Joonas Lahtinen [Mon, 23 Oct 2017 15:32:09 +0000 (18:32 +0300)]
drm/i915: Disable lazy PPGTT page table optimization for vGPU
When running under virtualization (vGPU active), we must disable
the lazy PPGTT page table initialization optimization introduced by
commit 14826673247e ("drm/i915: Only initialize partially filled
pagetables").
We must do this because GVT-g makes unduly assumptions about guest
behaviour, which this optimization breaks. This results in following
looking errors in the host:
ERROR gvt: guest page write error -22, gfn 0x7ada8, pa 0x7ada89a8, var 0x6, len 1
The real fix is to not to depend on i915 driver behaviour, but instead
either rely on only the contracts that i915 has with the hardware, or
add some paravirtualization. While the real fix is en route, it won't
be finished in time for 4.15, so the best option is to disable the
optimization for now when vGPU is active to avoid breaking 4.15 guests
in existing VM environments.
Fixes: 14826673247e ("drm/i915: Only initialize partially filled pagetables") Suggested-by: Xiaolin Zhang <xiaolin.zhang@intel.com> Signed-off-by: Xiaolin Zhang <xiaolin.zhang@intel.com>
[Joonas: Rewrote the commit message and added tags.] Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Zhenyu Wang <zhenyuw@linux.intel.com> Cc: Zhi Wang <zhi.a.wang@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Acked-by: Chris Wilson <chris@chris-wilson.co.uk> Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171023153209.10527-1-joonas.lahtinen@linux.intel.com
In file included from drivers/gpu/drm/i915/i915_sw_fence.c:517:0:
drivers/gpu/drm/i915/selftests/lib_sw_fence.c: In function ‘timed_fence_init’:
drivers/gpu/drm/i915/selftests/lib_sw_fence.c:63:2: error: implicit declaration of function ‘timer_setup_on_stack’; did you mean ‘hrtimer_init_on_stack’? [-Werror=implicit-function-declaration]
timer_setup_on_stack(&tf->timer, timed_fence_wake, 0);
Kees Cook [Tue, 24 Oct 2017 15:13:44 +0000 (08:13 -0700)]
drm/i915/selftests: Convert timers to use timer_setup()
In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.
Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: David Airlie <airlied@linux.ie> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: intel-gfx@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org Signed-off-by: Kees Cook <keescook@chromium.org> Link: https://patchwork.freedesktop.org/patch/msgid/20171024151344.GA104417@beast Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Tue, 24 Oct 2017 22:08:55 +0000 (23:08 +0100)]
drm/i915: Use same test for eviction and submitting kernel context
During evict, we wish to idle the GPU if we see that the GGTT is full.
However, our test for idle in i915_gem_evict_something() and in
i915_gem_switch_to_kernel_context() do not match leading to
disappointment - we never believe that we are idle and keep trying to
flush the GGTT ad infinitum.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103438 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171024220855.30155-2-chris@chris-wilson.co.uk Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Tue, 24 Oct 2017 22:08:54 +0000 (23:08 +0100)]
drm/i915/selftests: Don't try to queue a request with zero delay
Instead of trying to create a timer with zero delay (i.e. with expires
set to the current jiffies and not the future, an already expired
timer), execute that request immediately.
v2: Refactor list_del_init+signal into its own little function.
v3: Reorder testing so as not to immediately signal a delayed request.
Chris Wilson [Tue, 24 Oct 2017 20:50:53 +0000 (21:50 +0100)]
drm/i915: Call cond_resched() before repeating i915_gem_evict_something()
Insert a breakpoint, a chance to escape back to the scheduler and run
something else for a bit, if we find that the GGTT is full and needs to
be idled in order to make some room. In practice, this should only be an
issue in stress tests as the wait itself will normally give the chance
for the scheduler to intervene and make progress.
Ville Syrjälä [Tue, 24 Oct 2017 09:52:16 +0000 (12:52 +0300)]
drm/i915: Perform a central cdclk state sanity check
WARN if the cdclk state doesn't match what we expect after programming.
And let's remove the WARN from bdw_set_cdclk() that's trying to achieve
the same thing in a more limite fashion.
Also take the opportunity to refactor the code to use a common function
for dumping out a cdclk state.
Ville Syrjälä [Tue, 24 Oct 2017 09:52:14 +0000 (12:52 +0300)]
drm/i915: Adjust system agent voltage on CNL if required by DDI ports
On CNL we may need to bump up the system agent voltage not only due
to CDCLK but also when driving DDI port with a sufficiently high clock.
To that end start tracking the minimum acceptable voltage for each crtc.
