Having allowed the user to define a set of engines that they will want
to only use, we go one step further and allow them to bind those engines
into a single virtual instance. Submitting a batch to the virtual engine
will then forward it to any one of the set in a manner as best to
distribute load. The virtual engine has a single timeline across all
engines (it operates as a single queue), so it is not able to concurrently
run batches across multiple engines by itself; that is left up to the user
to submit multiple concurrent batches to multiple queues. Multiple users
will be load balanced across the system.
The mechanism used for load balancing in this patch is a late greedy
balancer. When a request is ready for execution, it is added to each
engine's queue, and when an engine is ready for its next request it
claims it from the virtual engine. The first engine to do so, wins, i.e.
the request is executed at the earliest opportunity (idle moment) in the
system.
As not all HW is created equal, the user is still able to skip the
virtual engine and execute the batch on a specific engine, all within the
same queue. It will then be executed in order on the correct engine,
with execution on other virtual engines being moved away due to the load
detection.
A couple of areas for potential improvement left!
- The virtual engine always take priority over equal-priority tasks.
Mostly broken up by applying FQ_CODEL rules for prioritising new clients,
and hopefully the virtual and real engines are not then congested (i.e.
all work is via virtual engines, or all work is to the real engine).
- We require the breadcrumb irq around every virtual engine request. For
normal engines, we eliminate the need for the slow round trip via
interrupt by using the submit fence and queueing in order. For virtual
engines, we have to allow any job to transfer to a new ring, and cannot
coalesce the submissions, so require the completion fence instead,
forcing the persistent use of interrupts.
- We only drip feed single requests through each virtual engine and onto
the physical engines, even if there was enough work to fill all ELSP,
leaving small stalls with an idle CS event at the end of every request.
Could we be greedy and fill both slots? Being lazy is virtuous for load
distribution on less-than-full workloads though.
Other areas of improvement are more general, such as reducing lock
contention, reducing dispatch overhead, looking at direct submission
rather than bouncing around tasklets etc.
sseu: Lift the restriction to allow sseu to be reconfigured on virtual
engines composed of RENDER_CLASS (rcs).
v2: macroize check_user_mbz()
v3: Cancel virtual engines on wedging
v4: Commence commenting
v5: Replace 64b sibling_mask with a list of class:instance
v6: Drop the one-element array in the uabi
v7: Assert it is an virtual engine in to_virtual_engine()
v8: Skip over holes in [class][inst] so we can selftest with (vcs0, vcs2)
Link: https://github.com/intel/media-driver/pull/283
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/20190521211134.16117-6-chris@chris-wilson.co.uk
A usecase arose out of handling context recovery in mesa, whereby they
wish to recreate a context with fresh logical state but preserving all
other details of the original. Currently, they create a new context and
iterate over which bits they want to copy across, but it would much more
convenient if they were able to just pass in a target context to clone
during creation. This essentially extends the setparam during creation
to pull the details from a target context instead of the user supplied
parameters.
The ideal here is that we don't expose control over anything more than
can be obtained via CONTEXT_PARAM. That is userspace retains explicit
control over all features, and this api is just convenience.
For example, you could replace
struct context_param p = { .param = CONTEXT_PARAM_VM };
param.ctx_id = old_id;
gem_context_get_param(&p.param);
new_id = gem_context_create();
param.ctx_id = new_id;
gem_context_set_param(&p.param);
gem_vm_destroy(param.value); /* drop the ref to VM_ID handle */
with
struct create_ext_param p = {
{ .name = CONTEXT_CREATE_CLONE },
.clone_id = old_id,
.flags = CLONE_FLAGS_VM
}
new_id = gem_context_create_ext(&p);
and not have to worry about stray namespace pollution etc.
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/20190521211134.16117-5-chris@chris-wilson.co.uk
Allow the user to specify a local engine index (as opposed to
class:index) that they can use to refer to a preset engine inside the
ctx->engine[] array defined by an earlier I915_CONTEXT_PARAM_ENGINES.
This will be useful for setting SSEU parameters on virtual engines that
are local to the context and do not have a valid global class:instance
lookup.
