linux/drivers/gpu/drm/i915/i915_gem_context.c
Chris Wilson 93c6e966b4 drm/i915: Remove i915.semaphores modparam
Having disabled the broken semaphores on Sandybridge, there is no need
for a modparam any more, so remove it in favour of a simple
HAS_LEGACY_SEMAPHORES() guard.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171120205504.21892-5-chris@chris-wilson.co.uk
2017-11-20 21:59:09 +00:00

1126 lines
30 KiB
C

/*
* 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 <linux/log2.h>
#include <drm/drmP.h>
#include <drm/i915_drm.h>
#include "i915_drv.h"
#include "i915_trace.h"
#define ALL_L3_SLICES(dev) (1 << NUM_L3_SLICES(dev)) - 1
static void lut_close(struct i915_gem_context *ctx)
{
struct i915_lut_handle *lut, *ln;
struct radix_tree_iter iter;
void __rcu **slot;
list_for_each_entry_safe(lut, ln, &ctx->handles_list, ctx_link) {
list_del(&lut->obj_link);
kmem_cache_free(ctx->i915->luts, lut);
}
rcu_read_lock();
radix_tree_for_each_slot(slot, &ctx->handles_vma, &iter, 0) {
struct i915_vma *vma = rcu_dereference_raw(*slot);
radix_tree_iter_delete(&ctx->handles_vma, &iter, slot);
__i915_gem_object_release_unless_active(vma->obj);
}
rcu_read_unlock();
}
static void i915_gem_context_free(struct i915_gem_context *ctx)
{
int i;
lockdep_assert_held(&ctx->i915->drm.struct_mutex);
GEM_BUG_ON(!i915_gem_context_is_closed(ctx));
i915_ppgtt_put(ctx->ppgtt);
for (i = 0; i < I915_NUM_ENGINES; i++) {
struct intel_context *ce = &ctx->engine[i];
if (!ce->state)
continue;
WARN_ON(ce->pin_count);
if (ce->ring)
intel_ring_free(ce->ring);
__i915_gem_object_release_unless_active(ce->state->obj);
}
kfree(ctx->name);
put_pid(ctx->pid);
list_del(&ctx->link);
ida_simple_remove(&ctx->i915->contexts.hw_ida, ctx->hw_id);
kfree_rcu(ctx, rcu);
}
static void contexts_free(struct drm_i915_private *i915)
{
struct llist_node *freed = llist_del_all(&i915->contexts.free_list);
struct i915_gem_context *ctx, *cn;
lockdep_assert_held(&i915->drm.struct_mutex);
llist_for_each_entry_safe(ctx, cn, freed, free_link)
i915_gem_context_free(ctx);
}
static void contexts_free_first(struct drm_i915_private *i915)
{
struct i915_gem_context *ctx;
struct llist_node *freed;
lockdep_assert_held(&i915->drm.struct_mutex);
freed = llist_del_first(&i915->contexts.free_list);
if (!freed)
return;
ctx = container_of(freed, typeof(*ctx), free_link);
i915_gem_context_free(ctx);
}
static void contexts_free_worker(struct work_struct *work)
{
struct drm_i915_private *i915 =
container_of(work, typeof(*i915), contexts.free_work);
mutex_lock(&i915->drm.struct_mutex);
contexts_free(i915);
mutex_unlock(&i915->drm.struct_mutex);
}
void i915_gem_context_release(struct kref *ref)
{
struct i915_gem_context *ctx = container_of(ref, typeof(*ctx), ref);
struct drm_i915_private *i915 = ctx->i915;
trace_i915_context_free(ctx);
if (llist_add(&ctx->free_link, &i915->contexts.free_list))
queue_work(i915->wq, &i915->contexts.free_work);
}
static void context_close(struct i915_gem_context *ctx)
{
i915_gem_context_set_closed(ctx);
/*
* The LUT uses the VMA as a backpointer to unref the object,
* so we need to clear the LUT before we close all the VMA (inside
* the ppgtt).
