Chris Wilson [Fri, 1 Nov 2019 13:04:06 +0000 (13:04 +0000)]
drm/i915/gt: Pull timeline initialise to intel_gt_init_early
Our timelines are currently contained within an intel_gt, and we only
need to perform list/spinlock initialisation, so we can pull the
intel_timelines_init() into our intel_gt_init_early().
drm/i915: Remove special case slave handling during hw programming, v3.
Now that we split plane_state which I didn't want to do yet, we can
program the slave plane without requiring the master plane.
This is useful for programming bigjoiner slave planes as well. We
will no longer need the master's plane_state.
Changes since v1:
- set src/dst rectangles after copy_uapi_to_hw_state.
Changes since v2:
- Use the correct color_plane for pre-gen11 by using planar_linked_plane != NULL.
- Use drm_format_info_is_yuv_semiplanar in skl_plane_check() to fix gen11+.
Splitting plane state is easier than splitting crtc_state,
before plane check we copy the drm properties to hw so we can
do the same in bigjoiner later on.
We copy the state after we did all the modeset handling, but fortunately
i915 seems to be split correctly and nothing during modeset looks
at plane_state.
Changes since v1:
- Do not clear hw state on duplication.
drm/i915: Add aliases for uapi and hw to plane_state
Prepare to split up hw and uapi machinally, by adding a uapi and
hw alias. We will remove the base in a bit. This is a split from the
original uapi/hw patch, which did it all in one go.
Now that we separated everything into uapi and hw, it's
time to make the split definitive. Remove the union and
make a copy of the hw state on modeset and fastset.
Color blobs are copied in crtc atomic_check(), right
before color management is checked.
Changes since v1:
- Copy all blobs immediately after drm_atomic_helper_check_modeset().
- Clear crtc_state->hw on disable, instead of using clear_intel_crtc_state().
Changes since v2:
- Use intel_crtc_free_hw_state + clear in intel_crtc_disable_noatomic().
- Make a intel_crtc_prepare_state() function that clears the crtc_state
and copies hw members.
- Remove setting uapi.adjusted_mode, we now have a direct call to
drm_calc_timestamping_constants().
Changes since v3:
- Rename prefix copy_hw_to_uapi_state() with intel_crtc.
- Copy color blobs to uapi as well.
- Add a intel_crtc_copy_uapi_to_hw_state_nomodeset() function for clarity.
Changes since v4:
- Copy hw.adjusted_mode back to uapi.adjusted_mode, to shut up
the call to drm_calc_timestamping_constants() in
drm_atomic_helper_update_legacy_modeset_state().
- Use drm_property_replace_blob (Ville).
Changes since v5:
- Use hw->mode in intel_modeset_readout_hw_state(). (Ville)
- Copy to uapi.mode using drm_atomic_set_mode_for_crtc(). (Ville)
drm/i915: Add aliases for uapi and hw to crtc_state
Prepare to split up hw and uapi machinally, by adding a uapi and
hw alias. We will remove the base in a bit. This is a split from the
original uapi/hw patch, which did it all in one go.
Chris Wilson [Fri, 1 Nov 2019 10:15:28 +0000 (10:15 +0000)]
drm/i915/selftests: Spin on all engines simultaneously
Vanshidhar Konda asked for the simplest test "to verify that the kernel
can submit and hardware can execute batch buffers on all the command
streamers in parallel." We have a number of tests in userspace that
submit load to each engine and verify that it is present, but strictly
we have no selftest to prove that the kernel can _simultaneously_
execute on all known engines. (We have tests to demonstrate that we can
submit to HW in parallel, but we don't insist that they execute in
parallel.)
v2: Improve the igt_spinner support for older gen.
Chris Wilson [Fri, 1 Nov 2019 08:49:40 +0000 (08:49 +0000)]
drm/i915/selftests: Start kthreads before stopping
An interesting observation made with our parallel selftests was that on
our small/single cpu systems we would call kthread_stop() before the
kthreads were spawned. If this happens, the kthread is never run at all;
completely bypassing the test.
A simple yield() from the parent will ensure that all children have the
opportunity to start before we reap them.
