context, but it also provides a foundation for
breaking up the complex request allocation to handle different scenarios
inside execbuf.

For the purpose of emitting a request from inside retirement (see the
next patch for engine power management), we also need to lift control
over the timeline mutex to the caller.

v2: Note that the request carries the active reference upon construction.

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/20190424200717.1686-4-chris@chris-wilson.co.uk
5 years agodrm/i915: Introduce context->enter() and context->exit()
Chris Wilson [Wed, 24 Apr 2019 20:07:15 +0000 (21:07 +0100)]
drm/i915: Introduce context->enter() and context->exit()

We wish to start segregating the power management into different control
domains, both with respect to the hardware and the user interface. The
first step is that at the lowest level flow of requests, we want to
process a context event (and not a global GEM operation). In this patch,
we introduce the context callbacks that in future patches will be
redirected to per-engine interfaces leading to global operations as
required.

The intent is that this will be guarded by the timeline->mutex, except
that retiring has not quite finished transitioning over from being
guarded by struct_mutex. So at the moment it is protected by
struct_mutex with a reminded to switch.

v2: Rename default handlers to intel_context_enter_engine.

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/20190424200717.1686-3-chris@chris-wilson.co.uk
5 years agodrm/i915: Pull the GEM powermangement coupling into its own file
Chris Wilson [Wed, 24 Apr 2019 20:07:14 +0000 (21:07 +0100)]
drm/i915: Pull the GEM powermangement coupling into its own file

Split out the powermanagement portion (GT wakeref, suspend/resume) of
GEM from i915_gem.c into its own file.

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/20190424200717.1686-2-chris@chris-wilson.co.uk
5 years agodrm/i915: Introduce struct intel_wakeref
Chris Wilson [Wed, 24 Apr 2019 20:07:13 +0000 (21:07 +0100)]
drm/i915: Introduce struct intel_wakeref

For controlling runtime pm of the GT and engines, we would like to have
a callback to do extra work the first time we wake up and the last time
we drop the wakeref. This first/last access needs serialisation and so
we encompass a mutex with the regular intel_wakeref_t tracker.

v2: Drop the _once naming and report the errors.

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/20190424200717.1686-1-chris@chris-wilson.co.uk
5 years agodrm/i915: Move GraphicsTechnology files under gt/
Chris Wilson [Wed, 24 Apr 2019 17:48:39 +0000 (18:48 +0100)]
drm/i915: Move GraphicsTechnology files under gt/

Start partitioning off the code that talks to the hardware (GT) from the
uapi layers and move the device facing code under gt/

One casualty is s/intel_ringbuffer.h/intel_engine.h/ with the plan to
subdivide that header and body further (and split out the submission
code from the ringbuffer and logical context handling). This patch aims
to be simple motion so git can fixup inflight patches with little mess.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190424174839.7141-1-chris@chris-wilson.co.uk
5 years agodrm/i915/selftests: Verify whitelist of context registers
Chris Wilson [Wed, 24 Apr 2019 11:09:41 +0000 (12:09 +0100)]
drm/i915/selftests: Verify whitelist of context registers

The RING_NONPRIV allows us to add registers to a whitelist that allows
userspace to modify them. Ideally such registers should be safe and
saved within the context such that they do not impact system behaviour
for other users. This selftest verifies that those registers we do add
are (a) then writable by userspace and (b) only affect a single client.

Opens:
- Is GEN9_SLICE_COMMON_ECO_CHICKEN1 really write-only?

v2: Remove the blatant copy-paste.
v3: Emulate userspace register writes via the batch again.

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/20190424110941.9869-1-chris@chris-wilson.co.uk
5 years agodrm/i915: Store the default sseu setup on the engine
Chris Wilson [Wed, 24 Apr 2019 09:51:34 +0000 (10:51 +0100)]
drm/i915: Store the default sseu setup on the engine

As we push for better compartmentalisation, it is more convenient to
copy the default sseu configuration from the engine into the derived
logical context, than it is to dig it out from i915->runtime_info.

v2: Use intel_sseu_from_device_info() to describe the converter

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/20190424095134.30249-1-chris@chris-wilson.co.uk
5 years agodrm/fb-helper: Fix drm_fb_helper_firmware_config() NULL pointer deref
Noralf Trønnes [Tue, 23 Apr 2019 14:53:53 +0000 (16:53 +0200)]
drm/fb-helper: Fix drm_fb_helper_firmware_config() NULL pointer deref

Non-atomic drivers like ast doesn't have connector->state set resulting
in a NULL pointer deref:

[   29.609593] BUG: unable to handle kernel NULL pointer dereference at 0000000000000010
[   29.609619] Call Trace:
[   29.609630]  ? drm_helper_probe_single_connector_modes+0x27f/0x680
[   29.609640]  drm_setup_crtcs+0x431/0xd80 [drm_kms_helper]
[   29.753065]  __drm_fb_helper_initial_config_and_unlock+0x6f/0x6a0
[   29.753160]  ? drm_modeset_unlock_all+0x31/0x50 [drm]
[   29.765758]  ast_fbdev_init+0xa8/0xc0 [ast]
[   29.765762]  ast_driver_load.cold.7+0x2b3/0xe11 [ast]
[   29.765775]  drm_dev_register+0x111/0x150 [drm]

Fix by bailing out if the driver does not support atomic modesetting.

Fixes: 09ded8af57bc ("drm/i915/fbdev: Move intel_fb_initial_config() to fbdev helper")
Reported-by: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Tested-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20190423145353.30158-1-noralf@tronnes.org
5 years agodrm/atomic: -EACCESS for lease-denied crtc lookup
Daniel Vetter [Thu, 28 Feb 2019 14:49:10 +0000 (15:49 +0100)]
drm/atomic: -EACCESS for lease-denied crtc lookup

With the previous patch drm_crtc_find will return NULL when the crtc
isn't in our lease, which will then disable the plane/connector. No
longer an issue since the lessor can't escape their lease terms
anymore, but not quite great semantics yet either.

Catch this and return -EACCES, so that at least evil test cases have a
better chance of making sure the kernel works correctly.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228144910.26488-8-daniel.vetter@ffwll.ch
5 years agodrm/atomic: Wire file_priv through for property changes
Daniel Vetter [Thu, 28 Feb 2019 14:49:09 +0000 (15:49 +0100)]
drm/atomic: Wire file_priv through for property changes

We need this to make sure lessees can only connect their
plane/connectors to crtc objects they own. And note that this is
irrespective of whether the lessor is atomic or not, lessor cannot
prevent lessees from enabling atomic.

