diff options
Diffstat (limited to '')
21 files changed, 1032 insertions, 191 deletions
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h index ba10bd374cee..fe5d7d261797 100644 --- a/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h +++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h @@ -144,6 +144,7 @@ enum intel_guc_action { INTEL_GUC_ACTION_DEREGISTER_CONTEXT_DONE = 0x4600, INTEL_GUC_ACTION_REGISTER_CONTEXT_MULTI_LRC = 0x4601, INTEL_GUC_ACTION_RESET_CLIENT = 0x5507, + INTEL_GUC_ACTION_SET_ENG_UTIL_BUFF = 0x550A, INTEL_GUC_ACTION_LIMIT }; diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc.h b/drivers/gpu/drm/i915/gt/uc/intel_guc.h index 31cf9fb48c7e..f9240d4baa69 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc.h @@ -95,6 +95,11 @@ struct intel_guc { */ struct ida guc_ids; /** + * @num_guc_ids: Number of guc_ids, selftest feature to be able + * to reduce this number while testing. + */ + int num_guc_ids; + /** * @guc_ids_bitmap: used to allocate new guc_ids, multi-lrc */ unsigned long *guc_ids_bitmap; @@ -138,6 +143,8 @@ struct intel_guc { u32 ads_regset_size; /** @ads_golden_ctxt_size: size of the golden contexts in the ADS */ u32 ads_golden_ctxt_size; + /** @ads_engine_usage_size: size of engine usage in the ADS */ + u32 ads_engine_usage_size; /** @lrc_desc_pool: object allocated to hold the GuC LRC descriptor pool */ struct i915_vma *lrc_desc_pool; @@ -172,6 +179,41 @@ struct intel_guc { /** @send_mutex: used to serialize the intel_guc_send actions */ struct mutex send_mutex; + + /** + * @timestamp: GT timestamp object that stores a copy of the timestamp + * and adjusts it for overflow using a worker. + */ + struct { + /** + * @lock: Lock protecting the below fields and the engine stats. + */ + spinlock_t lock; + + /** + * @gt_stamp: 64 bit extended value of the GT timestamp. + */ + u64 gt_stamp; + + /** + * @ping_delay: Period for polling the GT timestamp for + * overflow. + */ + unsigned long ping_delay; + + /** + * @work: Periodic work to adjust GT timestamp, engine and + * context usage for overflows. + */ + struct delayed_work work; + } timestamp; + +#ifdef CONFIG_DRM_I915_SELFTEST + /** + * @number_guc_id_stolen: The number of guc_ids that have been stolen + */ + int number_guc_id_stolen; +#endif }; static inline struct intel_guc *log_to_guc(struct intel_guc_log *log) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c index 621c893a009f..1a1edae67e4e 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c @@ -26,6 +26,8 @@ * | guc_policies | * +---------------------------------------+ * | guc_gt_system_info | + * +---------------------------------------+ + * | guc_engine_usage | * +---------------------------------------+ <== static * | guc_mmio_reg[countA] (engine 0.0) | * | guc_mmio_reg[countB] (engine 0.1) | @@ -47,6 +49,7 @@ struct __guc_ads_blob { struct guc_ads ads; struct guc_policies policies; struct guc_gt_system_info system_info; + struct guc_engine_usage engine_usage; /* From here on, location is dynamic! Refer to above diagram. */ struct guc_mmio_reg regset[0]; } __packed; @@ -628,3 +631,21 @@ void intel_guc_ads_reset(struct intel_guc *guc) guc_ads_private_data_reset(guc); } + +u32 intel_guc_engine_usage_offset(struct intel_guc *guc) +{ + struct __guc_ads_blob *blob = guc->ads_blob; + u32 base = intel_guc_ggtt_offset(guc, guc->ads_vma); + u32 offset = base + ptr_offset(blob, engine_usage); + + return offset; +} + +struct guc_engine_usage_record *intel_guc_engine_usage(struct intel_engine_cs *engine) +{ + struct intel_guc *guc = &engine->gt->uc.guc; + struct __guc_ads_blob *blob = guc->ads_blob; + u8 guc_class = engine_class_to_guc_class(engine->class); + + return &blob->engine_usage.engines[guc_class][ilog2(engine->logical_mask)]; +} diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.h index 3d85051d57e4..e74c110facff 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.h @@ -6,8 +6,11 @@ #ifndef _INTEL_GUC_ADS_H_ #define _INTEL_GUC_ADS_H_ +#include <linux/types.h> + struct intel_guc; struct drm_printer; +struct intel_engine_cs; int intel_guc_ads_create(struct intel_guc *guc); void intel_guc_ads_destroy(struct intel_guc *guc); @@ -15,5 +18,7 @@ void intel_guc_ads_init_late(struct intel_guc *guc); void intel_guc_ads_reset(struct intel_guc *guc); void intel_guc_ads_print_policy_info(struct intel_guc *guc, struct drm_printer *p); +struct guc_engine_usage_record *intel_guc_engine_usage(struct intel_engine_cs *engine); +u32 intel_guc_engine_usage_offset(struct intel_guc *guc); #endif diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c index a0cc34be7b56..aa6dd6415202 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c @@ -523,6 +523,15 @@ static inline bool ct_deadlocked(struct intel_guc_ct *ct) CT_ERROR(ct, "Communication stalled for %lld ms, desc status=%#x,%#x\n", ktime_ms_delta(ktime_get(), ct->stall_time), send->status, recv->status); + CT_ERROR(ct, "H2G Space: %u (Bytes)\n", + atomic_read(&ct->ctbs.send.space) * 4); + CT_ERROR(ct, "Head: %u (Dwords)\n", ct->ctbs.send.desc->head); + CT_ERROR(ct, "Tail: %u (Dwords)\n", ct->ctbs.send.desc->tail); + CT_ERROR(ct, "G2H Space: %u (Bytes)\n", + atomic_read(&ct->ctbs.recv.space) * 4); + CT_ERROR(ct, "Head: %u\n (Dwords)", ct->ctbs.recv.desc->head); + CT_ERROR(ct, "Tail: %u\n (Dwords)", ct->ctbs.recv.desc->tail); + ct->ctbs.send.broken = true; } @@ -582,12 +591,19 @@ static inline bool h2g_has_room(struct intel_guc_ct *ct, u32 len_dw) static int has_room_nb(struct intel_guc_ct *ct, u32 h2g_dw, u32 g2h_dw) { + bool h2g = h2g_has_room(ct, h2g_dw); + bool g2h = g2h_has_room(ct, g2h_dw); + lockdep_assert_held(&ct->ctbs.send.lock); - if (unlikely(!h2g_has_room(ct, h2g_dw) || !g2h_has_room(ct, g2h_dw))) { + if (unlikely(!h2g || !g2h)) { if (ct->stall_time == KTIME_MAX) ct->stall_time = ktime_get(); + /* Be paranoid and kick G2H tasklet to free credits */ + if (!g2h) + tasklet_hi_schedule(&ct->receive_tasklet); + if (unlikely(ct_deadlocked(ct))) return -EPIPE; else diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c index 196424be0998..31420ce1ce6b 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c @@ -40,9 +40,8 @@ static void guc_prepare_xfer(struct intel_uncore *uncore) } } -/* Copy RSA signature from the fw image to HW for verification */ -static int guc_xfer_rsa(struct intel_uc_fw *guc_fw, - struct intel_uncore *uncore) +static int guc_xfer_rsa_mmio(struct intel_uc_fw *guc_fw, + struct intel_uncore *uncore) { u32 rsa[UOS_RSA_SCRATCH_COUNT]; size_t copied; @@ -58,6 +57,27 @@ static int guc_xfer_rsa(struct intel_uc_fw *guc_fw, return 0; } +static int guc_xfer_rsa_vma(struct intel_uc_fw *guc_fw, + struct intel_uncore *uncore) +{ + struct intel_guc *guc = container_of(guc_fw, struct intel_guc, fw); + + intel_uncore_write(uncore, UOS_RSA_SCRATCH(0), + intel_guc_ggtt_offset(guc, guc_fw->rsa_data)); + + return 0; +} + +/* Copy RSA signature from the fw image to HW for verification */ +static int guc_xfer_rsa(struct intel_uc_fw *guc_fw, + struct intel_uncore *uncore) +{ + if (guc_fw->rsa_data) + return guc_xfer_rsa_vma(guc_fw, uncore); + else + return guc_xfer_rsa_mmio(guc_fw, uncore); +} + /* * Read the GuC status register (GUC_STATUS) and store it in the * specified location; then return a boolean