From b7cfe79f06d673fccd388896ff67f305b8378716 Mon Sep 17 00:00:00 2001 From: "Dr. David Alan Gilbert" Date: Sun, 3 Nov 2024 14:49:36 +0000 Subject: [PATCH 01/12] drm/i915/gt: Remove unused execlists_unwind_incomplete_requests execlists_unwind_incomplete_requests() is unused since 2021's commit eb5e7da736f3 ("drm/i915/guc: Reset implementation for new GuC interface") Remove it. Signed-off-by: Dr. David Alan Gilbert Link: https://patchwork.freedesktop.org/patch/msgid/20241103144936.238116-1-linux@treblig.org Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/intel_engine.h | 3 --- drivers/gpu/drm/i915/gt/intel_execlists_submission.c | 9 --------- 2 files changed, 12 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_engine.h b/drivers/gpu/drm/i915/gt/intel_engine.h index 40269e4c1e31..325da0414d94 100644 --- a/drivers/gpu/drm/i915/gt/intel_engine.h +++ b/drivers/gpu/drm/i915/gt/intel_engine.h @@ -126,9 +126,6 @@ execlists_active(const struct intel_engine_execlists *execlists) return active; } -struct i915_request * -execlists_unwind_incomplete_requests(struct intel_engine_execlists *execlists); - static inline u32 intel_read_status_page(const struct intel_engine_cs *engine, int reg) { diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c index 72090f52fb85..4a80ffa1b962 100644 --- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c +++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c @@ -405,15 +405,6 @@ __unwind_incomplete_requests(struct intel_engine_cs *engine) return active; } -struct i915_request * -execlists_unwind_incomplete_requests(struct intel_engine_execlists *execlists) -{ - struct intel_engine_cs *engine = - container_of(execlists, typeof(*engine), execlists); - - return __unwind_incomplete_requests(engine); -} - static void execlists_context_status_change(struct i915_request *rq, unsigned long status) { From c62018a002dd5da0262e005a89fe691ca8d57cf6 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Mon, 4 Nov 2024 13:35:09 -0800 Subject: [PATCH 02/12] drm/i915/pmu: Rename cpuhp_slot to cpuhp_state Both the documentation and most of other users call the return of cpuhp_setup_state_multi() as "state". Follow that. Reviewed-by: Matt Roper Link: https://patchwork.freedesktop.org/patch/msgid/20241104213512.2314930-2-lucas.demarchi@intel.com Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/i915/i915_pmu.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c index 93fbf53578da..8706957ddc0a 100644 --- a/drivers/gpu/drm/i915/i915_pmu.c +++ b/drivers/gpu/drm/i915/i915_pmu.c @@ -1218,7 +1218,7 @@ static int i915_pmu_cpu_offline(unsigned int cpu, struct hlist_node *node) return 0; } -static enum cpuhp_state cpuhp_slot = CPUHP_INVALID; +static enum cpuhp_state cpuhp_state = CPUHP_INVALID; int i915_pmu_init(void) { @@ -1232,28 +1232,28 @@ int i915_pmu_init(void) pr_notice("Failed to setup cpuhp state for i915 PMU! (%d)\n", ret); else - cpuhp_slot = ret; + cpuhp_state = ret; return 0; } void i915_pmu_exit(void) { - if (cpuhp_slot != CPUHP_INVALID) - cpuhp_remove_multi_state(cpuhp_slot); + if (cpuhp_state != CPUHP_INVALID) + cpuhp_remove_multi_state(cpuhp_state); } static int i915_pmu_register_cpuhp_state(struct i915_pmu *pmu) { - if (cpuhp_slot == CPUHP_INVALID) + if (cpuhp_state == CPUHP_INVALID) return -EINVAL; - return cpuhp_state_add_instance(cpuhp_slot, &pmu->cpuhp.node); + return cpuhp_state_add_instance(cpuhp_state, &pmu->cpuhp.node); } static void i915_pmu_unregister_cpuhp_state(struct i915_pmu *pmu) { - cpuhp_state_remove_instance(cpuhp_slot, &pmu->cpuhp.node); + cpuhp_state_remove_instance(cpuhp_state, &pmu->cpuhp.node); } void