*/
lut_close(ctx);
if (ctx->ppgtt)
i915_ppgtt_close(&ctx->ppgtt->base);
ctx->file_priv = ERR_PTR(-EBADF);
i915_gem_context_put(ctx);
}
static int assign_hw_id(struct drm_i915_private *dev_priv, unsigned *out)
{
int ret;
ret = ida_simple_get(&dev_priv->contexts.hw_ida,
0, MAX_CONTEXT_HW_ID, GFP_KERNEL);
if (ret < 0) {
/* Contexts are only released when no longer active.
* Flush any pending retires to hopefully release some
* stale contexts and try again.
*/
i915_gem_retire_requests(dev_priv);
ret = ida_simple_get(&dev_priv->contexts.hw_ida,
0, MAX_CONTEXT_HW_ID, GFP_KERNEL);
if (ret < 0)
return ret;
}
*out = ret;
return 0;
}
static u32 default_desc_template(const struct drm_i915_private *i915,
const struct i915_hw_ppgtt *ppgtt)
{
u32 address_mode;
u32 desc;
desc = GEN8_CTX_VALID | GEN8_CTX_PRIVILEGE;
address_mode = INTEL_LEGACY_32B_CONTEXT;
if (ppgtt && i915_vm_is_48bit(&ppgtt->base))
address_mode = INTEL_LEGACY_64B_CONTEXT;
desc |= address_mode << GEN8_CTX_ADDRESSING_MODE_SHIFT;
if (IS_GEN8(i915))
desc |= GEN8_CTX_L3LLC_COHERENT;
/* TODO: WaDisableLiteRestore when we start using semaphore
* signalling between Command Streamers
* ring->ctx_desc_template |= GEN8_CTX_FORCE_RESTORE;
*/
return desc;
}
static struct i915_gem_context *
__create_hw_context(struct drm_i915_private *dev_priv,
struct drm_i915_file_private *file_priv)
{
struct i915_gem_context *ctx;
int ret;
ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
if (ctx == NULL)
return ERR_PTR(-ENOMEM);
ret = assign_hw_id(dev_priv, &ctx->hw_id);
if (ret) {
kfree(ctx);
return ERR_PTR(ret);
}
kref_init(&ctx->ref);
list_add_tail(&ctx->link, &dev_priv->contexts.list);
ctx->i915 = dev_priv;
ctx->priority = I915_PRIORITY_NORMAL;
INIT_RADIX_TREE(&ctx->handles_vma, GFP_KERNEL);
INIT_LIST_HEAD(&ctx->handles_list);
/* Default context will never have a file_priv */
ret = DEFAULT_CONTEXT_HANDLE;
if (file_priv) {
ret = idr_alloc(&file_priv->context_idr, ctx,
DEFAULT_CONTEXT_HANDLE, 0, GFP_KERNEL);
if (ret < 0)
goto err_lut;
}
ctx->user_handle = ret;
ctx->file_priv = file_priv;
if (file_priv) {
ctx->pid = get_task_pid(current, PIDTYPE_PID);
ctx->name = kasprintf(GFP_KERNEL, "%s[%d]/%x",
current->comm,
pid_nr(ctx->pid),
ctx->user_handle);
if (!ctx->name) {
ret = -ENOMEM;
goto err_pid;
}
}
/* 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 = ALL_L3_SLICES(dev_priv);
i915_gem_context_set_bannable(ctx);
ctx->ring_size = 4 * PAGE_SIZE;
ctx->desc_template =
default_desc_template(dev_priv, dev_priv->mm.aliasing_ppgtt);
/* GuC requires the ring to be placed above GUC_WOPCM_TOP. If GuC is not
* present or not in use we still need a small bias as ring wraparound
* at offset 0 sometimes hangs. No idea why.
