diff options
Diffstat (limited to 'drivers/gpu/drm/i915/i915_perf.c')
| -rw-r--r-- | drivers/gpu/drm/i915/i915_perf.c | 1374 |
1 files changed, 898 insertions, 476 deletions
diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 231388d06c82..d2ac51fe4f04 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -196,7 +196,9 @@ #include <linux/uuid.h> #include "gem/i915_gem_context.h" -#include "gem/i915_gem_pm.h" +#include "gt/intel_engine_pm.h" +#include "gt/intel_engine_user.h" +#include "gt/intel_gt.h" #include "gt/intel_lrc_reg.h" #include "i915_drv.h" @@ -342,11 +344,14 @@ static const struct i915_oa_format gen8_plus_oa_formats[I915_OA_FORMAT_MAX] = { * struct perf_open_properties - for validated properties given to open a stream * @sample_flags: `DRM_I915_PERF_PROP_SAMPLE_*` properties are tracked as flags * @single_context: Whether a single or all gpu contexts should be monitored + * @hold_preemption: Whether the preemption is disabled for the filtered + * context * @ctx_handle: A gem ctx handle for use with @single_context * @metrics_set: An ID for an OA unit metric set advertised via sysfs * @oa_format: An OA unit HW report format * @oa_periodic: Whether to enable periodic OA unit sampling * @oa_period_exponent: The OA unit sampling period is derived from this + * @engine: The engine (typically rcs0) being monitored by the OA unit * * As read_properties_unlocked() enumerates and validates the properties given * to open a stream of metrics the configuration is built up in the structure @@ -356,6 +361,7 @@ struct perf_open_properties { u32 sample_flags; u64 single_context:1; + u64 hold_preemption:1; u64 ctx_handle; /* OA sampling state */ @@ -363,69 +369,66 @@ struct perf_open_properties { int oa_format; bool oa_periodic; int oa_period_exponent; + + struct intel_engine_cs *engine; +}; + +struct i915_oa_config_bo { + struct llist_node node; + + struct i915_oa_config *oa_config; + struct i915_vma *vma; }; static enum hrtimer_restart oa_poll_check_timer_cb(struct hrtimer *hrtimer); -static void free_oa_config(struct drm_i915_private *dev_priv, - struct i915_oa_config *oa_config) +void i915_oa_config_release(struct kref *ref) { - if (!PTR_ERR(oa_config->flex_regs)) - kfree(oa_config->flex_regs); - if (!PTR_ERR(oa_config->b_counter_regs)) - kfree(oa_config->b_counter_regs); - if (!PTR_ERR(oa_config->mux_regs)) - kfree(oa_config->mux_regs); - kfree(oa_config); -} + struct i915_oa_config *oa_config = + container_of(ref, typeof(*oa_config), ref); -static void put_oa_config(struct drm_i915_private *dev_priv, - struct i915_oa_config *oa_config) -{ - if (!atomic_dec_and_test(&oa_config->ref_count)) - return; + kfree(oa_config->flex_regs); + kfree(oa_config->b_counter_regs); + kfree(oa_config->mux_regs); - free_oa_config(dev_priv, oa_config); + kfree_rcu(oa_config, rcu); } -static int get_oa_config(struct drm_i915_private *dev_priv, - int metrics_set, - struct i915_oa_config **out_config) +struct i915_oa_config * +i915_perf_get_oa_config(struct i915_perf *perf, int metrics_set) { - int ret; - - if (metrics_set == 1) { - *out_config = &dev_priv->perf.test_config; - atomic_inc(&dev_priv->perf.test_config.ref_count); - return 0; - } - - ret = mutex_lock_interruptible(&dev_priv->perf.metrics_lock); - if (ret) - return ret; + struct i915_oa_config *oa_config; - *out_config = idr_find(&dev_priv->perf.metrics_idr, metrics_set); - if (!*out_config) - ret = -EINVAL; + rcu_read_lock(); + if (metrics_set == 1) + oa_config = &perf->test_config; else - atomic_inc(&(*out_config)->ref_count); + oa_config = idr_find(&perf->metrics_idr, metrics_set); + if (oa_config) + oa_config = i915_oa_config_get(oa_config); + rcu_read_unlock(); - mutex_unlock(&dev_priv->perf.metrics_lock); + return oa_config; +} - return ret; +static void free_oa_config_bo(struct i915_oa_config_bo *oa_bo) +{ + i915_oa_config_put(oa_bo->oa_config); + i915_vma_put(oa_bo->vma); + kfree(oa_bo); } static u32 gen8_oa_hw_tail_read(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; - return I915_READ(GEN8_OATAILPTR) & GEN8_OATAILPTR_MASK; + return intel_uncore_read(uncore, GEN8_OATAILPTR) & GEN8_OATAILPTR_MASK; } static u32 gen7_oa_hw_tail_read(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; - u32 oastatus1 = I915_READ(GEN7_OASTATUS1); + struct intel_uncore *uncore = stream->uncore; + u32 oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1); return oastatus1 & GEN7_OASTATUS1_TAIL_MASK; } @@ -456,7 +459,6 @@ static u32 gen7_oa_hw_tail_read(struct i915_perf_stream *stream) */ static bool oa_buffer_check_unlocked(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; int report_size = stream->oa_buffer.format_size; unsigned long flags; unsigned int aged_idx; @@ -479,7 +481,7 @@ static bool oa_buffer_check_unlocked(struct i915_perf_stream *stream) aged_tail = stream->oa_buffer.tails[aged_idx].offset; aging_tail = stream->oa_buffer.tails[!aged_idx].offset; - hw_tail = dev_priv->perf.ops.oa_hw_tail_read(stream); + hw_tail = stream->perf->ops.oa_hw_tail_read(stream); /* The tail pointer increases in 64 byte increments, * not in report_size steps... @@ -655,7 +657,7 @@ static int gen8_append_oa_reports(struct i915_perf_stream *stream, size_t count, size_t *offset) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; int report_size = stream->oa_buffer.format_size; u8 *oa_buf_base = stream->oa_buffer.vaddr; u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma); @@ -740,7 +742,7 @@ static int gen8_append_oa_reports(struct i915_perf_stream *stream, reason = ((report32[0] >> OAREPORT_REASON_SHIFT) & OAREPORT_REASON_MASK); if (reason == 0) { - if (__ratelimit(&dev_priv->perf.spurious_report_rs)) + if (__ratelimit(&stream->perf->spurious_report_rs)) DRM_NOTE("Skipping spurious, invalid OA report\n"); continue; } @@ -755,7 +757,7 @@ static int gen8_append_oa_reports(struct i915_perf_stream *stream, * Note: that we don't clear the valid_ctx_bit so userspace can * understand that the ID has been squashed by the kernel. */ - if (!(report32[0] & dev_priv->perf.gen8_valid_ctx_bit)) + if (!(report32[0] & stream->perf->gen8_valid_ctx_bit)) ctx_id = report32[2] = INVALID_CTX_ID; /* @@ -789,7 +791,7 @@ static int gen8_append_oa_reports(struct i915_perf_stream *stream, * switches since it's not-uncommon for periodic samples to * identify a switch before any 'context switch' report. */ - if (!dev_priv->perf.exclusive_stream->ctx || + if (!stream->perf->exclusive_stream->ctx || stream->specific_ctx_id == ctx_id || stream->oa_buffer.last_ctx_id == stream->specific_ctx_id || reason & OAREPORT_REASON_CTX_SWITCH) { @@ -798,7 +800,7 @@ static int gen8_append_oa_reports(struct i915_perf_stream *stream, * While filtering for a single context we avoid * leaking the IDs of other contexts. */ - if (dev_priv->perf.exclusive_stream->ctx && + if (stream->perf->exclusive_stream->ctx && stream->specific_ctx_id != ctx_id) { report32[2] = INVALID_CTX_ID; } @@ -830,7 +832,8 @@ static int gen8_append_oa_reports(struct i915_perf_stream *stream, */ head += gtt_offset; - I915_WRITE(GEN8_OAHEADPTR, head & GEN8_OAHEADPTR_MASK); + intel_uncore_write(uncore, GEN8_OAHEADPTR, + head & GEN8_OAHEADPTR_MASK); stream->oa_buffer.head = head; spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags); @@ -864,14 +867,14 @@ static int gen8_oa_read(struct i915_perf_stream *stream, size_t count, size_t *offset) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; u32 oastatus; int ret; if (WARN_ON(!stream->oa_buffer.vaddr)) return -EIO; - oastatus = I915_READ(GEN8_OASTATUS); + oastatus = intel_uncore_read(uncore, GEN8_OASTATUS); /* * We treat OABUFFER_OVERFLOW as a significant error: @@ -896,14 +899,14 @@ static int gen8_oa_read(struct i915_perf_stream *stream, DRM_DEBUG("OA buffer overflow (exponent = %d): force restart\n", stream->period_exponent); - dev_priv->perf.ops.oa_disable(stream); - dev_priv->perf.ops.oa_enable(stream); + stream->perf->ops.oa_disable(stream); + stream->perf->ops.oa_enable(stream); /* * Note: .oa_enable() is expected to re-init the oabuffer and * reset GEN8_OASTATUS for us */ - oastatus = I915_READ(GEN8_OASTATUS); + oastatus = intel_uncore_read(uncore, GEN8_OASTATUS); } if (oastatus & GEN8_OASTATUS_REPORT_LOST) { @@ -911,8 +914,8 @@ static int gen8_oa_read(struct i915_perf_stream *stream, DRM_I915_PERF_RECORD_OA_REPORT_LOST); if (ret) return ret; - I915_WRITE(GEN8_OASTATUS, - oastatus & ~GEN8_OASTATUS_REPORT_LOST); + intel_uncore_write(uncore, GEN8_OASTATUS, + oastatus & ~GEN8_OASTATUS_REPORT_LOST); } return gen8_append_oa_reports(stream, buf, count, offset); @@ -943,7 +946,7 @@ static int gen7_append_oa_reports(struct i915_perf_stream *stream, size_t count, size_t *offset) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; int report_size = stream->oa_buffer.format_size; u8 *oa_buf_base = stream->oa_buffer.vaddr; u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma); @@ -1017,7 +1020,7 @@ static int gen7_append_oa_reports(struct i915_perf_stream *stream, * copying it to userspace... */ if (report32[0] == 0) { - if (__ratelimit(&dev_priv->perf.spurious_report_rs)) + if (__ratelimit(&stream->perf->spurious_report_rs)) DRM_NOTE("Skipping spurious, invalid OA report\n"); continue; } @@ -1043,9 +1046,9 @@ static int gen7_append_oa_reports(struct i915_perf_stream *stream, */ head += gtt_offset; - I915_WRITE(GEN7_OASTATUS2, - ((head & GEN7_OASTATUS2_HEAD_MASK) | - GEN7_OASTATUS2_MEM_SELECT_GGTT)); + intel_uncore_write(uncore, GEN7_OASTATUS2, + (head & GEN7_OASTATUS2_HEAD_MASK) | + GEN7_OASTATUS2_MEM_SELECT_GGTT); stream->oa_buffer.head = head; spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags); @@ -1075,21 +1078,21 @@ static int gen7_oa_read(struct i915_perf_stream *stream, size_t count, size_t *offset) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; u32 oastatus1; int ret; if (WARN_ON(!stream->oa_buffer.vaddr)) return -EIO; - oastatus1 = I915_READ(GEN7_OASTATUS1); + oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1); /* XXX: On Haswell we don't have a safe way to clear oastatus1 * bits while the OA unit is enabled (while the tail pointer * may be updated asynchronously) so we ignore status bits * that have already been reported to userspace. */ - oastatus1 &= ~dev_priv->perf.gen7_latched_oastatus1; + oastatus1 &= ~stream->perf->gen7_latched_oastatus1; /* We treat OABUFFER_OVERFLOW as a significant error: * @@ -1120,10 +1123,10 @@ static int gen7_oa_read(struct i915_perf_stream *stream, DRM_DEBUG("OA buffer overflow (exponent = %d): force restart\n", stream->period_exponent); - dev_priv->perf.ops.oa_disable(stream); - dev_priv->perf.ops.oa_enable(stream); + stream->perf->ops.oa_disable(stream); + stream->perf->ops.oa_enable(stream); - oastatus1 = I915_READ(GEN7_OASTATUS1); + oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1); } if (unlikely(oastatus1 & GEN7_OASTATUS1_REPORT_LOST)) { @@ -1131,7 +1134,7 @@ static int gen7_oa_read(struct i915_perf_stream *stream, DRM_I915_PERF_RECORD_OA_REPORT_LOST); if (ret) return ret; - dev_priv->perf.gen7_latched_oastatus1 |= + stream->perf->gen7_latched_oastatus1 |= GEN7_OASTATUS1_REPORT_LOST; } @@ -1196,9 +1199,7 @@ static int i915_oa_read(struct i915_perf_stream *stream, size_t count, size_t *offset) { - struct drm_i915_private *dev_priv = stream->dev_priv; - - return dev_priv->perf.ops.read(stream, buf, count, offset); + return stream->perf->ops.read(stream, buf, count, offset); } static struct intel_context *oa_pin_context(struct i915_perf_stream *stream) @@ -1209,7 +1210,7 @@ static struct intel_context *oa_pin_context(struct i915_perf_stream *stream) int err; for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) { - if (ce->engine->class != RENDER_CLASS) + if (ce->engine != stream->engine) /* first match! */ continue; /* @@ -1239,14 +1240,13 @@ static struct intel_context *oa_pin_context(struct i915_perf_stream *stream) */ static int oa_get_render_ctx_id(struct i915_perf_stream *stream) { - struct drm_i915_private *i915 = stream->dev_priv; struct intel_context *ce; ce = oa_pin_context(stream); if (IS_ERR(ce)) return PTR_ERR(ce); - switch (INTEL_GEN(i915)) { + switch (INTEL_GEN(ce->engine->i915)) { case 7: { /* * On Haswell we don't do any post processing of the reports @@ -1260,7 +1260,7 @@ static int oa_get_render_ctx_id(struct i915_perf_stream *stream) case 8: case 9: case 10: - if (USES_GUC_SUBMISSION(i915)) { + if (USES_GUC_SUBMISSION(ce->engine->i915)) { /* * When using GuC, the context descriptor we write in * i915 is read by GuC and rewritten before it's @@ -1296,7 +1296,7 @@ static int oa_get_render_ctx_id(struct i915_perf_stream *stream) } default: - MISSING_CASE(INTEL_GEN(i915)); + MISSING_CASE(INTEL_GEN(ce->engine->i915)); } ce->tag = stream->specific_ctx_id_mask; @@ -1338,40 +1338,55 @@ free_oa_buffer(struct i915_perf_stream *stream) stream->oa_buffer.vaddr = NULL; } +static void +free_oa_configs(struct i915_perf_stream *stream) +{ + struct i915_oa_config_bo *oa_bo, *tmp; + + i915_oa_config_put(stream->oa_config); + llist_for_each_entry_safe(oa_bo, tmp, stream->oa_config_bos.first, node) + free_oa_config_bo(oa_bo); +} + +static void +free_noa_wait(struct i915_perf_stream *stream) +{ + i915_vma_unpin_and_release(&stream->noa_wait, 0); +} + static void i915_oa_stream_destroy(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct i915_perf *perf = stream->perf; - BUG_ON(stream != dev_priv->perf.exclusive_stream); + BUG_ON(stream != perf->exclusive_stream); /* * Unset exclusive_stream first, it will be checked while disabling * the metric set on gen8+. */ - mutex_lock(&dev_priv->drm.struct_mutex); - dev_priv->perf.exclusive_stream = NULL; - dev_priv->perf.ops.disable_metric_set(stream); - mutex_unlock(&dev_priv->drm.struct_mutex); + perf->exclusive_stream = NULL; + perf->ops.disable_metric_set(stream); free_oa_buffer(stream); - intel_uncore_forcewake_put(&dev_priv->uncore, FORCEWAKE_ALL); - intel_runtime_pm_put(&dev_priv->runtime_pm, stream->wakeref); + intel_uncore_forcewake_put(stream->uncore, FORCEWAKE_ALL); + intel_engine_pm_put(stream->engine); if (stream->ctx) oa_put_render_ctx_id(stream); - put_oa_config(dev_priv, stream->oa_config); + free_oa_configs(stream); + free_noa_wait(stream); - if (dev_priv->perf.spurious_report_rs.missed) { + if (perf->spurious_report_rs.missed) { DRM_NOTE("%d spurious OA report notices suppressed due to ratelimiting\n", - dev_priv->perf.spurious_report_rs.missed); + perf->spurious_report_rs.missed); } } static void gen7_init_oa_buffer(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma); unsigned long flags; @@ -1380,13 +1395,14 @@ static void gen7_init_oa_buffer(struct i915_perf_stream *stream) /* Pre-DevBDW: OABUFFER must be set with counters off, * before OASTATUS1, but after OASTATUS2 */ - I915_WRITE(GEN7_OASTATUS2, - gtt_offset | GEN7_OASTATUS2_MEM_SELECT_GGTT); /* head */ + intel_uncore_write(uncore, GEN7_OASTATUS2, /* head */ + gtt_offset | GEN7_OASTATUS2_MEM_SELECT_GGTT); stream->oa_buffer.head = gtt_offset; - I915_WRITE(GEN7_OABUFFER, gtt_offset); + intel_uncore_write(uncore, GEN7_OABUFFER, gtt_offset); - I915_WRITE(GEN7_OASTATUS1, gtt_offset | OABUFFER_SIZE_16M); /* tail */ + intel_uncore_write(uncore, GEN7_OASTATUS1, /* tail */ + gtt_offset | OABUFFER_SIZE_16M); /* Mark that we need updated tail pointers to read from... */ stream->oa_buffer.tails[0].offset = INVALID_TAIL_PTR; @@ -1398,7 +1414,7 @@ static void gen7_init_oa_buffer(struct i915_perf_stream *stream) * already seen since they can't be cleared while periodic * sampling is enabled. */ - dev_priv->perf.gen7_latched_oastatus1 = 0; + stream->perf->gen7_latched_oastatus1 = 0; /* NB: although the OA buffer will initially be allocated * zeroed via shmfs (and so this memset is redundant when @@ -1413,25 +1429,22 @@ static void gen7_init_oa_buffer(struct i915_perf_stream *stream) */ memset(stream->oa_buffer.vaddr, 0, OA_BUFFER_SIZE); - /* Maybe make ->pollin per-stream state if we support multiple - * concurrent streams in the future. - */ stream->pollin = false; } static void gen8_init_oa_buffer(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma); unsigned long flags; spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags); - I915_WRITE(GEN8_OASTATUS, 0); - I915_WRITE(GEN8_OAHEADPTR, gtt_offset); + intel_uncore_write(uncore, GEN8_OASTATUS, 0); + intel_uncore_write(uncore, GEN8_OAHEADPTR, gtt_offset); stream->oa_buffer.head = gtt_offset; - I915_WRITE(GEN8_OABUFFER_UDW, 0); + intel_uncore_write(uncore, GEN8_OABUFFER_UDW, 0); /* * PRM says: @@ -1441,9 +1454,9 @@ static void gen8_init_oa_buffer(struct i915_perf_stream *stream) * to enable proper functionality of the overflow * bit." */ - I915_WRITE(GEN8_OABUFFER, gtt_offset | + intel_uncore_write(uncore, GEN8_OABUFFER, gtt_offset | OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT); - I915_WRITE(GEN8_OATAILPTR, gtt_offset & GEN8_OATAILPTR_MASK); + intel_uncore_write(uncore, GEN8_OATAILPTR, gtt_offset & GEN8_OATAILPTR_MASK); /* Mark that we need updated tail pointers to read from... */ stream->oa_buffer.tails[0].offset = INVALID_TAIL_PTR; @@ -1472,17 +1485,12 @@ static void gen8_init_oa_buffer(struct i915_perf_stream *stream) */ memset(stream->oa_buffer.vaddr, 0, OA_BUFFER_SIZE); - /* - * Maybe make ->pollin per-stream state if we support multiple - * concurrent streams in the future. - */ stream->pollin = false; } static int alloc_oa_buffer(struct i915_perf_stream *stream) { struct drm_i915_gem_object *bo; - struct drm_i915_private *dev_priv = stream->dev_priv; struct i915_vma *vma; int ret; @@ -1492,7 +1500,7 @@ static int alloc_oa_buffer(struct i915_perf_stream *stream) BUILD_BUG_ON_NOT_POWER_OF_2(OA_BUFFER_SIZE); BUILD_BUG_ON(OA_BUFFER_SIZE < SZ_128K || OA_BUFFER_SIZE > SZ_16M); - bo = i915_gem_object_create_shmem(dev_priv, OA_BUFFER_SIZE); + bo = i915_gem_object_create_shmem(stream->perf->i915, OA_BUFFER_SIZE); if (IS_ERR(bo)) { DRM_ERROR("Failed to allocate OA buffer\n"); return PTR_ERR(bo); @@ -1515,10 +1523,6 @@ static int alloc_oa_buffer(struct i915_perf_stream *stream) goto err_unpin; } - DRM_DEBUG_DRIVER("OA Buffer initialized, gtt offset = 0x%x, vaddr = %p\n", - i915_ggtt_offset(stream->oa_buffer.vma), - stream->oa_buffer.vaddr); - return 0; err_unpin: @@ -1533,50 +1537,381 @@ err_unref: return ret; } -static void config_oa_regs(struct drm_i915_private *dev_priv, - const struct i915_oa_reg *regs, - u32 n_regs) +static u32 *save_restore_register(struct i915_perf_stream *stream, u32 *cs, + bool save, i915_reg_t reg, u32 offset, + u32 dword_count) +{ + u32 cmd; + u32 d; + + cmd = save ? MI_STORE_REGISTER_MEM : MI_LOAD_REGISTER_MEM; + if (INTEL_GEN(stream->perf->i915) >= 8) + cmd++; + + for (d = 0; d < dword_count; d++) { + *cs++ = cmd; + *cs++ = i915_mmio_reg_offset(reg) + 4 * d; + *cs++ = intel_gt_scratch_offset(stream->engine->gt, + offset) + 4 * d; + *cs++ = 0; + } + + return cs; +} + +static int alloc_noa_wait(struct i915_perf_stream *stream) +{ + struct drm_i915_private *i915 = stream->perf->i915; + struct drm_i915_gem_object *bo; + struct i915_vma *vma; + const u64 delay_ticks = 0xffffffffffffffff - + DIV64_U64_ROUND_UP( + atomic64_read(&stream->perf->noa_programming_delay) * + RUNTIME_INFO(i915)->cs_timestamp_frequency_khz, + 1000000ull); + const u32 base = stream->engine->mmio_base; +#define CS_GPR(x) GEN8_RING_CS_GPR(base, x) + u32 *batch, *ts0, *cs, *jump; + int ret, i; + enum { + START_TS, + NOW_TS, + DELTA_TS, + JUMP_PREDICATE, + DELTA_TARGET, + N_CS_GPR + }; + + bo = i915_gem_object_create_internal(i915, 4096); + if (IS_ERR(bo)) { + DRM_ERROR("Failed to allocate NOA wait batchbuffer\n"); + return PTR_ERR(bo); + } + + /* + * We pin in GGTT because we jump into this buffer now because + * multiple OA config BOs will have a jump to this address and it + * needs to be fixed during the lifetime of the i915/perf stream. + */ + vma = i915_gem_object_ggtt_pin(bo, NULL, 0, 0, PIN_HIGH); + if (IS_ERR(vma)) { + ret = PTR_ERR(vma); + goto