Chris Wilson [Thu, 31 Oct 2019 09:01:04 +0000 (09:01 +0000)]
drm/i915/gem: Refine occupancy test in kill_context()
Don't just look at the very last request in a queue when deciding if we
need to evict the context from the GPU, as that request may still be in
the submission queue while the rest of the context is running!
Instead, walk back along the queued requests looking for the active
request and checking that.
Fixes: 2e0986a58cc4 ("drm/i915/gem: Cancel contexts when hangchecking is disabled")
Testcase: igt/gem_ctx_persistence/queued Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Reviewed-by: Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191031090104.22245-1-chris@chris-wilson.co.uk
Matthew Auld [Wed, 30 Oct 2019 17:33:20 +0000 (17:33 +0000)]
drm/i915/lmem: add the fake lmem region
Intended for upstream testing so that we can still exercise the LMEM
plumbing and !i915_ggtt_has_aperture paths. Smoke tested on Skull Canyon
device. This works by allocating an intel_memory_region for a reserved
portion of system memory, which we treat like LMEM. For the LMEMBAR we
steal the aperture and 1:1 it map to the stolen region.
To enable simply set the i915 modparam fake_lmem_start= on the kernel
cmdline with the start of reserved region(see memmap=). The size of the
region we can use is determined by the size of the mappable aperture, so
the size of reserved region should be >= mappable_end. For now we only
enable for the selftests. Depends on CONFIG_DRM_I915_UNSTABLE being
enabled.
Signed-off-by: Matthew Auld <matthew.auld@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Abdiel Janulgue <abdiel.janulgue@linux.intel.com> Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> 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/20191030173320.8850-1-matthew.auld@intel.com
drm/i915/dp: Do not switch aux to TBT mode for non-TC ports
Non-TC ports always have tc_mode == TC_PORT_TBT_ALT so it was
switching aux to TBT mode for all combo-phy ports, happily this did
not caused any issue but is better follow BSpec.
Also this is reserved bit before ICL.
Recent GuC doesn't require the shared area. We still have one user in
i915 (engine reset via guc) because we haven't updated the command to
match the current guc submission flow [1]. Since the flow in guc is
about to change again, just disable the command for now and add a note
that we'll implement it as part of the new flow.
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: John Harrison <John.C.Harrison@Intel.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: Fernando Pacheco <fernando.pacheco@intel.com> Acked-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Matthew Brost <matthew.brost@intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Link: https://patchwork.freedesktop.org/patch/msgid/20191031013040.25803-2-daniele.ceraolospurio@intel.com
Recent GuC binaries (including all the ones we're currently using)
don't require this shared area anymore, having moved the relevant
entries into the stage pool instead. i915 itself doesn't write
anything into it either, so we can safely drop it.
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: John Harrison <John.C.Harrison@Intel.com> Cc: Matthew Brost <matthew.brost@intel.com> Acked-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Matthew Brost <matthew.brost@intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Link: https://patchwork.freedesktop.org/patch/msgid/20191031013040.25803-1-daniele.ceraolospurio@intel.com
Chris Wilson [Wed, 30 Oct 2019 19:21:49 +0000 (19:21 +0000)]
drm/i915: Split detaching and removing the vma
In order to keep the assert_bind_count() valid, we need to hold the vma
page reference until after we drop the bind count. However, we must also
keep the drm_mm_remove_node() as the last action of i915_vma_unbind() so
that it serialises with the unlocked check inside i915_vma_destroy(). So
we need to split up i915_vma_remove() so that we order the detach, drop
pages and remove as required during unbind.
Chris Wilson [Thu, 31 Oct 2019 09:42:59 +0000 (09:42 +0000)]
drm/i915/selftests: Assert that the idle_pulse is sent
When checking the heartbeat pulse, we expect it to have been sent by the
time we have slept. We can verify this by checking the engine serial
number to see if that matches the predicted pulse serial. It will always
be true if, and only if, the pulse was sent by itself (as designed by
the test).
Ville Syrjälä [Wed, 2 Oct 2019 16:25:04 +0000 (19:25 +0300)]
drm/i915: Simplify pipe_mask setup even further
Just set pipe_mask=~0 for the non-special cases where any pipe
will do. intel_encoder_possible_crtcs() will anyway drop out
anything that doesn't exist.