We do the tracking via crtcs because we don't have any kind of encoder
state. Also there's no downside to doing it this way, and it matches how
we track cdclk requirements on account of pixel rate.
v2: Allow disabled crtcs to use the min voltage
Add IS_CNL check to intel_ddi_compute_min_voltage() since
we're using CNL specific values there
s/intel_compute_min_voltage/cnl_compute_min_voltage/ since
the function makes hw specific assumptions about the voltage
values
v3: Drop the test hack leftovers from skl_modeset_calc_cdclk()
v4: s/voltage/voltage_level/ (Rodrigo)
Replace DPLL DVFS FIXMEs with an explanation why we don't
do anything there (Rodrigo)
Ville Syrjälä [Tue, 24 Oct 2017 09:52:13 +0000 (12:52 +0300)]
drm/i915: Use cdclk_state->voltage on CNL
Track the system agent voltage we request from pcode in the cdclk state
on CNL. Annoyingly we can't actually read out the current value since
there's no pcode command to do that, so we'll have to just assume that
it worked.
Ville Syrjälä [Tue, 24 Oct 2017 09:52:12 +0000 (12:52 +0300)]
drm/i915: Use cdclk_state->voltage on BXT/GLK
Track the system agent voltage we request from pcode in the cdclk state
on BXT/GLK. Annoyingly we can't actually read out the current value since
there's no pcode command to do that, so we'll have to just assume that
it worked.
Ville Syrjälä [Tue, 24 Oct 2017 09:52:11 +0000 (12:52 +0300)]
drm/i915: Use cdclk_state->voltage on SKL/KBL/CFL
Track the system agent voltage we request from pcode in the cdclk state
on SKL/KBL/CFL. Annoyingly we can't actually read out the current value since
there's no pcode command to do that, so we'll have to just assume that
it worked.
Ville Syrjälä [Tue, 24 Oct 2017 09:52:10 +0000 (12:52 +0300)]
drm/i915: Use cdclk_state->voltage on BDW
Track the system agent voltage we request from pcode in the cdclk state
on BDW. Annoyingly we can't actually read out the current value since
there's no pcode command to do that, so we'll have to just assume that
it worked.
v2: Keep the WARN_ON (Rodrigo)
v3: s/voltage/voltage_level/ (Rodrigo)
Ville Syrjälä [Tue, 24 Oct 2017 09:52:09 +0000 (12:52 +0300)]
drm/i915: Use cdclk_state->voltage on VLV/CHV
Store the punit DSPFREQUAR value into cdclk_state->voltage on
VLV/CHV. Since we can actually read that out from the hardware
this can give us a bit more cross checking between the hardware
and software state.
v2: Don't break waiting for cdclk change on VLV/CHV
v3: Split out the cdclk sanity check in vlv_set_cdclk() (Rodrigo)
v4: s/voltage/voltage_level/ (Rodrigo)
Ville Syrjälä [Tue, 24 Oct 2017 09:52:08 +0000 (12:52 +0300)]
drm/i915: Start tracking voltage level in the cdclk state
For CNL we'll need to start considering the port clocks when we select
the voltage level for the system agent. To that end start tracking the
voltage in the cdclk state (since that already has to adjust it).
v2: - Tune Render and Media wake rate values according to some extra
info I got from HW engineers. Value can be tuned, but for now
these are the recommended values.
- Fix typos pointed by James.
Cc: Nathan Ciobanu <nathan.d.ciobanu@intel.com> Cc: Wayne Boyer <wayne.boyer@intel.com> Cc: Joe Konno <joe.konno@linux.intel.com> Cc: David Weinehall <david.weinehall@linux.intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by: James Ausmus <james.ausmus@intel.com> Reviewed-by: David Weinehall <david.weinehall@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171023224612.27208-1-rodrigo.vivi@intel.com
Rodrigo Vivi [Mon, 23 Oct 2017 17:39:20 +0000 (10:39 -0700)]
drm/i915/cnl: Force DDI_A_4_LANES when needed.
As we faced in BXT, on CNL DDI_A_4_LANES is not
set as expected when system is boot with multiple
monitors connected. This result in wrong lane
setup impacting the max data rate available and
consequently blocking modeset on eDP, resulting
in a blank screen.
Most of CNL SKUs don't support DDI-E.
The only SKU that supports DDI-E is the same
that supports the full A/E split called DDI-F.
Also when DDI-F is used DDI-E cannot be used because
they share Interrupts. So DDI-E is almost useless.
Anyways let's consider this is possible and rely on
VBT for that.
This patch was initialy start by Clint, but required
many changes including full commit message. So
Credits entirely to Clint for finding this.
v2: Extract all messy conditions into a helper function
as suggested by Ville.
Along with simplification I removed the debug
message on the working case since now all conditions
are grouped.
v3: Split the conditions even more as suggested by Ville.
Get's cleaner and easier to add new cases in the
future.
Suggested-by: Clint Taylor <clinton.a.taylor@intel.com> Cc: Clint Taylor <clinton.a.taylor@intel.com> Cc: Mika Kahola <mika.kahola@intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171023173920.22890-1-rodrigo.vivi@intel.com
drm/i915/perf: fix perf enable/disable ioctls with 32bits userspace
The compat callback was missing and triggered failures in 32bits
userspace when enabling/disable the perf stream. We don't require any
particular processing here as these ioctls don't take any argument.