*/
if (HAS_GUC(dev_priv) && i915_modparams.enable_guc_loading)
ctx->ggtt_offset_bias = GUC_WOPCM_TOP;
else
ctx->ggtt_offset_bias = I915_GTT_PAGE_SIZE;
return ctx;
err_pid:
put_pid(ctx->pid);
idr_remove(&file_priv->context_idr, ctx->user_handle);
err_lut:
context_close(ctx);
return ERR_PTR(ret);
}
static void __destroy_hw_context(struct i915_gem_context *ctx,
struct drm_i915_file_private *file_priv)
{
idr_remove(&file_priv->context_idr, ctx->user_handle);
context_close(ctx);
}
/**
* The default context needs to exist per ring that uses contexts. It stores the
* context state of the GPU for applications that don't utilize HW contexts, as
* well as an idle case.
*/
static struct i915_gem_context *
i915_gem_create_context(struct drm_i915_private *dev_priv,
struct drm_i915_file_private *file_priv)
{
struct i915_gem_context *ctx;
lockdep_assert_held(&dev_priv->drm.struct_mutex);
/* Reap the most stale context */
contexts_free_first(dev_priv);
ctx = __create_hw_context(dev_priv, file_priv);
if (IS_ERR(ctx))
return ctx;
if (USES_FULL_PPGTT(dev_priv)) {
struct i915_hw_ppgtt *ppgtt;
ppgtt = i915_ppgtt_create(dev_priv, file_priv, ctx->name);
if (IS_ERR(ppgtt)) {
DRM_DEBUG_DRIVER("PPGTT setup failed (%ld)\n",
PTR_ERR(ppgtt));
__destroy_hw_context(ctx, file_priv);
return ERR_CAST(ppgtt);
}
ctx->ppgtt = ppgtt;
ctx->desc_template = default_desc_template(dev_priv, ppgtt);
}
trace_i915_context_create(ctx);
return ctx;
}
/**
* i915_gem_context_create_gvt - create a GVT GEM context
* @dev: drm device *
*
* This function is used to create a GVT specific GEM context.
*
* Returns:
* pointer to i915_gem_context on success, error pointer if failed
*
*/
struct i915_gem_context *
i915_gem_context_create_gvt(struct drm_device *dev)
{
struct i915_gem_context *ctx;
int ret;
if (!IS_ENABLED(CONFIG_DRM_I915_GVT))
return ERR_PTR(-ENODEV);
ret = i915_mutex_lock_interruptible(dev);
if (ret)
return ERR_PTR(ret);
ctx = __create_hw_context(to_i915(dev), NULL);
if (IS_ERR(ctx))
goto out;
ctx->file_priv = ERR_PTR(-EBADF);
i915_gem_context_set_closed(ctx); /* not user accessible */
i915_gem_context_clear_bannable(ctx);
i915_gem_context_set_force_single_submission(ctx);
if (!i915_modparams.enable_guc_submission)
ctx->ring_size = 512 * PAGE_SIZE; /* Max ring buffer size */
GEM_BUG_ON(i915_gem_context_is_kernel(ctx));
out:
mutex_unlock(&dev->struct_mutex);
return ctx;
}
struct i915_gem_context *
i915_gem_context_create_kernel(struct drm_i915_private *i915, int prio)
{
struct i915_gem_context *ctx;
ctx = i915_gem_create_context(i915, NULL);
if (IS_ERR(ctx))
return ctx;
i915_gem_context_clear_bannable(ctx);
ctx->priority = prio;
ctx->ring_size = PAGE_SIZE;
GEM_BUG_ON(!i915_gem_context_is_kernel(ctx));
return ctx;
}
static void
destroy_kernel_context(struct i915_gem_context **ctxp)
{
struct i915_gem_context *ctx;
/* Keep the context ref so that we can free it immediately ourselves */
ctx = i915_gem_context_get(fetch_and_zero(ctxp));
GEM_BUG_ON(!i915_gem_context_is_kernel(ctx));
context_close(ctx);
i915_gem_context_free(ctx);
}
int i915_gem_contexts_init(struct drm_i915_private *dev_priv)
{
struct i915_gem_context *ctx;
int err;
GEM_BUG_ON(dev_priv->kernel_context);
INIT_LIST_HEAD(&dev_priv->contexts.list);
INIT_WORK(&dev_priv->contexts.free_work, contexts_free_worker);
init_llist_head(&dev_priv->contexts.free_list);
/* Using the simple ida interface, the max is limited by sizeof(int) */
BUILD_BUG_ON(MAX_CONTEXT_HW_ID > INT_MAX);
ida_init(&dev_priv->contexts.hw_ida);
/* lowest priority; idle task */
ctx = i915_gem_context_create_kernel(dev_priv, I915_PRIORITY_MIN);
if (IS_ERR(ctx)) {
DRM_ERROR("Failed to create default global context\n");
err = PTR_ERR(ctx);
goto err;
}
/*
* For easy recognisablity, we want the kernel context to be 0 and then
* all user contexts will have non-zero hw_id.