err_unref; + } + + batch = cs = i915_gem_object_pin_map(bo, I915_MAP_WB); + if (IS_ERR(batch)) { + ret = PTR_ERR(batch); + goto err_unpin; + } + + /* Save registers. */ + for (i = 0; i < N_CS_GPR; i++) + cs = save_restore_register( + stream, cs, true /* save */, CS_GPR(i), + INTEL_GT_SCRATCH_FIELD_PERF_CS_GPR + 8 * i, 2); + cs = save_restore_register( + stream, cs, true /* save */, MI_PREDICATE_RESULT_1, + INTEL_GT_SCRATCH_FIELD_PERF_PREDICATE_RESULT_1, 1); + + /* First timestamp snapshot location. */ + ts0 = cs; + + /* + * Initial snapshot of the timestamp register to implement the wait. + * We work with 32b values, so clear out the top 32b bits of the + * register because the ALU works 64bits. + */ + *cs++ = MI_LOAD_REGISTER_IMM(1); + *cs++ = i915_mmio_reg_offset(CS_GPR(START_TS)) + 4; + *cs++ = 0; + *cs++ = MI_LOAD_REGISTER_REG | (3 - 2); + *cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(base)); + *cs++ = i915_mmio_reg_offset(CS_GPR(START_TS)); + + /* + * This is the location we're going to jump back into until the + * required amount of time has passed. + */ + jump = cs; + + /* + * Take another snapshot of the timestamp register. Take care to clear + * up the top 32bits of CS_GPR(1) as we're using it for other + * operations below. + */ + *cs++ = MI_LOAD_REGISTER_IMM(1); + *cs++ = i915_mmio_reg_offset(CS_GPR(NOW_TS)) + 4; + *cs++ = 0; + *cs++ = MI_LOAD_REGISTER_REG | (3 - 2); + *cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(base)); + *cs++ = i915_mmio_reg_offset(CS_GPR(NOW_TS)); + + /* + * Do a diff between the 2 timestamps and store the result back into + * CS_GPR(1). + */ + *cs++ = MI_MATH(5); + *cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(NOW_TS)); + *cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(START_TS)); + *cs++ = MI_MATH_SUB; + *cs++ = MI_MATH_STORE(MI_MATH_REG(DELTA_TS), MI_MATH_REG_ACCU); + *cs++ = MI_MATH_STORE(MI_MATH_REG(JUMP_PREDICATE), MI_MATH_REG_CF); + + /* + * Transfer the carry flag (set to 1 if ts1 < ts0, meaning the + * timestamp have rolled over the 32bits) into the predicate register + * to be used for the predicated jump. + */ + *cs++ = MI_LOAD_REGISTER_REG | (3 - 2); + *cs++ = i915_mmio_reg_offset(CS_GPR(JUMP_PREDICATE)); + *cs++ = i915_mmio_reg_offset(MI_PREDICATE_RESULT_1); + + /* Restart from the beginning if we had timestamps roll over. */ + *cs++ = (INTEL_GEN(i915) < 8 ? + MI_BATCH_BUFFER_START : + MI_BATCH_BUFFER_START_GEN8) | + MI_BATCH_PREDICATE; + *cs++ = i915_ggtt_offset(vma) + (ts0 - batch) * 4; + *cs++ = 0; + + /* + * Now add the diff between to previous timestamps and add it to : + * (((1 * << 64) - 1) - delay_ns) + * + * When the Carry Flag contains 1 this means the elapsed time is + * longer than the expected delay, and we can exit the wait loop. + */ + *cs++ = MI_LOAD_REGISTER_IMM(2); + *cs++ = i915_mmio_reg_offset(CS_GPR(DELTA_TARGET)); + *cs++ = lower_32_bits(delay_ticks); + *cs++ = i915_mmio_reg_offset(CS_GPR(DELTA_TARGET)) + 4; + *cs++ = upper_32_bits(delay_ticks); + + *cs++ = MI_MATH(4); + *cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(DELTA_TS)); + *cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(DELTA_TARGET)); + *cs++ = MI_MATH_ADD; + *cs++ = MI_MATH_STOREINV(MI_MATH_REG(JUMP_PREDICATE), MI_MATH_REG_CF); + + /* + * Transfer the result into the predicate register to be used for the + * predicated jump. + */ + *cs++ = MI_LOAD_REGISTER_REG | (3 - 2); + *cs++ = i915_mmio_reg_offset(CS_GPR(JUMP_PREDICATE)); + *cs++ = i915_mmio_reg_offset(MI_PREDICATE_RESULT_1); + + /* Predicate the jump. */ + *cs++ = (INTEL_GEN(i915) < 8 ? + MI_BATCH_BUFFER_START : + MI_BATCH_BUFFER_START_GEN8) | + MI_BATCH_PREDICATE; + *cs++ = i915_ggtt_offset(vma) + (jump - batch) * 4; + *cs++ = 0; + + /* Restore registers. */ + for (i = 0; i < N_CS_GPR; i++) + cs = save_restore_register( + stream, cs, false /* restore */, CS_GPR(i), + INTEL_GT_SCRATCH_FIELD_PERF_CS_GPR + 8 * i, 2); + cs = save_restore_register( + stream, cs, false /* restore */, MI_PREDICATE_RESULT_1, + INTEL_GT_SCRATCH_FIELD_PERF_PREDICATE_RESULT_1, 1); + + /* And return to the ring. */ + *cs++ = MI_BATCH_BUFFER_END; + + GEM_BUG_ON(cs - batch > PAGE_SIZE / sizeof(*batch)); + + i915_gem_object_flush_map(bo); + i915_gem_object_unpin_map(bo); + + stream->noa_wait = vma; + return 0; + +err_unpin: + i915_vma_unpin_and_release(&vma, 0); +err_unref: + i915_gem_object_put(bo); + return ret; +} + +static u32 *write_cs_mi_lri(u32 *cs, + const struct i915_oa_reg *reg_data, + u32 n_regs) { u32 i; for (i = 0; i < n_regs; i++) { - const struct i915_oa_reg *reg = regs + i; + if ((i % MI_LOAD_REGISTER_IMM_MAX_REGS) == 0) { + u32 n_lri = min_t(u32, + n_regs - i, + MI_LOAD_REGISTER_IMM_MAX_REGS); - I915_WRITE(reg->addr, reg->value); + *cs++ = MI_LOAD_REGISTER_IMM(n_lri); + } + *cs++ = i915_mmio_reg_offset(reg_data[i].addr); + *cs++ = reg_data[i].value; } + + return cs; } -static void delay_after_mux(void) +static int num_lri_dwords(int num_regs) { + int count = 0; + + if (num_regs > 0) { + count += DIV_ROUND_UP(num_regs, MI_LOAD_REGISTER_IMM_MAX_REGS); + count += num_regs * 2; + } + + return count; +} + +static struct i915_oa_config_bo * +alloc_oa_config_buffer(struct i915_perf_stream *stream, + struct i915_oa_config *oa_config) +{ + struct drm_i915_gem_object *obj; + struct i915_oa_config_bo *oa_bo; + size_t config_length = 0; + u32 *cs; + int err; + + oa_bo = kzalloc(sizeof(*oa_bo), GFP_KERNEL); + if (!oa_bo) + return ERR_PTR(-ENOMEM); + + config_length += num_lri_dwords(oa_config->mux_regs_len); + config_length += num_lri_dwords(oa_config->b_counter_regs_len); + config_length += num_lri_dwords(oa_config->flex_regs_len); + config_length++; /* MI_BATCH_BUFFER_END */ + config_length = ALIGN(sizeof(u32) * config_length, I915_GTT_PAGE_SIZE); + + obj = i915_gem_object_create_shmem(stream->perf->i915, config_length); + if (IS_ERR(obj)) { + err = PTR_ERR(obj); + goto err_free; + } + + cs = i915_gem_object_pin_map(obj, I915_MAP_WB); + if (IS_ERR(cs)) { + err = PTR_ERR(cs); + goto err_oa_bo; + } + + cs = write_cs_mi_lri(cs, + oa_config->mux_regs, + oa_config->mux_regs_len); + cs = write_cs_mi_lri(cs, + oa_config->b_counter_regs, + oa_config->b_counter_regs_len); + cs = write_cs_mi_lri(cs, + oa_config->flex_regs, + oa_config->flex_regs_len); + + *cs++ = MI_BATCH_BUFFER_END; + + i915_gem_object_flush_map(obj); + i915_gem_object_unpin_map(obj); + + oa_bo->vma = i915_vma_instance(obj, + &stream->engine->gt->ggtt->vm, + NULL); + if (IS_ERR(oa_bo->vma)) { + err = PTR_ERR(oa_bo->vma); + goto err_oa_bo; + } + + oa_bo->oa_config = i915_oa_config_get(oa_config); + llist_add(&oa_bo->node, &stream->oa_config_bos); + + return oa_bo; + +err_oa_bo: + i915_gem_object_put(obj); +err_free: + kfree(oa_bo); + return ERR_PTR(err); +} + +static struct i915_vma * +get_oa_vma(struct i915_perf_stream *stream, struct i915_oa_config *oa_config) +{ + struct i915_oa_config_bo *oa_bo; + /* - * It apparently takes a fairly long time for a new MUX - * configuration to be be applied after these register writes. - * This delay duration was derived empirically based on the - * render_basic config but hopefully it covers the maximum - * configuration latency. - * - * As a fallback, the checks in _append_oa_reports() to skip - * invalid OA reports do also seem to work to discard reports - * generated before this config has completed - albeit not - * silently. - * - * Unfortunately this is essentially a magic number, since we - * don't currently know of a reliable mechanism for predicting - * how long the MUX config will take to apply and besides - * seeing invalid reports we don't know of a reliable way to - * explicitly check that the MUX config has landed. - * - * It's even possible we've miss characterized the underlying - * problem - it just seems like the simplest explanation why - * a delay at this location would mitigate any invalid reports. + * Look for the buffer in the already allocated BOs attached + * to the stream. */ - usleep_range(15000, 20000); + llist_for_each_entry(oa_bo, stream->oa_config_bos.first, node) { + if (oa_bo->oa_config == oa_config && + memcmp(oa_bo->oa_config->uuid, + oa_config->uuid, + sizeof(oa_config->uuid)) == 0) + goto out; + } + + oa_bo = alloc_oa_config_buffer(stream, oa_config); + if (IS_ERR(oa_bo)) + return ERR_CAST(oa_bo); + +out: + return i915_vma_get(oa_bo->vma); +} + +static int emit_oa_config(struct i915_perf_stream *stream, + struct i915_oa_config *oa_config, + struct intel_context *ce) +{ + struct i915_request *rq; + struct i915_vma *vma; + int err; + + vma = get_oa_vma(stream, oa_config); + if (IS_ERR(vma)) + return PTR_ERR(vma); + + err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL | PIN_HIGH); + if (err) + goto err_vma_put; + + rq = i915_request_create(ce); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto err_vma_unpin; + } + + i915_vma_lock(vma); + err = i915_request_await_object(rq, vma->obj, 0); + if (!err) + err = i915_vma_move_to_active(vma, rq, 0); + i915_vma_unlock(vma); + if (err) + goto err_add_request; + + err = rq->engine->emit_bb_start(rq, + vma->node.start, 0, + I915_DISPATCH_SECURE); +err_add_request: + i915_request_add(rq); +err_vma_unpin: + i915_vma_unpin(vma); +err_vma_put: + i915_vma_put(vma); + return err; +} + +static struct intel_context *oa_context(struct i915_perf_stream *stream) +{ + return stream->pinned_ctx ?: stream->engine->kernel_context; } static int hsw_enable_metric_set(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; - const struct i915_oa_config *oa_config = stream->oa_config; + struct intel_uncore *uncore = stream->uncore; /* * PRM: @@ -1588,31 +1923,24 @@ static int hsw_enable_metric_set(struct i915_perf_stream *stream) * count the events from non-render domain. Unit level clock * gating for RCS should also be disabled. */ - I915_WRITE(GEN7_MISCCPCTL, (I915_READ(GEN7_MISCCPCTL) & - ~GEN7_DOP_CLOCK_GATE_ENABLE)); - I915_WRITE(GEN6_UCGCTL1, (I915_READ(GEN6_UCGCTL1) | - GEN6_CSUNIT_CLOCK_GATE_DISABLE)); + intel_uncore_rmw(uncore, GEN7_MISCCPCTL, + GEN7_DOP_CLOCK_GATE_ENABLE, 