Ville Syrjälä [Wed, 2 Oct 2019 16:25:01 +0000 (19:25 +0300)]
drm/i915: Simplify LVDS crtc_mask setup
We don't need to special case PCH vs. gen4 when setting up the LVDS
crtc_mask. Just claim pipes A|B|C work and
intel_encoder_possible_crtcs() will drop out any crtc that doesn't
exist.
v2: Put the special case first to match what most other encoders do
GuC 35.2.0 and HuC 7.0.3 are the first production releases for TGL.
GuC 35.2 for Gen12 is interface-compatible with 33.0 on older Gens,
because the differences are related to additional blocks/commands in
the interface to support new Gen12 features. These parts of the
interface will be added when the relevant features are enabled.
Ville Syrjälä [Tue, 29 Oct 2019 14:55:26 +0000 (16:55 +0200)]
drm/i915: Stop frobbing crtc->base.mode
The core no longer uses drm_crtc_state::mode with atomic drivers,
so let's stop frobbing it in the driver. For the user mode readout
we'll just use an on stack mode.
Imre Deak [Mon, 28 Oct 2019 18:15:17 +0000 (20:15 +0200)]
drm/i915: Avoid HPD poll detect triggering a new detect cycle
For the HPD interrupt functionality the HW depends on power wells in the
display core domain to be on. Accordingly when enabling these power
wells the HPD polling logic will force an HPD detection cycle to account
for hotplug events that may have happened when such a power well was
off.
Thus a detect cycle started by polling could start a new detect cycle if
a power well in the display core domain gets enabled during detect and
stays enabled after detect completes. That in turn can lead to a
detection cycle runaway.
To prevent re-triggering a poll-detect cycle make sure we drop all power
references we acquired during detect synchronously by the end of detect.
This will let the poll-detect logic continue with polling (matching the
off state of the corresponding power wells) instead of scheduling a new
detection cycle.
Fixes: 6cfe7ec02e85 ("drm/i915: Remove the unneeded AUX power ref from intel_dp_detect()")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112125 Reported-and-tested-by: Val Kulkov <val.kulkov@gmail.com> Reported-and-tested-by: wangqr <wqr.prg@gmail.com> Cc: Val Kulkov <val.kulkov@gmail.com> Cc: wangqr <wqr.prg@gmail.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191028181517.22602-1-imre.deak@intel.com
Chris Wilson [Wed, 30 Oct 2019 10:38:23 +0000 (10:38 +0000)]
drm/i915/gt: Always track callers to intel_rps_mark_interactive()
During startup, we may find ourselves in an interesting position where
we haven't fully enabled RPS before the display starts trying to use it.
This may lead to an imbalance in our "interactive" counter:
The solution employed is to not worry about RPS state and keep the tally
of the interactive counter separate. When we do enable RPS, we will then
take the display activity into account.
Lucas De Marchi [Tue, 29 Oct 2019 03:50:49 +0000 (20:50 -0700)]
drm/i915/tgl: add support to one DP-MST stream
This is the minimum change to support 1 (and only 1) DP-MST monitor
connected on Tiger Lake. This change was isolated from previous patch
from José. In order to support more streams we will need to create a
master-slave relation on the transcoders and that is not currently
working yet.
v2: remove unused macro and use REG_FIELD_PREP() (Ville)
Chris Wilson [Tue, 29 Oct 2019 20:23:38 +0000 (20:23 +0000)]
drm/i915/gem: Make context persistence optional
Our existing behaviour is to allow contexts and their GPU requests to
persist past the point of closure until the requests are complete. This
allows clients to operate in a 'fire-and-forget' manner where they can
setup a rendering pipeline and hand it over to the display server and
immediately exit. As the rendering pipeline is kept alive until
completion, the display server (or other consumer) can use the results
in the future and present them to the user.
The compute model is a little different. They have little to no buffer
sharing between processes as their kernels tend to operate on a
continuous stream, feeding the results back to the client application.
These kernels operate for an indeterminate length of time, with many
clients wishing that the kernel was always running for as long as they
keep feeding in the data, i.e. acting like a DSP.