Note that due to the ambiguity in using class:instance with
ctx->engines[], if a user supplied engine map is active the user must
specify the engine to alter by its index into the ctx->engines[].
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/20190521211134.16117-3-chris@chris-wilson.co.uk
Over the last few years, we have debated how to extend the user API to
support an increase in the number of engines, that may be sparse and
even be heterogeneous within a class (not all video decoders created
equal). We settled on using (class, instance) tuples to identify a
specific engine, with an API for the user to construct a map of engines
to capabilities. Into this picture, we then add a challenge of virtual
engines; one user engine that maps behind the scenes to any number of
physical engines. To keep it general, we want the user to have full
control over that mapping. To that end, we allow the user to constrain a
context to define the set of engines that it can access, order fully
controlled by the user via (class, instance). With such precise control
in context setup, we can continue to use the existing execbuf uABI of
specifying a single index; only now it doesn't automagically map onto
the engines, it uses the user defined engine map from the context.
v2: Fixup freeing of local on success of get_engines()
v3: Allow empty engines[]
v4: s/nengine/num_engines/
v5: Replace 64 limit on num_engines with a note that execbuf is
currently limited to only using the first 64 engines.
v6: Actually use the engines_mutex to guard the ctx->engines.
Testcase: igt/gem_ctx_engines
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/20190521211134.16117-2-chris@chris-wilson.co.uk
Having hid the partially exposed new ABI from the PR, put it back again
for completion of context recovery. A significant part of context
recovery is the ability to reuse as much of the old context as is
feasible (to avoid expensive reconstruction). The biggest chunk kept
hidden at the moment is fine-control over the ctx->ppgtt (the GPU page
tables and associated translation tables and kernel maps), so make
control over the ctx->ppgtt explicit.
This allows userspace to create and share virtual memory address spaces
(within the limits of a single fd) between contexts they own, along with
the ability to query the contexts for the vm state.
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/20190521211134.16117-1-chris@chris-wilson.co.uk
After "5918045c4ed4 drm/scheduler: rework job destruction", jobs are
only deleted when the timeout handler is able to be cancelled
successfully.
In case no timeout handler is running (timeout == MAX_SCHEDULE_TIMEOUT),
job cleanup would be skipped which may result in memory leaks.
Add the handling for the (timeout == MAX_SCHEDULE_TIMEOUT) case in
drm_sched_cleanup_jobs.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/306025/?series=60878&rev=2
After "5918045c4ed4 drm/scheduler: rework job destruction", lima started
to leak memory due to buffers not being destroyed after job execution in
the drm scheduler.
This started happening because the drm scheduler only destroyed buffers
after cancelling the job timeout handler, and for lima this handler was
never started as lima specified a MAX_SCHEDULE_TIMEOUT timeout.
Lima seems to run well in its current state with a real timeout, so to
make it more aligned with the other drivers from now on, let's use a
real default timeout.
This also fixes the observed memory leaks.
The 500ms value was chosen as it is the current value for all other
embedded gpu drivers using drm sched.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190520224229.21111-1-nunes.erico@gmail.com
Based on 2 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license as published by
the free software foundation either version 2 of the license or at
your option any later version this program is distributed in the
hope that it will be useful but without any warranty without even
the implied warranty of merchantability or fitness for a particular
purpose see the gnu general public license for more details you
should have received a copy of the gnu general public license along
with this program if not see http www gnu org licenses
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license as published by
the free software foundation either version 2 of the license or at
your option any later version this program is distributed in the
hope that it will be useful but without any warranty without even
the implied warranty of merchantability or fitness for a particular
purpose see the gnu general public license for more details [based]
[from] [clk] [highbank] [c] you should have received a copy of the
gnu general public license along with this program if not see http
www gnu org licenses
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-or-later
has been chosen to replace the boilerplate/reference in 355 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org>
Reviewed-by: Jilayne Lovejoy <opensource@jilayne.com>
Reviewed-by: Steve Winslow <swinslow@gmail.com>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190519154041.837383322@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add SPDX license identifiers to all Make/Kconfig files which:
- Have no license information of any form
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add SPDX license identifiers to all files which:
- Have no license information of any form
- Have MODULE_LICENCE("GPL*") inside which was used in the initial
scan/conversion to ignore the file
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add SPDX license identifiers to all files which:
- Have no license information of any form
- Have EXPORT_.*_SYMBOL_GPL inside which was used in the
initial scan/conversion to ignore the file
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If SVGA_3D_CMD_DX_SET_SHADER is called with a shader ID
of SVGA3D_INVALID_ID, and a shader type of
SVGA3D_SHADERTYPE_INVALID, the calculated binding.shader_slot
will be 4294967295, leading to an out-of-bounds read in vmw_binding_loc()
when the offset is calculated.