Cc: stable@vger.kernel.org
Cc: Keith Packard <keithp@keithp.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228144910.26488-7-daniel.vetter@ffwll.ch
5 years agodrm/lease: Make sure implicit planes are leased
Daniel Vetter [Thu, 28 Feb 2019 14:49:08 +0000 (15:49 +0100)]
drm/lease: Make sure implicit planes are leased

If userspace doesn't enable universal planes, then we automatically
add the primary and cursor planes. But for universal userspace there's
no such check (and maybe we only want to give the lessee one plane,
maybe not even the primary one), hence we need to check for the
implied plane.

v2: don't forget setcrtc ioctl.

v3: Still allow disabling of the crtc in SETCRTC.

Cc: stable@vger.kernel.org
Cc: Keith Packard <keithp@keithp.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228144910.26488-6-daniel.vetter@ffwll.ch
5 years agodrm/lease: Check for lessor outside of locks
Daniel Vetter [Thu, 28 Feb 2019 14:49:07 +0000 (15:49 +0100)]
drm/lease: Check for lessor outside of locks

The lessor is invariant over a lifetime of a lease, we don't have to
grab any locks for that. Speeds up the common case of not being a lease.

Cc: Keith Packard <keithp@keithp.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228144910.26488-5-daniel.vetter@ffwll.ch
5 years agodrm/leases: Don't init to 0 in drm_master_create
Daniel Vetter [Thu, 28 Feb 2019 14:49:06 +0000 (15:49 +0100)]
drm/leases: Don't init to 0 in drm_master_create

We kzalloc.

Cc: Keith Packard <keithp@keithp.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228144910.26488-4-daniel.vetter@ffwll.ch
5 years agodrm/lease: Drop recursive leads checks
Daniel Vetter [Thu, 28 Feb 2019 14:49:05 +0000 (15:49 +0100)]
drm/lease: Drop recursive leads checks

We disallow subleasing, so no point checking whether the master holds
all the leases - it will.

Spotted while typing exhaustive igt coverage for all these corner
cases.

Cc: Keith Packard <keithp@keithp.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228144910.26488-3-daniel.vetter@ffwll.ch
5 years agodrm/leases: Drop object_id validation for negative ids
Daniel Vetter [Thu, 28 Feb 2019 14:49:04 +0000 (15:49 +0100)]
drm/leases: Drop object_id validation for negative ids

Not exactly sure what's the aim here, but the canonical nil object has
id == 0, we don't use negative object ids for anything. Plus all
object_id are valided by the object_idr, there's nothing we need to do
on top of that ENOENT check a bit further down.

Spotted while typing exhaustive igt coverage for all these
corner-cases.

Cc: Keith Packard <keithp@keithp.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228144910.26488-2-daniel.vetter@ffwll.ch
5 years agoMerge branch 'drm-legacy-cleanup' of git://people.freedesktop.org/~airlied/linux...
Daniel Vetter [Wed, 24 Apr 2019 07:31:31 +0000 (09:31 +0200)]
Merge branch 'drm-legacy-cleanup' of git://people.freedesktop.org/~airlied/linux into drm-next

Pull legacy cleanups from Dave.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Dave Airlie <airlied@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/CAPM=9tztr1GoR0gr1CXPv8FsAXE4iuoRZDWKYovtnb6oDGF-Lg@mail.gmail.com
5 years agodrm/fb: revert the i915 Actually configure untiled displays from master
Dave Airlie [Wed, 24 Apr 2019 06:33:29 +0000 (16:33 +1000)]
drm/fb: revert the i915 Actually configure untiled displays from master

This code moved in here in master, so revert it the same way.

This is the same revert as 9fa246256e09 ("Revert "drm/i915/fbdev:
Actually configure untiled displays"") in drm-fixes.

Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/i915/icl: Fix MG_DP_MODE() register programming
Imre Deak [Fri, 19 Apr 2019 07:10:26 +0000 (10:10 +0300)]
drm/i915/icl: Fix MG_DP_MODE() register programming

Fix the order of lane, port parameters passed to the register macro.

Note that this was already partly fixed by commit
37fc7845df7b6 ("drm/i915: Call MG_DP_MODE() macro with the right parameters order")

While at it simplify things by using the macro directly instead of an
unnecessary redirection via an array.

v2:
- Add a note the commit message about simplifying things. (José)

Fixes: 58106b7d816e1 ("drm/i915: Make MG PHY macros semantically consistent")
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Aditya Swarup <aditya.swarup@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190419071026.32370-1-imre.deak@intel.com
(cherry picked from commit 9c11b12184bb01d8ba2c48e655509b184f02c769)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
5 years agodrm/i915: Avoid use-after-free in reporting create.size
Chris Wilson [Wed, 17 Apr 2019 13:25:07 +0000 (14:25 +0100)]
drm/i915: Avoid use-after-free in reporting create.size

We have to avoid chasing after a userspace race!

<3>[  473.114328] BUG: KASAN: use-after-free in i915_gem_create+0x1d2/0x1f0 [i915]
<3>[  473.114389] Read of size 8 at addr ffff88815bf1d840 by task gem_flink_race/1541

<4>[  473.114464] CPU: 1 PID: 1541 Comm: gem_flink_race Tainted: G     U            5.1.0-rc4-g7d07e025e786-kasan_88+ #1
<4>[  473.114469] Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./J4205-ITX, BIOS P1.10 09/29/2016
<4>[  473.114474] Call Trace:
<4>[  473.114488]  dump_stack+0x7c/0xbb
<4>[  473.114612]  ? i915_gem_create+0x1d2/0x1f0 [i915]
<4>[  473.114621]  print_address_description+0x65/0x270
<4>[  473.114728]  ? i915_gem_create+0x1d2/0x1f0 [i915]
<4>[  473.114839]  ? i915_gem_create+0x1d2/0x1f0 [i915]
<4>[  473.114848]  kasan_report+0x149/0x18d
<4>[  473.114962]  ? i915_gem_create+0x1d2/0x1f0 [i915]
<4>[  473.115069]  i915_gem_create+0x1d2/0x1f0 [i915]
<4>[  473.115176]  ? i915_gem_object_create.part.28+0x4b0/0x4b0 [i915]
<4>[  473.115289]  ? i915_gem_dumb_create+0x1a0/0x1a0 [i915]
<4>[  473.115297]  drm_ioctl_kernel+0x192/0x260
<4>[  473.115306]  ? drm_ioctl_permit+0x280/0x280
<4>[  473.115326]  drm_ioctl+0x67c/0x960
<4>[  473.115438]  ? i915_gem_dumb_create+0x1a0/0x1a0 [i915]
<4>[  473.115448]  ? drm_getstats+0x20/0x20
<4>[  473.115459]  ? __lock_acquire+0xa66/0x3fe0
<4>[  473.115474]  ? _raw_spin_unlock_irqrestore+0x39/0x60
<4>[  473.115485]  ? debug_object_active_state+0x2ea/0x4e0
<4>[  473.115496]  ? debug_show_all_locks+0x2d0/0x2d0
<4>[  473.115513]  do_vfs_ioctl+0x18d/0xfa0
<4>[  473.115522]  ? check_flags.part.27+0x440/0x440
<4>[  473.115532]  ? ioctl_preallocate+0x1a0/0x1a0
<4>[  473.115547]  ? __fget+0x2ac/0x410
<4>[  473.115561]  ? __ia32_sys_dup3+0xb0/0xb0
<4>[  473.115569]  ? rwlock_bug.part.0+0x90/0x90
<4>[  473.115590]  ksys_ioctl+0x35/0x70
<4>[  473.115597]  ? lockdep_hardirqs_off+0x1cb/0x2b0
<4>[  473.115608]  __x64_sys_ioctl+0x6a/0xb0
<4>[  473.115614]  ? lockdep_hardirqs_on+0x342/0x590
<4>[  473.115623]  do_syscall_64+0x97/0x400
<4>[  473.115633]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
<4>[  473.115641] RIP: 0033:0x7fce590d55d7
<4>[  473.115649] Code: b3 66 90 48 8b 05 b1 48 2d 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 81 48 2d 00 f7 d8 64 89 01 48
<4>[  473.115655] RSP: 002b:00007fce4d525ba8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
<4>[  473.115662] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fce590d55d7
<4>[  473.115667] RDX: 00007fce4d525c10 RSI: 00000000c010645b RDI: 0000000000000007
<4>[  473.115672] RBP: 00007fce4d525c10 R08: 00007fce4d526700 R09: 00007fce4d526700
<4>[  473.115677] R10: 0000000000000054 R11: 0000000000000246 R12: 00000000c010645b
<4>[  473.115682] R13: 0000000000000007 R14: 0000000000000000 R15: 00007ffe0e4a7450

