/*
* Copyright © 2011-2012 Intel Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*
* Authors:
* Ben Widawsky <ben@bwidawsk.net>
*
*/
/*
* This file implements HW context support. On gen5+ a HW context consists of an
* opaque GPU object which is referenced at times of context saves and restores.
* With RC6 enabled, the context is also referenced as the GPU enters and exists
* from RC6 (GPU has it's own internal power context, except on gen5). Though
* something like a context does exist for the media ring, the code only
* supports contexts for the render ring.
*
* In software, there is a distinction between contexts created by the user,
* and the default HW context. The default HW context is used by GPU clients
* that do not request setup of their own hardware context. The default
* context's state is never restored to help prevent programming errors. This
* would happen if a client ran and piggy-backed off another clients GPU state.
* The default context only exists to give the GPU some offset to load as the
* current to invoke a save of the context we actually care about. In fact, the
* code could likely be constructed, albeit in a more complicated fashion, to
* never use the default context, though that limits the driver's ability to
* swap out, and/or destroy other contexts.
*
* All other contexts are created as a request by the GPU client. These contexts
* store GPU state, and thus allow GPU clients to not re-emit state (and
* potentially query certain state) at any time. The kernel driver makes
* certain that the appropriate commands are inserted.
*
* The context life cycle is semi-complicated in that context BOs may live
* longer than the context itself because of the way the hardware, and object
* tracking works. Below is a very crude representation of the state machine
* describing the context life.
* refcount pincount active
* S0: initial state 0 0 0
* S1: context created 1 0 0
* S2: context is currently running 2 1 X
* S3: GPU referenced, but not current 2 0 1
* S4: context is current, but destroyed 1 1 0
* S5: like S3, but destroyed 1 0 1
*
* The most common (but not all) transitions:
* S0->S1: client creates a context
* S1->S2: client submits execbuf with context
* S2->S3: other clients submits execbuf with context
* S3->S1: context object was retired
* S3->S2: clients submits another execbuf
* S2->S4: context destroy called with current context
* S3->S5->S0: destroy path
* S4->S5->S0: destroy path on current context
*
* There are two confusing terms used above:
* The "current context" means the context which is currently running on the
* GPU. The GPU has loaded its state already and has stored away the gtt
* offset of the BO. The GPU is not actively referencing the data at this
* offset, but it will on the next context switch. The only way to avoid this
* is to do a GPU reset.
*
* An "active context' is one which was previously the "current context" and is
* on the active list waiting for the next context switch to occur. Until this
* happens, the object must remain at the same gtt offset. It is therefore
* possible to destroy a context, but it is still active.
*
*/
#include <drm/drmP.h>
#include <drm/i915_drm.h>
#include "i915_drv.h"
#include "i915_trace.h"
/* This is a HW constraint. The value below is the largest known requirement
* I've seen in a spec to date, and that was a workaround for a non-shipping
* part. It should be safe to decrease this, but it's more future proof as is.
*/
#define GEN6_CONTEXT_ALIGN (64<<10)
#define GEN7_CONTEXT_ALIGN 4096
static size_t get_context_alignment(struct drm_device *dev)
{
if (IS_GEN6(dev))
return GEN6_CONTEXT_ALIGN;
return GEN7_CONTEXT_ALIGN;
}
static int get_context_size(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int ret;
u32 reg;
switch (INTEL_INFO(dev)->gen) {
case 6:
reg = I915_READ(CXT_SIZE);
ret = GEN6_CXT_TOTAL_SIZE(reg) * 64;
break;
case 7:
reg = I915_READ(GEN7_CXT_SIZE);
if (IS_HASWELL(dev))
ret = HSW_CXT_TOTAL_SIZE;
else
ret = GEN7_CXT_TOTAL_SIZE(reg) * 64;
break;
case 8:
ret = GEN8_CXT_TOTAL_SIZE;
break;
default:
BUG();
}
return ret;
}
void i915_gem_context_free(struct kref *ctx_ref)
{
struct intel_context *ctx = container_of(ctx_ref,
typeof(*ctx), ref);
trace_i915_context_free(ctx);
if (i915.enable_execlists)
intel_lr_context_free(ctx);
i915_ppgtt_put(ctx->ppgtt);
if (ctx->legacy_hw_ctx.rcs_state)
drm_gem_object_unreference(&ctx->legacy_hw_ctx.rcs_state->base);
list_del(&ctx->link);
kfree(ctx);
}
struct drm_i915_gem_object *
i915_gem_alloc_context_obj(struct drm_device *dev, size_t size)
{
struct drm_i915_gem_object *obj;
int ret;
obj = i915_gem_alloc_object(dev, size);
if (obj == NULL)
return ERR_PTR(-ENOMEM);
/*
* Try to make the context utilize L3 as well as LLC.
*
* On VLV we don't have L3 controls in the PTEs so we
* shouldn't touch the cache level, especially as that
* would make the object snooped which might have a
* negative performance impact.
*/
if (INTEL_INFO(dev)->gen >= 7 && !IS_VALLEYVIEW(dev)) {
ret = i915_gem_object_set_cache_level(obj, I915_CACHE_L3_LLC);
/* Failure shouldn't ever happen this early */
if (WARN_ON(ret)) {
drm_gem_object_unreference(&obj->base);
return ERR_PTR(ret);
}
}
return obj;
}
static struct intel_context *
__create_hw_context(struct drm_device *dev,
struct drm_i915_file_private *file_priv)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_context *ctx;
int ret;
ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
if (ctx == NULL)
return ERR_PTR(-ENOMEM);
kref_init(&ctx->ref);
list_add_tail(&ctx->link, &dev_priv->context_list);
if (dev_priv->hw_context_size) {
struct drm_i915_gem_object *obj =
i915_gem_alloc_context_obj(dev, dev_priv->hw_context_size);
if (IS_ERR(obj)) {
ret = PTR_ERR(obj);
goto err_out;
}
ctx->legacy_hw_ctx.rcs_state = obj;
}
/* Default context will never have a file_priv */
if (file_priv != NULL) {
ret = idr_alloc(&file_priv->context_idr, ctx,
DEFAULT_CONTEXT_HANDLE, 0, GFP_KERNEL);
if (ret < 0)
goto err_out;
} else
ret = DEFAULT_CONTEXT_HANDLE;
ctx->file_priv = file_priv;
ctx->user_handle = ret;
/* NB: Mark all slices as needing a remap so that when the context first
* loads it will restore whatever remap state already exists. If there
* is no remap info, it will be a NOP. */
ctx->remap_slice = (1 << NUM_L3_SLICES(dev)) - 1;
return ctx;
err_out:
i915_gem_context_unreference(ctx);
return ERR_PTR(ret);
}
/**
* The