0); + intel_uncore_rmw(uncore, GEN6_UCGCTL1, + 0, GEN6_CSUNIT_CLOCK_GATE_DISABLE); - config_oa_regs(dev_priv, oa_config->mux_regs, oa_config->mux_regs_len); - delay_after_mux(); - - config_oa_regs(dev_priv, oa_config->b_counter_regs, - oa_config->b_counter_regs_len); - - return 0; + return emit_oa_config(stream, stream->oa_config, oa_context(stream)); } static void hsw_disable_metric_set(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; - I915_WRITE(GEN6_UCGCTL1, (I915_READ(GEN6_UCGCTL1) & - ~GEN6_CSUNIT_CLOCK_GATE_DISABLE)); - I915_WRITE(GEN7_MISCCPCTL, (I915_READ(GEN7_MISCCPCTL) | - GEN7_DOP_CLOCK_GATE_ENABLE)); + intel_uncore_rmw(uncore, GEN6_UCGCTL1, + GEN6_CSUNIT_CLOCK_GATE_DISABLE, 0); + intel_uncore_rmw(uncore, GEN7_MISCCPCTL, + 0, GEN7_DOP_CLOCK_GATE_ENABLE); - I915_WRITE(GDT_CHICKEN_BITS, (I915_READ(GDT_CHICKEN_BITS) & - ~GT_NOA_ENABLE)); + intel_uncore_rmw(uncore, GDT_CHICKEN_BITS, GT_NOA_ENABLE, 0); } static u32 oa_config_flex_reg(const struct i915_oa_config *oa_config, @@ -1647,9 +1975,8 @@ static void gen8_update_reg_state_unlocked(const struct intel_context *ce, const struct i915_perf_stream *stream) { - struct drm_i915_private *i915 = ce->engine->i915; - u32 ctx_oactxctrl = i915->perf.ctx_oactxctrl_offset; - u32 ctx_flexeu0 = i915->perf.ctx_flexeu0_offset; + u32 ctx_oactxctrl = stream->perf->ctx_oactxctrl_offset; + u32 ctx_flexeu0 = stream->perf->ctx_flexeu0_offset; /* The MMIO offsets for Flex EU registers aren't contiguous */ i915_reg_t flex_regs[] = { EU_PERF_CNTL0, @@ -1672,7 +1999,8 @@ gen8_update_reg_state_unlocked(const struct intel_context *ce, reg_state[ctx_flexeu0 + i * 2 + 1] = oa_config_flex_reg(stream->oa_config, flex_regs[i]); - reg_state[CTX_R_PWR_CLK_STATE] = intel_sseu_make_rpcs(i915, &ce->sseu); + reg_state[CTX_R_PWR_CLK_STATE] = + intel_sseu_make_rpcs(ce->engine->i915, &ce->sseu); } struct flex { @@ -1827,9 +2155,9 @@ static int gen8_configure_context(struct i915_gem_context *ctx, static int gen8_configure_all_contexts(struct i915_perf_stream *stream, const struct i915_oa_config *oa_config) { - struct drm_i915_private *i915 = stream->dev_priv; + struct drm_i915_private *i915 = stream->perf->i915; /* The MMIO offsets for Flex EU registers aren't contiguous */ - const u32 ctx_flexeu0 = i915->perf.ctx_flexeu0_offset; + const u32 ctx_flexeu0 = stream->perf->ctx_flexeu0_offset; #define ctx_flexeuN(N) (ctx_flexeu0 + 2 * (N) + 1) struct flex regs[] = { { @@ -1838,7 +2166,7 @@ static int gen8_configure_all_contexts(struct i915_perf_stream *stream, }, { GEN8_OACTXCONTROL, - i915->perf.ctx_oactxctrl_offset + 1, + stream->perf->ctx_oactxctrl_offset + 1, ((stream->period_exponent << GEN8_OA_TIMER_PERIOD_SHIFT) | (stream->periodic ? GEN8_OA_TIMER_ENABLE : 0) | GEN8_OA_COUNTER_RESUME) @@ -1859,7 +2187,7 @@ static int gen8_configure_all_contexts(struct i915_perf_stream *stream, for (i = 2; i < ARRAY_SIZE(regs); i++) regs[i].value = oa_config_flex_reg(oa_config, regs[i].reg); - lockdep_assert_held(&i915->drm.struct_mutex); + lockdep_assert_held(&stream->perf->lock); /* * The OA register config is setup through the context image. This image @@ -1922,8 +2250,8 @@ static int gen8_configure_all_contexts(struct i915_perf_stream *stream, static int gen8_enable_metric_set(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; - const struct i915_oa_config *oa_config = stream->oa_config; + struct intel_uncore *uncore = stream->uncore; + struct i915_oa_config *oa_config = stream->oa_config; int ret; /* @@ -1949,10 +2277,10 @@ static int gen8_enable_metric_set(struct i915_perf_stream *stream) * be read back from automatically triggered reports, as part of the * RPT_ID field. */ - if (IS_GEN_RANGE(dev_priv, 9, 11)) { - I915_WRITE(GEN8_OA_DEBUG, - _MASKED_BIT_ENABLE(GEN9_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS | - GEN9_OA_DEBUG_INCLUDE_CLK_RATIO)); + if (IS_GEN_RANGE(stream->perf->i915, 9, 11)) { + intel_uncore_write(uncore, GEN8_OA_DEBUG, + _MASKED_BIT_ENABLE(GEN9_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS | + GEN9_OA_DEBUG_INCLUDE_CLK_RATIO)); } /* @@ -1964,41 +2292,33 @@ static int gen8_enable_metric_set(struct i915_perf_stream *stream) if (ret) return ret; - config_oa_regs(dev_priv, oa_config->mux_regs, oa_config->mux_regs_len); - delay_after_mux(); - - config_oa_regs(dev_priv, oa_config->b_counter_regs, - oa_config->b_counter_regs_len); - - return 0; + return emit_oa_config(stream, oa_config, oa_context(stream)); } static void gen8_disable_metric_set(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; /* Reset all contexts' slices/subslices configurations. */ gen8_configure_all_contexts(stream, NULL); - I915_WRITE(GDT_CHICKEN_BITS, (I915_READ(GDT_CHICKEN_BITS) & - ~GT_NOA_ENABLE)); + intel_uncore_rmw(uncore, GDT_CHICKEN_BITS, GT_NOA_ENABLE, 0); } static void gen10_disable_metric_set(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; /* Reset all contexts' slices/subslices configurations. */ gen8_configure_all_contexts(stream, NULL); /* Make sure we disable noa to save power. */ - I915_WRITE(RPM_CONFIG1, - I915_READ(RPM_CONFIG1) & ~GEN10_GT_NOA_ENABLE); + intel_uncore_rmw(uncore, RPM_CONFIG1, GEN10_GT_NOA_ENABLE, 0); } static void gen7_oa_enable(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; struct i915_gem_context *ctx = stream->ctx; u32 ctx_id = stream->specific_ctx_id; bool periodic = stream->periodic; @@ -2016,19 +2336,19 @@ static void gen7_oa_enable(struct i915_perf_stream *stream) */ gen7_init_oa_buffer(stream); - I915_WRITE(GEN7_OACONTROL, - (ctx_id & GEN7_OACONTROL_CTX_MASK) | - (period_exponent << - GEN7_OACONTROL_TIMER_PERIOD_SHIFT) | - (periodic ? GEN7_OACONTROL_TIMER_ENABLE : 0) | - (report_format << GEN7_OACONTROL_FORMAT_SHIFT) | - (ctx ? GEN7_OACONTROL_PER_CTX_ENABLE : 0) | - GEN7_OACONTROL_ENABLE); + intel_uncore_write(uncore, GEN7_OACONTROL, + (ctx_id & GEN7_OACONTROL_CTX_MASK) | + (period_exponent << + GEN7_OACONTROL_TIMER_PERIOD_SHIFT) | + (periodic ? GEN7_OACONTROL_TIMER_ENABLE : 0) | + (report_format << GEN7_OACONTROL_FORMAT_SHIFT) | + (ctx ? GEN7_OACONTROL_PER_CTX_ENABLE : 0) | + GEN7_OACONTROL_ENABLE); } static void gen8_oa_enable(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct intel_uncore *uncore = stream->uncore; u32 report_format = stream->oa_buffer.format; /* @@ -2047,9 +2367,9 @@ static void gen8_oa_enable(struct i915_perf_stream *stream) * filtering and instead filter on the cpu based on the context-id * field of reports */ - I915_WRITE(GEN8_OACONTROL, (report_format << - GEN8_OA_REPORT_FORMAT_SHIFT) | - GEN8_OA_COUNTER_ENABLE); + intel_uncore_write(uncore, GEN8_OACONTROL, + (report_format << GEN8_OA_REPORT_FORMAT_SHIFT) | + GEN8_OA_COUNTER_ENABLE); } /** @@ -2063,9 +2383,7 @@ static void gen8_oa_enable(struct i915_perf_stream *stream) */ static void i915_oa_stream_enable(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; - - dev_priv->perf.ops.oa_enable(stream); + stream->perf->ops.oa_enable(stream); if (stream->periodic) hrtimer_start(&stream->poll_check_timer, @@ -2075,7 +2393,7 @@ static void i915_oa_stream_enable(struct i915_perf_stream *stream) static void gen7_oa_disable(struct i915_perf_stream *stream) { - struct intel_uncore *uncore = &stream->dev_priv->uncore; + struct intel_uncore *uncore = stream->uncore; intel_uncore_write(uncore, GEN7_OACONTROL, 0); if (intel_wait_for_register(uncore, @@ -2086,7 +2404,7 @@ static void gen7_oa_disable(struct i915_perf_stream *stream) static void gen8_oa_disable(struct i915_perf_stream *stream) { - struct intel_uncore *uncore = &stream->dev_priv->uncore; + struct intel_uncore *uncore = stream->uncore; intel_uncore_write(uncore, GEN8_OACONTROL, 0); if (intel_wait_for_register(uncore, @@ -2105,9 +2423,7 @@ static void gen8_oa_disable(struct i915_perf_stream *stream) */ static void i915_oa_stream_disable(struct i915_perf_stream *stream) { - struct drm_i915_private *dev_priv = stream->dev_priv; - - dev_priv->perf.ops.oa_disable(stream); + stream->perf->ops.oa_disable(stream); if (stream->periodic) hrtimer_cancel(&stream->poll_check_timer); @@ -2144,15 +2460,21 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, struct drm_i915_perf_open_param *param, struct perf_open_properties *props) { - struct drm_i915_private *dev_priv = stream->dev_priv; + struct i915_perf *perf = stream->perf; int format_size; int ret; - /* If the sysfs metrics/ directory wasn't registered for some + if (!props->engine) { + DRM_DEBUG("OA engine not specified\n"); + return -EINVAL; + } + + /* + * If the sysfs metrics/ directory wasn't registered for some * reason then don't let userspace try their luck with config * IDs */ - if (!dev_priv->perf.metrics_kobj) { + if (!perf->metrics_kobj) { DRM_DEBUG("OA metrics weren't advertised via sysfs\n"); return -EINVAL; } @@ -2162,16 +2484,17 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, return -EINVAL; } - if (!dev_priv->perf.ops.enable_metric_set) { + if (!perf->ops.enable_metric_set) { DRM_DEBUG("OA unit not supported\n"); return -ENODEV; } - /* To avoid the complexity of having to accurately filter + /* + * To avoid the complexity of having to accurately filter * counter reports and marshal to the appropriate client * we currently only allow exclusive access */ - if (dev_priv->perf.exclusive_stream) { + if (perf->exclusive_stream) { DRM_DEBUG("OA unit already in use\n"); return -EBUSY; } @@ -2181,9 +2504,12 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, return -EINVAL; } + stream->engine = props->engine; + stream->uncore = stream->engine->gt->uncore; + stream->sample_size = sizeof(struct drm_i915_perf_record_header); - format_size = dev_priv->perf.oa_formats[props->oa_format].size; + format_size = perf->oa_formats[props->oa_format].size; stream->sample_flags |= SAMPLE_OA_REPORT; stream->sample_size += format_size; @@ -2192,8 +2518,10 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, if (WARN_ON(stream->oa_buffer.format_size == 0)) return -EINVAL; + stream->hold_preemption = props->hold_preemption; + stream->oa_buffer.format = - dev_priv->perf.oa_formats[props->oa_format].format; + perf->oa_formats[props->oa_format].format; stream->periodic = props->oa_periodic; if (stream->periodic) @@ -2207,9 +2535,16 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, } } - ret = get_oa_config(dev_priv, props->metrics_set, &stream->oa_config); + ret = alloc_noa_wait(stream); if (ret) { + DRM_DEBUG("Unable to allocate NOA wait batch buffer\n"); + goto err_noa_wait_alloc; + } + + stream->oa_config = i915_perf_get_oa_config(perf, props->metrics_set); + if (!stream->oa_config) { DRM_DEBUG("Invalid OA config id=%i\n", props->metrics_set); + ret = -EINVAL; goto err_config; } @@ -2225,27 +2560,24 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, * In our case we are expecting that taking pm + FORCEWAKE * references will effectively disable RC6. */ - stream->wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm); - intel_uncore_forcewake_get(&dev_priv->uncore, FORCEWAKE_ALL); + intel_engine_pm_get(stream->engine); + intel_uncore_forcewake_get(stream->uncore, FORCEWAKE_ALL); ret = alloc_oa_buffer(stream); if (ret) goto err_oa_buf_alloc; - ret = i915_mutex_lock_interruptible(&dev_priv->drm); - if (ret) - goto err_lock; - stream->ops = &i915_oa_stream_ops; - dev_priv->perf.exclusive_stream = stream; + perf->exclusive_stream = stream; - ret = dev_priv->perf.ops.enable_metric_set(stream); + ret = perf->ops.enable_metric_set(stream); if (ret) { DRM_DEBUG("Unable to enable metric set\n"); goto err_enable; } - mutex_unlock(&dev_priv->drm.struct_mutex); + DRM_DEBUG("opening stream oa config uuid=%s\n", + stream->oa_config->uuid); hrtimer_init(&stream->poll_check_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); @@ -2256,20 +2588,21 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, return 0; err_enable: - dev_priv->perf.exclusive_stream = NULL; - dev_priv->perf.ops.disable_metric_set(stream); - mutex_unlock(&dev_priv->drm.struct_mutex); + perf->exclusive_stream = NULL; + perf->ops.disable_metric_set(stream); -err_lock: free_oa_buffer(stream); err_oa_buf_alloc: - put_oa_config(dev_priv, stream->oa_config); + free_oa_configs(stream); - intel_uncore_forcewake_put(&dev_priv->uncore, FORCEWAKE_ALL); - intel_runtime_pm_put(&dev_priv->runtime_pm, stream->wakeref); + intel_uncore_forcewake_put(stream->uncore, FORCEWAKE_ALL); + intel_engine_pm_put(stream->engine); err_config: + free_noa_wait(stream); + +err_noa_wait_alloc: if (stream->ctx) oa_put_render_ctx_id(stream); @@ -2359,7 +2692,7 @@ static ssize_t i915_perf_read(struct file *file, loff_t *ppos) { struct i915_perf_stream *stream = file->private_data; - struct drm_i915_private *dev_priv = stream->dev_priv; + struct i915_perf *perf = stream->perf; ssize_t ret; /* To ensure it's handled consistently we simply treat all reads of a @@ -2382,15 +2715,15 @@ static ssize_t i915_perf_read(struct file *file, if (ret) return ret; - mutex_lock(&dev_priv->perf.lock); + mutex_lock(&perf->lock); ret = i915_perf_read_locked(stream, file, buf, count, ppos); - mutex_unlock(&dev_priv->perf.lock); + mutex_unlock(&perf->lock); } while (ret == -EAGAIN); } else { - mutex_lock(&dev_priv->perf.lock); + mutex_lock(&perf->lock); ret = i915_perf_read_locked(stream, file, buf, count, ppos); - mutex_unlock(&dev_priv->perf.lock); + mutex_unlock(&perf->lock); } /* We allow the poll checking to sometimes report false positive EPOLLIN @@ -2428,7 +2761,6 @@ static enum hrtimer_restart oa_poll_check_timer_cb(struct hrtimer *hrtimer) /** * i915_perf_poll_locked - poll_wait() with a suitable wait queue for stream - * @dev_priv: i915 device instance * @stream: An i915 perf stream * @file: An i915 perf stream file * @wait: poll() state table @@ -2437,15 +2769,14 @@ static enum hrtimer_restart oa_poll_check_timer_cb(struct hrtimer *hrtimer) * &i915_perf_stream_ops->poll_wait to call poll_wait() with a wait queue that * will be woken for new stream data. * - * Note: The &drm_i915_private->perf.lock mutex has been taken to serialize + * Note: The &perf->lock mutex has been taken to serialize * with any non-file-operation driver hooks. * * Returns: any poll events that are ready without sleeping */ -static __poll_t i915_perf_poll_locked(struct drm_i915_private *dev_priv, - struct i915_perf_stream *stream, - struct file *file, - poll_table *wait) +static __poll_t i915_perf_poll_locked(struct i915_perf_stream *stream, + struct file *file, + poll_table *wait) { __poll_t events = 0; @@ -2479,12 +2810,12 @@ static __poll_t i915_perf_poll_locked(struct drm_i915_private *dev_priv, static __poll_t i915_perf_poll(struct file *file, poll_table *wait) { struct i915_perf_stream *stream = file->private_data; - struct drm_i915_private *dev_priv = stream->dev_priv; + struct i915_perf *perf = stream->perf; __poll_t ret; - mutex_lock(&dev_priv->perf.lock); - ret = i915_perf_poll_locked(dev_priv, stream, file, wait); - mutex_unlock(&dev_priv->perf.lock); + mutex_lock(&perf->lock); + ret = i915_perf_poll_locked(stream, file, wait); + mutex_unlock(&perf->lock); return ret; } @@ -2509,6 +2840,9 @@ static void i915_perf_enable_locked(struct i915_perf_stream *stream) if (stream->ops->enable) stream->ops->enable(stream); + + if (stream->hold_preemption) + i915_gem_context_set_nopreempt(stream->ctx); } /** @@ -2533,17 +2867,54 @@ static void i915_perf_disable_locked(struct i915_perf_stream *stream) /* Allow stream->ops->disable() to refer to this */ stream->enabled = false; + if (stream->hold_preemption) + i915_gem_context_clear_nopreempt(stream->ctx); + if (stream->ops->disable) stream->ops->disable(stream); } +static long i915_perf_config_locked(struct i915_perf_stream *stream, + unsigned long metrics_set) +{ + struct i915_oa_config *config; + long ret = stream->oa_config->id; + + config = i915_perf_get_oa_config(stream->perf, metrics_set); + if (!config) + return -EINVAL; + + if (config != stream->oa_config) { + int err; + + /* + * If OA is bound to a specific context, emit the + * reconfiguration inline from that context. The update + * will then be ordered with respect to submission on that + * context. + * + * When set globally, we use a low priority kernel context, + * so it will effectively take effect when idle. + */ + err = emit_oa_config(stream, config, oa_context(stream)); + if (err == 0) + config = xchg(&stream->oa_config, config); + else + ret = err; + } + + i915_oa_config_put(config); + + return ret; +} + /** * i915_perf_ioctl - support ioctl() usage with i915 perf stream FDs * @stream: An i915 perf stream * @cmd: the ioctl request * @arg: the ioctl data * - * Note: The &drm_i915_private->perf.lock mutex has been taken to serialize + * Note: The &perf->lock mutex has been taken to serialize * with any non-file-operation driver hooks. * * Returns: zero on success or a negative error code. Returns -EINVAL for @@ -2560,6 +2931,8 @@ static long i915_perf_ioctl_locked(struct i915_perf_stream *stream, case I915_PERF_IOCTL_DISABLE: i915_perf_disable_locked(stream); return 0; + case I915_PERF_IOCTL_CONFIG: + return i915_perf_config_locked(stream, arg); } return -EINVAL; @@ -2581,12 +2954,12 @@ static long i915_perf_ioctl(struct file *file, unsigned long arg) { struct i915_perf_stream *stream = file->private_data; - struct drm_i915_private *dev_priv = stream->dev_priv; + struct i915_perf *perf = stream->perf; long ret; - mutex_lock(&dev_priv->perf.lock); + mutex_lock(&perf->lock); ret = i915_perf_ioctl_locked(stream, cmd, arg); - mutex_unlock(&dev_priv->perf.lock); + mutex_unlock(&perf->lock); return ret; } @@ -2598,7 +2971,7 @@ static long i915_perf_ioctl(struct file *file, * Frees all resources associated with the given i915 perf @stream, disabling * any associated data capture in the process. * - * Note: The &drm_i915_private->perf.lock mutex has been taken to serialize + * Note: The &perf->lock mutex has been taken to serialize * with any non-file-operation driver hooks. */ static void i915_perf_destroy_locked(struct i915_perf_stream *stream) @@ -2609,8 +2982,6 @@ static void i915_perf_destroy_locked(struct i915_perf_stream *stream) if (stream->ops->destroy) stream->ops->destroy(stream); - list_del(&stream->link); - if (stream->ctx) i915_gem_context_put(stream->ctx); @@ -2631,14 +3002,14 @@ static void i915_perf_destroy_locked(struct i915_perf_stream *stream) static int i915_perf_release(struct inode *inode, struct file *file) { struct i915_perf_stream *stream = file->private_data; - struct drm_i915_private *dev_priv = stream->dev_priv; + struct i915_perf *perf = stream->perf; - mutex_lock(&dev_priv->perf.lock); + mutex_lock(&perf->lock); i915_perf_destroy_locked(stream); - mutex_unlock(&dev_priv->perf.lock); + mutex_unlock(&perf->lock); /* Release the reference the perf stream kept on the driver. */ - drm_dev_put(&dev_priv->drm); + drm_dev_put(&perf->i915->drm); return 0; } @@ -2660,7 +3031,7 @@ static const struct file_operations fops = { /** * i915_perf_open_ioctl_locked - DRM ioctl() for userspace to open a stream FD - * @dev_priv: i915 device instance + * @perf: i915 perf instance * @param: The open parameters passed to 'DRM_I915_PERF_OPEN` * @props: individually validated u64 property value pairs * @file: drm file @@ -2668,7 +3039,7 @@ static const struct file_operations fops = { * See i915_perf_ioctl_open() for interface details. * * Implements further stream config validation and stream initialization on - * behalf of i915_perf_open_ioctl() with the &drm_i915_private->perf.lock mutex + * behalf of i915_perf_open_ioctl() with the &perf->lock mutex * taken to serialize with any non-file-operation driver hooks. * * Note: at this point the @props have only been validated in isolation and @@ -2683,7 +3054,7 @@ static const struct file_operations fops = { * Returns: zero on success or a negative error code. */ static int -i915_perf_open_ioctl_locked(struct