*/
GEM_BUG_ON(ctx->hw_id);
dev_priv->kernel_context = ctx;
/* highest priority; preempting task */
ctx = i915_gem_context_create_kernel(dev_priv, INT_MAX);
if (IS_ERR(ctx)) {
DRM_ERROR("Failed to create default preempt context\n");
err = PTR_ERR(ctx);
goto err_kernel_context;
}
dev_priv->preempt_context = ctx;
DRM_DEBUG_DRIVER("%s context support initialized\n",
dev_priv->engine[RCS]->context_size ? "logical" :
"fake");
return 0;
err_kernel_context:
destroy_kernel_context(&dev_priv->kernel_context);
err:
return err;
}
void i915_gem_contexts_lost(struct drm_i915_private *dev_priv)
{
struct intel_engine_cs *engine;
enum intel_engine_id id;
lockdep_assert_held(&dev_priv->drm.struct_mutex);
for_each_engine(engine, dev_priv, id) {
engine->legacy_active_context = NULL;
if (!engine->last_retired_context)
continue;
engine->context_unpin(engine, engine->last_retired_context);
engine->last_retired_context = NULL;
}
}
void i915_gem_contexts_fini(struct drm_i915_private *i915)
{
lockdep_assert_held(&i915->drm.struct_mutex);
destroy_kernel_context(&i915->preempt_context);
destroy_kernel_context(&i915->kernel_context);
/* Must free all deferred contexts (via flush_workqueue) first */
ida_destroy(&i915->contexts.hw_ida);
}
static int context_idr_cleanup(int id, void *p, void *data)
{
struct i915_gem_context *ctx = p;
context_close(ctx);
return 0;
}
int i915_gem_context_open(struct drm_i915_private *i915,
struct drm_file *file)
{
struct drm_i915_file_private *file_priv = file->driver_priv;
struct i915_gem_context *ctx;
idr_init(&file_priv->context_idr);
mutex_lock(&i915->drm.struct_mutex);
ctx = i915_gem_create_context(i915, file_priv);
mutex_unlock(&i915->drm.struct_mutex);
if (IS_ERR(ctx)) {
idr_destroy(&file_priv->context_idr);
return PTR_ERR(ctx);
}
GEM_BUG_ON(i915_gem_context_is_kernel(ctx));
return 0;
}
void i915_gem_context_close(struct drm_file *file)
{
struct drm_i915_file_private *file_priv = file->driver_priv;
lockdep_assert_held(&file_priv->dev_priv->drm.struct_mutex);
idr_for_each(&file_priv->context_idr, context_idr_cleanup, NULL);
idr_destroy(&file_priv->context_idr);
}
static inline int
mi_set_context(struct drm_i915_gem_request *req, u32 flags)
{
struct drm_i915_private *dev_priv = req->i915;
struct intel_engine_cs *engine = req->engine;
enum intel_engine_id id;
const int num_rings =
/* Use an extended w/a on gen7 if signalling from other rings */
(HAS_LEGACY_SEMAPHORES(dev_priv) && IS_GEN7(dev_priv)) ?