i915_pmu_register(struct drm_i915_private *i915) From 9116b5760e615336b0c5060a85b25b2ec7d7c48b Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Mon, 4 Nov 2024 13:35:10 -0800 Subject: [PATCH 03/12] drm/i915/pmu: Stop setting event_init to NULL Setting event_init to NULL is mostly done to detect when the driver is partially working: i915 probed, but pmu is not registered. However, checking for event_init is odd as it was supposed to always be set and kernel/events/ would just crash if it found it set to NULL. Since there's already a "closed" boolean, use that instead and extend it's meaning to unregistered/unregistering. Reviewed-by: Matt Roper Link: https://patchwork.freedesktop.org/patch/msgid/20241104213512.2314930-3-lucas.demarchi@intel.com Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/i915/i915_pmu.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c index 8706957ddc0a..86faa5705fd8 100644 --- a/drivers/gpu/drm/i915/i915_pmu.c +++ b/drivers/gpu/drm/i915/i915_pmu.c @@ -302,7 +302,7 @@ void i915_pmu_gt_parked(struct intel_gt *gt) { struct i915_pmu *pmu = >->i915->pmu; - if (!pmu->base.event_init) + if (pmu->closed) return; spin_lock_irq(&pmu->lock); @@ -324,7 +324,7 @@ void i915_pmu_gt_unparked(struct intel_gt *gt) { struct i915_pmu *pmu = >->i915->pmu; - if (!pmu->base.event_init) + if (pmu->closed) return; spin_lock_irq(&pmu->lock); @@ -1177,8 +1177,6 @@ static int i915_pmu_cpu_online(unsigned int cpu, struct hlist_node *node) { struct i915_pmu *pmu = hlist_entry_safe(node, typeof(*pmu), cpuhp.node); - GEM_BUG_ON(!pmu->base.event_init); - /* Select the first online CPU as a designated reader. */ if (cpumask_empty(&i915_pmu_cpumask)) cpumask_set_cpu(cpu, &i915_pmu_cpumask); @@ -1191,8 +1189,6 @@ static int i915_pmu_cpu_offline(unsigned int cpu, struct hlist_node *node) struct i915_pmu *pmu = hlist_entry_safe(node, typeof(*pmu), cpuhp.node); unsigned int target = i915_pmu_target_cpu; - GEM_BUG_ON(!pmu->base.event_init); - /* * Unregistering an instance generates a CPU offline event which we must * ignore to avoid incorrectly modifying the shared i915_pmu_cpumask. @@ -1265,9 +1261,10 @@ void i915_pmu_register(struct drm_i915_private *i915) &i915_pmu_cpumask_attr_group, NULL }; - int ret = -ENOMEM; + pmu->closed = true; + spin_lock_init(&pmu->lock); hrtimer_init(&pmu->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); pmu->timer.function = i915_sample; @@ -1316,6 +1313,8 @@ void i915_pmu_register(struct drm_i915_private *i915) if (ret) goto err_unreg; + pmu->closed = false; + return; err_unreg: @@ -1323,7 +1322,6 @@ void i915_pmu_register(struct drm_i915_private *i915) err_groups: kfree(pmu->base.attr_groups); err_attr: - pmu->base.event_init = NULL; free_event_attributes(pmu); err_name: if (IS_DGFX(i915)) @@ -1336,9 +1334,6 @@ void i915_pmu_unregister(struct drm_i915_private *i915) { struct i915_pmu *pmu = &i915->pmu; - if (!pmu->base.event_init) - return; - /* * "Disconnect" the PMU callbacks - since all are atomic synchronize_rcu * ensures all currently executing ones will have exited before we @@ -1352,7 +1347,6 @@ void i915_pmu_unregister(struct drm_i915_private *i915) i915_pmu_unregister_cpuhp_state(pmu); perf_pmu_unregister(&pmu->base); - pmu->base.event_init = NULL; kfree(pmu->base.attr_groups); if (IS_DGFX(i915)) kfree(pmu->name); From 6ba29f1352482b815e2414b718bbd6de8d884d10 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Mon, 4 Nov 2024 13:35:11 -0800 Subject: [PATCH 04/12] drm/i915/pmu: Replace closed with registered Since i915 calls perf_pmu_register/perf_pmu_unregister, let's call the variable "registered" so we can flip the logic and rely on it being false by default. Looking at other drivers, it's also more common. Examples: arch/x86/events/intel/uncore.c and drivers/powercap/intel_rapl_common.c. Reviewed-by: Matt Roper Link: https://patchwork.freedesktop.org/patch/msgid/20241104213512.2314930-4-lucas.demarchi@intel.com Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/i915/i915_pmu.c | 25 +++++++++++++------------ drivers/gpu/drm/i915/i915_pmu.h | 4 ++-- 2 files changed, 15 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c index 86faa5705fd8..bd476297ed0a 100644 --- a/drivers/gpu/drm/i915/i915_pmu.c +++ b/drivers/gpu/drm/i915/i915_pmu.c @@ -302,7 +302,7 @@ void i915_pmu_gt_parked(struct intel_gt *gt) { struct i915_pmu *pmu = >->i915->pmu; - if (pmu->closed) + if (!pmu->registered) return; spin_lock_irq(&pmu->lock); @@ -324,7 +324,7 @@ void i915_pmu_gt_unparked(struct intel_gt *gt) { struct i915_pmu *pmu = >->i915->pmu; - if (pmu->closed) + if (!pmu->registered) return; spin_lock_irq(&pmu->lock); @@ -626,7 +626,7 @@ static int i915_pmu_event_init(struct perf_event *event) struct drm_i915_private *i915 = pmu_to_i915(pmu); int ret; - if (pmu->closed) + if (!pmu->registered) return -ENODEV; if (event->attr.type != event->pmu->type) @@ -724,7 +724,7 @@ static void i915_pmu_event_read(struct perf_event *event) struct hw_perf_event *hwc = &event->hw; u64 prev, new; - if (pmu->closed) { + if (!pmu->registered) { event->hw.state = PERF_HES_STOPPED; return; } @@ -850,7 +850,7 @@ static void i915_pmu_event_start(struct perf_event *event, int flags) { struct i915_pmu *pmu = event_to_pmu(event); - if (pmu->closed) + if (!pmu->registered) return; i915_pmu_enable(event); @@ -861,7 +861,7 @@ static void i915_pmu_event_stop(struct perf_event *event, int flags) { struct i915_pmu *pmu = event_to_pmu(event); - if (pmu->closed) + if (!pmu->registered) goto out; if (flags & PERF_EF_UPDATE) @@ -877,7 +877,7 @@ static int i915_pmu_event_add(struct perf_event *event, int flags) { struct i915_pmu *pmu = event_to_pmu(event); - if (pmu->closed) + if (!pmu->registered) return -ENODEV; if (flags & PERF_EF_START) @@ -1193,7 +1193,7 @@ static int i915_pmu_cpu_offline(unsigned int cpu, struct hlist_node *node) * Unregistering an instance generates a CPU offline event which we must * ignore to avoid incorrectly modifying the shared i915_pmu_cpumask. */ - if (pmu->closed) + if (!pmu->registered) return 0; if (cpumask_test_and_clear_cpu(cpu, &i915_pmu_cpumask)) { @@ -1263,8 +1263,6 @@ void i915_pmu_register(struct drm_i915_private *i915) }; int ret = -ENOMEM; - pmu->closed = true; - spin_lock_init(&pmu->lock); hrtimer_init(&pmu->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); pmu->timer.function = i915_sample; @@ -1313,7 +1311,7 @@ void i915_pmu_register(struct drm_i915_private *i915) if (ret) goto err_unreg; - pmu->closed = false; + pmu->registered = true; return; @@ -1334,12 +1332,15 @@ void i915_pmu_unregister(struct drm_i915_private *i915) { struct i915_pmu *pmu = &i915->pmu; + if (!pmu->registered) + return; + /* * "Disconnect" the PMU callbacks - since all are atomic synchronize_rcu * ensures all currently executing ones will have exited before we * proceed with unregistration. */ - pmu->closed = true; + pmu->registered = false; synchronize_rcu(); hrtimer_cancel(&pmu->timer); diff --git a/drivers/gpu/drm/i915/i915_pmu.h b/drivers/gpu/drm/i915/i915_pmu.h index 41af038c3738..8e66d63d0c9f 100644 --- a/drivers/gpu/drm/i915/i915_pmu.h +++ b/drivers/gpu/drm/i915/i915_pmu.h @@ -68,9 +68,9 @@ struct i915_pmu { */ struct pmu base; /** - * @closed: i915 is