Not all clients want this persistent "desktop" behaviour and would prefer
that the contexts are cleaned up immediately upon closure. This ensures
that when clients are run without hangchecking (e.g. for compute kernels
of indeterminate runtime), any GPU hang or other unexpected workloads
are terminated with the process and does not continue to hog resources.
The default behaviour for new contexts is the legacy persistence mode,
as some desktop applications are dependent upon the existing behaviour.
New clients will have to opt in to immediate cleanup on context
closure. If the hangchecking modparam is disabled, so is persistent
context support -- all contexts will be terminated on closure.
We expect this behaviour change to be welcomed by compute users, who
have often been caught between a rock and a hard place. They disable
hangchecking to avoid their kernels being "unfairly" declared hung, but
have also experienced true hangs that the system was then unable to
clean up. Naturally, this leads to bug reports.
Testcase: igt/gem_ctx_persistence Link: https://github.com/intel/compute-runtime/pull/228 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Reviewed-by: Jon Bloomfield <jon.bloomfield@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Acked-by: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20191029202338.8841-1-chris@chris-wilson.co.uk
Matt Roper [Tue, 29 Oct 2019 17:31:02 +0000 (10:31 -0700)]
drm/i915: Provide more information on DP AUX failures
We're seeing some failures where an aux transaction still shows as
'busy' well after the timeout limit that the hardware is supposed to
enforce. Improve the error message so that we can see exactly which aux
channel this error happened on and what the status bits were during this
case that isn't supposed to happen.
v2:
- Make timeout a const variable so that the timeout & message will
match if we decide to change it in the future. (Lucas)
- Don't bother testing intel_dp->aux.name for NULL. (Lucas)
HDCP could be fused off, so not all GEN9+ platforms will support it.
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Martin Peres <martin.peres@linux.intel.com> Reviewed-by: Ramalingam C <ramalingam.c@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: José Roberto de Souza <jose.souza@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191026001323.216052-2-jose.souza@intel.com
drm/i915: Add two spaces before the SKL_DFSM registers
The next patches are going to touch this registers so here already
fixing it for older registers and make it consistent with most of
the other registers in this file.
Matt Roper [Fri, 25 Oct 2019 23:06:23 +0000 (16:06 -0700)]
drm/i915/tgl: Add AUX B & C to DC_OFF_POWER_DOMAINS
Our TGL CI platforms are running into cases where aux transactions have
failed to complete or declare a timeout well after the timeout limit
that the hardware is supposed to enforce. From the logs it appears that
these failures arise when aux transactions happen after we've entered
DC6:
<7> [622.523650] [drm:skl_enable_dc6 [i915]] Enabling DC6
<7> [622.523685] [drm:gen9_set_dc_state [i915]] Setting DC state from 00 to 02
...
<3> [622.535753] [drm:intel_dp_aux_xfer [i915]] *ERROR* dp aux hw did not signal timeout!
<3> [622.547745] [drm:intel_dp_aux_xfer [i915]] *ERROR* dp aux hw did not signal timeout!
<3> [622.559746] [drm:intel_dp_aux_xfer [i915]] *ERROR* dp aux hw did not signal timeout!
<3> [622.571744] [drm:intel_dp_aux_xfer [i915]] *ERROR* dp aux hw did not signal timeout!
<3> [622.583743] [drm:intel_dp_aux_xfer [i915]] *ERROR* dp aux hw did not signal timeout!
<3> [622.583780] [drm:intel_dp_aux_xfer [i915]] *ERROR* dp_aux_ch not done status 0xad400bff
<7> [622.863725] [drm:drm_dp_dpcd_access] Too many retries, giving up. First error: -110
On TGL AUX B & C are in PG1 (managed by the DMC firmware) rather
than PG3 as they were on ICL, so allowing DC6 means the DMC firmware
might shut off the power wells behind our backs when we're trying to use
them.
Matt Roper [Sat, 26 Oct 2019 05:12:26 +0000 (22:12 -0700)]
drm/i915: Drop unused AUX register offsets
We reference DP AUX registers via the DP_AUX_CH_CTL() and
DP_AUX_CH_DATA() macros that calculate all the register offsets for us
automatically; there's no need to explicitly define every offset in
i915_reg.h if they're never going to be used by the driver code.
v2: Apparently GVT was directly using these raw definitions in a couple
places. Switch GVT code over to using our preferred macros.