<3>[  473.115731] Allocated by task 1541:
<4>[  473.115766]  kmem_cache_alloc+0xce/0x290
<4>[  473.115895]  i915_gem_object_create.part.28+0x1c/0x4b0 [i915]
<4>[  473.116000]  i915_gem_create+0xe3/0x1f0 [i915]
<4>[  473.116008]  drm_ioctl_kernel+0x192/0x260
<4>[  473.116013]  drm_ioctl+0x67c/0x960
<4>[  473.116020]  do_vfs_ioctl+0x18d/0xfa0
<4>[  473.116026]  ksys_ioctl+0x35/0x70
<4>[  473.116032]  __x64_sys_ioctl+0x6a/0xb0
<4>[  473.116038]  do_syscall_64+0x97/0x400
<4>[  473.116044]  entry_SYSCALL_64_after_hwframe+0x49/0xbe

<3>[  473.116071] Freed by task 1542:
<4>[  473.116101]  kmem_cache_free+0xb7/0x2f0
<4>[  473.116205]  __i915_gem_free_objects+0x7d4/0xe10 [i915]
<4>[  473.116311]  i915_gem_create_ioctl+0xaa/0xd0 [i915]
<4>[  473.116318]  drm_ioctl_kernel+0x192/0x260
<4>[  473.116323]  drm_ioctl+0x67c/0x960
<4>[  473.116330]  do_vfs_ioctl+0x18d/0xfa0
<4>[  473.116335]  ksys_ioctl+0x35/0x70
<4>[  473.116341]  __x64_sys_ioctl+0x6a/0xb0
<4>[  473.116347]  do_syscall_64+0x97/0x400
<4>[  473.116354]  entry_SYSCALL_64_after_hwframe+0x49/0xbe

Testcase: igt/gem_flink_race/flink_close
Fixes: e163484afa8d ("drm/i915: Update size upon return from GEM_CREATE")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190417132507.27133-1-chris@chris-wilson.co.uk
(cherry picked from commit 99534023490686ce4453c45e5cb813535b9bff95)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
5 years agodrm/udl: move to embedding drm device inside udl device.
Dave Airlie [Fri, 5 Apr 2019 03:17:15 +0000 (13:17 +1000)]
drm/udl: move to embedding drm device inside udl device.

This should help with some of the lifetime issues, and move us away
from load/unload.

Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190405031715.5959-4-airlied@gmail.com
5 years agodrm/udl: introduce a macro to convert dev to udl.
Dave Airlie [Fri, 5 Apr 2019 03:17:14 +0000 (13:17 +1000)]
drm/udl: introduce a macro to convert dev to udl.

This just makes it easier to later embed drm into udl.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190405031715.5959-3-airlied@gmail.com
5 years agoMerge tag 'exynos-drm-next-for-v5.2' of git://git.kernel.org/pub/scm/linux/kernel...
Dave Airlie [Wed, 24 Apr 2019 03:09:50 +0000 (13:09 +1000)]
Merge tag 'exynos-drm-next-for-v5.2' of git://git./linux/kernel/git/daeinki/drm-exynos into drm-next

Log cleanups
- Correct the use of log macro in error case.
- Drop unnecessary messages.
- Replace DRM_ERROR/DEBUG with DRM_DEV_ERROR/DEBUG.
- Print out debug messages with correct device name in vidi and ipp drivers.

One trivial cleanup
- Just fix checkpatch error, "foo* bar" to "foo *bar" in g2d driver.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Inki Dae <inki.dae@samsung.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1556073313-9923-1-git-send-email-inki.dae@samsung.com
5 years agodrm/legacy: remove some legacy lock struct members
Dave Airlie [Tue, 23 Apr 2019 00:01:50 +0000 (10:01 +1000)]
drm/legacy: remove some legacy lock struct members

This removes these unless legacy is enabled.

The lock count init is unneeded anyways since it's kzalloc.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/legacy: place all drm legacy members under DRM_LEGACY.
Dave Airlie [Mon, 22 Apr 2019 23:56:01 +0000 (09:56 +1000)]
drm/legacy: place all drm legacy members under DRM_LEGACY.

This places a bunch of the legacy members of drm_device into
only being there when legacy is enabled.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm: allow removal of legacy codepaths (v4.1)
Dave Airlie [Thu, 18 Apr 2019 07:10:40 +0000 (17:10 +1000)]
drm: allow removal of legacy codepaths (v4.1)

If you don't want the legacy drivers, then lets get rid of all the
legacy codepaths from the core module.