INTEL_INFO(dev_priv)->num_rings - 1 :
0;
int len;
u32 *cs;
flags |= MI_MM_SPACE_GTT;
if (IS_HASWELL(dev_priv))
/* These flags are for resource streamer on HSW+ */
flags |= HSW_MI_RS_SAVE_STATE_EN | HSW_MI_RS_RESTORE_STATE_EN;
else
flags |= MI_SAVE_EXT_STATE_EN | MI_RESTORE_EXT_STATE_EN;
len = 4;
if (IS_GEN7(dev_priv))
len += 2 + (num_rings ? 4*num_rings + 6 : 0);
cs = intel_ring_begin(req, len);
if (IS_ERR(cs))
return PTR_ERR(cs);
/* WaProgramMiArbOnOffAroundMiSetContext:ivb,vlv,hsw,bdw,chv */
if (IS_GEN7(dev_priv)) {
*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
if (num_rings) {
struct intel_engine_cs *signaller;
*cs++ = MI_LOAD_REGISTER_IMM(num_rings);
for_each_engine(signaller, dev_priv, id) {
if (signaller == engine)
continue;
*cs++ = i915_mmio_reg_offset(
RING_PSMI_CTL(signaller->mmio_base));
*cs++ = _MASKED_BIT_ENABLE(
GEN6_PSMI_SLEEP_MSG_DISABLE);
}
}
}
*cs++ = MI_NOOP;
*cs++ = MI_SET_CONTEXT;
*cs++ = i915_ggtt_offset(req->ctx->engine[RCS].state) | flags;
/*
* w/a: MI_SET_CONTEXT must always be followed by MI_NOOP
* WaMiSetContext_Hang:snb,ivb,vlv
*/
*cs++ = MI_NOOP;
if (IS_GEN7(dev_priv)) {
if (num_rings) {
struct intel_engine_cs *signaller;
i915_reg_t last_reg = {}; /* keep gcc quiet */
*cs++ = MI_LOAD_REGISTER_IMM(num_rings);
for_each_engine(signaller, dev_priv, id) {
if (signaller == engine)
continue;
last_reg = RING_PSMI_CTL(signaller->mmio_base);
*cs++ = i915_mmio_reg_offset(last_reg);
*cs++ = _MASKED_BIT_DISABLE(
GEN6_PSMI_SLEEP_MSG_DISABLE);
}
/* Insert a delay before the next switch! */
*cs++ = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT;
*cs++ = i915_mmio_reg_offset(last_reg);
*cs++ = i915_ggtt_offset(engine->scratch);
*cs++ = MI_NOOP;
}
*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
}
intel_ring_advance(req, cs);
return 0;
}
static int remap_l3(struct drm_i915_gem_request *req, int slice)
{
u32 *cs, *remap_info = req->i915->l3_parity.remap_info[slice];
int i;
if (!remap_info)
return 0;
cs = intel_ring_begin(req, GEN7_L3LOG_SIZE/4 * 2 + 2);
if (IS_ERR(cs))
return PTR_ERR(cs);
/*
* Note: We do not worry about the concurrent register cacheline hang
* here because no other code should access these registers other than
* at initialization time.