Chris Wilson [Tue, 24 Oct 2017 11:55:01 +0000 (12:55 +0100)]
drm/i915/execlists: Remove the priority "optimisation"
Originally we set the priority to max upon inserting the request into
the execlists queue (and removing it from the scheduler lists). We could
then use the prio==INT_MAX as a shortcut within execlists_schedule() to
detect the end of the dependency chain. Since commit 1f181225f8ec
("drm/i915/execlists: Keep request->priority for its lifetime") this is
no longer true as we use the request completion as an indicator the
schedule dependency chain is complete instead. (This allows us to then
reschedule requests even when its context is in flight.) However, this
makes the GEM_BUG_ON() inside execlists_schedule() racy as we may change
the rq->prio at the same time. As the assertion is useful, let's keep
the assertion and remove the micro-optimisation.
Fixes: 1f181225f8ec ("drm/i915/execlists: Keep request->priority for its lifetime") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171024115501.21033-1-chris@chris-wilson.co.uk Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Chris Wilson [Mon, 23 Oct 2017 21:32:36 +0000 (22:32 +0100)]
drm/i915: Filter out spurious execlists context-switch interrupts
Back in commit a4b2b01523a8 ("drm/i915: Don't mark an execlists
context-switch when idle") we noticed the presence of late
context-switch interrupts. We were able to filter those out by looking
at whether the ELSP remained active, but in commit beecec901790
("drm/i915/execlists: Preemption!") that became problematic as we now
anticipate receiving a context-switch event for preemption while ELSP
may be empty. To restore the spurious interrupt suppression, add a
counter for the expected number of pending context-switches and skip if
we do not need to handle this interrupt to make forward progress.
v2: Don't forget to switch on for preempt.
v3: Reduce the counter to a on/off boolean tracker. Declare the HW as
active when we first submit, and idle after the final completion event
(with which we confirm the HW says it is idle), and track each source
of activity separately. With a finite number of sources, it should aide
us in debugging which gets stuck.
Fixes: beecec901790 ("drm/i915/execlists: Preemption!") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michal Winiarski <michal.winiarski@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171023213237.26536-3-chris@chris-wilson.co.uk Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Chris Wilson [Mon, 23 Oct 2017 21:32:35 +0000 (22:32 +0100)]
drm/i915: Synchronize irq before parking each engine
When we park the engine (upon idling), we kill the irq tasklet. However,
to be sure that it is not restarted by a spurious interrupt after doing so,
flush the interrupt handler before parking. As we only park the engines
when we believe the system is idle, there should not be any interrupts to
distrub us; so flushing the final in-flight interrupt should be sufficient.
(However, we are still dependent on the HW behaving in an orderly and
timely fashion, which we shall endeavour to improve upon later.)
Chris Wilson [Mon, 23 Oct 2017 21:32:34 +0000 (22:32 +0100)]
drm/i915: Bump wait-times for the final CS interrupt before parking
In the idle worker we drop the prolonged GT wakeref used to cover such
essentials as interrupt delivery. (When a CS interrupt arrives, we also
assert that the GT is awake.) However, it turns out that 10ms is not
long enough to be assured that the last CS interrupt has been delivered,
so bump that to 200ms, and move the entirety of that wait to before we
take the struct_mutex to avoid blocking. As this is now a potentially
long wait, restore the earlier behaviour of bailing out early when a new
request arrives.
v2: Break out the repeated check for new requests into its own little
helper to try and improve the self-commentary.
v2: Move defines to a better place.
This is actually CNL_PCH not CNL only.
v3: Accepting Ville's suggestions: enums and array to
to make this future proof.
v4: Protect the array access as Ville suggested.
Also accepting all Jani's suggestions:
- use already defined gmbus pin definitions.
- use map_ddc_pin for disambiguation.
- Add /* sic */ comment on inverted values
so people can easily see it it nos a mistake
we have the map 3 -> 4 and 4 -> 3 :/
Cc: Jani Nikula <jani.nikula@intel.com> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Anusha Srivatsa <anusha.srivatsa@intel.com> Cc: Clinton Taylor <clinton.a.taylor@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171020172641.16029-1-rodrigo.vivi@intel.com
Madhav Chauhan [Fri, 13 Oct 2017 12:45:00 +0000 (18:15 +0530)]
drm/i915: Use existing DSI backlight ports info
This patch re-use already parsed DSI backlight/cabc ports
info for saving it inside struct intel_dsi rather than
parsing it at the time of DSI initialization.
V2: Remove backlight and cabc variable initialization (Jani N).
Madhav Chauhan [Fri, 13 Oct 2017 12:44:59 +0000 (18:14 +0530)]
drm/i915: Parse DSI backlight/cabc ports.
This patch parse DSI backlight/cabc ports info from
VBT and save them inside local structure. This saved info
can be directly used while initializing DSI for different
platforms instead of parsing for each platform.