Cc: <stable@vger.kernel.org>
Fixes: d80efd5cb3 ("drm/vmwgfx: Initial DX support")
Signed-off-by: Murray McAllister <murray.mcallister@gmail.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
If SVGA_3D_CMD_DX_DEFINE_RENDERTARGET_VIEW is called with a surface
ID of SVGA3D_INVALID_ID, the srf struct will remain NULL after
vmw_cmd_res_check(), leading to a null pointer dereference in
vmw_view_add().
Cc: <stable@vger.kernel.org>
Fixes: d80efd5cb3 ("drm/vmwgfx: Initial DX support")
Signed-off-by: Murray McAllister <murray.mcallister@gmail.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Use struct sg_dma_page_iter in favour struct of sg_page_iter, which fairly
recently was declared useless for obtaining dma addresses.
With a struct sg_dma_page_iter we can't call sg_page_iter_page() so
when the page is needed, use the same page lookup mechanism as for the
non-sg dma modes instead of calling sg_dma_page_iter.
Note, the fixes tag doesn't really point to a commit introducing a
failure / regression, but rather to a commit that implemented a simple
workaround for this problem.
Cc: Jason Gunthorpe <jgg@mellanox.com>
Fixes: d901b2760d ("lib/scatterlist: Provide a DMA page iterator")
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jason Gunthorpe <jgg@mellanox.com>
In compat mode, we allowed host-backed user-space with guest-backed
kernel / device. In this mode, set shader commands was broken since
no relocations were emitted. Fix this.
Cc: <stable@vger.kernel.org>
Fixes: e8c66efbfe ("drm/vmwgfx: Make user resource lookups reference-free during validation")
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
User-space handles equal to zero are interpreted as uninitialized or
illegal by some drm systems (most notably kms). This means that a
dumb buffer or surface with a zero user-space handle can never be
used as a kms frame-buffer.
Cc: <stable@vger.kernel.org>
Fixes: c7eae62666 ("drm/vmwgfx: Make the object handles idr-generated")
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Deepak Rawat <drawat@vmware.com>
This may confuse user-space clients like plymouth that opens a drm
file descriptor as a result of a hotplug event and then generates a
new event...
Cc: <stable@vger.kernel.org>
Fixes: 5ea1734827 ("drm/vmwgfx: Send a hotplug event at master_set")
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Deepak Rawat <drawat@vmware.com>
"ret" is uninitialized on this path but it should be -EINVAL.
Fixes: 930c8dfea4 ("drm/i915/gvt: Check if get_next_pt_type() always returns a valid value")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
the vGPU write on TRTTE and 0x4dfc is now write to vreg first. their
values all be restored hardware when context switching.
Fixes: e39c5add32 ("drm/i915/gvt: vGPU MMIO virtualization")
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
0x4dfc is in-context mmio for gen9+, but each vm have different settings
need to add it to save-restore list along with other trtt registers
Fixes: 1786571393 ("drm/i915/gvt: vGPU context switch")
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
for restore-inhibit context, hardware will not load in-context mmios
(engine context part) to hardware, but hardware will save the mmio
values in hardware back to context image. So, in order to save correct
values of vGPU back to context image, values of vGPU mmios have to be
loaded into hardware first for restore-inhibit context.
In this patch, the mechanism is applied to all gen9 platform.