drm_i915_private *dev_priv, +i915_perf_open_ioctl_locked(struct i915_perf *perf, struct drm_i915_perf_open_param *param, struct perf_open_properties *props, struct drm_file *file) @@ -2708,6 +3079,15 @@ i915_perf_open_ioctl_locked(struct drm_i915_private *dev_priv, } } + if (props->hold_preemption) { + if (!props->single_context) { + DRM_DEBUG("preemption disable with no context\n"); + ret = -EINVAL; + goto err; + } + privileged_op = true; + } + /* * On Haswell the OA unit supports clock gating off for a specific * context and in this mode there's no visibility of metrics for the @@ -2722,7 +3102,7 @@ i915_perf_open_ioctl_locked(struct drm_i915_private *dev_priv, * MI_REPORT_PERF_COUNT commands and so consider it a privileged op to * enable the OA unit by default. */ - if (IS_HASWELL(dev_priv) && specific_ctx) + if (IS_HASWELL(perf->i915) && specific_ctx && !props->hold_preemption) privileged_op = false; /* Similar to perf's kernel.perf_paranoid_cpu sysctl option @@ -2732,7 +3112,7 @@ i915_perf_open_ioctl_locked(struct drm_i915_private *dev_priv, */ if (privileged_op && i915_perf_stream_paranoid && !capable(CAP_SYS_ADMIN)) { - DRM_DEBUG("Insufficient privileges to open system-wide i915 perf stream\n"); + DRM_DEBUG("Insufficient privileges to open i915 perf stream\n"); ret = -EACCES; goto err_ctx; } @@ -2743,7 +3123,7 @@ i915_perf_open_ioctl_locked(struct drm_i915_private *dev_priv, goto err_ctx; } - stream->dev_priv = dev_priv; + stream->perf = perf; stream->ctx = specific_ctx; ret = i915_oa_stream_init(stream, param, props); @@ -2759,8 +3139,6 @@ i915_perf_open_ioctl_locked(struct drm_i915_private *dev_priv, goto err_flags; } - list_add(&stream->link, &dev_priv->perf.streams); - if (param->flags & I915_PERF_FLAG_FD_CLOEXEC) f_flags |= O_CLOEXEC; if (param->flags & I915_PERF_FLAG_FD_NONBLOCK) @@ -2769,7 +3147,7 @@ i915_perf_open_ioctl_locked(struct drm_i915_private *dev_priv, stream_fd = anon_inode_getfd("[i915_perf]", &fops, stream, f_flags); if (stream_fd < 0) { ret = stream_fd; - goto err_open; + goto err_flags; } if (!(param->flags & I915_PERF_FLAG_DISABLED)) @@ -2778,12 +3156,10 @@ i915_perf_open_ioctl_locked(struct drm_i915_private *dev_priv, /* Take a reference on the driver that will be kept with stream_fd * until its release. */ - drm_dev_get(&dev_priv->drm); + drm_dev_get(&perf->i915->drm); return stream_fd; -err_open: - list_del(&stream->link); err_flags: if (stream->ops->destroy) stream->ops->destroy(stream); @@ -2796,15 +3172,15 @@ err: return ret; } -static u64 oa_exponent_to_ns(struct drm_i915_private *dev_priv, int exponent) +static u64 oa_exponent_to_ns(struct i915_perf *perf, int exponent) { return div64_u64(1000000000ULL * (2ULL << exponent), - 1000ULL * RUNTIME_INFO(dev_priv)->cs_timestamp_frequency_khz); + 1000ULL * RUNTIME_INFO(perf->i915)->cs_timestamp_frequency_khz); } /** * read_properties_unlocked - validate + copy userspace stream open properties - * @dev_priv: i915 device instance + * @perf: i915 perf instance * @uprops: The array of u64 key value pairs given by userspace * @n_props: The number of key value pairs expected in @uprops * @props: The stream configuration built up while validating properties @@ -2817,7 +3193,7 @@ static u64 oa_exponent_to_ns(struct drm_i915_private *dev_priv, int exponent) * we shouldn't validate or assume anything about ordering here. This doesn't * rule out defining new properties with ordering requirements in the future. */ -static int read_properties_unlocked(struct drm_i915_private *dev_priv, +static int read_properties_unlocked(struct i915_perf *perf, u64 __user *uprops, u32 n_props, struct perf_open_properties *props) @@ -2832,6 +3208,15 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv, return -EINVAL; } + /* At the moment we only support using i915-perf on the RCS. */ + props->engine = intel_engine_lookup_user(perf->i915, + I915_ENGINE_CLASS_RENDER, + 0); + if (!props->engine) { + DRM_DEBUG("No RENDER-capable engines\n"); + return -EINVAL; + } + /* Considering that ID = 0 is reserved and assuming that we don't * (currently) expect any configurations to ever specify duplicate * values for a particular property ID then the last _PROP_MAX value is @@ -2883,7 +3268,7 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv, value); return -EINVAL; } - if (!dev_priv->perf.oa_formats[value].size) { + if (!perf->oa_formats[value].size) { DRM_DEBUG("Unsupported OA report format %llu\n", value); return -EINVAL; @@ -2904,7 +3289,7 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv, */ BUILD_BUG_ON(sizeof(oa_period) != 8); - oa_period = oa_exponent_to_ns(dev_priv, value); + oa_period = oa_exponent_to_ns(perf, value); /* This check is primarily to ensure that oa_period <= * UINT32_MAX (before passing to do_div which only @@ -2929,6 +3314,9 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv, props->oa_periodic = true; props->oa_period_exponent = value; break; + case DRM_I915_PERF_PROP_HOLD_PREEMPTION: + props->hold_preemption = !!value; + break; case DRM_I915_PERF_PROP_MAX: MISSING_CASE(id); return -EINVAL; @@ -2958,7 +3346,7 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv, * mutex to avoid an awkward lockdep with mmap_sem. * * Most of the implementation details are handled by - * i915_perf_open_ioctl_locked() after taking the &drm_i915_private->perf.lock + * i915_perf_open_ioctl_locked() after taking the &perf->lock * mutex for serializing with any non-file-operation driver hooks. * * Return: A newly opened i915 Perf stream file descriptor or negative @@ -2967,13 +3355,13 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv, int i915_perf_open_ioctl(struct drm_device *dev, void *data, struct drm_file *file) { - struct drm_i915_private *dev_priv = dev->dev_private; + struct i915_perf *perf = &to_i915(dev)->perf; struct drm_i915_perf_open_param *param = data; struct perf_open_properties props; u32 known_open_flags; int ret; - if (!dev_priv->perf.initialized) { + if (!perf->i915) { DRM_DEBUG("i915 perf interface not available for this system\n"); return -ENOTSUPP; } @@ -2986,124 +3374,128 @@ int i915_perf_open_ioctl(struct drm_device *dev, void *data, return -EINVAL; } - ret = read_properties_unlocked(dev_priv, + ret = read_properties_unlocked(perf, u64_to_user_ptr(param->properties_ptr), param->num_properties, &props); if (ret) return ret; - mutex_lock(&dev_priv->perf.lock); - ret = i915_perf_open_ioctl_locked(dev_priv, param, &props, file); - mutex_unlock(&dev_priv->perf.lock); + mutex_lock(&perf->lock); + ret = i915_perf_open_ioctl_locked(perf, param, &props, file); + mutex_unlock(&perf->lock); return ret; } /** * i915_perf_register - exposes i915-perf to userspace - * @dev_priv: i915 device instance + * @i915: i915 device instance * * In particular OA metric sets are advertised under a sysfs metrics/ * directory allowing userspace to enumerate valid IDs that can be * used to open an i915-perf stream. */ -void i915_perf_register(struct drm_i915_private *dev_priv) +void i915_perf_register(struct drm_i915_private *i915) { + struct i915_perf *perf = &i915->perf; int ret; - if (!dev_priv->perf.initialized) + if (!perf->i915) return; /* To be sure we're synchronized with an attempted * i915_perf_open_ioctl(); considering that we register after * being exposed to userspace. */ - mutex_lock(&dev_priv->perf.lock); + mutex_lock(&perf->lock); - dev_priv->perf.metrics_kobj = + perf->metrics_kobj = kobject_create_and_add("metrics", - &dev_priv->drm.primary->kdev->kobj); - if (!dev_priv->perf.metrics_kobj) + &i915->drm.primary->kdev->kobj); + if (!perf->metrics_kobj) goto exit; - sysfs_attr_init(&dev_priv->perf.test_config.sysfs_metric_id.attr); - - if (INTEL_GEN(dev_priv) >= 11) { - i915_perf_load_test_config_icl(dev_priv); - } else if (IS_CANNONLAKE(dev_priv)) { - i915_perf_load_test_config_cnl(dev_priv); - } else if (IS_COFFEELAKE(dev_priv)) { - if (IS_CFL_GT2(dev_priv)) - i915_perf_load_test_config_cflgt2(dev_priv); - if (IS_CFL_GT3(dev_priv)) - i915_perf_load_test_config_cflgt3(dev_priv); - } else if (IS_GEMINILAKE(dev_priv)) { - i915_perf_load_test_config_glk(dev_priv); - } else if (IS_KABYLAKE(dev_priv)) { - if (IS_KBL_GT2(dev_priv)) - i915_perf_load_test_config_kblgt2(dev_priv); - else if (IS_KBL_GT3(dev_priv)) - i915_perf_load_test_config_kblgt3(dev_priv); - } else if (IS_BROXTON(dev_priv)) { - i915_perf_load_test_config_bxt(dev_priv); - } else if (IS_SKYLAKE(dev_priv)) { - if (IS_SKL_GT2(dev_priv)) - i915_perf_load_test_config_sklgt2(dev_priv); - else if (IS_SKL_GT3(dev_priv)) - i915_perf_load_test_config_sklgt3(dev_priv); - else if (IS_SKL_GT4(dev_priv)) - i915_perf_load_test_config_sklgt4(dev_priv); - } else if (IS_CHERRYVIEW(dev_priv)) { - i915_perf_load_test_config_chv(dev_priv); - } else if (IS_BROADWELL(dev_priv)) { - i915_perf_load_test_config_bdw(dev_priv); - } else if (IS_HASWELL(dev_priv)) { - i915_perf_load_test_config_hsw(dev_priv); -} - - if (dev_priv->perf.test_config.id == 0) + sysfs_attr_init(&perf->test_config.sysfs_metric_id.attr); + + if (INTEL_GEN(i915) >= 11) { + i915_perf_load_test_config_icl(i915); + } else if (IS_CANNONLAKE(i915)) { + i915_perf_load_test_config_cnl(i915); + } else if (IS_COFFEELAKE(i915)) { + if (IS_CFL_GT2(i915)) + i915_perf_load_test_config_cflgt2(i915); + if (IS_CFL_GT3(i915)) + i915_perf_load_test_config_cflgt3(i915); + } else if (IS_GEMINILAKE(i915)) { + i915_perf_load_test_config_glk(i915); + } else if (IS_KABYLAKE(i915)) { + if (IS_KBL_GT2(i915)) + i915_perf_load_test_config_kblgt2(i915); + else if (IS_KBL_GT3(i915)) + i915_perf_load_test_config_kblgt3(i915); + } else if (IS_BROXTON(i915)) { + i915_perf_load_test_config_bxt(i915); + } else if (IS_SKYLAKE(i915)) { + if (IS_SKL_GT2(i915)) + i915_perf_load_test_config_sklgt2(i915); + else if (IS_SKL_GT3(i915)) + i915_perf_load_test_config_sklgt3(i915); + else if (IS_SKL_GT4(i915)) + i915_perf_load_test_config_sklgt4(i915); + } else if (IS_CHERRYVIEW(i915)) { + i915_perf_load_test_config_chv(i915); + } else