V2: Changes:
- Typo fix in commit message.
- Move up newly added port variables (Jani N)
- Remove redundant initialization (Jani N)
- Don't parse CABC ports if not supported (Jani N)
V3: Patch restructure (Suggested by Jani N)
Chris Wilson [Thu, 19 Oct 2017 06:37:33 +0000 (07:37 +0100)]
drm/i915: Skip waking the device to service pwrite
If the device is in runtime suspend, resuming takes time and reduces our
powersaving. If this was for a small write into an object, that resume
will take longer than any savings in using the indirect GGTT access to
avoid the cpu cache.
Ville Syrjälä [Wed, 18 Oct 2017 18:19:58 +0000 (21:19 +0300)]
drm/i915: Drop the redundant hdmi prefix/suffix from a lot of variables
A bunch of functions are now exclusively used for HDMI, so naming the
variables with hdmi prefix/suffix is redundant. Also use int rather
than u32 for the translation level consistently.
Ville Syrjälä [Wed, 18 Oct 2017 18:19:34 +0000 (21:19 +0300)]
drm/i915: Unify error handling for missing DDI buf trans tables
Handle missing buf trans tables, or out of bounds buf trans levels
the same way everywhere. These should never be hit under normal
conditions, but let's play it safe for now.
Ville Syrjälä [Mon, 16 Oct 2017 14:57:03 +0000 (17:57 +0300)]
drm/i915: Centralize the SKL DDI A/E vs. B/C/D buf trans handling
SKL DDI B/C/D only have 9 usable buf trans registers for DP/eDP. That
matches the normal DP buf trans tables, but the low vswing eDP tables
have 10 entries. Thus the eDP tables can only be used safely with DDI A
and E.
We try to catch cases where DDI B/C/D gets used with the wrong number of
entires in some parts of the code, but not everywhere. Let's move the
code to deal with that deeper into intel_ddi_get_buf_trans_edp(). And
for sake of symmetry do the same in intel_ddi_get_buf_trans_dp(). That
would also avoid explosions in the rather unlikely case that the DP
tables would get revised to 10 entries as well.
Ville Syrjälä [Mon, 16 Oct 2017 14:57:02 +0000 (17:57 +0300)]
drm/i915: Kill off the BXT buf_trans default_index
default_index contained in the BXT buf_trans tables is actually useless.
For DP we should always have a valid level selected (otherwise the link
training logic would be buggy), and for HDMI we can just do what the
other platforms do and pick the correct entry in intel_ddi_hdmi_level().
We'll want to use the intel_ddi_get_buf_trans_*() functions a bit
earlier in the file, so move them up. While at it start using them
in the iboost setup to get rid of the platform checks there.
Chris Wilson [Wed, 18 Oct 2017 12:16:21 +0000 (13:16 +0100)]
drm/i915: Flush the idle-worker for debugfs/i915_drop_caches
After being requested to idle the GPU, flush the idle worker to drop the
residual active state, and any internal object caches.
v2: By popular demand, introduce DROP_IDLE for fine-grained control from
userspace, though it should be used as part of a
DROP_ACTIVE | DROP_RETIRE | DROP_IDLE | DROP_FREED
sequence.
v3: Convert to BIT() to sell it to Joonas.
drm/i915: adjust get_crtc_fence_y_offset() to use base.y instead of crtc.y
This is to use clipped y coordinate here. I left get_crtc_fence_y_offset()
function itself in place as oneliner to maintain comment above it why this
is done.
Kees Cook [Tue, 17 Oct 2017 06:53:04 +0000 (09:53 +0300)]
drm/i915: Convert timers to use timer_setup()
In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.
Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: David Airlie <airlied@linux.ie> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Oscar Mateo <oscar.mateo@intel.com> Cc: intel-gfx@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org Signed-off-by: Kees Cook <keescook@chromium.org> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171017065304.3358-1-joonas.lahtinen@linux.intel.com
Harsha Sharma [Sat, 14 Oct 2017 18:36:44 +0000 (00:06 +0530)]
drm/i915: Replace *_reference/unreference() or *_ref/unref with _get/put()
Replace instances of drm_framebuffer_reference/unreference() with
*_get/put() suffixes and drm_dev_unref with *_put() suffix
because get/put is shorter and consistent with the
kernel use of *_get/put suffixes.
