Before adding yet another copy of struct live_test and its handler,
refactor the existing code into a common framework for live selftests.
For many live selftests, we want to know if the GPU hung or otherwise
misbehaved during the execution of the test (beyond any infraction in
the behaviour under test), live_test provides this by comparing the
GPU state before and after, alerting if it unexpectedly changed (e.g.
the reset counter changed). It also ensures that the GPU is idle before
and after the test, so that residual code running on the GPU is flushed
before testing.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190121222117.23305-5-chris@chris-wilson.co.uk
Some tests (e.g. igt_vma_pin1) presume that we have a completely clean
GGTT so that it can probe boundaries without fear that something is
already allocated there. However, the mock device is starting to get
complicated and following similar rules to the live device, i.e. we
can't guarantee that i915->ggtt remains clean, so create a temporary
address_space equivalent to the mock ggtt for the purpose.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190121222117.23305-7-chris@chris-wilson.co.uk
There are 3 problems with the dsi code's pipe_bpp handling for 6 bpc
pixel-formats which this commit addresses:
1) It assumes that the pipe_bpp is the same as the bpp going over the dsi
lanes. This assumption is not valid for MIPI_DSI_FMT_RGB666, where pipe_bpp
should be 18 so that we do proper dithering but we actually send 24 bpp
over the dsi lanes (MIPI_DSI_FMT_RGB666_PACKED sends 18 bpp).
This assumption is enforced by an assert in *_dsi_get_pclk(). This assert
triggers on the initial hw-state readback on BYT/CHT devices which use
MIPI_DSI_FMT_RGB666, such as the Prowise PT301 tablet. PIPECONF is set to
6BPC / 18 bpp by the GOP, while mipi_dsi_pixel_format_to_bpp() returns 24.
This commits switches the calculations in *_dsi_get_pclk() to use the bpp
from mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format) which
returns the bpp going over the mipi lanes and drops the assert.
2) On BXT bxt_dsi_get_pipe_config() wrongly overrides the pipe_bpp which
i9xx_get_pipe_config() reads from PIPECONF with the return value from
mipi_dsi_pixel_format_to_bpp(). This avoids the assert from 1. but is wrong
since the pipe is actually running at the value configured in PIPECONF.
This commit drops the override of pipe_bpp from bxt_dsi_get_pipe_config().
3) The dsi encoder's compute_config() never assigns a value to pipe_bpp,
unlike most other encoders. Falling back on compute_baseline_pipe_bpp()
which always picks 24. 24 is only correct for MIPI_DSI_FMT_RGB88 for the
others we should use 18 bpp so that we correctly do 6bpc color dithering.
This commit adds code to intel_dsi_compute_config() to properly set
pipe_bpp based on intel_dsi->pixel_format.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181201113148.23184-1-hdegoede@redhat.com
Since commit d4ccceb055 ("drm/i915/icl: Ringbuffer interrupt handling")
we have required a mechanism to avoid touching the interrupt hardware
for breadcrumbs, superseding our mock interface for selftests.
The residual problem (ideas welcome) is in probing the mock ring
registers for ring_is_idle. Hmm, maybe we should just install
mock handlers for i915->uncore.mmio__write and friends? Only problem
being is that we would to truly mock some expected reads. :(
References: d4ccceb055 ("drm/i915/icl: Ringbuffer interrupt handling")
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/20190118112225.13780-1-chris@chris-wilson.co.uk
The motivation for introducing the check that we only enable breadcrumb
irqs if the device's irq was installed was once upon a time we waited
during suspend after disabling interrupts (which was quite slow until
the bug was discovered). Since then we have the notion of pinning the
breadcrumb irq, broadening it from the sole purpose of user interrupt
notification and waiting, and more importantly decoupling it from a very
defined time period during which enabling the irq was expected. So stop
insisting the irq is installed before we setup our IMR masks, if the IER
isn't yet enabled, nothing will happen and we will timeout instead,
revealing the lack of irq in the hang debug messages.
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/20190117233126.30165-1-chris@chris-wilson.co.uk
Something that I completely missed when implementing the new MST VCPI
atomic helpers is that with those helpers, there's technically a chance
of us having to grab additional modeset locks in ->compute_config() and
furthermore, that means we have the potential to hit a normal modeset
deadlock. However, because ->compute_config() only returns a bool this
means we can't return -EDEADLK when we need to drop locks and try again
which means we end up just failing the atomic check permanently. Whoops.
So, fix this by modifying ->compute_config() to pass down an actual
error code instead of a bool so that the atomic check can be restarted
on modeset deadlocks.
Thanks to Ville Syrjälä for pointing this out!
Changes since v1:
* Add some newlines
* Return only -EINVAL from hsw_crt_compute_config()
* Propogate return code from intel_dp_compute_dsc_params()
* Change all of the intel_dp_compute_link_config*() variants
* Don't miss if (hdmi_port_clock_valid()) branch in
intel_hdmi_compute_config()
[Cherry-picked from drm-misc-next to drm-intel-next-queued to fix
linux-next & drm-tip conflict, while waiting for proper propagation of
the DP MST series that this commit fixes. In hindsight, a topic branch
might have been a better approach for it.]