if (IS_BROADWELL(i915)) { + i915_perf_load_test_config_bdw(i915); + } else if (IS_HASWELL(i915)) { + i915_perf_load_test_config_hsw(i915); + } + + if (perf->test_config.id == 0) goto sysfs_error; - ret = sysfs_create_group(dev_priv->perf.metrics_kobj, - &dev_priv->perf.test_config.sysfs_metric); + ret = sysfs_create_group(perf->metrics_kobj, + &perf->test_config.sysfs_metric); if (ret) goto sysfs_error; - atomic_set(&dev_priv->perf.test_config.ref_count, 1); + perf->test_config.perf = perf; + kref_init(&perf->test_config.ref); goto exit; sysfs_error: - kobject_put(dev_priv->perf.metrics_kobj); - dev_priv->perf.metrics_kobj = NULL; + kobject_put(perf->metrics_kobj); + perf->metrics_kobj = NULL; exit: - mutex_unlock(&dev_priv->perf.lock); + mutex_unlock(&perf->lock); } /** * i915_perf_unregister - hide i915-perf from userspace - * @dev_priv: i915 device instance + * @i915: i915 device instance * * i915-perf state cleanup is split up into an 'unregister' and * 'deinit' phase where the interface is first hidden from * userspace by i915_perf_unregister() before cleaning up * remaining state in i915_perf_fini(). */ -void i915_perf_unregister(struct drm_i915_private *dev_priv) +void i915_perf_unregister(struct drm_i915_private *i915) { - if (!dev_priv->perf.metrics_kobj) + struct i915_perf *perf = &i915->perf; + + if (!perf->metrics_kobj) return; - sysfs_remove_group(dev_priv->perf.metrics_kobj, - &dev_priv->perf.test_config.sysfs_metric); + sysfs_remove_group(perf->metrics_kobj, + &perf->test_config.sysfs_metric); - kobject_put(dev_priv->perf.metrics_kobj); - dev_priv->perf.metrics_kobj = NULL; + kobject_put(perf->metrics_kobj); + perf->metrics_kobj = NULL; } -static bool gen8_is_valid_flex_addr(struct drm_i915_private *dev_priv, u32 addr) +static bool gen8_is_valid_flex_addr(struct i915_perf *perf, u32 addr) { static const i915_reg_t flex_eu_regs[] = { EU_PERF_CNTL0, @@ -3123,7 +3515,7 @@ static bool gen8_is_valid_flex_addr(struct drm_i915_private *dev_priv, u32 addr) return false; } -static bool gen7_is_valid_b_counter_addr(struct drm_i915_private *dev_priv, u32 addr) +static bool gen7_is_valid_b_counter_addr(struct i915_perf *perf, u32 addr) { return (addr >= i915_mmio_reg_offset(OASTARTTRIG1) && addr <= i915_mmio_reg_offset(OASTARTTRIG8)) || @@ -3133,7 +3525,7 @@ static bool gen7_is_valid_b_counter_addr(struct drm_i915_private *dev_priv, u32 addr <= i915_mmio_reg_offset(OACEC7_1)); } -static bool gen7_is_valid_mux_addr(struct drm_i915_private *dev_priv, u32 addr) +static bool gen7_is_valid_mux_addr(struct i915_perf *perf, u32 addr) { return addr == i915_mmio_reg_offset(HALF_SLICE_CHICKEN2) || (addr >= i915_mmio_reg_offset(MICRO_BP0_0) && @@ -3144,34 +3536,34 @@ static bool gen7_is_valid_mux_addr(struct drm_i915_private *dev_priv, u32 addr) addr <= i915_mmio_reg_offset(OA_PERFMATRIX_HI)); } -static bool gen8_is_valid_mux_addr(struct drm_i915_private *dev_priv, u32 addr) +static bool gen8_is_valid_mux_addr(struct i915_perf *perf, u32 addr) { - return gen7_is_valid_mux_addr(dev_priv, addr) || + return gen7_is_valid_mux_addr(perf, addr) || addr == i915_mmio_reg_offset(WAIT_FOR_RC6_EXIT) || (addr >= i915_mmio_reg_offset(RPM_CONFIG0) && addr <= i915_mmio_reg_offset(NOA_CONFIG(8))); } -static bool gen10_is_valid_mux_addr(struct drm_i915_private *dev_priv, u32 addr) +static bool gen10_is_valid_mux_addr(struct i915_perf *perf, u32 addr) { - return gen8_is_valid_mux_addr(dev_priv, addr) || + return gen8_is_valid_mux_addr(perf, addr) || addr == i915_mmio_reg_offset(GEN10_NOA_WRITE_HIGH) || (addr >= i915_mmio_reg_offset(OA_PERFCNT3_LO) && addr <= i915_mmio_reg_offset(OA_PERFCNT4_HI)); } -static bool hsw_is_valid_mux_addr(struct drm_i915_private *dev_priv, u32 addr) +static bool hsw_is_valid_mux_addr(struct i915_perf *perf, u32 addr) { - return gen7_is_valid_mux_addr(dev_priv, addr) || + return gen7_is_valid_mux_addr(perf, addr) || (addr >= 0x25100 && addr <= 0x2FF90) || (addr >= i915_mmio_reg_offset(HSW_MBVID2_NOA0) && addr <= i915_mmio_reg_offset(HSW_MBVID2_NOA9)) || addr == i915_mmio_reg_offset(HSW_MBVID2_MISR0); } -static bool chv_is_valid_mux_addr(struct drm_i915_private *dev_priv, u32 addr) +static bool chv_is_valid_mux_addr(struct i915_perf *perf, u32 addr) { - return gen7_is_valid_mux_addr(dev_priv, addr) || + return gen7_is_valid_mux_addr(perf, addr) || (addr >= 0x182300 && addr <= 0x1823A4); } @@ -3194,8 +3586,8 @@ static u32 mask_reg_value(u32 reg, u32 val) return val; } -static struct i915_oa_reg *alloc_oa_regs(struct drm_i915_private *dev_priv, - bool (*is_valid)(struct drm_i915_private *dev_priv, u32 addr), +static struct i915_oa_reg *alloc_oa_regs(struct i915_perf *perf, + bool (*is_valid)(struct i915_perf *perf, u32 addr), u32 __user *regs, u32 n_regs) { @@ -3225,7 +3617,7 @@ static struct i915_oa_reg *alloc_oa_regs(struct drm_i915_private *dev_priv, if (err) goto addr_err; - if (!is_valid(dev_priv, addr)) { + if (!is_valid(perf, addr)) { DRM_DEBUG("Invalid oa_reg address: %X\n", addr); err = -EINVAL; goto addr_err; @@ -3258,7 +3650,7 @@ static ssize_t show_dynamic_id(struct device *dev, return sprintf(buf, "%d\n", oa_config->id); } -static int create_dynamic_oa_sysfs_entry(struct drm_i915_private *dev_priv, +static int create_dynamic_oa_sysfs_entry(struct i915_perf *perf, struct i915_oa_config *oa_config) { sysfs_attr_init(&oa_config->sysfs_metric_id.attr); @@ -3273,7 +3665,7 @@ static int create_dynamic_oa_sysfs_entry(struct drm_i915_private *dev_priv, oa_config->sysfs_metric.name = oa_config->uuid; oa_config->sysfs_metric.attrs = oa_config->attrs; - return sysfs_create_group(dev_priv->perf.metrics_kobj, + return sysfs_create_group(perf->metrics_kobj, &oa_config->sysfs_metric); } @@ -3293,17 +3685,18 @@ static int create_dynamic_oa_sysfs_entry(struct drm_i915_private *dev_priv, int i915_perf_add_config_ioctl(struct drm_device *dev, void *data, struct drm_file *file) { - struct drm_i915_private *dev_priv = dev->dev_private; + struct i915_perf *perf = &to_i915(dev)->perf; struct drm_i915_perf_oa_config *args = data; struct i915_oa_config *oa_config, *tmp; + static struct i915_oa_reg *regs; int err, id; - if (!dev_priv->perf.initialized) { + if (!perf->i915) { DRM_DEBUG("i915 perf interface not available for this system\n"); return -ENOTSUPP; } - if (!dev_priv->perf.metrics_kobj) { + if (!perf->metrics_kobj) { DRM_DEBUG("OA metrics weren't advertised via sysfs\n"); return -EINVAL; } @@ -3326,7 +3719,8 @@ int i915_perf_add_config_ioctl(struct drm_device *dev, void *data, return -ENOMEM; } - atomic_set(&oa_config->ref_count, 1); + oa_config->perf = perf; + kref_init(&oa_config->ref); if (!uuid_is_valid(args->uuid)) { DRM_DEBUG("Invalid uuid format for OA config\n"); @@ -3340,59 +3734,59 @@ int i915_perf_add_config_ioctl(struct drm_device *dev, void *data, memcpy(oa_config->uuid, args->uuid, sizeof(args->uuid)); oa_config->mux_regs_len = args->n_mux_regs; - oa_config->mux_regs = - alloc_oa_regs(dev_priv, - dev_priv->perf.ops.is_valid_mux_reg, - u64_to_user_ptr(args->mux_regs_ptr), - args->n_mux_regs); + regs = alloc_oa_regs(perf, + perf->ops.is_valid_mux_reg, + u64_to_user_ptr(args->mux_regs_ptr), + args->n_mux_regs); - if (IS_ERR(oa_config->mux_regs)) { + if (IS_ERR(regs)) { DRM_DEBUG("Failed to create OA config for mux_regs\n"); - err = PTR_ERR(oa_config->mux_regs); + err = PTR_ERR(regs); goto reg_err; } + oa_config->mux_regs = regs; oa_config->b_counter_regs_len = args->n_boolean_regs; - oa_config->b_counter_regs = - alloc_oa_regs(dev_priv, - dev_priv->perf.ops.is_valid_b_counter_reg, - u64_to_user_ptr(args->boolean_regs_ptr), - args->n_boolean_regs); + regs = alloc_oa_regs(perf, + perf->ops.is_valid_b_counter_reg, + u64_to_user_ptr(args->boolean_regs_ptr), + args->n_boolean_regs); - if (IS_ERR(oa_config->b_counter_regs)) { + if (IS_ERR(regs)) { DRM_DEBUG("Failed to create OA config for b_counter_regs\n"); - err = PTR_ERR(oa_config->b_counter_regs); + err = PTR_ERR(regs); goto reg_err; } + oa_config->b_counter_regs = regs; - if (INTEL_GEN(dev_priv) < 8) { + if (INTEL_GEN(perf->i915) < 8) { if (args->n_flex_regs != 0) { err = -EINVAL; goto reg_err; } } else { oa_config->flex_regs_len = args->n_flex_regs; - oa_config->flex_regs = - alloc_oa_regs(dev_priv, - dev_priv->perf.ops.is_valid_flex_reg, - u64_to_user_ptr(args->flex_regs_ptr), - args->n_flex_regs); + regs = alloc_oa_regs(perf, + perf->ops.is_valid_flex_reg, + u64_to_user_ptr(args->flex_regs_ptr), + args->n_flex_regs); - if (IS_ERR(oa_config->flex_regs)) { + if (IS_ERR(regs)) { DRM_DEBUG("Failed to create OA config for flex_regs\n"); - err = PTR_ERR(oa_config->flex_regs); + err = PTR_ERR(regs); goto reg_err; } + oa_config->flex_regs = regs; } - err = mutex_lock_interruptible(&dev_priv->perf.metrics_lock); + err = mutex_lock_interruptible(&perf->metrics_lock); if (err) goto reg_err; /* We shouldn't have too many configs, so this iteration shouldn't be * too costly. */ - idr_for_each_entry(&dev_priv->perf.metrics_idr, tmp, id) { + idr_for_each_entry(&perf->metrics_idr, tmp, id) { if (!strcmp(tmp->uuid, oa_config->uuid)) { DRM_DEBUG("OA config already exists with this uuid\n"); err = -EADDRINUSE; @@ -3400,14 +3794,14 @@ int i915_perf_add_config_ioctl(struct drm_device *dev, void *data, } } - err = create_dynamic_oa_sysfs_entry(dev_priv, oa_config); + err = create_dynamic_oa_sysfs_entry(perf, oa_config); if (err) { DRM_DEBUG("Failed to create sysfs entry for OA config\n"); goto sysfs_err; } /* Config id 0 is invalid, id 1 for kernel stored test config. */ - oa_config->id = idr_alloc(&dev_priv->perf.metrics_idr, + oa_config->id = idr_alloc(&perf->metrics_idr, oa_config, 2, 0, GFP_KERNEL); if (oa_config->id < 0) { @@ -3416,16 +3810,16 @@ int i915_perf_add_config_ioctl(struct drm_device *dev, void *data, goto sysfs_err; } - mutex_unlock(&dev_priv->perf.metrics_lock); + mutex_unlock(&perf->metrics_lock); DRM_DEBUG("Added config %s id=%i\n", oa_config->uuid, oa_config->id); return oa_config->id; sysfs_err: - mutex_unlock(&dev_priv->perf.metrics_lock); + mutex_unlock(&perf->metrics_lock); reg_err: - put_oa_config(dev_priv, oa_config); + i915_oa_config_put(oa_config); DRM_DEBUG("Failed to add new OA config\n"); return err; } @@ -3444,12 +3838,12 @@ reg_err: int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data, struct drm_file *file) { - struct drm_i915_private *dev_priv = dev->dev_private; + struct i915_perf *perf = &to_i915(dev)->perf; u64 *arg = data; struct i915_oa_config *oa_config; int ret; - if (!dev_priv->perf.initialized) { + if (!perf->i915) { DRM_DEBUG("i915 perf interface not available for this system\n"); return -ENOTSUPP; } @@ -3459,31 +3853,33 @@ int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data, return -EACCES; } - ret = mutex_lock_interruptible(&dev_priv->perf.metrics_lock); + ret = mutex_lock_interruptible(&perf->metrics_lock); if (ret) - goto lock_err; + return ret; - oa_config = idr_find(&dev_priv->perf.metrics_idr, *arg); + oa_config = idr_find(&perf->metrics_idr, *arg); if (!oa_config) { DRM_DEBUG("Failed to remove unknown OA config\n"); ret = -ENOENT; - goto config_err; + goto err_unlock; } GEM_BUG_ON(*arg != oa_config->id); - sysfs_remove_group(dev_priv->perf.metrics_kobj, - &oa_config->sysfs_metric); + sysfs_remove_group(perf->metrics_kobj, &oa_config->sysfs_metric); - idr_remove(&dev_priv->perf.metrics_idr, *arg); + idr_remove(&perf->metrics_idr, *arg); + + mutex_unlock(&perf->metrics_lock); DRM_DEBUG("Removed config %s id=%i\n", oa_config->uuid, oa_config->id); - put_oa_config(dev_priv, oa_config); + i915_oa_config_put(oa_config); -config_err: - mutex_unlock(&dev_priv->perf.metrics_lock); -lock_err: + return 0; + +err_unlock: + mutex_unlock(&perf->metrics_lock); return ret; } @@ -3531,103 +3927,103 @@ static struct ctl_table dev_root[] = { /** * i915_perf_init - initialize i915-perf state on module load - * @dev_priv: i915 device instance + * @i915: i915 device instance * * Initializes i915-perf state without exposing anything to userspace. * * Note: i915-perf initialization is split into an 'init' and 'register' * phase with the i915_perf_register() exposing state to userspace. */ -void i915_perf_init(struct drm_i915_private *dev_priv) -{ - if (IS_HASWELL(dev_priv)) { - dev_priv->perf.ops.is_valid_b_counter_reg = - gen7_is_valid_b_counter_addr; - dev_priv->perf.ops.is_valid_mux_reg = - hsw_is_valid_mux_addr; - dev_priv->perf.ops.is_valid_flex_reg = NULL; - dev_priv->perf.ops.enable_metric_set = hsw_enable_metric_set; - dev_priv->perf.ops.disable_metric_set = hsw_disable_metric_set; - dev_priv->perf.ops.oa_enable = gen7_oa_enable; - dev_priv->perf.ops.oa_disable = gen7_oa_disable; - dev_priv->perf.ops.read = gen7_oa_read; - dev_priv->perf.ops.oa_hw_tail_read = - gen7_oa_hw_tail_read; - - dev_priv->perf.oa_formats = hsw_oa_formats; - } else if (HAS_LOGICAL_RING_CONTEXTS(dev_priv)) { +void i915_perf_init(struct drm_i915_private *i915) +{ + struct i915_perf *perf = &i915->perf; + + /* XXX const struct i915_perf_ops! */ + + if (IS_HASWELL(i915)) { + perf->ops.is_valid_b_counter_reg = gen7_is_valid_b_counter_addr; + perf->ops.is_valid_mux_reg = hsw_is_valid_mux_addr; + perf->ops.is_valid_flex_reg = NULL; + perf->ops.enable_metric_set = hsw_enable_metric_set; + perf->ops.disable_metric_set = hsw_disable_metric_set; + perf->ops.oa_enable = gen7_oa_enable; + perf->ops.oa_disable = gen7_oa_disable; + perf->ops.read = gen7_oa_read; + perf->ops.oa_hw_tail_read = gen7_oa_hw_tail_read; + + perf->oa_formats = hsw_oa_formats; + } else if (HAS_LOGICAL_RING_CONTEXTS(i915)) { /* Note: that although we could theoretically also support the * legacy ringbuffer mode on BDW (and earlier iterations of * this driver, before upstreaming did this) it didn't seem * worth the complexity to maintain now that BDW+ enable * execlist mode by default. */ - dev_priv->perf.oa_formats = gen8_plus_oa_formats; + perf->oa_formats = gen8_plus_oa_formats; - dev_priv->perf.ops.oa_enable = gen8_oa_enable; - dev_priv->perf.ops.oa_disable = gen8_oa_disable; - dev_priv->perf.ops.read = gen8_oa_read; - dev_priv->perf.ops.oa_hw_tail_read = gen8_oa_hw_tail_read; + perf->ops.oa_enable = gen8_oa_enable; + perf->ops.oa_disable = gen8_oa_disable; + perf->ops.read = gen8_oa_read; + perf->ops.oa_hw_tail_read = gen8_oa_hw_tail_read; - if (IS_GEN_RANGE(dev_priv, 8, 9)) { - dev_priv->perf.ops.is_valid_b_counter_reg = + if (IS_GEN_RANGE(i915, 8, 9)) { + perf->ops.is_valid_b_counter_reg = gen7_is_valid_b_counter_addr; - dev_priv->perf.ops.is_valid_mux_reg = + perf->ops.is_valid_mux_reg = gen8_is_valid_mux_addr; - dev_priv->perf.ops.is_valid_flex_reg = + perf->ops.is_valid_flex_reg = gen8_is_valid_flex_addr; - if (IS_CHERRYVIEW(dev_priv)) { - dev_priv->perf.ops.is_valid_mux_reg = + if (IS_CHERRYVIEW(i915)) { + perf->ops.is_valid_mux_reg = chv_is_valid_mux_addr; } - dev_priv->perf.ops.enable_metric_set = gen8_enable_metric_set; - dev_priv->perf.ops.disable_metric_set = gen8_disable_metric_set; + perf->ops.enable_metric_set = gen8_enable_metric_set; + perf->ops.disable_metric_set = gen8_disable_metric_set; - if (IS_GEN(dev_priv, 8)) { - dev_priv->perf.ctx_oactxctrl_offset = 0x120; - dev_priv->perf.ctx_flexeu0_offset = 0x2ce; + if (IS_GEN(i915, 8)) { + perf->ctx_oactxctrl_offset = 0x120; + perf->ctx_flexeu0_offset = 0x2ce; - dev_priv->perf.gen8_valid_ctx_bit = BIT(25); + perf->gen8_valid_ctx_bit = BIT(25); } else { - dev_priv->perf.ctx_oactxctrl_offset = 0x128; - dev_priv->perf.ctx_flexeu0_offset = 0x3de; + perf->ctx_oactxctrl_offset = 0x128; + perf->ctx_flexeu0_offset = 0x3de; - dev_priv->perf.gen8_valid_ctx_bit = BIT(16); + perf->gen8_valid_ctx_bit = BIT(16); } - } else if (IS_GEN_RANGE(dev_priv, 10, 11)) { - dev_priv->perf.ops.is_valid_b_counter_reg = + } else if (IS_GEN_RANGE(i915, 10, 11)) { + perf->ops.is_valid_b_counter_reg = gen7_is_valid_b_counter_addr; - dev_priv->perf.ops.is_valid_mux_reg = + perf->ops.is_valid_mux_reg = gen10_is_valid_mux_addr; - dev_priv->perf.ops.is_valid_flex_reg = + perf->ops.is_valid_flex_reg = gen8_is_valid_flex_addr; - dev_priv->perf.ops.enable_metric_set = gen8_enable_metric_set; - dev_priv->perf.ops.disable_metric_set = gen10_disable_metric_set; + perf->ops.enable_metric_set = gen8_enable_metric_set; + perf->ops.disable_metric_set = gen10_disable_metric_set; - if (IS_GEN(dev_priv, 10)) { - dev_priv->perf.ctx_oactxctrl_offset = 0x128; - dev_priv->perf.ctx_flexeu0_offset = 0x3de; + if (IS_GEN(i915, 10)) { + perf->ctx_oactxctrl_offset = 0x128; + perf->ctx_flexeu0_offset = 0x3de; } else { - dev_priv->perf.ctx_oactxctrl_offset = 0x124; - dev_priv->perf.ctx_flexeu0_offset = 0x78e; + perf->ctx_oactxctrl_offset = 0x124; + perf->ctx_flexeu0_offset = 0x78e; } - dev_priv->perf.gen8_valid_ctx_bit = BIT(16); + perf->gen8_valid_ctx_bit = BIT(16); } } - if (dev_priv->perf.ops.enable_metric_set) { - INIT_LIST_HEAD(&dev_priv->perf.streams); - mutex_init(&dev_priv->perf.lock); + if (perf->ops.enable_metric_set) { + mutex_init(&perf->lock); oa_sample_rate_hard_limit = 1000 * - (RUNTIME_INFO(dev_priv)->cs_timestamp_frequency_khz / 2); - dev_priv->perf.sysctl_header = register_sysctl_table(dev_root); + (RUNTIME_INFO(i915)->cs_timestamp_frequency_khz / 2); + perf->sysctl_header = register_sysctl_table(dev_root); - mutex_init(&dev_priv->perf.metrics_lock); - idr_init(&dev_priv->perf.metrics_idr); + mutex_init(&perf->metrics_lock); + idr_init(&perf->metrics_idr); /* We set up some ratelimit state to potentially throttle any * _NOTES about spurious, invalid OA reports which we don't @@ -3639,44 +4035,70 @@ void i915_perf_init(struct drm_i915_private *dev_priv) * * Using the same limiting factors as printk_ratelimit() */ - ratelimit_state_init(&dev_priv->perf.spurious_report_rs, - 5 * HZ, 10); + ratelimit_state_init(&perf->spurious_report_rs, 5 * HZ, 10); /* Since we use a DRM_NOTE for spurious reports it would be * inconsistent to let __ratelimit() automatically print a * warning for throttling. */ - ratelimit_set_flags(&dev_priv->perf.spurious_report_rs, + ratelimit_set_flags(&perf->spurious_report_rs, RATELIMIT_MSG_ON_RELEASE); - dev_priv->perf.initialized = true; + atomic64_set(&perf->noa_programming_delay, + 500 * 1000 /* 500us */); + + perf->i915 = i915; } } static int destroy_config(int id, void *p, void *data) { - struct drm_i915_private *dev_priv = data; - struct i915_oa_config *oa_config = p; - - put_oa_config(dev_priv, oa_config); - + i915_oa_config_put(p); return 0; } /** * i915_perf_fini - Counter part to i915_perf_init() - * @dev_priv: i915 device instance + * @i915: i915 device instance */ -void i915_perf_fini(struct drm_i915_private *dev_priv) +void i915_perf_fini(struct drm_i915_private *i915) { - if (!dev_priv->perf.initialized) + struct i915_perf *perf = &i915->perf; + + if (!perf->i915) return; - idr_for_each(&dev_priv->perf.metrics_idr, destroy_config, dev_priv); - idr_destroy(&dev_priv->perf.metrics_idr); + idr_for_each(&perf->metrics_idr, destroy_config, perf); + idr_destroy(&perf->metrics_idr); - unregister_sysctl_table(dev_priv->perf.sysctl_header); + unregister_sysctl_table(perf->sysctl_header); - memset(&dev_priv->perf.ops, 0, sizeof(dev_priv->perf.ops)); + memset(&perf->ops, 0, sizeof(perf->ops)); + perf->i915 = NULL; +} - dev_priv->perf.initialized = false; +/** + * i915_perf_ioctl_version - Version of the i915-perf subsystem + * + * This version number is used by userspace to detect available features. + */ +int i915_perf_ioctl_version(void) +{ + /* + * 1: Initial version + * I915_PERF_IOCTL_ENABLE + * I915_PERF_IOCTL_DISABLE + * + * 2: Added runtime modification of OA config. + * I915_PERF_IOCTL_CONFIG + * + * 3: Add DRM_I915_PERF_PROP_HOLD_PREEMPTION parameter to hold + * preemption on a particular context so that performance data is + * accessible from a delta of MI_RPC reports without looking at the + * OA buffer. + */ + return 3; } + +#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) +#include "selftests/i915_perf.c" +#endif |