This drop the size of drm.ko for me by about 10%.
 380515    7422    4192  392129   5fbc1 ../../drm-next-build/drivers/gpu/drm/drm.ko
 351736    7298    4192  363226   58ada ../../drm-next-build/drivers/gpu/drm/drm.ko

v2: drop drm_lock as well, fix some DMA->DRM typos
v3: avoid ifdefs in mainline code
v4: rework ioctl defs
v4.1: fix nouveau Kconfig

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/legacy: don't include any of ati_pcigart in legacy. (v2)
Dave Airlie [Mon, 22 Apr 2019 23:55:20 +0000 (09:55 +1000)]
drm/legacy: don't include any of ati_pcigart in legacy. (v2)

This could probably be done with Kconfig somehow, but I failed in my
first 2 minute attempt.

v2: use Kconfig better.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/legacy: move legacy dev reinit into legacy misc
Dave Airlie [Mon, 22 Apr 2019 23:48:13 +0000 (09:48 +1000)]
drm/legacy: move legacy dev reinit into legacy misc

This moves the legacy dev reinit into the legacy misc file.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/legacy: move init/destroy of struct members into legacy file
Dave Airlie [Mon, 22 Apr 2019 23:47:27 +0000 (09:47 +1000)]
drm/legacy: move init/destroy of struct members into legacy file

This introduces drm_legacy_misc.c as a place for some misc legacy code,
eventually I want to give the option to remove this from the build.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/legacy: move map_hash create/destroy into inlines
Dave Airlie [Mon, 22 Apr 2019 23:40:54 +0000 (09:40 +1000)]
drm/legacy: move map_hash create/destroy into inlines

This allows them to be removed later.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/legacy: move lock cleanup for master into lock file (v2)
Dave Airlie [Mon, 22 Apr 2019 23:36:54 +0000 (09:36 +1000)]
drm/legacy: move lock cleanup for master into lock file (v2)

This makes it easier to remove legacy code later.

v2: move check into lock file as well.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/radeon: drop unused ati pcigart include.
Dave Airlie [Mon, 22 Apr 2019 23:33:56 +0000 (09:33 +1000)]
drm/radeon: drop unused ati pcigart include.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/legacy: move map cleanups into drm_bufs.c
Dave Airlie [Mon, 22 Apr 2019 22:45:12 +0000 (08:45 +1000)]
drm/legacy: move map cleanups into drm_bufs.c

This makes it easier to clean this up later.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/legacy: move drm_legacy_master_rmmaps to non-driver legacy header.
Dave Airlie [Mon, 22 Apr 2019 22:42:23 +0000 (08:42 +1000)]
drm/legacy: move drm_legacy_master_rmmaps to non-driver legacy header.

This isn't used by drivers, and won't be in the future.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/nouveau: add kconfig option to turn off nouveau legacy contexts. (v3)
Dave Airlie [Thu, 18 Apr 2019 06:45:15 +0000 (16:45 +1000)]
drm/nouveau: add kconfig option to turn off nouveau legacy contexts. (v3)

There was a nouveau DDX that relied on legacy context ioctls to work,
but we fixed it years ago, give distros that have a modern DDX the
option to break the uAPI and close the mess of holes that legacy
context support is.

Full context of the story:

commit 0e975980d435d58df2d430d688b8c18778b42218
Author: Peter Antoine <peter.antoine@intel.com>
Date:   Tue Jun 23 08:18:49 2015 +0100

    drm: Turn off Legacy Context Functions

    The context functions are not used by the i915 driver and should not
    be used by modeset drivers. These driver functions contain several bugs
    and security holes. This change makes these functions optional can be
    turned on by a setting, they are turned off by default for modeset
    driver with the exception of the nouvea driver that may require them with
    an old version of libdrm.

    The previous attempt was

    commit 7c510133d93dd6f15ca040733ba7b2891ed61fd1
    Author: Daniel Vetter <daniel.vetter@ffwll.ch>
    Date:   Thu Aug 8 15:41:21 2013 +0200

        drm: mark context support as a legacy subsystem

    but this had to be reverted

    commit c21eb21cb50d58e7cbdcb8b9e7ff68b85cfa5095
    Author: Dave Airlie <airlied@redhat.com>
    Date:   Fri Sep 20 08:32:59 2013 +1000

        Revert "drm: mark context support as a legacy subsystem"

    v2: remove returns from void function, and formatting (Daniel Vetter)

    v3:
    - s/Nova/nouveau/ in the commit message, and add references to the
      previous attempts
    - drop the part touching the drm hw lock, that should be a separate
      patch.

Signed-off-by: Peter Antoine <peter.antoine@intel.com> (v2)
Cc: Peter Antoine <peter.antoine@intel.com> (v2)
Reviewed-by: Peter Antoine <peter.antoine@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
v2: move DRM_VM dependency into legacy config.
v3: fix missing dep (kbuild robot)

Cc: stable@vger.kernel.org
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agodrm/ipp: clean up debug messages
Inki Dae [Mon, 15 Apr 2019 08:13:38 +0000 (17:13 +0900)]
drm/ipp: clean up debug messages

Print out debug messages with correct device name.

As for this, this patch adds device pointer to exynos_drm_ipp structure,
and in case of exynos_drm_ipp_task structure, replace drm_device pointer
with device one. This will make each ipp driver to print out debug
messages with correct device name.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
5 years agodrm/vidi: replace platform_device pointer with device one
Inki Dae [Mon, 15 Apr 2019 07:35:08 +0000 (16:35 +0900)]
drm/vidi: replace platform_device pointer with device one

Add device pointer to vidi_context and remove platform_device pointer.

It doesn't need for vidi_context to contain platform_device object.
Instead, this patch makes this driver more simply by replacing platform_device
pointer with device one.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
5 years agodrm/exynos: use DRM_DEV_DEBUG* instead of DRM_DEBUG macro
Inki Dae [Mon, 15 Apr 2019 07:25:12 +0000 (16:25 +0900)]
drm/exynos: use DRM_DEV_DEBUG* instead of DRM_DEBUG macro

Use DRM_DEV_DEBUG* instead of DRM_DEBUG macro to print out
debug messages.

This patch just cleans up the use of debug log macro, which changes
the log macro to DRM_DEV_DEBUG*.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
5 years agodrm/exynos: use DRM_DEV_ERROR to print out error message
Inki Dae [Mon, 15 Apr 2019 05:24:36 +0000 (14:24 +0900)]
drm/exynos: use DRM_DEV_ERROR to print out error message

This patch just cleans up the use of error log macro, which changes
the log macro to DRM_DEV_ERROR.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
5 years agodrm/exynos: remove unnecessary messages
Inki Dae [Mon, 15 Apr 2019 00:41:15 +0000 (09:41 +0900)]
drm/exynos: remove unnecessary messages

This patch removes unnecessary messages from fimd_clear_channels
and decon_clear_channels functions which print out just function
name.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
5 years agodrm/fimd: use DRM_ERROR instead of DRM_INFO in error case
Inki Dae [Mon, 15 Apr 2019 00:32:53 +0000 (09:32 +0900)]
drm/fimd: use DRM_ERROR instead of DRM_INFO in error case

This patch makes error messages to be printed out using DRM_ERROR
instead of DRM_INFO.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
5 years agodrm/exynos: g2d: remove style error
Seung-Woo Kim [Wed, 27 Mar 2019 05:07:09 +0000 (14:07 +0900)]
drm/exynos: g2d: remove style error

Remove checkpatch error, "foo* bar" should be "foo *bar".

Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
5 years agoMerge tag 'drm-msm-next-2019-04-21' of https://gitlab.freedesktop.org/drm/msm into...
Dave Airlie [Wed, 24 Apr 2019 01:54:26 +0000 (11:54 +1000)]
Merge tag 'drm-msm-next-2019-04-21' of https://gitlab.freedesktop.org/drm/msm into drm-next

This time around it is a bunch of cleanup and fixes, expanding gpu
"zap" shader support (so we can take the GPU out of secure mode on
boot) to a6xx, and small UABI extension to support robustness (see
mesa MR 673).

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rob Clark <robdclark@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/CAF6AEGsHwsEfi4y2LYKSqeqDEYvffwVgKhiP8jHcHpxp13J5LQ@mail.gmail.com
5 years agoMerge branch 'drm-next-5.2' of git://people.freedesktop.org/~agd5f/linux into drm...
Dave Airlie [Wed, 24 Apr 2019 01:24:22 +0000 (11:24 +1000)]
Merge branch 'drm-next-5.2' of git://people.freedesktop.org/~agd5f/linux into drm-next

- Add the amdgpu specific bits for timeline support
- Add internal interfaces for xgmi pstate support
- DC Z ordering fixes for planes
- Add support for NV12 planes in DC
- Add colorspace properties for planes in DC
- eDP optimizations if the GOP driver already initialized eDP
- DC bandwidth validation tracing support

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190419150034.3473-1-alexander.deucher@amd.com
5 years agoMerge tag 'drm/tegra/for-5.2-rc1' of git://anongit.freedesktop.org/tegra/linux into...
Dave Airlie [Wed, 24 Apr 2019 00:30:41 +0000 (10:30 +1000)]
Merge tag 'drm/tegra/for-5.2-rc1' of git://anongit.freedesktop.org/tegra/linux into drm-next

drm/tegra: Changes for v5.2-rc1

This contains a fix for the usage of shared resets that previously
generated a WARN on boot. In addition, there's a fix for CPU cache
maintenance of GEM buffers allocated using get_pages().

(airlied: contains a merge from a shared tegra tree)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thierry Reding <thierry.reding@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190418151447.9430-1-thierry.reding@gmail.com
5 years agoMerge tag 'drm-misc-next-2019-04-18' of git://anongit.freedesktop.org/drm/drm-misc...
Dave Airlie [Wed, 24 Apr 2019 00:12:34 +0000 (10:12 +1000)]
Merge tag 'drm-misc-next-2019-04-18' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

drm-misc-next for v5.2:

UAPI Changes:
- Document which feature flags belong to which command in virtio_gpu.h
- Make the FB_DAMAGE_CLIPS available for atomic userspace only, it's useless for legacy.

Cross-subsystem Changes:
- Add device tree bindings for lg,acx467akm-7 panel and ST-Ericsson Multi Channel Display Engine MCDE
- Add parameters to the device tree bindings for tfp410
- iommu/io-pgtable: Add ARM Mali midgard MMU page table format
- dma-buf: Only do a 64-bits seqno compare when driver explicitly asks for it, else wraparound.
- Use the 64-bits compare for dma-fence-chains

Core Changes:
- Make the fb conversion functions use __iomem dst.
- Rename drm_client_add to drm_client_register
- Move intel_fb_initial_config to core.
- Add a drm_gem_objects_lookup helper
- Add drm_gem_fence_array helpers, and use it in lima.
- Add drm_format_helper.c to kerneldoc.

Driver Changes:
- Add panfrost driver for mali midgard/bitfrost.
- Converts bochs to use the simple display type.
- Small fixes to sun4i, tinydrm, ti-fp410.
- Fid aspeed's Kconfig options.
- Make some symbols/functions static in lima, sun4i and meson.
- Add a driver for the lg,acx467akm-7 panel.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/737ad994-213d-45b5-207a-b99d795acd21@linux.intel.com
5 years agoMerge tag 'drm-intel-next-2019-04-17' of git://anongit.freedesktop.org/drm/drm-intel...
Dave Airlie [Wed, 24 Apr 2019 00:02:20 +0000 (10:02 +1000)]
Merge tag 'drm-intel-next-2019-04-17' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

UAPI Changes:

- uAPI "Fixes:" patch for the upcoming kernel 5.1, included here too

  We have an Ack from the media folks (only current user) for this
  late tweak

Cross-subsystem Changes:

- ALSA: hda: Fix racy display power access (Takashi, Chris)

Driver Changes:

- DDI and MIPI-DSI clocks fixes for Icelake (Vandita)
- Fix Icelake frequency change/locking (RPS) (Mika)
- Temporarily disable ppGTT read-only bit on Icelake (Mika)
- Add missing Icelake W/As (Mika)
- Enable 12 deep CSB status FIFO on Icelake (Mika)
- Inherit more Icelake code for Elkhartlake (Bob, Jani)

- Handle catastrophic error on engine reset (Mika)
- Shortcut readiness to reset check (Mika)
- Regression fix for GEM_BUSY causing us to report a mixed uabi-class request as not busy (Chris)
- Revert back to max link rate and lane count on eDP (Jani)
- Fix pipe BPP readout for BXT/GLK DSI (Ville)
- Set DP min_bpp to 8*3 for non-RGB output formats (Ville)
- Enable coarse preemption boundaries for Gen8 (Chris)
- Do not enable FEC without DSC (Ville)
- Restore correct BXT DDI latency optim setting calculation (Ville)
- Always reset context's RING registers to avoid running workload twice during reset (Chris)
- Set GPU wedged on driver unload (Janusz)
- Consolidate two similar barries from timeline into one (Chris)
- Only reset the pinned kernel contexts on resume (Chris)
- Wakeref tracking improvements (Chris, Imre)
- Lockdep fixes for shrinker interactions (Chris)
- Bump ready tasks ahead of busywaits in prep of semaphore use (Chris)