indicating whether @@ -142,7 +162,10 @@ int intel_guc_fw_upload(struct intel_guc *guc) /* * Note that GuC needs the CSS header plus uKernel code to be copied * by the DMA engine in one operation, whereas the RSA signature is - * loaded via MMIO. + * loaded separately, either by copying it to the UOS_RSA_SCRATCH + * register (if key size <= 256) or through a ggtt-pinned vma (if key + * size > 256). The RSA size and therefore the way we provide it to the + * HW is fixed for each platform and hard-coded in the bootrom. */ ret = guc_xfer_rsa(&guc->fw, uncore); if (ret) @@ -164,6 +187,6 @@ int intel_guc_fw_upload(struct intel_guc *guc) return 0; out: - intel_uc_fw_change_status(&guc->fw, INTEL_UC_FIRMWARE_FAIL); + intel_uc_fw_change_status(&guc->fw, INTEL_UC_FIRMWARE_LOAD_FAIL); return ret; } diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h index 722933e26347..7072e30e99f4 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h @@ -294,6 +294,19 @@ struct guc_ads { u32 reserved[15]; } __packed; +/* Engine usage stats */ +struct guc_engine_usage_record { + u32 current_context_index; + u32 last_switch_in_stamp; + u32 reserved0; + u32 total_runtime; + u32 reserved1[4]; +} __packed; + +struct guc_engine_usage { + struct guc_engine_usage_record engines[GUC_MAX_ENGINE_CLASSES][GUC_MAX_INSTANCES_PER_CLASS]; +} __packed; + /* GuC logging structures */ enum guc_log_buffer_type { diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_log.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_log.h index ac1ee1d5ce10..fe6ab7550a14 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_log.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_log.h @@ -15,9 +15,12 @@ struct intel_guc; -#ifdef CONFIG_DRM_I915_DEBUG_GUC +#if defined(CONFIG_DRM_I915_DEBUG_GUC) #define CRASH_BUFFER_SIZE SZ_2M #define DEBUG_BUFFER_SIZE SZ_16M +#elif defined(CONFIG_DRM_I915_DEBUG_GEM) +#define CRASH_BUFFER_SIZE SZ_1M +#define DEBUG_BUFFER_SIZE SZ_2M #else #define CRASH_BUFFER_SIZE SZ_8K #define DEBUG_BUFFER_SIZE SZ_64K diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.c index 46026c2c1722..ddfbe334689f 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.c @@ -10,28 +10,80 @@ #include "intel_guc.h" #include "intel_guc_log.h" #include "intel_guc_log_debugfs.h" +#include "intel_uc.h" + +static u32 obj_to_guc_log_dump_size(struct drm_i915_gem_object *obj) +{ + u32 size; + + if (!obj) + return PAGE_SIZE; + + /* "0x%08x 0x%08x 0x%08x 0x%08x\n" => 16 bytes -> 44 chars => x2.75 */ + size = ((obj->base.size * 11) + 3) / 4; + + /* Add padding for final blank line, any extra header info, etc. */ + size = PAGE_ALIGN(size + PAGE_SIZE); + + return size; +} + +static u32 guc_log_dump_size(struct intel_guc_log *log) +{ + struct intel_guc *guc = log_to_guc(log); + + if (!intel_guc_is_supported(guc)) + return PAGE_SIZE; + + if (!log->vma) + return PAGE_SIZE; + + return obj_to_guc_log_dump_size(log->vma->obj); +} static int guc_log_dump_show(struct seq_file *m, void *data) { struct drm_printer p = drm_seq_file_printer(m); + int ret; - return intel_guc_log_dump(m->private, &p, false); + ret = intel_guc_log_dump(m->private, &p, false); + + if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) && seq_has_overflowed(m)) + pr_warn_once("preallocated size:%zx for %s exceeded\n", + m->size, __func__); + + return ret; +} +DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE_WITH_SIZE(guc_log_dump, guc_log_dump_size); + +static u32 guc_load_err_dump_size(struct intel_guc_log *log) +{ + struct intel_guc *guc = log_to_guc(log); + struct intel_uc *uc = container_of(guc, struct intel_uc, guc); + + if (!intel_guc_is_supported(guc)) + return PAGE_SIZE; + + return obj_to_guc_log_dump_size(uc->load_err_log); } -DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(guc_log_dump); static int guc_load_err_log_dump_show(struct seq_file *m, void *data) { struct drm_printer p = drm_seq_file_printer(m); + if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) && seq_has_overflowed(m)) + pr_warn_once("preallocated size:%zx for %s exceeded\n", + m->size, __func__); + return intel_guc_log_dump(m->private, &p, true); } -DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(guc_load_err_log_dump); +DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE_WITH_SIZE(guc_load_err_log_dump, guc_load_err_dump_size); static int guc_log_level_get(void *data, u64 *val) { struct intel_guc_log *log = data; - if (!intel_guc_is_used(log_to_guc(log))) + if (!log->vma) return -ENODEV; *val = intel_guc_log_get_level(log); @@ -43,7 +95,7 @@ static int guc_log_level_set(void *data, u64 val) { struct intel_guc_log *log = data; - if (!intel_guc_is_used(log_to_guc(log))) + if (!log->vma) return -ENODEV; return intel_guc_log_set_level(log, val); diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.c index 65a3e7fdb2b2..13b27b8ff74e 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.c @@ -79,29 +79,6 @@ static void slpc_mem_set_disabled(struct slpc_shared_data *data, slpc_mem_set_param(data, enable_id, 0); } -int intel_guc_slpc_init(struct intel_guc_slpc *slpc) -{ - struct intel_guc *guc = slpc_to_guc(slpc); - struct drm_i915_private *i915 = slpc_to_i915(slpc); - u32 size = PAGE_ALIGN(sizeof(struct slpc_shared_data)); - int err; - - GEM_BUG_ON(slpc->vma); - - err = intel_guc_allocate_and_map_vma(guc, size, &slpc->vma, (void **)&slpc->vaddr); - if (unlikely(err)) { - drm_err(&i915->drm, - "Failed to allocate SLPC struct (err=%pe)\n", - ERR_PTR(err)); - return err; - } - - slpc->max_freq_softlimit = 0; - slpc->min_freq_softlimit = 0; - - return err; -} - static u32 slpc_get_state(struct intel_guc_slpc *slpc) { struct slpc_shared_data *data; @@ -203,6 +180,86 @@ static int slpc_unset_param(struct intel_guc_slpc *slpc, return guc_action_slpc_unset_param(guc, id); } +static int slpc_force_min_freq(struct intel_guc_slpc *slpc, u32 freq) +{ + struct drm_i915_private *i915 = slpc_to_i915(slpc); + struct intel_guc *guc = slpc_to_guc(slpc); + intel_wakeref_t wakeref; + int ret = 0; + + lockdep_assert_held(&slpc->lock); + + if (!intel_guc_is_ready(guc)) + return -ENODEV; + + /* + * This function is a little different as compared to + * intel_guc_slpc_set_min_freq(). Softlimit will not be updated + * here since this is used to temporarily change min freq, + * for example, during a waitboost. Caller is responsible for + * checking bounds. + */ + + with_intel_runtime_pm(&i915->runtime_pm, wakeref) { + ret = slpc_set_param(slpc, + SLPC_PARAM_GLOBAL_MIN_GT_UNSLICE_FREQ_MHZ, + freq); + if (ret) + drm_err(&i915->drm, "Unable to force min freq to %u: %d", + freq, ret); + } + + return ret; +} + +static void slpc_boost_work(struct work_struct *work) +{ + struct intel_guc_slpc *slpc = container_of(work, typeof(*slpc), boost_work); + + /* + * Raise min freq to boost. It's possible that + * this is greater than current max. But it will + * certainly be limited by RP0. An error setting + * the min param is not fatal. + */ + mutex_lock(&slpc->lock); + if (atomic_read(&slpc->num_waiters)) { + slpc_force_min_freq(slpc, slpc->boost_freq); + slpc->num_boosts++; + } + mutex_unlock(&slpc->lock); +} + +int intel_guc_slpc_init(struct intel_guc_slpc *slpc) +{ + struct intel_guc *guc = slpc_to_guc(slpc); + struct