Done with following coccinelle semantic patch
Oscar Mateo [Tue, 17 Oct 2017 20:25:45 +0000 (13:25 -0700)]
drm/i915: Use a mask when applying WaProgramL3SqcReg1Default
Otherwise we are blasting other bits in GEN8_L3SQCREG1 that might be important
(although we probably aren't at the moment because 0 seems to be the default
for all the other bits).
v2: Extra parentheses (Michel)
Fixes: 050fc46 ("drm/i915:bxt: implement WaProgramL3SqcReg1DefaultForPerf") Fixes: 450174f ("drm/i915/chv: Tune L3 SQC credits based on actual latencies") Signed-off-by: Oscar Mateo <oscar.mateo@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Imre Deak <imre.deak@intel.com> Reviewed-by: Michel Thierry <michel.thierry@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/1508271945-14961-1-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Oscar Mateo [Tue, 17 Oct 2017 20:27:51 +0000 (13:27 -0700)]
drm/i915: No need for RING_MAX_NONPRIV_SLOTS space
Now that we write RING_FORCE_TO_NONPRIV registers directly to hardware,
[commit 32ced39 ("drm/i915: Transform whitelisting WAs into a simple reg
write")] there is no need to save space for them in the list of context
workarounds.
v2: Refer to previous commit in commit message (Michel)
Chris Wilson [Tue, 17 Oct 2017 10:37:23 +0000 (11:37 +0100)]
drm/i915/selftests: Silence the compiler for impossible errors
It should be impossible for these tests not to run due to an empty
ppgtt, but if it should happen, let's report ENODEV (our typical
internal error for impossible events).
In file included from drivers/gpu/drm/i915/i915_gem.c:5415:
drivers/gpu/drm/i915/selftests/huge_pages.c: In function 'igt_mock_ppgtt_huge_fill':
>> drivers/gpu/drm/i915/selftests/huge_pages.c:612: error: 'err' may be used uninitialized in this function
drivers/gpu/drm/i915/selftests/huge_pages.c: In function 'igt_ppgtt_exhaust_huge':
drivers/gpu/drm/i915/selftests/huge_pages.c:1159: error: 'err' may be used uninitialized in this function
Chris Wilson [Mon, 16 Oct 2017 20:27:32 +0000 (21:27 +0100)]
drm/i915: Report -EFAULT before pwrite fast path into shmemfs
When pwriting into shmemfs, the fast path pagecache_write does not
notice when it is writing to beyond the end of the truncated shmemfs
inode. Report -EFAULT directly when we try to use pwrite into the
!I915_MADV_WILLNEED object.
Fixes: 7c55e2c5772d ("drm/i915: Use pagecache write to prepopulate shmemfs from pwrite-ioctl")
Testcase: igt/gem_madvise/dontneed-before-pwrite Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.william.auld@gmail.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171016202732.25459-1-chris@chris-wilson.co.uk
Ville Syrjälä [Fri, 13 Oct 2017 19:40:51 +0000 (22:40 +0300)]
drm/i915: Remove mostly duplicated video DIP handling from PSR code
Now that the infoframe hooks are part of the intel_dig_port, we can use
the normal .write_infoframe() hook to update the VSC SDP. We do need to
deal with the size difference between the VSC DIP and the others though.
Another minor snag is that the compiler will complain to use if we keep
using enum hdmi_infoframe_type type and passing in the DP define instead,
so et's just change to unsigned int all over for the inforframe type.
Dave Airlie [Thu, 12 Oct 2017 23:52:01 +0000 (00:52 +0100)]
drm/plane: drop num_overlay_planes (v3)
In order to implement plane leasing we need to count things,
just make the code consistent with the counting code currently
used for counting crtcs/encoders/connectors and drop the need
for num_overlay_planes.
v2: don't forget to assign plane_ptr. (keithp)
v3: use correct bounds check, found by igt.
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Tue, 17 Oct 2017 00:13:47 +0000 (10:13 +1000)]
Merge tag 'tilcdc-4.15' of https://github.com/jsarha/linux into drm-next
tilcdc changes for v4.15
* tag 'tilcdc-4.15' of https://github.com/jsarha/linux:
drm/tilcdc: Remove redundant OF_DETACHED flag setting
drm/tilcdc: Precalculate total frametime in tilcdc_crtc_set_mode()
drm/tilcdc: Use tilcdc_crtc_shutdown() in tilcdc_crtc_destroy()
drm/tilcdc: Remove WARN_ON(!drm_modeset_is_locked(&crtc->mutex)) checks
drm/tilcdc: Turn raster off in crtc reset, if it was on in the HW
drm/tilcdc: switch to drm_*{get,put} helpers
drm/tilcdc: tilcdc_tfp410: make of_device_ids const.
drm/tilcdc: tilcdc_panel: make of_device_ids const.
Dave Airlie [Tue, 17 Oct 2017 00:10:17 +0000 (10:10 +1000)]
Merge tag 'drm-misc-next-2017-10-16' of git://anongit.freedesktop.org/drm/drm-misc into drm-next
Quick 4.15 misc pull for the build fix:
Cross-subsystem Changes:
- piles an piles of misc/trivial patches all over, some more from
outreachy applicants
Core Changes:
- build fix for the bridge/of cleanup (Maarten)
- fix vblank count in arm_vblank_event (Ville)
- some kerneldoc typo fixes from Thierry
Driver Changes:
- vc4: Fix T-format tiling scanout, cleanup clock divider w/a (Anholt)
- sun4i: small cleanups and improved code comments all over (Chen-Yu
Tsai)
* tag 'drm-misc-next-2017-10-16' of git://anongit.freedesktop.org/drm/drm-misc: (21 commits)
drm/via: use ARRAY_SIZE
drm/gma500: use ARRAY_SIZE
drm/sun4i: hdmi: Move PAD_CTRL1 setting to mode_set function
drm/sun4i: hdmi: Document PAD_CTRL1 output invert bits
drm/sun4i: backend: Add comment explaining why registers are cleared
drm/sun4i: backend: Use drm_fb_cma_get_gem_addr() to get display memory
drm/sun4i: backend: Create regmap after access is possible
drm/sun4i: don't add components that are already in the queue
drm/vc4: Fix pitch setup for T-format scanout.