Signed-off-by: Lyude Paul <lyude@redhat.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Fixes: eceae14724 ("drm/dp_mst: Start tracking per-port VCPI allocations")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=109320
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190115200800.3121-1-lyude@redhat.com
(cherry picked from commit 96550555a7)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Acked-by: Daniel Vetter <daniel@ffwll.ch>
Use the new PM-runtime interface to get the accounted suspended time:
pm_runtime_suspended_time().
This new interface helps to simplify and cleanup the code that computes
__I915_SAMPLE_RC6_ESTIMATED and to remove direct access to internals of
PM-runtime.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Something that I completely missed when implementing the new MST VCPI
atomic helpers is that with those helpers, there's technically a chance
of us having to grab additional modeset locks in ->compute_config() and
furthermore, that means we have the potential to hit a normal modeset
deadlock. However, because ->compute_config() only returns a bool this
means we can't return -EDEADLK when we need to drop locks and try again
which means we end up just failing the atomic check permanently. Whoops.
So, fix this by modifying ->compute_config() to pass down an actual
error code instead of a bool so that the atomic check can be restarted
on modeset deadlocks.
Thanks to Ville Syrjälä for pointing this out!
Changes since v1:
* Add some newlines
* Return only -EINVAL from hsw_crt_compute_config()
* Propogate return code from intel_dp_compute_dsc_params()
* Change all of the intel_dp_compute_link_config*() variants
* Don't miss if (hdmi_port_clock_valid()) branch in
intel_hdmi_compute_config()
Signed-off-by: Lyude Paul <lyude@redhat.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Fixes: eceae14724 ("drm/dp_mst: Start tracking per-port VCPI allocations")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=109320
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190115200800.3121-1-lyude@redhat.com
Since commit 93065ac753 ("mm, oom: distinguish blockable mode for mmu
notifiers") we have been able to report failure from
mmu_invalidate_range_start which allows us to use a trylock on the
struct_mutex to avoid potential recursion and report -EBUSY instead.
Furthermore, this allows us to pull the work into the main callback and
avoid the sleight-of-hand in using a workqueue to avoid lockdep.
However, not all paths to mmu_invalidate_range_start are prepared to
handle failure, so instead of reporting the recursion, deal with it by
propagating the failure upwards, who can decide themselves to handle it
or report it.
v2: Mark up the recursive lock behaviour and comment on the various weak
points.
v3: Follow commit 3824e41975 ("drm/i915: Use mutex_lock_killable() from
inside the shrinker") and also use mutex_lock_killable().
v3.1: No leak on EINTR.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108375
References: 93065ac753 ("mm, oom: distinguish blockable mode for mmu notifiers")
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/20190115124442.3500-1-chris@chris-wilson.co.uk
This is to fix missed mmap range check on vGPU bar2 region
and only allow to map vGPU allocated GMADDR range, which means
user space should support sparse mmap to get proper offset for
mmap vGPU aperture. And this takes care of actual pgoff in mmap
request as original code always does from beginning of vGPU
aperture.
Fixes: 659643f7d8 ("drm/i915/gvt/kvmgt: add vfio/mdev support to KVMGT")
Cc: "Monroy, Rodrigo Axel" <rodrigo.axel.monroy@intel.com>
Cc: "Orrala Contreras, Alfredo" <alfredo.orrala.contreras@intel.com>
Cc: stable@vger.kernel.org # v4.10+
Reviewed-by: Hang Yuan <hang.yuan@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
VFIO region space is allocated when one region is registered for
one vgpu. So free the space when destroy the vgpu.
Also change the parameter of detach_vgpu callback to use vgpu directly.
Fixes: b851adeac0 ("drm/i915/gvt: Add opregion support")
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Hang Yuan <hang.yuan@linux.intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
On Braswell, under heavy stress, if we update the GGTT while
simultaneously accessing another region inside the GTT, we are returned
the wrong values. To prevent this we stop the machine to update the GGTT
entries so that no memory traffic can occur at the same time.
This was first spotted in
commit 5bab6f60cb
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri Oct 23 18:43:32 2015 +0100
drm/i915: Serialise updates to GGTT with access through GGTT on Braswell
but removed again in forlorn hope with
commit 4509276ee8
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon Feb 20 12:47:18 2017 +0000
drm/i915: Remove Braswell GGTT update w/a
However, gem_concurrent_blit is once again only stable with the patch
applied and CI is detecting the odd failure in forked gem_mmap_gtt tests
(which smell like the same issue). Fwiw, a wide variety of CPU memory
barriers (around GGTT flushing, fence updates, PTE updates) and GPU
flushes/invalidates (between requests, after PTE updates) were tried as
part of the investigation to find an alternate cause, nothing comes
close to serialised GGTT updates.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105591
Testcase: igt/gem_concurrent_blit
Testcase: igt/gem_mmap_gtt/*forked*
References: 5bab6f60cb ("drm/i915: Serialise updates to GGTT with access through GGTT on Braswell")
References: 4509276ee8 ("drm/i915: Remove Braswell GGTT update w/a")
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/20190114211729.30352-1-chris@chris-wilson.co.uk