- Huge step in splitting display code into fine grained files (Jani)
- Refactor the IRQ init/reset macros for code saving (Paulo)
- Convert IRQ initialization code to uncore MMIO access (Paulo)
- Convert workarounds code to use uncore MMIO access (Chris)
- Nuke drm_crtc_state and use intel_atomic_state instead (Manasi)
- Update SKL clock-gating WA (Radhakrishna, Ville)
- Isolate GuC reset code flow (Chris)
- Expose force_dsc_enable through debugfs (Manasi)
- Header standalone compile testing framework (Jani)
- Code cleanups to reduce driver footprint (Chris)
- PSR code fixes and cleanups (Jose)
- Sparse and kerneldoc updates (Chris)
- Suppress spurious combo PHY B warning (Vile)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190418080426.GA6409@jlahtine-desk.ger.corp.intel.com
5 years agoMerge branch 'vmwgfx-next' of https://gitlab.freedesktop.org/drawat/linux into drm...
Dave Airlie [Tue, 23 Apr 2019 21:21:30 +0000 (07:21 +1000)]
Merge branch 'vmwgfx-next' of https://gitlab.freedesktop.org/drawat/linux into drm-next

Resource dirtying improvement by Thomas,
user-space error logging improvement and
some other minor fixes.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Deepak Rawat <drawat@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190423211630.61874-1-drawat@vmware.com
5 years agodrm/i915/icl: Fix clockgating issue when using scalers
Radhakrishna Sripada [Wed, 17 Apr 2019 18:59:01 +0000 (11:59 -0700)]
drm/i915/icl: Fix clockgating issue when using scalers

Fixes the clock-gating issue when pipe scaling is enabled.
(Lineage #2006604312)

V2: Fix typo in headline(Chris)
    Handle the non double buffered nature of the register(Ville)
V3: Fix checkpatch warning. BAT failure for V2 on gen3 looks unrelated.
V4: Split the icl and skl wa's(Ville)
V5: Split the checks for icl and skl(Ville)
V6: Correct the flipped checks in intel_pre_plane_update(Ville)
V7: Use enum for pipe and extend the WA for plane scalers(Ville)
V8: Eliminate the redundant use of pch_pfit(Ville)

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Clint Taylor <clinton.a.taylor@intel.com>
Cc: Aditya Swarup <aditya.swarup@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190417185901.14833-1-radhakrishna.sripada@intel.com
5 years agodrm/i915: Fix skl+ max plane width
Ville Syrjälä [Thu, 18 Apr 2019 19:59:07 +0000 (22:59 +0300)]
drm/i915: Fix skl+ max plane width

The spec has changed since skl_max_plane_width() was written.
Now the SKL limits are lower than what they were initially, and
GLK and ICL have different limits. Update the code to match the
spec.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190418195907.23912-1-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
5 years agodrm/i915/icl: Fix MG_DP_MODE() register programming
Imre Deak [Fri, 19 Apr 2019 07:10:26 +0000 (10:10 +0300)]
drm/i915/icl: Fix MG_DP_MODE() register programming

Fix the order of lane, port parameters passed to the register macro.

Note that this was already partly fixed by commit
37fc7845df7b6 ("drm/i915: Call MG_DP_MODE() macro with the right parameters order")

While at it simplify things by using the macro directly instead of an
unnecessary redirection via an array.

v2:
- Add a note the commit message about simplifying things. (José)

Fixes: 58106b7d816e1 ("drm/i915: Make MG PHY macros semantically consistent")
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Aditya Swarup <aditya.swarup@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190419071026.32370-1-imre.deak@intel.com
5 years agodrm/msm/a6xx: Don't enable GPU state code if dependencies are missing
Jordan Crouse [Wed, 10 Apr 2019 16:58:16 +0000 (10:58 -0600)]
drm/msm/a6xx: Don't enable GPU state code if dependencies are missing

Add CONFIG_DRM_MSM_GPU_STATE to conditionally compile Adreno GPU state
code depending on the availability of the dependencies.

Reported-by: Hulk Robot <hulkci@huawei.com>
Reported-by: YueHaibing <yuehaibing@huawei.com>
Fixes: 1707add81551 ("drm/msm/a6xx: Add a6xx gpu state")
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodt-bindings: drm/msm/gpu: Document a5xx / a6xx zap shader region
Jordan Crouse [Fri, 19 Apr 2019 19:46:16 +0000 (13:46 -0600)]
dt-bindings: drm/msm/gpu: Document a5xx / a6xx zap shader region

Describe the zap-shader node that defines a reserved memory region
to store the zap shader.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm/a6xx: Add zap shader load
Jordan Crouse [Fri, 19 Apr 2019 19:46:15 +0000 (13:46 -0600)]
drm/msm/a6xx: Add zap shader load

The a6xx GPU powers on in secure mode which restricts what memory it can
write to. To get out of secure mode the GPU driver can write to
REG_A6XX_RBBM_SECVID_TRUST_CNTL but on targets that are "secure" that
register region is blocked and writes will cause the system to go down.

For those targets we need to execute a special sequence that involves
loadinga special shader that clears the GPU registers and use a PM4
sequence to pull the GPU out of secure. Add support for loading the zap
shader and executing the secure sequence. For targets that do not support
SCM or the specific SCM sequence this should fail and we would fall back
to writing the register.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm/gpu: Move zap shader loading to adreno
Jordan Crouse [Fri, 19 Apr 2019 19:46:14 +0000 (13:46 -0600)]
drm/msm/gpu: Move zap shader loading to adreno

a5xx and a6xx both share (mostly) the same code to load the zap shader and
bring the GPU out of secure mode. Move the formerly 5xx specific code to
adreno to make it available for a6xx too.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodt-bindings: drm/msm/a6xx: Document interconnect properties for GPU
Jordan Crouse [Fri, 19 Apr 2019 19:56:27 +0000 (13:56 -0600)]
dt-bindings: drm/msm/a6xx: Document interconnect properties for GPU

Add documentation for the interconnect and interconnect-names bindings
for the GPU node as detailed by bindings/interconnect/interconnect.txt.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Georgi Djakov <georgi.djakov@linaro.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/i915: Start writeback from the shrinker
Chris Wilson [Sat, 20 Apr 2019 11:55:39 +0000 (12:55 +0100)]
drm/i915: Start writeback from the shrinker

When we are called to relieve mempressue via the shrinker, the only way
we can make progress is either by discarding unwanted pages (those
objects that userspace has marked MADV_DONTNEED) or by reclaiming the
dirty objects via swap. As we know that is the only way to make further
progress, we can initiate the writeback as we invalidate the objects.
This means the objects we put onto the inactive anon lru list are
already marked for reclaim+writeback and so will trigger a wait upon the
writeback inside direct reclaim, greatly improving the success rate of
direct reclaim on i915 objects.

The corollary is that we may start a slow swap on opportunistic
mempressure from the likes of the compaction + migration kthreads. This
is limited by those threads only being allowed to shrink idle pages, but
also that if we reactivate the page before it is swapped out by gpu
activity, we only page the cost of repinning the page. The cost is most
felt when an object is reused after mempressure, which hopefully
excludes the latency sensitive tasks (as we are just extending the
impact of swap thrashing to them).