*/
*cs++ = MI_LOAD_REGISTER_IMM(GEN7_L3LOG_SIZE/4);
for (i = 0; i < GEN7_L3LOG_SIZE/4; i++) {
*cs++ = i915_mmio_reg_offset(GEN7_L3LOG(slice, i));
*cs++ = remap_info[i];
}
*cs++ = MI_NOOP;
intel_ring_advance(req, cs);
return 0;
}
static inline bool skip_rcs_switch(struct i915_hw_ppgtt *ppgtt,
struct intel_engine_cs *engine,
struct i915_gem_context *to)
{
if (to->remap_slice)
return false;
if (ppgtt && (intel_engine_flag(engine) & ppgtt->pd_dirty_rings))
return false;
return to == engine->legacy_active_context;
}
static bool
needs_pd_load_pre(struct i915_hw_ppgtt *ppgtt, struct intel_engine_cs *engine)
{
struct i915_gem_context *from = engine->legacy_active_context;
if (!ppgtt)
return false;
/* Always load the ppgtt on first use */
if (!from)
return true;
/* Same context without new entries, skip */
if ((!from->ppgtt || from->ppgtt == ppgtt) &&
!(intel_engine_flag(engine) & ppgtt->pd_dirty_rings))
return false;
if (engine->id != RCS)
return true;
return true;
}
static int do_rcs_switch(struct drm_i915_gem_request *req)
{
struct i915_gem_context *to = req->ctx;
struct intel_engine_cs *engine = req->engine;
struct i915_hw_ppgtt *ppgtt = to->ppgtt ?: req->i915->mm.aliasing_ppgtt;
struct i915_gem_context *from = engine->legacy_active_context;
u32 hw_flags;
int ret, i;
GEM_BUG_ON(engine->id != RCS);
if (skip_rcs_switch(ppgtt, engine, to))
return 0;
if (needs_pd_load_pre(ppgtt, engine)) {
/* Older GENs and non render rings still want the load first,
* "PP_DCLV followed by PP_DIR_BASE register through Load
* Register Immediate commands in Ring Buffer before submitting
* a context."*/
trace_switch_mm(engine, to);
ret = ppgtt->switch_mm(ppgtt, req);
if (ret)
return ret;
}
if (i915_gem_context_is_kernel(to))
/*
* The kernel context(s) is treated as pure scratch and is not
* expected to retain any state (as we sacrifice it during
* suspend and on resume it may be corrupted). This is ok,
* as nothing actually executes using the kernel context; it
* is purely used for flushing user contexts.
*/
hw_flags = MI_RESTORE_INHIBIT;
else if (ppgtt && intel_engine_flag(engine) & ppgtt->pd_dirty_rings)
hw_flags = MI_FORCE_RESTORE;
else
hw_flags = 0;
if (to != from || (hw_flags & MI_FORCE_RESTORE)) {
ret = mi_set_context(req, hw_flags);
if (ret)
return ret;
engine->legacy_active_context = to;
}
if (ppgtt)
ppgtt->pd_dirty_rings &= ~intel_engine_flag(engine);
for (i = 0; i < MAX_L3_SLICES; i++) {
if (!(to->remap_slice & (1<<i)))
continue;
ret = remap_l3(req, i);
if (ret)
return ret;
to->remap_slice &= ~(1<<i);
}
return 0;
}
/**
* i915_switch_context() - perform a GPU context switch.
* @req: request for which we'll execute the context switch
*
* The context life cycle is simple. The context refcount is incremented and
* decremented by 1 and create and destroy. If the context is in use by the GPU,
* it will have a refcount > 1. This allows us to destroy the context abstract
* object while letting the normal object tracking destroy the backing BO.
*
* This function should not be used in execlists mode. Instead the context is
* switched by writing to the ELSP and requests keep a reference to their
* context.
*/
int i915_switch_context(struct drm_i915_gem_request *req)
{
struct intel_engine_cs *engine = req->engine;
lockdep_assert_held(&req->i915->drm.struct_mutex);
GEM_BUG_ON(HAS_EXECLISTS(req->i915));
if (!req->ctx->engine[engine->id].state) {
struct i915_gem_context *to = req->ctx;
struct i915_hw_ppgtt *ppgtt =
to->ppgtt ?: req->i915->mm.aliasing_ppgtt;
