Quoting Tvrtko Ursulin (2018-08-14 15:40:58)
> From: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
>
> We want to allow userspace to reconfigure the subslice configuration for
> its own use case. To do so, we expose a context parameter to allow
> adjustment of the RPCS register stored within the context image (and
> currently not accessible via LRI). If the context is adjusted before
> first use, the adjustment is for "free"; otherwise if the context is
> active we flush the context off the GPU (stalling all users) and forcing
> the GPU to save the context to memory where we can modify it and so
> ensure that the register is reloaded on next execution.
>
> The overhead of managing additional EU subslices can be significant,
> especially in multi-context workloads. Non-GPGPU contexts should
> preferably disable the subslices it is not using, and others should
> fine-tune the number to match their workload.
>
> We expose complete control over the RPCS register, allowing
> configuration of slice/subslice, via masks packed into a u64 for
> simplicity. For example,
>
> struct drm_i915_gem_context_param arg;
> struct drm_i915_gem_context_param_sseu sseu = { .class = 0,
> .instance = 0, };
>
> memset(&arg, 0, sizeof(arg));
> arg.ctx_id = ctx;
> arg.param = I915_CONTEXT_PARAM_SSEU;
> arg.value = (uintptr_t) &sseu;
> if (drmIoctl(fd, DRM_IOCTL_I915_GEM_CONTEXT_GETPARAM, &arg) == 0) {
> sseu.packed.subslice_mask = 0;
>
> drmIoctl(fd, DRM_IOCTL_I915_GEM_CONTEXT_SETPARAM, &arg);
> }
>
> could be used to disable all subslices where supported.
>
> v2: Fix offset of CTX_R_PWR_CLK_STATE in intel_lr_context_set_sseu() (Lionel)
>
> v3: Add ability to program this per engine (Chris)
>
> v4: Move most get_sseu() into i915_gem_context.c (Lionel)
>
> v5: Validate sseu configuration against the device's capabilities (Lionel)
>
> v6: Change context powergating settings through MI_SDM on kernel context (Chris)
>
> v7: Synchronize the requests following a powergating setting change using a global
> dependency (Chris)
> Iterate timelines through dev_priv.gt.active_rings (Tvrtko)
> Disable RPCS configuration setting for non capable users (Lionel/Tvrtko)
>
> v8: s/union intel_sseu/struct intel_sseu/ (Lionel)
> s/dev_priv/i915/ (Tvrtko)
> Change uapi class/instance fields to u16 (Tvrtko)
> Bump mask fields to 64bits (Lionel)
> Don't return EPERM when dynamic sseu is disabled (Tvrtko)
>
> v9: Import context image into kernel context's ppgtt only when
> reconfiguring powergated slice/subslices (Chris)
> Use aliasing ppgtt when needed (Michel)
>
> Tvrtko Ursulin:
>
> v10:
> * Update for upstream changes.
> * Request submit needs a RPM reference.
> * Reject on !FULL_PPGTT for simplicity.
> * Pull out get/set param to helpers for readability and less indent.
> * Use i915_request_await_dma_fence in add_global_barrier to skip waits
> on the same timeline and avoid GEM_BUG_ON.
> * No need to explicitly assign a NULL pointer to engine in legacy mode.
> * No need to move gen8_make_rpcs up.
> * Factored out global barrier as prep patch.
> * Allow to only CAP_SYS_ADMIN if !Gen11.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100899
> Issue: https://github.com/intel/media-driver/issues/267
> Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@xxxxxxxxx>
> Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@xxxxxxxxx>
> Cc: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxx>
> Cc: Zhipeng Gong <zhipeng.gong@xxxxxxxxx>
> Cc: Joonas Lahtinen <joonas.lahtinen@xxxxxxxxxxxxxxx>
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxx>
> ---
> drivers/gpu/drm/i915/i915_gem_context.c | 187 +++++++++++++++++++++++-
> drivers/gpu/drm/i915/intel_lrc.c | 55 +++++++
> drivers/gpu/drm/i915/intel_ringbuffer.h | 4 +
> include/uapi/drm/i915_drm.h | 43 ++++++
> 4 files changed, 288 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
> index 8a12984e7495..6d6220634e9e 100644
> --- a/drivers/gpu/drm/i915/i915_gem_context.c
> +++ b/drivers/gpu/drm/i915/i915_gem_context.c
> @@ -773,6 +773,91 @@ int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
> return 0;
> }
>
> +static int
> +i915_gem_context_reconfigure_sseu(struct i915_gem_context *ctx,
> + struct intel_engine_cs *engine,
> + struct intel_sseu sseu)
> +{
> + struct drm_i915_private *i915 = ctx->i915;
> + struct i915_request *rq;
> + struct intel_ring *ring;
> + int ret;
> +
> + lockdep_assert_held(&i915->drm.struct_mutex);
> +
> + /* Submitting requests etc needs the hw awake. */
> + intel_runtime_pm_get(i915);
> +
> + i915_retire_requests(i915);
?
> +
> + /* Now use the RCS to actually reconfigure. */
> + engine = i915->engine[RCS];
? Modifying registers stored in another engine's context image.
> +
> + rq = i915_request_alloc(engine, i915->kernel_context);
> + if (IS_ERR(rq)) {
> + ret = PTR_ERR(rq);
> + goto out_put;
> + }
> +
> + ret = engine->emit_rpcs_config(rq, ctx, sseu);
It's just an LRI, I'd rather we do it directly unless there's evidence
that there will be na explicit rpcs config instruction in future. It
just doesn't seem general enough.
> + if (ret)
> + goto out_add;
> +
> + /* Queue this switch after all other activity */
Only needs to be after the target ctx.
> + list_for_each_entry(ring, &i915->gt.active_rings, active_link) {
> + struct i915_request *prev;
> +
> + prev = last_request_on_engine(ring->timeline, engine);
As constructed above you need target-engine + RCS.
> + if (prev)
> + i915_sw_fence_await_sw_fence_gfp(&rq->submit,
> + &prev->submit,
> + I915_FENCE_GFP);
> + }
> +
> + i915_gem_set_global_barrier(i915, rq);
This is just for a link from ctx-engine to this rq. Overkill much?
Presumably this stems from using the wrong engine.
> +
> +out_add:
> + i915_request_add(rq);
And I'd still recommend not using indirect access if we can apply the
changes immediately.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