drm/vc4: Move the DSI clock divider workaround closer to the clock call.
drm: Replace kzalloc with kcalloc
drm/tinydrm: Remove explicit .best_encoder assignment
drm/tinydrm: Replace dev_error with DRM_DEV_ERROR
drm/drm_of: Move drm_of_panel_bridge_remove_function into header.
drm/atomic-helper: Fix reference to drm_crtc_send_vblank_event()
drm/atomic-helper: Fix typo
drm: Add missing __user annotation to drm_syncobj_array_find()
drm/rockchip: add PINCTRL dependency for LVDS
drm/kirin: Checking for IS_ERR() instead of NULL
driver:gpu: return -ENOMEM on allocation failure.
...
Rodrigo Vivi [Tue, 3 Oct 2017 22:08:59 +0000 (15:08 -0700)]
drm/i915/cnl: Fix PLL initialization for HDMI.
HDMI Mode selection on CNL is on CFGCR0 for that PLL, not
on in a global CTRL1 as it was on SKL.
The original patch addressed this difference, but leaving behind
this single entry here. So we were checking the wrong bits during
the PLL initialization and consequently avoiding the CFGCR1 setup
during HDMI initialization. Luckly when only HDMI was in use BIOS
had already setup this for us. But the dual display with hot plug
were messed up.
Rodrigo Vivi [Tue, 3 Oct 2017 22:08:58 +0000 (15:08 -0700)]
drm/i915/cnl: Fix PLL mapping.
On PLL Enable sequence we need to "Configure DPCLKA_CFGCR0 to turn on
the clock for the DDI and map the DPLL to the DDI"
So we first do the map and then we unset DDI_CLK_OFF to turn the clock
on. We do this in 2 separated steps.
However, on this second step where we should only unset the off bit we are
also unmapping the ddi from the pll. So we end up using the pll 0
for almost everything. Consequently breaking cases with more than one
display.
Fixes: 555e38d27317 ("drm/i915/cnl: DDI - PLL mapping") Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Kahola, Mika <mika.kahola@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by: James Ausmus <james.ausmus@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171003220859.21352-2-rodrigo.vivi@intel.com
Chris Wilson [Sun, 15 Oct 2017 20:43:10 +0000 (21:43 +0100)]
drm/i915: Add in-flight request details to intel_engine_dump()
In the intel_engine_cs dumper, we were showing the request details for
the request queue but not of those requests already passed to the hw
(just a summary of the seqno). If we show those details, we can then
eliminate the entirely redundant and forgotten debugfs/i915_gem_request
Chris Wilson [Sun, 15 Oct 2017 14:37:25 +0000 (15:37 +0100)]
drm/i915: Skip HW reinitialisation on resume if still wedged
If we fail to recover the HW state upon resume (i.e. our attempt to
clear the wedged bit and reset during i915_gem_sanitize() fails), then
skip the HW restart inside i915_gem_init_hw(). We will ultimately do the
HW restart when successfully unwedging and resetting the HW later,
but attempting to restore a wedged device upon resume is risky as the HW
is in an unknown state.
v2: Suppress the error message when detecting the already wedged HW.
Chris Wilson [Fri, 13 Oct 2017 20:26:21 +0000 (21:26 +0100)]
drm/i915: Trim struct_mutex hold duration for i915_gem_free_objects
We free objects in bulk after they wait for their RCU grace period.
Currently, we take struct_mutex and unbind all the objects. This can lead
to a long lock duration during which time those objects have their pages
unfreeable (i.e. the shrinker is prevented from reaping those pages). If
we only process a single object under the struct_mutex and then free the
pages, the number of objects locked away from the shrinker is minimal
and we allow regular clients better access to struct_mutex if they need
it.
Chris Wilson [Fri, 13 Oct 2017 20:26:20 +0000 (21:26 +0100)]
drm/i915: Only free the oldest stale object before a fresh allocation
Inspired by Tvrtko's critique of the reaping of the stale contexts
before allocating a new one, also limit the freed object reaping to the
oldest stale object before allocating a fresh object. Unlike contexts,
objects may have radically different sizes of backing storage, but
similar to contexts, while we want to prevent starvation due to
excessive freed lists, we also do not want to delay fresh allocations
for too long. Only freeing the oldest on the freed object list before
each allocation is a reasonable compromise.
v2: Only a single consumer of llist_del_first() is allowed (although
multiple llist_add are still allowed in parallel). Unlike
i915_gem_context, i915_gem_flush_free_objects() is itself not serialized
and so we need to add our own spinlock. Otherwise KASAN eventually spots
a use-after-free for the race on *first->next.