drm_i915_private *i915 = slpc_to_i915(slpc); + u32 size = PAGE_ALIGN(sizeof(struct slpc_shared_data)); + int err; + + GEM_BUG_ON(slpc->vma); + + err = intel_guc_allocate_and_map_vma(guc, size, &slpc->vma, (void **)&slpc->vaddr); + if (unlikely(err)) { + drm_err(&i915->drm, + "Failed to allocate SLPC struct (err=%pe)\n", + ERR_PTR(err)); + return err; + } + + slpc->max_freq_softlimit = 0; + slpc->min_freq_softlimit = 0; + + slpc->boost_freq = 0; + atomic_set(&slpc->num_waiters, 0); + slpc->num_boosts = 0; + + mutex_init(&slpc->lock); + INIT_WORK(&slpc->boost_work, slpc_boost_work); + + return err; +} + static const char *slpc_global_state_to_string(enum slpc_global_state state) { switch (state) { @@ -393,7 +450,11 @@ int intel_guc_slpc_set_min_freq(struct intel_guc_slpc *slpc, u32 val) val > slpc->max_freq_softlimit) return -EINVAL; + /* Need a lock now since waitboost can be modifying min as well */ + mutex_lock(&slpc->lock); + with_intel_runtime_pm(&i915->runtime_pm, wakeref) { + ret = slpc_set_param(slpc, SLPC_PARAM_GLOBAL_MIN_GT_UNSLICE_FREQ_MHZ, val); @@ -406,6 +467,8 @@ int intel_guc_slpc_set_min_freq(struct intel_guc_slpc *slpc, u32 val) if (!ret) slpc->min_freq_softlimit = val; + mutex_unlock(&slpc->lock); + return ret; } @@ -522,6 +585,9 @@ static void slpc_get_rp_values(struct intel_guc_slpc *slpc) GT_FREQUENCY_MULTIPLIER; slpc->min_freq = REG_FIELD_GET(RPN_CAP_MASK, rp_state_cap) * GT_FREQUENCY_MULTIPLIER; + + if (!slpc->boost_freq) + slpc->boost_freq = slpc->rp0_freq; } /* @@ -557,7 +623,7 @@ int intel_guc_slpc_enable(struct intel_guc_slpc *slpc) if (unlikely(ret < 0)) return ret; - intel_guc_pm_intrmsk_enable(&i915->gt); + intel_guc_pm_intrmsk_enable(to_gt(i915)); slpc_get_rp_values(slpc); @@ -588,6 +654,47 @@ int intel_guc_slpc_enable(struct intel_guc_slpc *slpc) return 0; } +int intel_guc_slpc_set_boost_freq(struct intel_guc_slpc *slpc, u32 val) +{ + int ret = 0; + + if (val < slpc->min_freq || val > slpc->rp0_freq) + return -EINVAL; + + mutex_lock(&slpc->lock); + + if (slpc->boost_freq != val) { + /* Apply only if there are active waiters */ + if (atomic_read(&slpc->num_waiters)) { + ret = slpc_force_min_freq(slpc, val); + if (ret) { + ret = -EIO; + goto done; + } + } + + slpc->boost_freq = val; + } + +done: + mutex_unlock(&slpc->lock); + return ret; +} + +void intel_guc_slpc_dec_waiters(struct intel_guc_slpc *slpc) +{ + /* + * Return min back to the softlimit. + * This is called during request retire, + * so we don't need to fail that if the + * set_param fails. + */ + mutex_lock(&slpc->lock); + if (atomic_dec_and_test(&slpc->num_waiters)) + slpc_force_min_freq(slpc, slpc->min_freq_softlimit); + mutex_unlock(&slpc->lock); +} + int intel_guc_slpc_print_info(struct intel_guc_slpc *slpc, struct drm_printer *p) { struct drm_i915_private *i915 = slpc_to_i915(slpc); @@ -611,6 +718,8 @@ int intel_guc_slpc_print_info(struct intel_guc_slpc *slpc, struct drm_printer *p slpc_decode_max_freq(slpc)); drm_printf(p, "\tMin freq: %u MHz\n", slpc_decode_min_freq(slpc)); + drm_printf(p, "\twaitboosts: %u\n", + slpc->num_boosts); } } diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.h index e45054d5b9b4..0caa8fee3c04 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.h @@ -34,9 +34,12 @@ int intel_guc_slpc_enable(struct intel_guc_slpc *slpc); void intel_guc_slpc_fini(struct intel_guc_slpc *slpc); int intel_guc_slpc_set_max_freq(struct intel_guc_slpc *slpc, u32 val); int intel_guc_slpc_set_min_freq(struct intel_guc_slpc *slpc, u32 val); +int intel_guc_slpc_set_boost_freq(struct intel_guc_slpc *slpc, u32 val); int intel_guc_slpc_get_max_freq(struct intel_guc_slpc *slpc, u32 *val); int intel_guc_slpc_get_min_freq(struct intel_guc_slpc *slpc, u32 *val); int intel_guc_slpc_print_info(struct intel_guc_slpc *slpc, struct drm_printer *p); void intel_guc_pm_intrmsk_enable(struct intel_gt *gt); +void intel_guc_slpc_boost(struct intel_guc_slpc *slpc); +void intel_guc_slpc_dec_waiters(struct intel_guc_slpc *slpc); #endif diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc_types.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc_types.h index 41d13527666f..bf5b9a563c09 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc_types.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc_types.h @@ -6,6 +6,9 @@ #ifndef _INTEL_GUC_SLPC_TYPES_H_ #define _INTEL_GUC_SLPC_TYPES_H_ +#include <linux/atomic.h> +#include <linux/workqueue.h> +#include <linux/mutex.h> #include <linux/types.h> #define SLPC_RESET_TIMEOUT_MS 5 @@ -20,10 +23,20 @@ struct intel_guc_slpc { u32 min_freq; u32 rp0_freq; u32 rp1_freq; + u32 boost_freq; /* frequency softlimits */ u32 min_freq_softlimit; u32 max_freq_softlimit; + + /* Protects set/reset of boost freq + * and value of num_waiters + */ + struct mutex lock; + + struct work_struct boost_work; + atomic_t num_waiters; + u32 num_boosts; }; #endif 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 38b47e73e35d..e7517206af82 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -13,6 +13,7 @@ #include "gt/intel_engine_heartbeat.h" #include "gt/intel_gpu_commands.h" #include "gt/intel_gt.h" +#include "gt/intel_gt_clock_utils.h" #include "gt/intel_gt_irq.h" #include "gt/intel_gt_pm.h" #include "gt/intel_gt_requests.h" @@ -21,6 +22,7 @@ #include "gt/intel_mocs.h" #include "gt/intel_ring.h" +#include "intel_guc_ads.h" #include "intel_guc_submission.h" #include "i915_drv.h" @@ -143,7 +145,8 @@ guc_create_parallel(struct intel_engine_cs **engines, * use should be low and 1/16 should be sufficient. Minimum of 32 guc_ids for * multi-lrc. */ -#define NUMBER_MULTI_LRC_GUC_ID (GUC_MAX_LRC_DESCRIPTORS / 16) +#define NUMBER_MULTI_LRC_GUC_ID(guc) \ + ((guc)->submission_state.num_guc_ids / 16) /* * Below is a set of functions which control the GuC scheduling state which @@ -1038,8 +1041,6 @@ static void scrub_guc_desc_for_outstanding_g2h(struct intel_guc *guc) spin_unlock(&ce->guc_state.lock); - GEM_BUG_ON(!do_put && !destroyed); - if (pending_enable || destroyed || deregister) { decr_outstanding_submission_g2h(guc); if (deregister) @@ -1077,6 +1078,271 @@ static void scrub_guc_desc_for_outstanding_g2h(struct intel_guc *guc) xa_unlock_irqrestore(&guc->context_lookup, flags); } +/* + * GuC stores busyness stats for each engine at context in/out boundaries. A + * context 'in' logs execution start time, 'out' adds in -> out delta to total. + * i915/kmd accesses 'start', 'total' and 'context id' from memory shared with + * GuC. + * + * __i915_pmu_event_read samples engine busyness. When sampling, if context id + * is valid (!