if (needs_pd_load_pre(ppgtt, engine)) {
int ret;
trace_switch_mm(engine, to);
ret = ppgtt->switch_mm(ppgtt, req);
if (ret)
return ret;
ppgtt->pd_dirty_rings &= ~intel_engine_flag(engine);
}
engine->legacy_active_context = to;
return 0;
}
return do_rcs_switch(req);
}
static bool engine_has_idle_kernel_context(struct intel_engine_cs *engine)
{
struct i915_gem_timeline *timeline;
list_for_each_entry(timeline, &engine->i915->gt.timelines, link) {
struct intel_timeline *tl;
if (timeline == &engine->i915->gt.global_timeline)
continue;
tl = &timeline->engine[engine->id];
if (i915_gem_active_peek(&tl->last_request,
&engine->i915->drm.struct_mutex))
return false;
}
return intel_engine_has_kernel_context(engine);
}
int i915_gem_switch_to_kernel_context(struct drm_i915_private *dev_priv)
{
struct intel_engine_cs *engine;
struct i915_gem_timeline *timeline;
enum intel_engine_id id;
lockdep_assert_held(&dev_priv->drm.struct_mutex);
i915_gem_retire_requests(dev_priv);
for_each_engine(engine, dev_priv, id) {
struct drm_i915_gem_request *req;
if (engine_has_idle_kernel_context(engine))
continue;
req = i915_gem_request_alloc(engine, dev_priv->kernel_context);
if (IS_ERR(req))
return PTR_ERR(req);
/* Queue this switch after all other activity */
list_for_each_entry(timeline, &dev_priv->gt.timelines, link) {
struct drm_i915_gem_request *prev;
struct intel_timeline *tl;
tl = &timeline->engine[engine->id];
prev = i915_gem_active_raw(&tl->last_request,
&dev_priv->drm.struct_mutex);
if (prev)
i915_sw_fence_await_sw_fence_gfp(&req->submit,
&prev->submit,
GFP_KERNEL);
}
i915_add_request(req);
}
return 0;
}
static bool client_is_banned(struct drm_i915_file_private *file_priv)
{
return atomic_read(&file_priv->context_bans) > I915_MAX_CLIENT_CONTEXT_BANS;
}
int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
struct drm_file *file)
{
struct drm_i915_private *dev_priv = to_i915(dev);
struct drm_i915_gem_context_create *args = data;
struct drm_i915_file_private *file_priv = file->driver_priv;
struct i915_gem_context *ctx;
int ret;
if (!dev_priv->engine[RCS]->context_size)
return -ENODEV;
if (args->pad != 0)
return -EINVAL;
if (client_is_banned(file_priv)) {
DRM_DEBUG("client %s[%d] banned from creating ctx\n",
current->comm,
pid_nr(get_task_pid(current, PIDTYPE_PID)));
return -EIO;
}
ret = i915_mutex_lock_interruptible(dev);
if (ret)
return ret;
ctx = i915_gem_create_context(dev_priv, file_priv);
mutex_unlock(&dev->struct_mutex);
if (IS_ERR(ctx))
return PTR_ERR(ctx);
GEM_BUG_ON(i915_gem_context_is_kernel(ctx));
args->ctx_id = ctx->user_handle;
DRM_DEBUG("HW context %d created\n", args->ctx_id);
return 0;
}
int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
struct drm_file *file)
{
struct drm_i915_gem_context_destroy *args = data;
struct drm_i915_file_private *file_priv = file->driver_priv;
struct i915_gem_context *ctx;
int ret;
if (args->pad != 0)
return -EINVAL;
if (args->ctx_id == DEFAULT_CONTEXT_HANDLE)
return -ENOENT;
ctx = i915_gem_context_lookup(file_priv, args->ctx_id);
if (!ctx)
return -ENOENT;
ret = mutex_lock_interruptible(&dev->struct_mutex);
if (ret)
goto out;
__destroy_hw_context(ctx, file_priv);
mutex_unlock(&dev->struct_mutex);
out:
i915_gem_context_put(ctx);
return 0;
}
int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data,
struct drm_file *file)
{
struct drm_i915_file_private *file_priv = file->driver_priv;
struct drm_i915_gem_context_param *args = data;
struct i915_gem_context *ctx;
int ret = 0;
ctx = i915_gem_context_lookup(file_priv, args->ctx_id);