The reason excluding gen8 platforms is only because of lacking of testing
on those platforms.
v3: for mocs registers, goto in-context mmios save-restore path for skl
platform as well (weinan li)
v2: update vreg when scanning indirect context for inhibit context for
gen9
Cc: Weinan Li <weinan.z.li@intel.com>
Acked-by: Weinan Li <weinan.z.li@intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
"To track whether a request has started on HW, we can emit a breadcrumb at
the beginning of the request and check its timeline's HWSP to see if the
breadcrumb has advanced past the start of this request." It means all the
request which timeline's has_init_breadcrumb is true, then the
emit_init_breadcrumb process must have before emitting the real commands,
otherwise, the scheduler might get a wrong state of this request during
reset. If the request is exactly the guilty one, the scheduler won't
terminate it with the wrong state. To avoid this, do emit_init_breadcrumb
for all the requests from gvt.
v2: cc to stable kernel
Fixes: 8547444137 ("drm/i915: Identify active requests")
Cc: stable@vger.kernel.org
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Weinan <weinan.z.li@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Fix compute profile switching on process termination.
Add a dedicated reference counter to keep track of entry/exit to/from
compute profile. This enables switching compute profiles for other
reasons than process creation or termination.
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Eric Huang <JinhuiEric.Huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
PSP fw primary buffer is not used under SRIOV.
Under SRIOV, VBIOS or hypervisor driver will load psp
sos and psp sysdrv. Therefore, we don't need to
allocate memory for it.
v2: remove superfluous check for amdgpu_bo_free_kernel().
Signed-off-by: Yintian Tao <yttao@amd.com>
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
There is a typo so the code unlocks twice instead of taking the lock and
then releasing it.
Fixes: f14a323db5 ("drm/amd/powerplay: implement update enabled feature state to smc for smu11")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
For SR-IOV, vram_width can't be read from ATOM as
RAVEN, and DF related registers is not readable, so hardcord
is the only way to set the correct vram_width.
Reviewed-by: Yintian Tao <yttao@amd.com>
Signed-off-by: Trigger Huang <Trigger.Huang@amd.com>
Signed-off-by: Yintian Tao <yttao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHY]
Some early Raven boards had a bad SBIOS that doesn't play nicely with
the DMCU FW. We thought the issues were fixed by ignoring errors on DMCU
load but that doesn't seem to be the case. We've still seen reports of
users unable to boot their systems at all.
[HOW]
Disable DMCU load on Raven 1. Only load it for Raven 2 and Picasso.
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
In commit b7404c7ecb ("drm/i915: Bump ready tasks ahead of
busywaits"), I tried cutting a corner in order to not install a signal
for each of our dependencies, and only listened to requests on which we
were intending to busywait. The compromise that was made was that
instead of then being able to promote the request with a full
NOSEMAPHORE like its non-busywaiting brethren, as we had not ensured we
had cleared the semaphore chain, we settled for only using the NEWCLIENT
boost. With an over saturated system with multiple NEWCLIENTS in flight
at any time, this was found to be an inadequate promotion and left us
with a much poorer scheduling order than prior to using semaphores.
The outcome of this patch, is that all requests have NOSEMAPHORE
priority when they have no dependencies and are ready to run and not
busywait, restoring the pre-semaphore ordering on saturated systems.