= ~0) and start is non-zero, the engine is considered to be + * active. For an active engine total busyness = total + (now - start), where + * 'now' is the time at which the busyness is sampled. For inactive engine, + * total busyness = total. + * + * All times are captured from GUCPMTIMESTAMP reg and are in gt clock domain. + * + * The start and total values provided by GuC are 32 bits and wrap around in a + * few minutes. Since perf pmu provides busyness as 64 bit monotonically + * increasing ns values, there is a need for this implementation to account for + * overflows and extend the GuC provided values to 64 bits before returning + * busyness to the user. In order to do that, a worker runs periodically at + * frequency = 1/8th the time it takes for the timestamp to wrap (i.e. once in + * 27 seconds for a gt clock frequency of 19.2 MHz). + */ + +#define WRAP_TIME_CLKS U32_MAX +#define POLL_TIME_CLKS (WRAP_TIME_CLKS >> 3) + +static void +__extend_last_switch(struct intel_guc *guc, u64 *prev_start, u32 new_start) +{ + u32 gt_stamp_hi = upper_32_bits(guc->timestamp.gt_stamp); + u32 gt_stamp_last = lower_32_bits(guc->timestamp.gt_stamp); + + if (new_start == lower_32_bits(*prev_start)) + return; + + if (new_start < gt_stamp_last && + (new_start - gt_stamp_last) <= POLL_TIME_CLKS) + gt_stamp_hi++; + + if (new_start > gt_stamp_last && + (gt_stamp_last - new_start) <= POLL_TIME_CLKS && gt_stamp_hi) + gt_stamp_hi--; + + *prev_start = ((u64)gt_stamp_hi << 32) | new_start; +} + +static void guc_update_engine_gt_clks(struct intel_engine_cs *engine) +{ + struct guc_engine_usage_record *rec = intel_guc_engine_usage(engine); + struct intel_engine_guc_stats *stats = &engine->stats.guc; + struct intel_guc *guc = &engine->gt->uc.guc; + u32 last_switch = rec->last_switch_in_stamp; + u32 ctx_id = rec->current_context_index; + u32 total = rec->total_runtime; + + lockdep_assert_held(&guc->timestamp.lock); + + stats->running = ctx_id != ~0U && last_switch; + if (stats->running) + __extend_last_switch(guc, &stats->start_gt_clk, last_switch); + + /* + * Instead of adjusting the total for overflow, just add the + * difference from previous sample stats->total_gt_clks + */ + if (total && total != ~0U) { + stats->total_gt_clks += (u32)(total - stats->prev_total); + stats->prev_total = total; + } +} + +static void guc_update_pm_timestamp(struct intel_guc *guc, + struct intel_engine_cs *engine, + ktime_t *now) +{ + u32 gt_stamp_now, gt_stamp_hi; + + lockdep_assert_held(&guc->timestamp.lock); + + gt_stamp_hi = upper_32_bits(guc->timestamp.gt_stamp); + gt_stamp_now = intel_uncore_read(engine->uncore, + RING_TIMESTAMP(engine->mmio_base)); + *now = ktime_get(); + + if (gt_stamp_now < lower_32_bits(guc->timestamp.gt_stamp)) + gt_stamp_hi++; + + guc->timestamp.gt_stamp = ((u64)gt_stamp_hi << 32) | gt_stamp_now; +} + +/* + * Unlike the execlist mode of submission total and active times are in terms of + * gt clocks. The *now parameter is retained to return the cpu time at which the + * busyness was sampled. + */ +static ktime_t guc_engine_busyness(struct intel_engine_cs *engine, ktime_t *now) +{ + struct intel_engine_guc_stats stats_saved, *stats = &engine->stats.guc; + struct i915_gpu_error *gpu_error = &engine->i915->gpu_error; + struct intel_gt *gt = engine->gt; + struct intel_guc *guc = >->uc.guc; + u64 total, gt_stamp_saved; + unsigned long flags; + u32 reset_count; + bool in_reset; + + spin_lock_irqsave(&guc->timestamp.lock, flags); + + /* + * If a reset happened, we risk reading partially updated engine + * busyness from GuC, so we just use the driver stored copy of busyness. + * Synchronize with gt reset using reset_count and the + * I915_RESET_BACKOFF flag. Note that reset flow updates the reset_count + * after I915_RESET_BACKOFF flag, so ensure that the reset_count is + * usable by checking the flag afterwards. + */ + reset_count = i915_reset_count(gpu_error); + in_reset = test_bit(I915_RESET_BACKOFF, >->reset.flags); + + *now = ktime_get(); + + /* + * The active busyness depends on start_gt_clk and gt_stamp. + * gt_stamp is updated by i915 only when gt is awake and the + * start_gt_clk is derived from GuC state. To get a consistent + * view of activity, we query the GuC state only if gt is awake. + */ + if (!in_reset && intel_gt_pm_get_if_awake(gt)) { + stats_saved = *stats; + gt_stamp_saved = guc->timestamp.gt_stamp; + guc_update_engine_gt_clks(engine); + guc_update_pm_timestamp(guc, engine, now); + intel_gt_pm_put_async(gt); + if (i915_reset_count(gpu_error) != reset_count) { + *stats = stats_saved; + guc->timestamp.gt_stamp = gt_stamp_saved; + } + } + + total = intel_gt_clock_interval_to_ns(gt, stats->total_gt_clks); + if (stats->running) { + u64 clk = guc->timestamp.gt_stamp - stats->start_gt_clk; + + total += intel_gt_clock_interval_to_ns(gt, clk); + } + + spin_unlock_irqrestore(&guc->timestamp.lock, flags); + + return ns_to_ktime(total); +} + +static void __reset_guc_busyness_stats(struct intel_guc *guc) +{ + struct intel_gt *gt = guc_to_gt(guc); + struct intel_engine_cs *engine; + enum intel_engine_id id; + unsigned long flags; + ktime_t unused; + + cancel_delayed_work_sync(&guc->timestamp.work); + + spin_lock_irqsave(&guc->timestamp.lock, flags); + + for_each_engine(engine, gt, id) { + guc_update_pm_timestamp(guc, engine, &unused); + guc_update_engine_gt_clks(engine); + engine->stats.guc.prev_total = 0; + } + + spin_unlock_irqrestore(&guc->timestamp.lock, flags); +} + +static void __update_guc_busyness_stats(struct intel_guc *guc) +{ + struct intel_gt *gt = guc_to_gt(guc); + struct intel_engine_cs *engine; + enum intel_engine_id id; + unsigned long flags; + ktime_t unused; + + spin_lock_irqsave(&guc->timestamp.lock, flags); + for_each_engine(engine, gt, id) { + guc_update_pm_timestamp(guc, engine, &unused); + guc_update_engine_gt_clks(engine); + } + spin_unlock_irqrestore(&guc->timestamp.lock, flags); +} + +static void guc_timestamp_ping(struct work_struct *wrk) +{ + struct intel_guc *guc = container_of(wrk, typeof(*guc), + timestamp.work.work); + struct intel_uc *uc = container_of(guc, typeof(*uc), guc); + struct intel_gt *gt = guc_to_gt(guc); + intel_wakeref_t wakeref; + int srcu, ret; + + /* + * Synchronize with gt reset to make sure the worker does not + * corrupt the engine/guc stats. + */ + ret = intel_gt_reset_trylock(gt, &srcu); + if (ret) + return; + + with_intel_runtime_pm(>->i915->runtime_pm, wakeref) + __update_guc_busyness_stats(guc); + + intel_gt_reset_unlock(gt, srcu); + + mod_delayed_work(system_highpri_wq, &guc->timestamp.work, + guc->timestamp.ping_delay); +} + +static int guc_action_enable_usage_stats(struct intel_guc *guc) +{ + u32 offset = intel_guc_engine_usage_offset(guc); + u32 action[] = { + INTEL_GUC_ACTION_SET_ENG_UTIL_BUFF, + offset, + 0, + }; + + return intel_guc_send(guc, action, ARRAY_SIZE(action)); +} + +static void guc_init_engine_stats(struct intel_guc *guc) +{ + struct intel_gt *gt = guc_to_gt(guc); + intel_wakeref_t wakeref; + + mod_delayed_work(system_highpri_wq, &guc->timestamp.work, + guc->timestamp.ping_delay); + + with_intel_runtime_pm(>->i915->runtime_pm, wakeref) { + int ret = guc_action_enable_usage_stats(guc); + + if (ret) + drm_err(>->i915->drm, + "Failed to enable usage stats: %d!