Chris Wilson [Tue, 29 Oct 2019 09:16:32 +0000 (09:16 +0000)]
drm/i915/gt: Make timeslice duration configurable
Execlists uses a scheduling quantum (a timeslice) to alternate execution
between ready-to-run contexts of equal priority. This ensures that all
users (though only if they of equal importance) have the opportunity to
run and prevents livelocks where contexts may have implicit ordering due
to userspace semaphores. However, not all workloads necessarily benefit
from timeslicing and in the extreme some sysadmin may want to disable or
reduce the timeslicing granularity.
The timeslicing mechanism can be compiled out^W^W disabled (but should
DCE!) with
Commit f2db53f14d3d ("drm/i915: Replace "_load" with "_probe"
consequently") deliberately left the name of the module parameter
unchanged as that would require a corresponding change on IGT size.
Now as the IGT side change has been submitted, complete the switch to
the "probe" nomenclature.
Suggested-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com> Cc: Michał Wajdeczko <michal.wajdeczko@intel.com> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Piotr Piórkowski <piotr.piorkowski@intel.com> Cc: Tomasz Lis <tomasz.lis@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.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/20191029102036.6326-3-janusz.krzysztofik@linux.intel.com
drm/i915: Fix i915_inject_load_error() name to read *_probe_*
Commit 50d84418f586 ("drm/i915: Add i915 to i915_inject_probe_failure")
introduced new functions unfortunately named incompatibly with rules
established by commit f2db53f14d3d ("drm/i915: Replace "_load" with
"_probe" consequently"). Fix it for consistency.
Suggested-by: Michał Wajdeczko <michal.wajdeczko@intel.com> Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com> Cc: Michał Wajdeczko <michal.wajdeczko@intel.com> Cc: Michał Winiarski <michal.winiarski@intel.com> Cc: Piotr Piórkowski <piotr.piorkowski@intel.com> Cc: Tomasz Lis <tomasz.lis@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.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/20191029102036.6326-2-janusz.krzysztofik@linux.intel.com
The design of the OA unit has been split into several units. We now
have a global unit (OAG) and a render specific unit (OAR). This leads
to some changes on how we program things. Some details :
OAR:
- has its own set of counter registers, they are per-context
saved/restored
- counters are not written to the circular OA buffer
- a snapshot of the counters can be acquired with
MI_RECORD_PERF_COUNT, or a single counter can be read with
MI_STORE_REGISTER_MEM.
OAG:
- has global counters that increment across context switches
- counters are written into the circular OA buffer (if requested)
v2: Fix checkpatch warnings on code style (Lucas)
v3: (Umesh)
- Update register from which tail, status and head are read
- Update logic to sample context reports
- Update whitelist mux and b counter regs
v4: Fix a bug when updating context image for new contexts (Umesh)
v5: Squash patch enabling save/restore of counters into context image
We want this so we can preempt performance queries and keep the
system responsive even when long running queries are ongoing. We
avoid doing it for all contexts.
- use LRI to modify context control (Chris)
- use MASKED_FIELD to program just the masked bits (Chris)
- disable save/restore of counters on cleanup (Chris)
v6: Do not use implicit parameters (Chris)
Michal Wajdeczko [Tue, 29 Oct 2019 09:58:54 +0000 (09:58 +0000)]
drm/i915: Don't try to place HWS in non-existing mappable region
HWS placement restrictions can't just rely on HAS_LLC flag.
Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com> Signed-off-by: Matthew Auld <matthew.auld@intel.com> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Acked-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/20191029095856.25431-5-matthew.auld@intel.com
If the aperture is not available in HW we can't use a ggtt slot and wc
copy, so fall back to regular kmap.
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by: Abdiel Janulgue <abdiel.janulgue@linux.intel.com> Signed-off-by: Matthew Auld <matthew.auld@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/20191029095856.25431-4-matthew.auld@intel.com
drm/i915: set num_fence_regs to 0 if there is no aperture
We can't fence anything without aperture.