if (!ctx)
return -ENOENT;
args->size = 0;
switch (args->param) {
case I915_CONTEXT_PARAM_BAN_PERIOD:
ret = -EINVAL;
break;
case I915_CONTEXT_PARAM_NO_ZEROMAP:
args->value = ctx->flags & CONTEXT_NO_ZEROMAP;
break;
case I915_CONTEXT_PARAM_GTT_SIZE:
if (ctx->ppgtt)
args->value = ctx->ppgtt->base.total;
else if (to_i915(dev)->mm.aliasing_ppgtt)
args->value = to_i915(dev)->mm.aliasing_ppgtt->base.total;
else
args->value = to_i915(dev)->ggtt.base.total;
break;
case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE:
args->value = i915_gem_context_no_error_capture(ctx);
break;
case I915_CONTEXT_PARAM_BANNABLE:
args->value = i915_gem_context_is_bannable(ctx);
break;
case I915_CONTEXT_PARAM_PRIORITY:
args->value = ctx->priority;
break;
default:
ret = -EINVAL;
break;
}
i915_gem_context_put(ctx);
return ret;
}
int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
struct drm_file *file)
{
struct drm_i915_file_private *file_priv = file->driver_priv;
struct drm_i915_gem_context_param *args = data;
struct i915_gem_context *ctx;
int ret;
ctx = i915_gem_context_lookup(file_priv, args->ctx_id);
if (!ctx)
return -ENOENT;
ret = i915_mutex_lock_interruptible(dev);
if (ret)
goto out;
switch (args->param) {
case I915_CONTEXT_PARAM_BAN_PERIOD:
ret = -EINVAL;
break;
case I915_CONTEXT_PARAM_NO_ZEROMAP:
if (args->size) {
ret = -EINVAL;
} else {
ctx->flags &= ~CONTEXT_NO_ZEROMAP;
ctx->flags |= args->value ? CONTEXT_NO_ZEROMAP : 0;
}
break;
case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE:
if (args->size)
ret = -EINVAL;
else if (args->value)
i915_gem_context_set_no_error_capture(ctx);
else
i915_gem_context_clear_no_error_capture(ctx);
break;
case I915_CONTEXT_PARAM_BANNABLE:
if (args->size)
ret = -EINVAL;
else if (!capable(CAP_SYS_ADMIN) && !args->value)
ret = -EPERM;
else if (args->value)
i915_gem_context_set_bannable(ctx);
else
i915_gem_context_clear_bannable(ctx);
break;
case I915_CONTEXT_PARAM_PRIORITY:
{
int priority = args->value;
if (args->size)
ret = -EINVAL;
else if (!to_i915(dev)->engine[RCS]->schedule)
ret = -ENODEV;
else if (priority > I915_CONTEXT_MAX_USER_PRIORITY ||
priority < I915_CONTEXT_MIN_USER_PRIORITY)
ret = -EINVAL;
else if (priority > I915_CONTEXT_DEFAULT_PRIORITY &&
!capable(CAP_SYS_NICE))
ret = -EPERM;
else
ctx->priority = priority;
}
break;
default:
ret = -EINVAL;
break;
}
mutex_unlock(&dev->struct_mutex);
out:
i915_gem_context_put(ctx);
return ret;
}
int i915_gem_context_reset_stats_ioctl(struct drm_device *dev,
void *data, struct drm_file *file)
{
struct drm_i915_private *dev_priv = to_i915(dev);
struct drm_i915_reset_stats *args = data;
struct i915_gem_context *ctx;
int ret;
if (args->flags || args->pad)
return -EINVAL;
ret = -ENOENT;
rcu_read_lock();
ctx = __i915_gem_context_lookup_rcu(file->driver_priv, args->ctx_id);
if (!ctx)
goto out;
/*
* We opt for unserialised reads here. This may result in tearing
* in the extremely unlikely event of a GPU hang on this context
* as we are querying them. If we need that extra layer of protection,
* we should wrap the hangstats with a seqlock.
*/
if (capable(CAP_SYS_ADMIN))
args->reset_count = i915_reset_count(&dev_priv->gpu_error);
else
args->reset_count = 0;
args->batch_active = atomic_read(&ctx->guilty_count);
args->batch_pending = atomic_read(&ctx->active_count);
ret = 0;
out:
rcu_read_unlock();
return ret;
}
#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
#include "selftests/mock_context.c"
#include "selftests/i915_gem_context.c"
#endif