Apparently this is not the first time we've had this idea. Back in
commit 5537252b6b6d ("drm/i915: Invalidate our pages under memory
pressure") we wanted to start writeback but settled on invalidate after
Hugh Dickins warned us about a possibility of a deadlock within shmemfs
if we started writeback from shrink_slab. Looking at the callchain,
using writeback from i915_gem_shrink should be equivalent to the pageout
also employed by shrink_slab, i.e. it should not be any riskier afaict.

v2: Leave mmapings intact. At this point, the only mmapings of our
objects will be via CPU mmaps on the shmemfs filp, which are
out-of-scope for our LRU tracking. Instead leave those pages to the
inactive anon LRU page list for aging and pageout as normal.

v3: Be selective on which paths trigger writeback, in particular
excluding paths shrinking just to reclaim vm space (e.g. mmap, vmap
reapers) and avoid starting writeback on the entire process space from
within the pm freezer.

References: https://bugs.freedesktop.org/show_bug.cgi?id=108686
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Michal Hocko <mhocko@suse.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> #v1
Link: https://patchwork.freedesktop.org/patch/msgid/20190420115539.29081-1-chris@chris-wilson.co.uk
5 years agodrm/i915/selftests: Check that gpu reset is usable from atomic context
Fernando Pacheco [Fri, 19 Apr 2019 23:00:15 +0000 (16:00 -0700)]
drm/i915/selftests: Check that gpu reset is usable from atomic context

GPU reset is now available with GuC enabled, so re-enable our check that
this reset is usable from atomic context.

Signed-off-by: Fernando Pacheco <fernando.pacheco@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/20190419230015.18121-6-fernando.pacheco@intel.com
5 years agoRevert "drm/i915/guc: Disable global reset"
Fernando Pacheco [Fri, 19 Apr 2019 23:00:14 +0000 (16:00 -0700)]
Revert "drm/i915/guc: Disable global reset"

We have now prepared the guc reset paths to avoid taking struct_mutex, or
any other lock, and so it is now safe to re-enable.

References: fe62365f9f80 ("drm/i915/guc: Disable global reset")
Signed-off-by: Fernando Pacheco <fernando.pacheco@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/20190419230015.18121-5-fernando.pacheco@intel.com
5 years agodrm/i915/uc: Place uC firmware in upper range of GGTT
Fernando Pacheco [Fri, 19 Apr 2019 23:00:13 +0000 (16:00 -0700)]
drm/i915/uc: Place uC firmware in upper range of GGTT

Currently we pin the GuC or HuC firmware image just before uploading.
Perma-pin during uC initialization instead and use the range reserved at
the top of the address space.

Moving the firmware resulted in needing to:
- use an additional pinning for the rsa signature which will be used
  during HuC auth as addresses above GUC_GGTT_TOP do not map through GTT.

v2: Remove call to set to gtt domain
    Do not restore fw gtt mapping unconditionally
    Separate out pin/unpin functions and drop usage of pin/unpin
    Use uc_fw init/fini functions to bind/unbind fw object

v3: Bind is only needed during xfer (Chris)
    Remove attempts to bind outside of xfer (Chris)
    Mark fw bind/unbind static

Signed-off-by: Fernando Pacheco <fernando.pacheco@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/20190419230015.18121-4-fernando.pacheco@intel.com
5 years agodrm/i915/uc: Reserve upper range of GGTT
Fernando Pacheco [Fri, 19 Apr 2019 23:00:12 +0000 (16:00 -0700)]
drm/i915/uc: Reserve upper range of GGTT

GuC and HuC depend on struct_mutex for device reinitialization. Moving
away from this dependency requires perma-pinning the firmware images in
GGTT.  The upper portion of the GuC address space has a sizeable hole
(several MB) that is inaccessible by GuC. Reserve this range within GGTT
as it can comfortably hold GuC/HuC firmware images.

v2: Reserve node rather than insert (Chris)
    Simpler determination of node start/size (Daniele)
    Move reserve/release out to intel_guc.* files

v3: Reserve starting at GUC_GGTT_TOP only and bail if this
    fails (Chris)

Signed-off-by: Fernando Pacheco <fernando.pacheco@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/20190419230015.18121-3-fernando.pacheco@intel.com
5 years agodrm/i915/uc: Rename uC firmware init/fini functions
Fernando Pacheco [Fri, 19 Apr 2019 23:00:11 +0000 (16:00 -0700)]
drm/i915/uc: Rename uC firmware init/fini functions

he uC firmware init function is called during GuC/HuC init early phases.
 Rename to include "_early" and properly reflect which phase we are at.

The uC firmware fini function is cleaning up the state set/created on
firmware fetch. Replace "_fini" with "_cleanup_fetch".

v2: also rename uC fw fini function

Signed-off-by: Fernando Pacheco <fernando.pacheco@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/20190419230015.18121-2-fernando.pacheco@intel.com
5 years agodrm/i915/gtt: Skip clearing the GGTT under gen6+ full-ppgtt
Chris Wilson [Fri, 19 Apr 2019 20:12:07 +0000 (21:12 +0100)]
drm/i915/gtt: Skip clearing the GGTT under gen6+ full-ppgtt

If we know that the user cannot access the GGTT, by virtue of having a
segregated memory area, we can skip clearing the unused entries as they
cannot be accessed.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
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/20190419201207.5477-1-chris@chris-wilson.co.uk
5 years agodrm/i915: remove DRM_AUTH from IOCTLs which also have DRM_RENDER_ALLOW
Christian König [Wed, 17 Apr 2019 11:25:24 +0000 (13:25 +0200)]
drm/i915: remove DRM_AUTH from IOCTLs which also have DRM_RENDER_ALLOW

This is to work around problems with libva and vainfo.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20190417112525.16848-1-christian.koenig@amd.com
5 years agodrm/i915: Expose the busyspin durations for i915_wait_request
Chris Wilson [Fri, 19 Apr 2019 18:26:25 +0000 (19:26 +0100)]
drm/i915: Expose the busyspin durations for i915_wait_request

An interesting discussion regarding "hybrid interrupt polling" for NVMe
came to the conclusion that the ideal busyspin before sleeping was half
of the expected request latency (and better if it was already halfway
through that request). This suggested that we too should look again at
our tradeoff between spinning and waiting. Currently, our spin simply
tries to hide the cost of enabling the interrupt, which is good to avoid
penalising nop requests (i.e. test throughput) and not much else.
Studying real world workloads suggests that a spin of upto 500us can
dramatically boost performance, but the suggestion is that this is not
from avoiding interrupt latency per-se, but from secondary effects of
sleeping such as allowing the CPU reduce cstate and context switch away.