We can demonstrate the effect of poor scheduling order by oversaturating
the system using gem_wsim on a system with multiple vcs engines
(i.e running the same workloads across more clients than required for
peak throughput, e.g. media_load_balance_17i7.wsim -c4 -b context):
x v5.1 (normalized)
+ tip
* fix
+------------------------------------------------------------------------+
| x |
| x |
| x |
| x |
| %x |
| %%x |
| %%x |
| %%x |
| %%x |
| %%x |
| %%x |
| %%x |
| %%x |
| %%x |
| %%x |
| %#x |
| %#x |
| %#x |
| %#x |
| %#x |
| + %#xx |
| + %#xx |
| + %%#xx |
| + %%#xx |
| + %%#xx |
| + %%#xx |
| + %%##x |
| +++ %%##x |
| +++ %%##x |
| +++ %%##x |
| ++++ %%##x |
| ++++ %%##x |
| ++++ %%##xx |
| ++++ %###xx |
| ++++ %###xx |
| ++++ %###xx |
| ++++ %###xx |
| ++++ + %#O#xx |
| ++++ + %#O#xx |
| ++++++ + %#O#xx |
| ++++++++++ %OOOxxx|
| ++++++++++ + %#OOO#xx|
| + ++++++++++++ ++ +++++ + ++ @@OOOO#xx|
| |A_| |
||__________M_______A____________________| |
| |A_| |
+------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 120 0.99456 1.00628 0.999985 1.0001545 0.0024387139
+ 120 0.873021 1.00037 0.884134 0.90148752 0.039190862
Difference at 99.5% confidence
-0.098667 +/- 0.0110762
-9.86517% +/- 1.10745%
(Student's t, pooled s = 0.0277657)
% 120 0.990207 1.00165 0.9970265 0.99699748 0.0021024
Difference at 99.5% confidence
-0.003157 +/- 0.000908245
-0.315651% +/- 0.0908105%
(Student's t, pooled s = 0.00227678)
Fixes: b7404c7ecb ("drm/i915: Bump ready tasks ahead of busywaits")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
Cc: Dmitry Ermilov <dmitry.ermilov@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190515130052.4475-2-chris@chris-wilson.co.uk
(cherry picked from commit 17db337f50)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Commit 1413b2bc07 ("drm/i915: Trim NEWCLIENT boosting") had the
intended consequence of not allowing a sequence of work that merely
crossed into a new engine the privilege to be promoted to NEWCLIENT
status. It also had the unintended consequence of actually making
NEWCLIENT effective on heavily oversubscribed transcode machines and
impacting upon their throughput.
If we consider a client packet composed of (rcsA, rcsB, vcs) and 30 of
those clients, using the NEWCLIENT boost that will be scheduled as
rcsA x 30, (rcsB, vcs) x 30
where as before it would have been
(rcsA, rcsB, vcs) x 30
That is with NEWCLIENT only boosting the first request of each client,
we would execute all rcsA requests prior to running on the vcs engines;
acruing a lot of dead time as compared to the previous case where the
vcs engine would be started in parallel to processing the second client.
The previous patch has the effect of delaying submission until it is
required by a third party (either the user with an explicit wait, or by
another client/engine). We reduce the NEWCLIENT bump to a mere WAIT,
which has the effect of removing its preemptive grant and reducing it to
the same level as any other user interaction -- that it will not be
promoted above the interengine dependencies, and so preventing NEWCLIENTS
from starving other engines. This a large nerf to the rrul properties of
the current NEWCLIENT, but it still does give prioritised submission to
new requests from light workloads.
References: b16c765122 ("drm/i915: Priority boost for new clients")
Fixes: 1413b2bc07 ("drm/i915: Trim NEWCLIENT boosting") # customer impact
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
Cc: Dmitry Ermilov <dmitry.ermilov@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190515130052.4475-4-chris@chris-wilson.co.uk
(cherry picked from commit 68fc728b01)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
The handling of the no-preemption priority level imposes the restriction
that we need to maintain the implied ordering even though preemption is
disabled. Otherwise we may end up with an AB-BA deadlock across multiple
engine due to a real preemption event reordering the no-preemption
WAITs. To resolve this issue we currently promote all requests to WAIT
on unsubmission, however this interferes with the timeslicing
requirement that we do not apply any implicit promotion that will defeat
the round-robin timeslice list. (If we automatically promote the active
request it will go back to the head of the queue and not the tail!)
So we need implicit promotion to prevent reordering around semaphores
where we are not allowed to preempt, and we must avoid implicit
promotion on unsubmission. So instead of at unsubmit, if we apply that
implicit promotion on adding the dependency, we avoid the semaphore
deadlock and we also reduce the gains made by the promotion for user
space waiting. Furthermore, by keeping the earlier dependencies at a
higher level, we reduce the search space for timeslicing without
altering runtime scheduling too badly (no dependencies at all will be
assigned a higher priority for rrul).
v2: Limit the bump to external edges (as originally intended) i.e.
between contexts and out to the user.