unregistering. + * @registered: PMU is registered and not in the unregistering process. */ - bool closed; + bool registered; /** * @name: Name as registered with perf core. */ From 79367b7a58c82d0b1c0a7b0ef748f7aafa91d048 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Mon, 4 Nov 2024 13:35:12 -0800 Subject: [PATCH 05/12] drm/i915/pmu: Remove pointless synchronize_rcu() call This is already done inside perf_pmu_unregister() - no need to do it before. Reviewed-by: Matt Roper Link: https://patchwork.freedesktop.org/patch/msgid/20241104213512.2314930-5-lucas.demarchi@intel.com Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/i915/i915_pmu.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c index bd476297ed0a..e55db036be1b 100644 --- a/drivers/gpu/drm/i915/i915_pmu.c +++ b/drivers/gpu/drm/i915/i915_pmu.c @@ -1335,13 +1335,8 @@ void i915_pmu_unregister(struct drm_i915_private *i915) if (!pmu->registered) return; - /* - * "Disconnect" the PMU callbacks - since all are atomic synchronize_rcu - * ensures all currently executing ones will have exited before we - * proceed with unregistration. - */ + /* Disconnect the PMU callbacks */ pmu->registered = false; - synchronize_rcu(); hrtimer_cancel(&pmu->timer); From b939a08bc378a7c716ad7a9486b48794b95d22f5 Mon Sep 17 00:00:00 2001 From: Zhanjun Dong Date: Mon, 4 Nov 2024 13:41:03 -0800 Subject: [PATCH 06/12] drm/i915/guc: Flush ct receive tasklet during reset preparation GuC to host communication is interrupt driven, the handling has 3 parts: interrupt context, tasklet and request queue worker. During GuC reset prepare, interrupt is disabled before destroy contexts steps start. The IRQ and worker are flushed to finish any outstanding in-progress message handling. But, the tasklet flush is missing, it might causes 2 race conditions: 1. Tasklet runs after IRQ flushed, add request to queue after worker flush started, causes unexpected G2H message request processing, meanwhile, reset prepare code already get the context destroyed. This will causes error reported about bad context state. (https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/11349 and https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12303) 2. Tasklet runs after intel_guc_submission_reset_prepare, ct_try_receive_message start to run, while intel_uc_reset_prepare already finished guc sanitize and set ct->enable to false. This will causes warning on incorrect ct->enable state. (https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12439) Add the missing tasklet flush to flush all 3 parts. Signed-off-by: Zhanjun Dong Reviewed-by: Alan Previn Signed-off-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20241104214103.214702-1-zhanjun.dong@intel.com --- drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c index ed979847187f..6a3d550f6ed1 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -1688,6 +1688,10 @@ void intel_guc_submission_reset_prepare(struct intel_guc *guc) spin_lock_irq(guc_to_gt(guc)->irq_lock); spin_unlock_irq(guc_to_gt(guc)->irq_lock); + /* Flush tasklet */ + tasklet_disable(&guc->ct.receive_tasklet); + tasklet_enable(&guc->ct.receive_tasklet); + guc_flush_submissions(guc); guc_flush_destroyed_contexts(guc); flush_work(&guc->ct.requests.worker); From 2e0438f9c3d25eea8bc8e9b4dcff7edfb64cb9e7 Mon Sep 17 00:00:00 2001 From: Krzysztof Karas Date: Mon, 18 Nov 2024 12:19:22 +0000 Subject: [PATCH 07/12] drm/i915: ensure segment offset never exceeds allowed max Commit 255fc1703e42 ("drm/i915/gem: Calculate object page offset for partial memory mapping") introduced a new offset, which accounts for userspace mapping