Chris Wilson [Fri, 13 Oct 2017 20:26:19 +0000 (21:26 +0100)]
drm/i915: Set our shrinker->batch to 4096 (~16MiB)
Prefer to defer activating our GEM shrinker until we have a few
megabytes to free; or we have accumulated sufficient mempressure by
deferring the reclaim to force a shrink. The intent is that because our
objects may typically be large, we are too effective at shrinking and
are not rewarded for freeing more pages than the batch. It will also
defer the initial shrinking to hopefully put it at a lower priority than
say the buffer cache (although it will balance out over a number of
reclaims, with GEM being more bursty).
v2: Give it a feedback system to try and tune the batch size towards
an effective size for the available objects.
v3: Start keeping track of shrinker stats in debugfs
v4: Protect against finding no shrinkable objects (div-by-zero)
Chris Wilson [Fri, 13 Oct 2017 20:26:18 +0000 (21:26 +0100)]
drm/i915: Wire up shrinkctl->nr_scanned
shrink_slab() allows us to report back the number of objects we
successfully scanned (out of the target shrinkctl->nr_to_scan). As
report the number of pages owned by each GEM object as a separate item
to the shrinker, we cannot precisely control the number of shrinker
objects we scan on each pass; and indeed may free more than requested.
If we fail to tell the shrinker about the number of objects we process,
it will continue to hold a grudge against us as any objects left
unscanned are added to the next reclaim -- and so we will keep on
"unfairly" shrinking our own slab in comparison to other slabs.
v2: fixup the misplaced addition, we want to count everything we scan
(to match the number we reported earlier) not just the objects we
successfully validated and freed.
Chris Wilson [Mon, 16 Oct 2017 11:40:37 +0000 (12:40 +0100)]
drm/i915: Move dev_priv->mm.[un]bound_list to its own lock
Remove the struct_mutex requirement around dev_priv->mm.bound_list and
dev_priv->mm.unbound_list by giving it its own spinlock. This reduces
one more requirement for struct_mutex and in the process gives us
slightly more accurate unbound_list tracking, which should improve the
shrinker - but the drawback is that we drop the retirement before
counting so i915_gem_object_is_active() may be stale and lead us to
underestimate the number of objects that may be shrunk (see commit bed50aea61df ("drm/i915/shrinker: Flush active on objects before
counting")).
v2: Crosslink the spinlock to the lists it protects, and btw this
changes s/obj->global_link/obj->mm.link/
v3: Fix decoupling of old links in i915_gem_object_attach_phys()
v3.1: Fix the fix, only unlink if it was linked
v3.2: Use a local for to_i915(obj->base.dev)->mm.obj_lock
Chris Wilson [Fri, 13 Oct 2017 20:26:16 +0000 (21:26 +0100)]
drm/i915: Remove walk over obj->vma_list for the shrinker
In the next patch, we want to reduce the lock coverage within the
shrinker, and one of the dangerous walks we have is over obj->vma_list.
We are only walking the obj->vma_list in order to check whether it has
been permanently pinned by HW access, typically via use on the scanout.
But we have a couple of other long term pins, the context objects for
which we currently have to check the individual vma pin_count. If we
instead mark these using obj->pin_display, we can forgo the dangerous
and sometimes slow list iteration.
v2: Rearrange code to try and avoid confusion from false associations
due to arrangement of whitespace along with rebasing on obj->pin_global.
Chris Wilson [Fri, 13 Oct 2017 20:26:15 +0000 (21:26 +0100)]
drm/i915: Drop debugfs/i915_gem_pin_display
It has now lost its meaning (it shows more than just pin_display), I do
not believe that we are using in preference to the complete listing from
i915_gem_gtt, or the listing from i915_gem_framebuffer, or the listing
of active display objects in i915_display_info.
Chris Wilson [Fri, 13 Oct 2017 20:26:14 +0000 (21:26 +0100)]
drm/i915: Rename obj->pin_display to obj->pin_global
In the next patch, we want to extend use of the global pin counter for
semi-permanent pinning of context/ring objects. Given that we plan to
extend the usage to encompass a disparate set of objects, we want a name
that reflects both and should entail less confusion.
Chris Wilson [Fri, 13 Oct 2017 20:26:13 +0000 (21:26 +0100)]
drm/i915: Refactor testing obj->mm.pages
Since we occasionally stuff an error pointer into obj->mm.pages for a
semi-permanent or even permanent failure, we have to be more careful and
not just test against NULL when deciding if the object has a complete
set of its concurrent pages.