In a truly hybrid interrupt polling scheme, we would aim to sleep until
just before the request completed and then wake up in advance of the
interrupt and do a quick poll to handle completion. This is tricky for
ourselves at the moment as we are not recording request times, and since
we allow preemption, our requests are not on as a nicely ordered
timeline as IO. However, the idea is interesting, for it will certainly
help us decide when busyspinning is worthwhile.

v2: Expose the spin setting via Kconfig options for easier adjustment
and testing.
v3: Don't get caught sneaking in a change to the busyspin parameters.
v4: Explain more about the "hybrid interrupt polling" scheme that we
want to migrate towards.

Suggested-by: Sagar Kamble <sagar.a.kamble@intel.com>
References: http://events.linuxfoundation.org/sites/events/files/slides/lemoal-nvme-polling-vault-2017-final_0.pdf
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Sagar Kamble <sagar.a.kamble@intel.com>
Cc: Eero Tamminen <eero.t.tamminen@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Ben Widawsky <ben@bwidawsk.net>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Sagar Kamble <sagar.a.kamble@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190419182625.11186-1-chris@chris-wilson.co.uk
5 years agodrm/msm: Split submit_lookup_objects() into two loops
Kristian H. Kristensen [Wed, 20 Mar 2019 17:09:10 +0000 (10:09 -0700)]
drm/msm: Split submit_lookup_objects() into two loops

First loop does copy_from_user() without the table lock held and
just stores the handle. Second loop looks up buffer objects with the
table_lock held without potentially blocking or faulting. This lets us
clean up a bunch of custom, non-faulting copy_from_user() code.

Signed-off-by: Kristian H. Kristensen <hoegsberg@chromium.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm: Stop dropping struct_mutex in recover_worker()
Kristian H. Kristensen [Wed, 20 Mar 2019 17:09:09 +0000 (10:09 -0700)]
drm/msm: Stop dropping struct_mutex in recover_worker()

Now that we don't have the mmap_sem lock inversion, we don't need to
jump through this particular hoop anymore.

Signed-off-by: Kristian H. Kristensen <hoegsberg@chromium.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm: Implement .gem_free_object_unlocked
Kristian H. Kristensen [Wed, 20 Mar 2019 17:09:08 +0000 (10:09 -0700)]
drm/msm: Implement .gem_free_object_unlocked

We use a llist and a worker to delay the object cleanup. This avoids
taking mmap_sem and struct_mutex in the wrong order when calling
drm_gem_object_put_unlocked() from drm_gem_mmap().

Fixes lockdep problem with copy_from_user() in msm_ioctl_gem_submit().

Signed-off-by: Kristian H. Kristensen <hoegsberg@chromium.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm/a6xx: Remove an unused struct member
Jordan Crouse [Mon, 4 Feb 2019 16:15:44 +0000 (09:15 -0700)]
drm/msm/a6xx: Remove an unused struct member

The HFI tasklet was removed in df0dff1 ("drm/msm/a6xx: Poll for HFI
responses") but the tasklet_struct was accidentally left behind.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agomsm/drm/a6xx: Turn off the GMU if resume fails
Jordan Crouse [Mon, 4 Feb 2019 16:15:43 +0000 (09:15 -0700)]
msm/drm/a6xx: Turn off the GMU if resume fails

Currently if the GMU resume function fails all we try to do is clear the
BOOT_SLUMBER oob which usually times out and ends up in a cycle of death.
If the resume function fails at any point remove any RPMh votes that might
have been added and try to shut down the GMU hardware cleanly.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm/a6xx: Make GMU reset useful
Jordan Crouse [Mon, 4 Feb 2019 16:15:42 +0000 (09:15 -0700)]
drm/msm/a6xx: Make GMU reset useful

Now that the GX domain is sorted we can wire up a working GMU reset.
IF a GMU hang was detected then try to forcefully shut down the GMU
in the power down sequence which should ensure that it can recover
normally on the next power up.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm/gpu: Attach to the GPU GX power domain
Jordan Crouse [Mon, 4 Feb 2019 16:15:41 +0000 (09:15 -0700)]
drm/msm/gpu: Attach to the GPU GX power domain

99.999% of the time during normal operation the GMU is responsible
for power and clock control on the GX domain and the CPU remains
blissfully unaware. However, there is one situation where the CPU
needs to get involved:

The power sequencing rules dictate that the GX needs to be turned
off before the CX so that the CX can be turned on before the GX
during power up. During normal operation when the CPU is taking
down the CX domain a stop command is sent to the GMU which turns
off the GX domain and then the CPU handles the CX domain.

But if the GMU happened to be unresponsive while the GX domain was
left then the CPU will need to step in and turn off the GX domain
before resetting the CX and rebooting the GMU. This unfortunately
means that the CPU needs to be marginally aware of the GX domain
even though it is expected to usually keep its hands off.

To support this we create a semi-disabled GX power domain that
does nothing to the hardware on power up but tries to shut it
down normally on power down. In this method the reference counting
is correct and we can step in with the pm_runtime_put() at the right
time during the failure path.

This patch sets up the connection to the GX power domain and does
the magic to "enable" and disable it at the right points.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodt-bindings: drm/msm/a6xx: Add GX power-domain for GMU bindings
Jordan Crouse [Mon, 4 Feb 2019 16:15:40 +0000 (09:15 -0700)]
dt-bindings: drm/msm/a6xx: Add GX power-domain for GMU bindings

The GMU should have two power domains defined: "cx" and "gx". "cx" is the
actual power domain for the device and "gx" will be attached at runtime
to manage reference counting on the GPU device in case of a GMU crash.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm/a6xx: Remove unwanted regulator code
Jordan Crouse [Mon, 4 Feb 2019 16:15:39 +0000 (09:15 -0700)]
drm/msm/a6xx: Remove unwanted regulator code

The GMU code currently has some misguided code to try to work around
a hardware quirk that requires the power domains on the GPU be
collapsed in a certain order. Upcoming patches will do this the
right way so get rid of the unused and unwanted regulator
code.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm/gpu: Add submit queue queries
Jordan Crouse [Fri, 22 Mar 2019 20:21:22 +0000 (14:21 -0600)]
drm/msm/gpu: Add submit queue queries

Add the capability to query information from a submit queue.
The first available parameter is for querying the number of GPU faults
(hangs) that can be attributed to the queue.

This is useful for implementing context robustness. A user context can
regularly query the number of faults to see if it is responsible for any
and if so it can invalidate itself.

This is also helpful for testing by confirming to the user  driver if a
particular command stream caused a fault (or not as the case may be).

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
5 years agodrm/msm: add param to retrieve # of GPU faults (global)
Rob Clark [Tue, 16 Apr 2019 23:13:28 +0000 (16:13 -0700)]
drm/msm: add param to retrieve # of GPU faults (global)

For KHR_robustness, userspace wants to know two things, the count of GPU
faults globally, and the count of faults attributed to a given context.
This patch providees the former, and the next patch provides the latter.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>