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by: Stuart Summers <stuart.summers@intel.com> Signed-off-by: Matthew Auld <matthew.auld@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/20191029095856.25431-3-matthew.auld@intel.com
Chris Wilson [Mon, 28 Oct 2019 20:30:12 +0000 (20:30 +0000)]
drm/i915/gem: Limit the blitter sizes to ensure low preemption latency
Currently we insert a arbitration point every 128MiB during a blitter
copy. At 8GiB/s, this is around 30ms. This is a little on the large side
if we need to inject a high priority work, so reduced it down to 8MiB or
roughly 1ms.
Michal Wajdeczko [Mon, 28 Oct 2019 16:45:20 +0000 (16:45 +0000)]
drm/i915/execlists: Use vfunc to check engine submission mode
While processing CSB there is no need to look at GuC submission
settings, just check if engine is configured for execlists mode.
While today GuC submission is disabled it's settings are still
based on modparam values that might not correctly reflect actual
submission status in case of any fallback. Until that is fully
fixed, use alternate method to confirm that engine really runs in
execlists mode by comparing set_default_submission vfunc.
v2: add other immediate use of new helper
Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com> Reviewed-by: Janusz Krzysztofik <janusz.krzysztofik@linux.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/20191028164520.31772-1-michal.wajdeczko@intel.com
Chris Wilson [Mon, 28 Oct 2019 14:26:52 +0000 (14:26 +0000)]
drm/i915/display: Mark conn as initialised by iterator
smatch complains about
drivers/gpu/drm/i915//display/intel_display.c:14403 intel_set_dp_tp_ctl_normal() error: uninitialized symbol 'conn'.
because it has no way to determine that the loop must have an entry.
Tell the static analysers to ignore the local, it will always be set.
Chris Wilson [Mon, 28 Oct 2019 12:41:25 +0000 (12:41 +0000)]
drm/i915/execlists: Simply walk back along request timeline on reset
The request's timeline will only contain requests from this context, in
order of execution. Therefore, we can simply look back along this
timeline to find the currently executing request.
If we do find that the current context has completed its last request,
that does not imply that all requests are completed in the context, so
only advance the ring->head up to the end of the known completions!
Chris Wilson [Sun, 27 Oct 2019 22:58:07 +0000 (22:58 +0000)]
drm/i915/selftests: Use a random engine for GEM coherency tests
Select a random user accessible engine for checking coherency results.
While we should check all engines, we use a random selection so that
over repeated runs we cover all.
Chris Wilson [Sun, 27 Oct 2019 15:43:14 +0000 (15:43 +0000)]
drm/i915: Put future HW and their uAPIs under STAGING & BROKEN
We would like some freedom to break the user API/ABI for future HW but
yet still expose the driver for upstream development on that HW.
Currently, we have the i915.force_probe module parameter to avoid binding
to HW while the driver is under development, but that is still a little
too soft with respect to the stringent no-regression rules if we also
plan to be redesigning the uAPI to go along with the new HW.
To allow the uAPI to be changed during development, only expose that API
and in development HW under STAGING (and BROKEN). Hopefully, making it
explicit that such interfaces to that HW are under development and not
to be blindly enabled by distributions.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Dave Airlie <airlied@redhat.com> Acked-by: Dave Airlie <airlied@redhat.com> Acked-by: Jani Nikula <jani.nikula@intel.com> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191027154314.11139-1-chris@chris-wilson.co.uk
Andi Shyti [Thu, 24 Oct 2019 21:16:41 +0000 (22:16 +0100)]
drm/i915: Extract GT render power state management
i915_irq.c is large. One reason for this is that has a large chunk of
the GT render power management stashed away in it. Extract that logic
out of i915_irq.c and intel_pm.c and put it under one roof.
Chris Wilson [Sat, 26 Oct 2019 08:22:20 +0000 (09:22 +0100)]
drm/i915/tgl: Adjust the location of RING_MI_MODE in the context image
The location of RING_MI_MODE (used to stop the ring across resets) moved
for Tigerlake. Fixup the new location and include a selftest to verify
the location in the default context image.
Avoid angering clang and smatch by using a constant value in a '&&' test,
by forcing that constant value into a boolean.