Michal Wajdeczko [Mon, 16 Oct 2017 14:47:20 +0000 (14:47 +0000)]
drm/i915: Update DMC firmware load error messages
Some of the error messages from DMC load were too generic and
may be confusing for the user. Lets explicitly add DMC words there.
Also as homepage of DMC firmware is same as for the GuC and Huc,
lets reuse URL definition.
Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Imre Deak <imre.deak@intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171016144724.17244-12-michal.wajdeczko@intel.com
Michal Wajdeczko [Mon, 16 Oct 2017 14:47:17 +0000 (14:47 +0000)]
drm/i915/guc: Pick better place for Guc final status message
GuC status message printed right after firmware upload may be too
optimistic, as we may fail on subsequent steps. Move that message
to the end of intel_uc_init_hw where we know the status for sure.
v2: use dev_info (Chris)
Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171016144724.17244-9-michal.wajdeczko@intel.com
Weinan Li [Sun, 15 Oct 2017 03:55:25 +0000 (11:55 +0800)]
drm/i915: enable to read CSB and CSB write pointer from HWSP in GVT-g VM
Let GVT-g VM read the CSB and CSB write pointer from virtual HWSP, not all
the host support this feature, need to check the BIT(3) of caps in PVINFO.
v3 : Remove unnecessary comments.
v4 : Separate VM enable patch with GVT-g implementation patch due to code
dependency.
v5 : Use inline for GVT virtual HWSP caps check function.
v6 : Comments refine.
Jérémy Lefaure [Mon, 16 Oct 2017 02:33:56 +0000 (22:33 -0400)]
drm/via: use ARRAY_SIZE
Using the ARRAY_SIZE macro improves the readability of the code.
Found with Coccinelle with the following semantic patch:
@r depends on (org || report)@
type T;
T[] E;
position p;
@@
(
(sizeof(E)@p /sizeof(*E))
|
(sizeof(E)@p /sizeof(E[...]))
|
(sizeof(E)@p /sizeof(T))
)
Jérémy Lefaure [Mon, 16 Oct 2017 02:30:46 +0000 (22:30 -0400)]
drm/gma500: use ARRAY_SIZE
Using the ARRAY_SIZE macro improves the readability of the code. Also,
it is useless to re-invent it.
Found with Coccinelle with the following semantic patch:
@r depends on (org || report)@
type T;
T[] E;
position p;
@@
(
(sizeof(E)@p /sizeof(*E))
|
(sizeof(E)@p /sizeof(E[...]))
|
(sizeof(E)@p /sizeof(T))
)
Chen-Yu Tsai [Sat, 14 Oct 2017 04:02:52 +0000 (12:02 +0800)]
drm/sun4i: hdmi: Move PAD_CTRL1 setting to mode_set function
Initially we configured the PAD_CTRL1 register at probe/bind time.
However it seems the HDMI controller will modify some of the bits
in this register by itself. On the A10 it is particularly annoying
as it toggles the output invert bits, which inverts the colors on
the display output.
The U-boot driver this driver is based on sets this register twice,
though it seems it's only needed for actual display output. Hence
we move it to the mode_set function.
While debugging inverted color from the HDMI output on the A10, I
found that the lowest 3 bits were set. These were cleared on A20
boards that had normal display output. By manually toggling these
bits the mapping of the color components to these bits was found.
While these are not used anywhere, it would be nice to document
them somewhere.
Chen-Yu Tsai [Sat, 14 Oct 2017 04:02:49 +0000 (12:02 +0800)]
drm/sun4i: backend: Add comment explaining why registers are cleared
Many of the backend's layer configuration registers have undefined
default values. This poses a risk as we use regmap_update_bits in
some places, and don't overwrite the whole register.
At probe/bind time we explicitly clear all the control registers
by writing 0 to them. This patch adds a more detailed explanation
on why we're doing this.
Chen-Yu Tsai [Sat, 14 Oct 2017 04:02:48 +0000 (12:02 +0800)]
drm/sun4i: backend: Use drm_fb_cma_get_gem_addr() to get display memory
Commit 4636ce93d5b2 ("drm/fb-cma-helper: Add drm_fb_cma_get_gem_addr()")
adds a new helper, which covers fetching a drm_framebuffer's GEM object
and calculating the buffer address for a given plane.
This patch uses this helper to replace our own open coded version of the
same function.
Chen-Yu Tsai [Sat, 14 Oct 2017 04:02:46 +0000 (12:02 +0800)]
drm/sun4i: don't add components that are already in the queue
Even though the components framework can handle duplicate entries,
the extra entries cause a lot more debug messages to be generated,
which would be confusing to developers not familiar with our driver
and the framework in general.
Instead, we can scan the relatively small queue and check if the
component to be added is already queued up. Since the display
pipelines are symmetrical (not considering the third display
pipeline on the A80), and we add components level by level, when
we get to the second instance at the same level, any shared downstream
components would already be in the queue.
projects / mirror_ubuntu-eoan-kernel.git / log