Testcase: igt/gem_concurrent_blit
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/20190515130052.4475-3-chris@chris-wilson.co.uk
(cherry picked from commit 6e7eb7a807)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
With gtt remapping in place we can use arbitrarily large
framebuffers. Let's bump the limits to 16kx16k on gen7+.
The limit was chosen to match the maximum 2D surface size
of the 3D engine.
With the remapping we could easily go higher than that for the
display engine. However the modesetting ddx will blindly assume
it can handle whatever is reported via kms. The oversized
buffer dimensions are not caught by glamor nor Mesa until
finally an assert will trip when genxml attempts to pack the
SURFACE_STATE. So we pick a safe limit to avoid the X server
from crashing (or potentially misbehaving if the genxml asserts
are compiled out).
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110187
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190509122159.24376-9-ville.syrjala@linux.intel.com
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
With gtt remapping plugged in we can simply raise the stride
limit on gen4+. Let's just pick the limit to match the render
engine max stride (256KiB on gen7+, 128KiB on gen4+).
No remapping CCS because the virtual address of each page actually
matters due to the new hash mode
(WaCompressedResourceDisplayNewHashMode:skl,kbl etc.), and no
remapping on gen2/3 due extra complications from fence alignment
and gen2 2KiB GTT tile size. Also no real benefit since the
display engine limits already match the other limits.
v2: Rebase due to is_ccs_modifier()
v3: Tweak the comment and commit msg
v4: Fix gen4+ stride limit to be 128KiB
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> #v3
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190509122159.24376-8-ville.syrjala@linux.intel.com
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
The display engine stride limits are getting in our way. On SKL+
we are limited to 8k pixels, which is easily exceeded with three
4k displays. To overcome this limitation we can remap the pages
in the GTT to provide the display engine with a view of memory
with a smaller stride.
The code is mostly already there as We already play tricks with
the plane surface address and x/y offsets.
A few caveats apply:
* linear buffers need the fb stride to be page aligned, as
otherwise the remapped lines wouldn't start at the same
spot
* compressed buffers can't be remapped due to the new
ccs hash mode causing the virtual address of the pages
to affect the interpretation of the compressed data. IIRC
the old hash was limited to the low 12 bits so if we were
using that mode we could remap. As it stands we just refuse
to remapp with compressed fbs.
* no remapping gen2/3 as we'd need a fence for the remapped
vma, which we currently don't have. Need to deal with the
fence POT requirements, and do something about the gen2
gtt page size vs tile size difference
v2: Rebase due to is_ccs_modifier()
Fix up the skl+ stride_mult mess
memset() the gtt_view because otherwise we could leave
junk in plane[1] when going from 2 plane to 1 plane format
v3: intel_check_plane_stride() was split out
v4: Drop the aligned viewport stuff, it was meant for ccs which
can't be remapped anyway
v5: Introduce intel_plane_can_remap()
Reorder the code so that plane_state->view gets filled
even for invisible planes, otherwise we'd keep using
stale values and could explode during remapping. The new
logic never remaps invisible planes since we don't have
a viewport, and instead pins the full fb instead
v6: Fix plane src coord checks after remapping by moving
plane_state->base.src to the final plane x/y offsets.
Allow intel_plane_check_stride() to fail even with
remapping (can happen at least with a linear 64bpp
fb with a 4k plane and a suitably inconvenient src
coordinates).
Improve aux plane FIXME (Daniel)
Move some code shuffling into a separate patch (Daniel)
Testcase: igt/kms_big_fb
Cc: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190509122159.24376-6-ville.syrjala@linux.intel.com
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Add a live selftest to excercise rotated/remapped vmas. We simply
write through the rotated/remapped vma, and confirm that the data
appears in the right page when read through the normal vma.
Not sure what the fallout of making all rotated/remapped vmas
mappable/fenceable would be, hence I just hacked it in the test.
v2: Grab rpm reference (Chris)
GEM_BUG_ON(view.type not as expected) (Chris)
Allow CAN_FENCE for rotated/remapped vmas (Chris)
Update intel_plane_uses_fence() to ask for a fence
only for normal vmas on gen4+
v3: Deal with intel_wakeref_t
v4: Rebase
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190509122159.24376-4-ville.syrjala@linux.intel.com
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