E.g.,
drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c:159:13: warning: use of logical '&&' with constant operand [-Wconstant-logical-operand]
if (!delay && CONFIG_DRM_I915_PREEMPT_TIMEOUT) {
^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Reported-by: kbuild test robot <lkp@intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Nathan Chancellor <natechancellor@gmail.com> Cc: Nick Desaulniers <ndesaulniers@google.com> Cc: Jani Nikula <jani.nikula@intel.com> Reviewed-by: Nathan Chancellor <natechancellor@gmail.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191025135943.12524-1-chris@chris-wilson.co.uk
Chris Wilson [Fri, 25 Oct 2019 16:54:42 +0000 (17:54 +0100)]
drm/i915/pmu: Initialise the spinlock before registering
As the GT may be running in parallel with the module initialisation
code, we may enter i915_pmu_gt_parked() as we are executing
i915_pmu_register(). We have to init the spinlock before we mark
pmu.event_init so that it is available for use by i915_pmu_gt_parked()
(which may run as soon as event_init is set).
Matthew Auld [Fri, 25 Oct 2019 17:25:11 +0000 (18:25 +0100)]
drm/i915/selftests/blt: add some kthreads into the mix
We can be more aggressive in our testing by launching a number of
kthreads, where each is submitting its own copy or fill batches on a set
of random sized objects. Also since the underlying fill and copy batches
can be pre-empted mid-batch(for particularly large objects), throw in a
random mixture of ctx priorities per thread to make pre-emption a
possibility.
Matthew Auld [Fri, 25 Oct 2019 15:37:28 +0000 (16:37 +0100)]
drm/i915/selftests: add sanity selftest for huge-GTT-pages
Now that for all the relevant backends we do randomised testing, we need
to make sure we still sanity check the obvious cases that might blow up,
such that introducing a temporary regression is less likely. Also
rather than do this for every backend, just limit to our two memory
types: system and local.
Matthew Auld [Fri, 25 Oct 2019 15:37:27 +0000 (16:37 +0100)]
drm/i915/selftests: prefer random sizes for the huge-GTT-page smoke tests
Ditch the dubious static list of sizes to enumerate, in favour of
choosing a random size within the limits of each backing store. With
repeated CI runs this should give us a wider range of object sizes, and
in turn more page-size combinations, while using less machine time.
Abdiel Janulgue [Fri, 25 Oct 2019 15:37:24 +0000 (16:37 +0100)]
drm/i915/lmem: support kernel mapping
We can create LMEM objects, but we also need to support mapping them
into kernel space for internal use.
Signed-off-by: Abdiel Janulgue <abdiel.janulgue@linux.intel.com> Signed-off-by: Matthew Auld <matthew.auld@intel.com> Signed-off-by: Steve Hampson <steven.t.hampson@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.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/20191025153728.23689-3-chris@chris-wilson.co.uk
Matthew Auld [Fri, 25 Oct 2019 15:37:22 +0000 (16:37 +0100)]
drm/i915: support creating LMEM objects
We currently define LMEM, or local memory, as just another memory
region, like system memory or stolen, which we can expose to userspace
and can be mapped to the CPU via some BAR.
Lucas De Marchi [Thu, 24 Oct 2019 19:51:22 +0000 (12:51 -0700)]
drm/i915: split gen11_irq_handler to make it shareable
Split gen11_irq_handler() to receive as parameter the function
pointers. This allows to share the interrupt handler even if the enable/disable
functions are different.
Make sure it's always inlined to avoid the extra indirect call on the
hot path. Checking with gcc 9 this produce the exact same code as of
now:
$ size drivers/gpu/drm/i915/i915_irq*.o
text data bss dec hex filename
47511 560 0 48071 bbc7 drivers/gpu/drm/i915/i915_irq.o
47511 560 0 48071 bbc7 drivers/gpu/drm/i915/i915_irq_new.o
Lucas De Marchi [Thu, 24 Oct 2019 19:51:21 +0000 (12:51 -0700)]
drm/i915: do not set MOCS control values on dgfx
On dgfx there's no LLC and eDRAM control table. Since now this
also means the device has global MOCS, just return early on the
initialization function.
L3 settings still apply and still need to be tweaked.
projects / kernel.git / log