not starting from the beginning of object's scatterlist. This works fine for cases where first object pte is larger than the new offset - "r->sgt.curr" counter is set to the offset to match the difference in the number of total pages. However, if object's first pte's size is equal to or smaller than the offset, then information about the offset in userspace is covered up by moving "r->sgt" pointer in remap_sg(): r->sgt.curr += PAGE_SIZE; if (r->sgt.curr >= r->sgt.max) r->sgt = __sgt_iter(__sg_next(r->sgt.sgp), use_dma(r->iobase)); This means that two or more pages from virtual memory are counted for only one page in object's memory, because after moving "r->sgt" pointer "r->sgt.curr" will be 0. We should account for this mismatch by moving "r->sgt" pointer to the next pte. For that we may use "r.sgt.max", which already holds the max allowed size. This change also eliminates possible confusion, when looking at i915_scatterlist.h and remap_io_sg() code: former has scatterlist pointer definition, which differentiates "s.max" value based on "dma" flag (sg_dma_len() is used only when the flag is enabled), while latter uses sg_dma_len() indiscriminately. This patch aims to resolve issue: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12031 v3: - instead of checking if r.sgt.curr would exceed allowed max, changed the value in the while loop to be aligned with `dma` value v4: - remove unnecessary parent relation v5: - update commit message with explanation about page counting mismatch and link to the issue Signed-off-by: Krzysztof Karas Reviewed-by: Andi Shyti Signed-off-by: Andi Shyti Link: https://patchwork.freedesktop.org/patch/msgid/upbjdavlbcxku63ns4vstp5kgbn2anxwewpmnppszgb67fn66t@tfclfgkqijue --- drivers/gpu/drm/i915/i915_mm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_mm.c b/drivers/gpu/drm/i915/i915_mm.c index f5c97a620962..76e2801619f0 100644 --- a/drivers/gpu/drm/i915/i915_mm.c +++ b/drivers/gpu/drm/i915/i915_mm.c @@ -143,8 +143,8 @@ int remap_io_sg(struct vm_area_struct *vma, /* We rely on prevalidation of the io-mapping to skip track_pfn(). */ GEM_BUG_ON((vma->vm_flags & EXPECTED_FLAGS) != EXPECTED_FLAGS); - while (offset >= sg_dma_len(r.sgt.sgp) >> PAGE_SHIFT) { - offset -= sg_dma_len(r.sgt.sgp) >> PAGE_SHIFT; + while (offset >= r.sgt.max >> PAGE_SHIFT) { + offset -= r.sgt.max >> PAGE_SHIFT; r.sgt = __sgt_iter(__sg_next(r.sgt.sgp), use_dma(iobase)); if (!r.sgt.sgp) return -EINVAL; From 630e03808a71c06407f5720b494fd76f5665ffc2 Mon Sep 17 00:00:00 2001 From: Sk Anirban Date: Tue, 3 Dec 2024 11:41:14 +0530 Subject: [PATCH 08/12] drm/i915/selftests: Add delay to stabilize frequency in live_rps_power Add delays to allow frequency stabilization before power measurement to fix sporadic power conservation issues in live_rps_power test. v2: - Move delay to respective function (Badal) Signed-off-by: Sk Anirban Reviewed-by: Badal Nilawar Signed-off-by: Anshuman Gupta Link: https://patchwork.freedesktop.org/patch/msgid/20241203061114.2790448-1-sk.anirban@intel.com --- drivers/gpu/drm/i915/gt/selftest_rps.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/gt/selftest_rps.c b/drivers/gpu/drm/i915/gt/selftest_rps.c index dcef8d498919..c207a4fb03bf 100644 --- a/drivers/gpu/drm/i915/gt/selftest_rps.c +++ b/drivers/gpu/drm/i915/gt/selftest_rps.c @@ -1125,6 +1125,7 @@ static u64 measure_power(struct intel_rps *rps, int *freq) static u64 measure_power_at(struct intel_rps *rps, int *freq) { *freq = rps_set_check(rps, *freq); + msleep(100); return measure_power(rps, freq); } From abd318237fa6556c1e5225529af145ef15d5ff0d Mon Sep 17 00:00:00 2001 From: Umesh Nerlige Ramappa Date: Wed, 27 Nov 2024 09:40:04 -0800 Subject: [PATCH 09/12] i915/guc: Reset engine utilization buffer before registration On GT reset, we store total busyness counts for all engines and re-register the utilization buffer with GuC. At that time we should reset the buffer, so that we don't get spurious busyness counts on subsequent queries. To repro this issue, run igt@perf_pmu@busy-hang followed by igt@perf_pmu@most-busy-idle-check-all for a couple iterations. Fixes: 77cdd054dd2c ("drm/i915/pmu: Connect engine busyness stats from GuC to pmu") Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20241127174006.190128-2-umesh.nerlige.ramappa@intel.com --- .../gpu/drm/i915/gt/uc/intel_guc_submission.c | 21 +++++++++++++++++++ 1 file changed, 21 insertions(+) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c index 6a3d550f6ed1..795058bb7141 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -1243,6 +1243,21 @@ static void __get_engine_usage_record(struct intel_engine_cs *engine, } while (++i < 6); } +static void __set_engine_usage_record(struct intel_engine_cs *engine, + u32 last_in, u32 id, u32 total) +{ + struct iosys_map rec_map = intel_guc_engine_usage_record_map(engine); + +#define record_write(map_, field_, val_) \ + iosys_map_wr_field(map_, 0, struct guc_engine_usage_record, field_, val_) + + record_write(&rec_map, last_switch_in_stamp, last_in); + record_write(&rec_map, current_context_index, id); + record_write(&rec_map, total_runtime, total); + +#undef record_write +} + static void guc_update_engine_gt_clks(struct intel_engine_cs *engine) { struct intel_engine_guc_stats *stats = &engine->stats.guc; @@ -1543,6 +1558,9 @@ static void guc_timestamp_ping(struct work_struct *wrk) static int guc_action_enable_usage_stats(struct intel_guc *guc) { + struct intel_gt *gt = guc_to_gt(guc); + struct intel_engine_cs *engine; + enum intel_engine_id id; u32 offset = intel_guc_engine_usage_offset(guc); u32 action[] = { INTEL_GUC_ACTION_SET_ENG_UTIL_BUFF, @@ -1550,6 +1568,9 @@ static int guc_action_enable_usage_stats(struct intel_guc *guc) 0, }; + for_each_engine(engine, gt, id) + __set_engine_usage_record(engine, 0, 0xffffffff, 0); + return intel_guc_send(guc, action, ARRAY_SIZE(action)); } From cf907f6d294217985e9dafd9985dce874e04ca37 Mon Sep 17 00:00:00 2001 From: Umesh Nerlige Ramappa Date: Wed, 27 Nov 2024 09:40:05 -0800 Subject: [PATCH 10/12] i915/guc: Ensure busyness counter increases motonically Active busyness of an engine is calculated using gt timestamp and the context switch in time. While capturing the gt timestamp, it's possible that the context switches out. This race could result in an active busyness value that is greater than the actual context runtime value by a small amount. This leads to a negative delta and throws off busyness calculations for the user. If a subsequent count is smaller than the previous one, just return the previous one, since we expect the busyness to catch up. Fixes: 77cdd054dd2c ("drm/i915/pmu: Connect engine busyness stats from GuC to pmu") Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20241127174006.190128-3-umesh.nerlige.ramappa@intel.com --- drivers/gpu/drm/i915/gt/intel_engine_types.h | 5 +++++ drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 5 ++++- 2 files changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h index ba55c059063d..fe1f85e5dda3 100644 --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h @@ -343,6 +343,11 @@ struct intel_engine_guc_stats { * @start_gt_clk: GT clock time of last idle to active transition. */ u64 start_gt_clk; + + /** + * @total: The last value of total returned + */ + u64 total; }; union intel_engine_tlb_inv_reg { diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c index 795058bb7141..43c9c5dc2eb5 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -1378,9 +1378,12 @@ static ktime_t guc_engine_busyness(struct intel_engine_cs *engine, ktime_t *now) total += intel_gt_clock_interval_to_ns(gt, clk); } + if (total > stats->total) + stats->total = total; + spin_unlock_irqrestore(&guc->timestamp.lock, flags); - return ns_to_ktime(total); + return ns_to_ktime(stats->total); } static void guc_enable_busyness_worker(struct intel_guc *guc) From 7ed047da59cfa1acb558b95169d347acc8d85da1 Mon Sep 17 00:00:00 2001 From: Umesh Nerlige Ramappa Date: Wed, 27 Nov 2024 09:40:06 -0800 Subject: [PATCH 11/12] i915/guc: Accumulate active runtime on gt reset On gt reset, if a context is running, then accumulate it's active time into the busyness counter since there will be no chance for the context to switch out and update it's run time. v2: Move comment right above the if (John) Fixes: 77cdd054dd2c ("drm/i915/pmu: Connect engine busyness stats from GuC to pmu") Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20241127174006.190128-4-umesh.nerlige.ramappa@intel.com --- drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c index 43c9c5dc2eb5..6e8e2a287c83 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -1449,8 +1449,21 @@ static void __reset_guc_busyness_stats(struct intel_guc *guc) guc_update_pm_timestamp(guc, &unused); for_each_engine(engine, gt, id) { + struct intel_engine_guc_stats *stats = &engine->stats.guc; + guc_update_engine_gt_clks(engine); - engine->stats.guc.prev_total = 0; + + /* + * If resetting a running context, accumulate the active + * time as well since there will be no context switch. + */ + if (stats->running) { + u64 clk = guc->timestamp.gt_stamp - stats->start_gt_clk; + + stats->total_gt_clks += clk; + } + stats->prev_total = 0; + stats->running = 0; } spin_unlock_irqrestore(&guc->timestamp.lock, flags); From f373ebec18a75d671908e81ed9925aebf279ec2f Mon Sep 17 00:00:00 2001 From: Jesus Narvaez Date: Fri, 13 Dec 2024 12:47:20 -0800 Subject: [PATCH 12/12] drm/i915/guc: Update guc_err message to show outstanding g2h responses Updating the guc_error message to show how many g2h responses are still outstanding, in order to help with future debugging. Signed-off-by: Jesus Narvaez Cc: Daniele Ceraolo Spurio Cc: John Harrison Reviewed-by: Jonathan Cavitt Signed-off-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20241213204720.3918056-1-jesus.narvaez@intel.com --- drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c index 6e8e2a287c83..a2812621625b 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -2046,6 +2046,8 @@ void intel_guc_submission_cancel_requests(struct intel_guc *guc) void intel_guc_submission_reset_finish(struct intel_guc *guc) { + int outstanding; + /* Reset called during driver load or during wedge? */ if (unlikely(!guc_submission_initialized(guc) || !intel_guc_is_fw_running(guc) || @@ -2059,8 +2061,10 @@ void intel_guc_submission_reset_finish(struct intel_guc *guc) * see in CI if this happens frequently / a precursor to taking down the * machine. */ - if (atomic_read(&guc->outstanding_submission_g2h)) - guc_err(guc, "Unexpected outstanding GuC to Host in reset finish\n"); + outstanding = atomic_read(&guc->outstanding_submission_g2h); + if (outstanding) + guc_err(guc, "Unexpected outstanding GuC to Host response(s) in reset finish: %d\n", + outstanding); atomic_set(&guc->outstanding_submission_g2h, 0); intel_guc_global_policies_update(guc);