\n", ret); + } +} + +void intel_guc_busyness_park(struct intel_gt *gt) +{ + struct intel_guc *guc = >->uc.guc; + + if (!guc_submission_initialized(guc)) + return; + + cancel_delayed_work(&guc->timestamp.work); + __update_guc_busyness_stats(guc); +} + +void intel_guc_busyness_unpark(struct intel_gt *gt) +{ + struct intel_guc *guc = >->uc.guc; + + if (!guc_submission_initialized(guc)) + return; + + mod_delayed_work(system_highpri_wq, &guc->timestamp.work, + guc->timestamp.ping_delay); +} + static inline bool submission_disabled(struct intel_guc *guc) { @@ -1138,6 +1404,7 @@ void intel_guc_submission_reset_prepare(struct intel_guc *guc) intel_gt_park_heartbeats(guc_to_gt(guc)); disable_submission(guc); guc->interrupts.disable(guc); + __reset_guc_busyness_stats(guc); /* Flush IRQ handler */ spin_lock_irq(&guc_to_gt(guc)->irq_lock); @@ -1484,6 +1751,7 @@ static void destroyed_worker_func(struct work_struct *w); */ int intel_guc_submission_init(struct intel_guc *guc) { + struct intel_gt *gt = guc_to_gt(guc); int ret; if (guc->lrc_desc_pool) @@ -1508,10 +1776,14 @@ int intel_guc_submission_init(struct intel_guc *guc) destroyed_worker_func); guc->submission_state.guc_ids_bitmap = - bitmap_zalloc(NUMBER_MULTI_LRC_GUC_ID, GFP_KERNEL); + bitmap_zalloc(NUMBER_MULTI_LRC_GUC_ID(guc), GFP_KERNEL); if (!guc->submission_state.guc_ids_bitmap) return -ENOMEM; + spin_lock_init(&guc->timestamp.lock); + INIT_DELAYED_WORK(&guc->timestamp.work, guc_timestamp_ping); + guc->timestamp.ping_delay = (POLL_TIME_CLKS / gt->clock_frequency + 1) * HZ; + return 0; } @@ -1598,13 +1870,13 @@ static int new_guc_id(struct intel_guc *guc, struct intel_context *ce) if (intel_context_is_parent(ce)) ret = bitmap_find_free_region(guc->submission_state.guc_ids_bitmap, - NUMBER_MULTI_LRC_GUC_ID, + NUMBER_MULTI_LRC_GUC_ID(guc), order_base_2(ce->parallel.number_children + 1)); else ret = ida_simple_get(&guc->submission_state.guc_ids, - NUMBER_MULTI_LRC_GUC_ID, - GUC_MAX_LRC_DESCRIPTORS, + NUMBER_MULTI_LRC_GUC_ID(guc), + guc->submission_state.num_guc_ids, GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN); if (unlikely(ret < 0)) @@ -1662,14 +1934,18 @@ static int steal_guc_id(struct intel_guc *guc, struct intel_context *ce) GEM_BUG_ON(intel_context_is_parent(cn)); list_del_init(&cn->guc_id.link); - ce->guc_id = cn->guc_id; + ce->guc_id.id = cn->guc_id.id; - spin_lock(&ce->guc_state.lock); + spin_lock(&cn->guc_state.lock); clr_context_registered(cn); - spin_unlock(&ce->guc_state.lock); + spin_unlock(&cn->guc_state.lock); set_context_guc_id_invalid(cn); +#ifdef CONFIG_DRM_I915_SELFTEST + guc->number_guc_id_stolen++; +#endif + return 0; } else { return -EAGAIN; @@ -2373,7 +2649,6 @@ static inline void guc_lrc_desc_unpin(struct intel_context *ce) unsigned long flags; bool disabled; - lockdep_assert_held(&guc->submission_state.lock); GEM_BUG_ON(!intel_gt_pm_is_awake(gt)); GEM_BUG_ON(!lrc_desc_registered(guc, ce->guc_id.id)); GEM_BUG_ON(ce != __get_context(guc, ce->guc_id.id)); @@ -2389,7 +2664,7 @@ static inline void guc_lrc_desc_unpin(struct intel_context *ce) } spin_unlock_irqrestore(&ce->guc_state.lock, flags); if (unlikely(disabled)) { - __release_guc_id(guc, ce); + release_guc_id(guc, ce); __guc_context_destroy(ce); return; } @@ -2423,36 +2698,48 @@ static void __guc_context_destroy(struct intel_context *ce) static void guc_flush_destroyed_contexts(struct intel_guc *guc) { - struct intel_context *ce, *cn; + struct intel_context *ce; unsigned long flags; GEM_BUG_ON(!submission_disabled(guc) && guc_submission_initialized(guc)); - spin_lock_irqsave(&guc->submission_state.lock, flags); - list_for_each_entry_safe(ce, cn, - &guc->submission_state.destroyed_contexts, - destroyed_link) { - list_del_init(&ce->destroyed_link); - __release_guc_id(guc, ce); + while (!list_empty(&guc->submission_state.destroyed_contexts)) { + spin_lock_irqsave(&guc->submission_state.lock, flags); + ce = list_first_entry_or_null(&guc->submission_state.destroyed_contexts, + struct intel_context, + destroyed_link); + if (ce) + list_del_init(&ce->destroyed_link); + spin_unlock_irqrestore(&guc->submission_state.lock, flags); + + if (!ce) + break; + + release_guc_id(guc, ce); __guc_context_destroy(ce); } - spin_unlock_irqrestore(&guc->submission_state.lock, flags); } static void deregister_destroyed_contexts(struct intel_guc *guc) { - struct intel_context *ce, *cn; + struct intel_context *ce; unsigned long flags; - spin_lock_irqsave(&guc->submission_state.lock, flags); - list_for_each_entry_safe(ce, cn, - &guc->submission_state.destroyed_contexts, - destroyed_link) { - list_del_init(&ce->destroyed_link); + while (!list_empty(&guc->submission_state.destroyed_contexts)) { + spin_lock_irqsave(&guc->submission_state.lock, flags); + ce = list_first_entry_or_null(&guc->submission_state.destroyed_contexts, + struct intel_context, + destroyed_link); + if (ce) + list_del_init(&ce->destroyed_link); + spin_unlock_irqrestore(&guc->submission_state.lock, flags); + + if (!ce) + break; + guc_lrc_desc_unpin(ce); } - spin_unlock_irqrestore(&guc->submission_state.lock, flags); } static void destroyed_worker_func(struct work_struct *w) @@ -3080,8 +3367,8 @@ guc_create_parallel(struct intel_engine_cs **engines, ce = intel_engine_create_virtual(siblings, num_siblings, FORCE_VIRTUAL); - if (!ce) { - err = ERR_PTR(-ENOMEM); + if (IS_ERR(ce)) { + err = ERR_CAST(ce); goto unwind; } @@ -3369,7 +3656,9 @@ static void guc_default_vfuncs(struct intel_engine_cs *engine) engine->emit_flush = gen12_emit_flush_xcs; } engine->set_default_submission = guc_set_default_submission; + engine->busyness = guc_engine_busyness; + engine->flags |= I915_ENGINE_SUPPORTS_STATS; engine->flags |= I915_ENGINE_HAS_PREEMPTION; engine->flags |= I915_ENGINE_HAS_TIMESLICES; @@ -3468,6 +3757,7 @@ int intel_guc_submission_setup(struct intel_engine_cs *engine) void intel_guc_submission_enable(struct intel_guc *guc) { guc_init_lrc_mapping(guc); + guc_init_engine_stats(guc); } void intel_guc_submission_disable(struct intel_guc *guc) @@ -3494,6 +3784,7 @@ static bool __guc_submission_selected(struct intel_guc *guc) void intel_guc_submission_init_early(struct intel_guc *guc) { + guc->submission_state.num_guc_ids = GUC_MAX_LRC_DESCRIPTORS; guc->submission_supported = __guc_submission_supported(guc); guc->submission_selected = __guc_submission_selected(guc); } @@ -3695,6 +3986,7 @@ int intel_guc_context_reset_process_msg(struct intel_guc *guc, const u32 *msg, u32 len) { struct intel_context *ce; + unsigned long flags; int desc_idx; if (unlikely(len != 1)) { @@ -3703,11 +3995,24 @@ int intel_guc_context_reset_process_msg(struct intel_guc *guc, } desc_idx = msg[0]; + + /* + * The context lookup uses the xarray but lookups only require an RCU lock + * not the full spinlock. So take the lock explicitly and keep it until the + * context has been reference count locked to ensure it can't be destroyed + * asynchronously until the reset is done. + */ + xa_lock_irqsave(&guc->context_lookup, flags); ce = g2h_context_lookup(guc, desc_idx); + if (ce) + intel_context_get(ce); + xa_unlock_irqrestore(&guc->context_lookup, flags); + if (unlikely(!ce)) return -EPROTO; guc_handle_context_reset(guc, ce); + intel_context_put(ce); return 0; } @@ -3728,11 +4033,12 @@ int intel_guc_engine_failure_process_msg(struct intel_guc *guc, const u32 *msg, u32 len) { struct intel_engine_cs *engine; + struct intel_gt *gt = guc_to_gt(guc); u8 guc_class, instance; u32 reason; if (unlikely(len != 3)) { - drm_err(&guc_to_gt(guc)->i915->drm, "Invalid length %u", len); + drm_err(>->i915->drm, "Invalid length %u", len); return -EPROTO; } @@ -3742,12 +4048,19 @@ int intel_guc_engine_failure_process_msg(struct intel_guc *guc, engine = guc_lookup_engine(guc, guc_class, instance); if (unlikely(!engine)) { - drm_err(&guc_to_gt(guc)->i915->drm, + drm_err(>->i915->drm, "Invalid engine %d:%d", guc_class, instance); return -EPROTO; } - intel_gt_handle_error(guc_to_gt(guc), engine->mask, + /* + * This is an unexpected failure of a hardware feature. So, log a real + * error message not just the informational that comes with the reset. + */ + drm_err(>->i915->drm, "GuC engine reset request failed on %d:%d (%s) because 0x%08X", + guc_class, instance, engine->name, reason); + + intel_gt_handle_error(gt, engine->mask, I915_ERROR_CAPTURE, "GuC failed to reset %s (reason=0x%08x)\n", engine->name, reason); diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h index c7ef44fa0c36..5a95a9f0a8e3 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h @@ -28,6 +28,8 @@ void intel_guc_submission_print_context_info(struct intel_guc *guc, void intel_guc_dump_active_requests(struct intel_engine_cs *engine, struct i915_request *hung_rq, struct drm_printer *m); +void intel_guc_busyness_park(struct intel_gt *gt); +void intel_guc_busyness_unpark(struct intel_gt *gt); bool intel_guc_virtual_engine_has_heartbeat(const struct intel_engine_cs *ve); diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc.c b/drivers/gpu/drm/i915/gt/uc/intel_huc.c index ff4b6869b80b..d10b227ac4aa 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_huc.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_huc.c @@ -54,65 +54,6 @@ void intel_huc_init_early(struct intel_huc *huc) } } -static int intel_huc_rsa_data_create(struct intel_huc *huc) -{ - struct intel_gt *gt = huc_to_gt(huc); - struct intel_guc *guc = >->uc.guc; - struct i915_vma *vma; - size_t copied; - void *vaddr; - int err; - - err = i915_inject_probe_error(gt->i915, -ENXIO); - if (err) - return err; - - /* - * HuC firmware will sit above GUC_GGTT_TOP and will not map - * through GTT. Unfortunately, this means GuC cannot perform - * the HuC auth. as the rsa offset now falls within the GuC - * inaccessible range. We resort to perma-pinning an additional - * vma within the accessible range that only contains the rsa - * signature. The GuC can use this extra pinning to perform - * the authentication since its GGTT offset will be GuC - * accessible. - */ - GEM_BUG_ON(huc->fw.rsa_size > PAGE_SIZE); - vma = intel_guc_allocate_vma(guc, PAGE_SIZE); - if (IS_ERR(vma)) - return PTR_ERR(vma); - - vaddr = i915_gem_object_pin_map_unlocked(vma->obj, - i915_coherent_map_type(gt->i915, - vma->obj, true)); - if (IS_ERR(vaddr)) { - i915_vma_unpin_and_release(&vma, 0); - err = PTR_ERR(vaddr); - goto unpin_out; - } - - copied = intel_uc_fw_copy_rsa(&huc->fw, vaddr, vma->size); - i915_gem_object_unpin_map(vma->obj); - - if (copied < huc->fw.rsa_size) { - err = -ENOMEM; - goto unpin_out; - } - - huc->rsa_data = vma; - - return 0; - -unpin_out: - i915_vma_unpin_and_release(&vma, 0); - return err; -} - -static void intel_huc_rsa_data_destroy(struct intel_huc *huc) -{ - i915_vma_unpin_and_release(&huc->rsa_data, 0); -} - int intel_huc_init(struct intel_huc *huc) { struct drm_i915_private *i915 = huc_to_gt(huc)->i915; @@ -122,21 +63,10 @@ int intel_huc_init(struct intel_huc *huc) if (err) goto out; - /* - * HuC firmware image is outside GuC accessible range. - * Copy the RSA signature out of the image into - * a perma-pinned region set aside for it - */ - err = intel_huc_rsa_data_create(huc); - if (err) - goto out_fini; - intel_uc_fw_change_status(&huc->fw, INTEL_UC_FIRMWARE_LOADABLE); return 0; -out_fini: - intel_uc_fw_fini(&huc->fw); out: i915_probe_error(i915, "failed with %d\n", err); return err; @@ -147,7 +77,6 @@ void intel_huc_fini(struct intel_huc *huc) if (!intel_uc_fw_is_loadable(&huc->fw)) return; - intel_huc_rsa_data_destroy(huc); intel_uc_fw_fini(&huc->fw); } @@ -177,7 +106,7 @@ int intel_huc_auth(struct intel_huc *huc) goto fail; ret = intel_guc_auth_huc(guc, - intel_guc_ggtt_offset(guc, huc->rsa_data)); + intel_guc_ggtt_offset(guc, huc->fw.rsa_data)); if (ret) { DRM_ERROR("HuC: GuC did not ack Auth request %d\n", ret); goto fail; @@ -199,7 +128,7 @@ int intel_huc_auth(struct intel_huc *huc) fail: i915_probe_error(gt->i915, "HuC: Authentication failed %d\n", ret); - intel_uc_fw_change_status(&huc->fw, INTEL_UC_FIRMWARE_FAIL); + intel_uc_fw_change_status(&huc->fw, INTEL_UC_FIRMWARE_LOAD_FAIL); return ret; } diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc.h b/drivers/gpu/drm/i915/gt/uc/intel_huc.h index daee43b661d4..ae8c8a6c8cc8 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_huc.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_huc.h @@ -15,8 +15,6 @@ struct intel_huc { struct intel_uc_fw fw; /* HuC-specific additions */ - struct i915_vma *rsa_data; - struct { i915_reg_t reg; u32 mask; diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_uc.c index 2fef3b0bbe95..09ed29df67bc 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_uc.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.c @@ -8,6 +8,7 @@ #include "intel_guc.h" #include "intel_guc_ads.h" #include "intel_guc_submission.h" +#include "gt/intel_rps.h" #include "intel_uc.h" #include "i915_drv.h" @@ -35,7 +36,7 @@ static void uc_expand_default_options(struct intel_uc *uc) } /* Intermediate platforms are HuC authentication only */ - if (IS_ALDERLAKE_S(i915)) { + if (IS_ALDERLAKE_S(i915) && !IS_ADLS_RPLS(i915)) { i915->params.enable_guc = ENABLE_GUC_LOAD_HUC; return; } @@ -462,6 +463,8 @@ static int __uc_init_hw(struct intel_uc *uc) else attempts = 1; + intel_rps_raise_unslice(&uc_to_gt(uc)->rps); + while (attempts--) { /* * Always reset the GuC just before (re)loading, so @@ -499,6 +502,9 @@ static int __uc_init_hw(struct intel_uc *uc) ret = intel_guc_slpc_enable(&guc->slpc); if (ret) goto err_submission; + } else { + /* Restore GT back to RPn for non-SLPC path */ + intel_rps_lower_unslice(&uc_to_gt(uc)->rps); } drm_info(&i915->drm, "%s firmware %s version %u.%u %s:%s\n", @@ -529,6 +535,9 @@ err_submission: err_log_capture: __uc_capture_load_err_log(uc); err_out: + /* Return GT back to RPn */ + intel_rps_lower_unslice(&uc_to_gt(uc)->rps); + __uc_sanitize(uc); if (!ret) { diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c index 3aa87be4f2e4..a5af05bde6f2 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c @@ -48,22 +48,39 @@ void intel_uc_fw_change_status(struct intel_uc_fw *uc_fw, * Note that RKL and ADL-S have the same GuC/HuC device ID's and use the same * firmware as TGL. */ -#define INTEL_UC_FIRMWARE_DEFS(fw_def, guc_def, huc_def) \ - fw_def(ALDERLAKE_P, 0, guc_def(adlp, 62, 0, 3), huc_def(tgl, 7, 9, 3)) \ - fw_def(ALDERLAKE_S, 0, guc_def(tgl, 62, 0, 0), huc_def(tgl, 7, 9, 3)) \ - fw_def(DG1, 0, guc_def(dg1, 62, 0, 0), huc_def(dg1, 7, 9, 3)) \ - fw_def(ROCKETLAKE, 0, guc_def(tgl, 62, 0, 0), huc_def(tgl, 7, 9, 3)) \ - fw_def(TIGERLAKE, 0, guc_def(tgl, 62, 0, 0), huc_def(tgl, 7, 9, 3)) \ - fw_def(JASPERLAKE, 0, guc_def(ehl, 62, 0, 0), huc_def(ehl, 9, 0, 0)) \ - fw_def(ELKHARTLAKE, 0, guc_def(ehl, 62, 0, 0), huc_def(ehl, 9, 0, 0)) \ - fw_def(ICELAKE, 0, guc_def(icl, 62, 0, 0), huc_def(icl, 9, 0, 0)) \ - fw_def(COMETLAKE, 5, guc_def(cml, 62, 0, 0), huc_def(cml, 4, 0, 0)) \ - fw_def(COMETLAKE, 0, guc_def(kbl, 62, 0, 0), huc_def(kbl, 4, 0, 0)) \ - fw_def(COFFEELAKE, 0, guc_def(kbl, 62, 0, 0), huc_def(kbl, 4, 0, 0)) \ - fw_def(GEMINILAKE, 0, guc_def(glk, 62, 0, 0), huc_def(glk, 4, 0, 0)) \ - fw_def(KABYLAKE, 0, guc_def(kbl, 62, 0, 0), huc_def(kbl, 4, 0, 0)) \ - fw_def(BROXTON, 0, guc_def(bxt, 62, 0, 0), huc_def(bxt, 2, 0, 0)) \ - fw_def(SKYLAKE, 0, guc_def(skl, 62, 0, 0), huc_def(skl, 2, 0, 0)) +#define INTEL_GUC_FIRMWARE_DEFS(fw_def, guc_def) \ + fw_def(ALDERLAKE_P, 0, guc_def(adlp, 62, 0, 3)) \ + fw_def(ALDERLAKE_S, 0, guc_def(tgl, 62, 0, 0)) \ + fw_def(DG1, 0, guc_def(dg1, 62, 0, 0)) \ + fw_def(ROCKETLAKE, 0, guc_def(tgl, 62, 0, 0)) \ + fw_def(TIGERLAKE, 0, guc_def(tgl, 62, 0, 0)) \ + fw_def(JASPERLAKE, 0, guc_def(ehl, 62, 0, 0)) \ + fw_def(ELKHARTLAKE, 0, guc_def(ehl, 62, 0, 0)) \ + fw_def(ICELAKE, 0, guc_def(icl, 62, 0, 0)) \ + fw_def(COMETLAKE, 5, guc_def(cml, 62, 0, 0)) \ + fw_def(COMETLAKE, 0, guc_def(kbl, 62, 0, 0)) \ + fw_def(COFFEELAKE, 0, guc_def(kbl, 62, 0, 0)) \ + fw_def(GEMINILAKE, 0, guc_def(glk, 62, 0, 0)) \ + fw_def(KABYLAKE, 0, guc_def(kbl, 62, 0, 0)) \ + fw_def(BROXTON, 0, guc_def(bxt, 62, 0, 0)) \ + fw_def(SKYLAKE, 0, guc_def(skl, 62, 0, 0)) + +#define INTEL_HUC_FIRMWARE_DEFS(fw_def, huc_def) \ + fw_def(ALDERLAKE_P, 0, huc_def(tgl, 7, 9, 3)) \ + fw_def(ALDERLAKE_S, 0, huc_def(tgl, 7, 9, 3)) \ + fw_def(DG1, 0, huc_def(dg1, 7, 9, 3)) \ + fw_def(ROCKETLAKE, 0, huc_def(tgl, 7, 9, 3)) \ + fw_def(TIGERLAKE, 0, huc_def(tgl, 7, 9, 3)) \ + fw_def(JASPERLAKE, 0, huc_def(ehl, 9, 0, 0)) \ + fw_def(ELKHARTLAKE, 0, huc_def(ehl, 9, 0, 0)) \ + fw_def(ICELAKE, 0, huc_def(icl, 9, 0, 0)) \ + fw_def(COMETLAKE, 5, huc_def(cml, 4, 0, 0)) \ + fw_def(COMETLAKE, 0, huc_def(kbl, 4, 0, 0)) \ + fw_def(COFFEELAKE, 0, huc_def(kbl, 4, 0, 0)) \ + fw_def(GEMINILAKE, 0, huc_def(glk, 4, 0, 0)) \ + fw_def(KABYLAKE, 0, huc_def(kbl, 4, 0, 0)) \ + fw_def(BROXTON, 0, huc_def(bxt, 2, 0, 0)) \ + fw_def(SKYLAKE, 0, huc_def(skl, 2, 0, 0)) #define __MAKE_UC_FW_PATH(prefix_, name_, major_, minor_, patch_) \ "i915/" \ @@ -79,11 +96,11 @@ void intel_uc_fw_change_status(struct intel_uc_fw *uc_fw, __MAKE_UC_FW_PATH(prefix_, "_huc_", major_, minor_, bld_num_) /* All blobs need to be declared via MODULE_FIRMWARE() */ -#define INTEL_UC_MODULE_FW(platform_, revid_, guc_, huc_) \ - MODULE_FIRMWARE(guc_); \ - MODULE_FIRMWARE(huc_); +#define INTEL_UC_MODULE_FW(platform_, revid_, uc_) \ + MODULE_FIRMWARE(uc_); -INTEL_UC_FIRMWARE_DEFS(INTEL_UC_MODULE_FW, MAKE_GUC_FW_PATH, MAKE_HUC_FW_PATH) +INTEL_GUC_FIRMWARE_DEFS(INTEL_UC_MODULE_FW, MAKE_GUC_FW_PATH) +INTEL_HUC_FIRMWARE_DEFS(INTEL_UC_MODULE_FW, MAKE_HUC_FW_PATH) /* The below structs and macros are used to iterate across the list of blobs */ struct __packed uc_fw_blob { @@ -106,31 +123,47 @@ struct __packed uc_fw_blob { struct __packed uc_fw_platform_requirement { enum intel_platform p; u8 rev; /* first platform rev using this FW */ - const struct uc_fw_blob blobs[INTEL_UC_FW_NUM_TYPES]; + const struct uc_fw_blob blob; }; -#define MAKE_FW_LIST(platform_, revid_, guc_, huc_) \ +#define MAKE_FW_LIST(platform_, revid_, uc_) \ { \ .p = INTEL_##platform_, \ .rev = revid_, \ - .blobs[INTEL_UC_FW_TYPE_GUC] = guc_, \ - .blobs[INTEL_UC_FW_TYPE_HUC] = huc_, \ + .blob = uc_, \ }, +struct fw_blobs_by_type { + const struct uc_fw_platform_requirement *blobs; + u32 count; +}; + static void __uc_fw_auto_select(struct drm_i915_private *i915, struct intel_uc_fw *uc_fw) { - static const struct uc_fw_platform_requirement fw_blobs[] = { - INTEL_UC_FIRMWARE_DEFS(MAKE_FW_LIST, GUC_FW_BLOB, HUC_FW_BLOB) + static const struct uc_fw_platform_requirement blobs_guc[] = { + INTEL_GUC_FIRMWARE_DEFS(MAKE_FW_LIST, GUC_FW_BLOB) + }; + static const struct uc_fw_platform_requirement blobs_huc[] = { + INTEL_HUC_FIRMWARE_DEFS(MAKE_FW_LIST, HUC_FW_BLOB) }; + static const struct fw_blobs_by_type blobs_all[INTEL_UC_FW_NUM_TYPES] = { + [INTEL_UC_FW_TYPE_GUC] = { blobs_guc, ARRAY_SIZE(blobs_guc) }, + [INTEL_UC_FW_TYPE_HUC] = { blobs_huc, ARRAY_SIZE(blobs_huc) }, + }; + static const struct uc_fw_platform_requirement *fw_blobs; enum intel_platform p = INTEL_INFO(i915)->platform; + u32 fw_count; u8 rev = INTEL_REVID(i915); int i; - for (i = 0; i < ARRAY_SIZE(fw_blobs) && p <= fw_blobs[i].p; i++) { + GEM_BUG_ON(uc_fw->type >= ARRAY_SIZE(blobs_all)); + fw_blobs = blobs_all[uc_fw->type].blobs; + fw_count = blobs_all[uc_fw->type].count; + + for (i = 0; i < fw_count && p <= fw_blobs[i].p; i++) { if (p == fw_blobs[i].p && rev >= fw_blobs[i].rev) { - const struct uc_fw_blob *blob = - &fw_blobs[i].blobs[uc_fw->type]; + const struct uc_fw_blob *blob = &fw_blobs[i].blob; uc_fw->path = blob->path; uc_fw->major_ver_wanted = blob->major; uc_fw->minor_ver_wanted = blob->minor; @@ -140,7 +173,7 @@ __uc_fw_auto_select(struct drm_i915_private *i915, struct intel_uc_fw *uc_fw) /* make sure the list is ordered as expected */ if (IS_ENABLED(CONFIG_DRM_I915_SELFTEST)) { - for (i = 1; i < ARRAY_SIZE(fw_blobs); i++) { + for (i = 1; i < fw_count; i++) { if (fw_blobs[i].p < fw_blobs[i - 1].p) continue; @@ -322,13 +355,6 @@ int intel_uc_fw_fetch(struct intel_uc_fw *uc_fw) uc_fw->ucode_size = (css->size_dw - css->header_size_dw) * sizeof(u32); /* now RSA */ - if (unlikely(css->key_size_dw != UOS_RSA_SCRATCH_COUNT)) { - drm_warn(&i915->drm, "%s firmware %s: unexpected key size: %u != %u\n", - intel_uc_fw_type_repr(uc_fw->type), uc_fw->path, - css->key_size_dw, UOS_RSA_SCRATCH_COUNT); - err = -EPROTO; - goto fail; - } uc_fw->rsa_size = css->key_size_dw * sizeof(u32); /* At least, it should have header, uCode and RSA. Size of all three. */ @@ -540,10 +566,79 @@ fail: i915_probe_error(gt->i915, "Failed to load %s firmware %s (%d)\n", intel_uc_fw_type_repr(uc_fw->type), uc_fw->path, err); - intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_FAIL); + intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_LOAD_FAIL); return err; } +static inline bool uc_fw_need_rsa_in_memory(struct intel_uc_fw *uc_fw) +{ + /* + * The HW reads the GuC RSA from memory if the key size is > 256 bytes, + * while it reads it from the 64 RSA registers if it is smaller. + * The HuC RSA is always read from memory. + */ + return uc_fw->type == INTEL_UC_FW_TYPE_HUC || uc_fw->rsa_size > 256; +} + +static int uc_fw_rsa_data_create(struct intel_uc_fw *uc_fw) +{ + struct intel_gt *gt = __uc_fw_to_gt(uc_fw); + struct i915_vma *vma; + size_t copied; + void *vaddr; + int err; + + err = i915_inject_probe_error(gt->i915, -ENXIO); + if (err) + return err; + + if (!uc_fw_need_rsa_in_memory(uc_fw)) + return 0; + + /* + * uC firmwares will sit above GUC_GGTT_TOP and will not map through + * GGTT. Unfortunately, this means that the GuC HW cannot perform the uC + * authentication from memory, as the RSA offset now falls within the + * GuC inaccessible range. We resort to perma-pinning an additional vma + * within the accessible range that only contains the RSA signature. + * The GuC HW can use this extra pinning to perform the authentication + * since its GGTT offset will be GuC accessible. + */ + GEM_BUG_ON(uc_fw->rsa_size > PAGE_SIZE); + vma = intel_guc_allocate_vma(>->uc.guc, PAGE_SIZE); + if (IS_ERR(vma)) + return PTR_ERR(vma); + + vaddr = i915_gem_object_pin_map_unlocked(vma->obj, + i915_coherent_map_type(gt->i915, vma->obj, true)); + if (IS_ERR(vaddr)) { + i915_vma_unpin_and_release(&vma, 0); + err = PTR_ERR(vaddr); + goto unpin_out; + } + + copied = intel_uc_fw_copy_rsa(uc_fw, vaddr, vma->size); + i915_gem_object_unpin_map(vma->obj); + + if (copied < uc_fw->rsa_size) { + err = -ENOMEM; + goto unpin_out; + } + + uc_fw->rsa_data = vma; + + return 0; + +unpin_out: + i915_vma_unpin_and_release(&vma, 0); + return err; +} + +static void uc_fw_rsa_data_destroy(struct intel_uc_fw *uc_fw) +{ + i915_vma_unpin_and_release(&uc_fw->rsa_data, 0); +} + int intel_uc_fw_init(struct intel_uc_fw *uc_fw) { int err; @@ -558,14 +653,29 @@ int intel_uc_fw_init(struct intel_uc_fw *uc_fw) if (err) { DRM_DEBUG_DRIVER("%s fw pin-pages err=%d\n", intel_uc_fw_type_repr(uc_fw->type), err); - intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_FAIL); + goto out; } + err = uc_fw_rsa_data_create(uc_fw); + if (err) { + DRM_DEBUG_DRIVER("%s fw rsa data creation failed, err=%d\n", + intel_uc_fw_type_repr(uc_fw->type), err); + goto out_unpin; + } + + return 0; + +out_unpin: + i915_gem_object_unpin_pages(uc_fw->obj); +out: + intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_INIT_FAIL); return err; } void intel_uc_fw_fini(struct intel_uc_fw *uc_fw) { + uc_fw_rsa_data_destroy(uc_fw); + if (i915_gem_object_has_pinned_pages(uc_fw->obj)) i915_gem_object_unpin_pages(uc_fw->obj); diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h index 1e00bf65639e..d9d1dc0b4cbb 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h @@ -32,11 +32,12 @@ struct intel_gt; * | | MISSING <--/ | \--> ERROR | * | fetch | V | * | | AVAILABLE | - * +------------+- | -+ + * +------------+- | \ -+ + * | | | \--> INIT FAIL | * | init | V | * | | /------> LOADABLE <----<-----------\ | * +------------+- \ / \ \ \ -+ - * | | FAIL <--< \--> TRANSFERRED \ | + * | | LOAD FAIL <--< \--> TRANSFERRED \ | * | upload | \ / \ / | * | | \---------/ \--> RUNNING | * +------------+---------------------------------------------------+ @@ -50,8 +51,9 @@ enum intel_uc_fw_status { INTEL_UC_FIRMWARE_MISSING, /* blob not found on the system */ INTEL_UC_FIRMWARE_ERROR, /* invalid format or version */ INTEL_UC_FIRMWARE_AVAILABLE, /* blob found and copied in mem */ + INTEL_UC_FIRMWARE_INIT_FAIL, /* failed to prepare fw objects for load */ INTEL_UC_FIRMWARE_LOADABLE, /* all fw-required objects are ready */ - INTEL_UC_FIRMWARE_FAIL, /* failed to xfer or init/auth the fw */ + INTEL_UC_FIRMWARE_LOAD_FAIL, /* failed to xfer or init/auth the fw */ INTEL_UC_FIRMWARE_TRANSFERRED, /* dma xfer done */ INTEL_UC_FIRMWARE_RUNNING /* init/auth done */ }; @@ -84,6 +86,7 @@ struct intel_uc_fw { * or during a GT reset (mutex guarantees single threaded). */ struct i915_vma dummy; + struct i915_vma *rsa_data; /* * The firmware build process will generate a version header file with major and @@ -130,10 +133,12 @@ const char *intel_uc_fw_status_repr(enum intel_uc_fw_status status) return "ERROR"; case INTEL_UC_FIRMWARE_AVAILABLE: return "AVAILABLE"; + case INTEL_UC_FIRMWARE_INIT_FAIL: + return "INIT FAIL"; case INTEL_UC_FIRMWARE_LOADABLE: return "LOADABLE"; - case INTEL_UC_FIRMWARE_FAIL: - return "FAIL"; + case INTEL_UC_FIRMWARE_LOAD_FAIL: + return "LOAD FAIL"; case INTEL_UC_FIRMWARE_TRANSFERRED: return "TRANSFERRED"; case INTEL_UC_FIRMWARE_RUNNING: @@ -155,7 +160,8 @@ static inline int intel_uc_fw_status_to_error(enum intel_uc_fw_status status) return -ENOENT; case INTEL_UC_FIRMWARE_ERROR: return -ENOEXEC; - case INTEL_UC_FIRMWARE_FAIL: + case INTEL_UC_FIRMWARE_INIT_FAIL: + case INTEL_UC_FIRMWARE_LOAD_FAIL: return -EIO; case INTEL_UC_FIRMWARE_SELECTED: return -ESTALE; diff --git a/drivers/gpu/drm/i915/gt/uc/selftest_guc.c b/drivers/gpu/drm/i915/gt/uc/selftest_guc.c index fb0e4a7bd8ca..d3327b802b76 100644 --- a/drivers/gpu/drm/i915/gt/uc/selftest_guc.c +++ b/drivers/gpu/drm/i915/gt/uc/selftest_guc.c @@ -3,8 +3,21 @@ * Copyright �� 2021 Intel Corporation */ +#include "selftests/igt_spinner.h" #include "selftests/intel_scheduler_helpers.h" +static int request_add_spin(struct i915_request *rq, struct igt_spinner *spin) +{ + int err = 0; + + i915_request_get(rq); + i915_request_add(rq); + if (spin && !igt_wait_for_spinner(spin, rq)) + err = -ETIMEDOUT; + + return err; +} + static struct i915_request *nop_user_request(struct intel_context *ce, struct i915_request *from) { @@ -110,12 +123,172 @@ err: return ret; } +/* + * intel_guc_steal_guc_ids - Test to exhaust all guc_ids and then steal one + * + * This test creates a spinner which is used to block all subsequent submissions + * until it completes. Next, a loop creates a context and a NOP request each + * iteration until the guc_ids are exhausted (request creation returns -EAGAIN). + * The spinner is ended, unblocking all requests created in the loop. At this + * point all guc_ids are exhausted but are available to steal. Try to create + * another request which should successfully steal a guc_id. Wait on last + * request to complete, idle GPU, verify a guc_id was stolen via a counter, and + * exit the test. Test also artificially reduces the number of guc_ids so the + * test runs in a timely manner. + */ +static int intel_guc_steal_guc_ids(void *arg) +{ + struct intel_gt *gt = arg; + struct intel_guc *guc = >->uc.guc; + int ret, sv, context_index = 0; + intel_wakeref_t wakeref; + struct intel_engine_cs *engine; + struct intel_context **ce; + struct igt_spinner spin; + struct i915_request *spin_rq = NULL, *rq, *last = NULL; + int number_guc_id_stolen = guc->number_guc_id_stolen; + + ce = kzalloc(sizeof(*ce) * GUC_MAX_LRC_DESCRIPTORS, GFP_KERNEL); + if (!ce) { + pr_err("Context array allocation failed\n"); + return -ENOMEM; + } + + wakeref = intel_runtime_pm_get(gt->uncore->rpm); + engine = intel_selftest_find_any_engine(gt); + sv = guc->submission_state.num_guc_ids; + guc->submission_state.num_guc_ids = 4096; + + /* Create spinner to block requests in below loop */ + ce[context_index] = intel_context_create(engine); + if (IS_ERR(ce[context_index])) { + ret = PTR_ERR(ce[context_index]); + ce[context_index] = NULL; + pr_err("Failed to create context: %d\n", ret); + goto err_wakeref; + } + ret = igt_spinner_init(&spin, engine->gt); + if (ret) { + pr_err("Failed to create spinner: %d\n", ret); + goto err_contexts; + } + spin_rq = igt_spinner_create_request(&spin, ce[context_index], + MI_ARB_CHECK); + if (IS_ERR(spin_rq)) { + ret = PTR_ERR(spin_rq); + pr_err("Failed to create spinner request: %d\n", ret); + goto err_contexts; + } + ret = request_add_spin(spin_rq, &spin); + if (ret) { + pr_err("Failed to add Spinner request: %d\n", ret); + goto err_spin_rq; + } + + /* Use all guc_ids */ + while (ret != -EAGAIN) { + ce[++context_index] = intel_context_create(engine); + if (IS_ERR(ce[context_index])) { + ret = PTR_ERR(ce[context_index--]); + ce[context_index] = NULL; + pr_err("Failed to create context: %d\n", ret); + goto err_spin_rq; + } + + rq = nop_user_request(ce[context_index], spin_rq); + if (IS_ERR(rq)) { + ret = PTR_ERR(rq); + rq = NULL; + if (ret != -EAGAIN) { + pr_err("Failed to create request, %d: %d\n", + context_index, ret); + goto err_spin_rq; + } + } else { + if (last) + i915_request_put(last); + last = rq; + } + } + + /* Release blocked requests */ + igt_spinner_end(&spin); + ret = intel_selftest_wait_for_rq(spin_rq); + if (ret) { + pr_err("Spin request failed to complete: %d\n", ret); + i915_request_put(last); + goto err_spin_rq; + } + i915_request_put(spin_rq); + igt_spinner_fini(&spin); + spin_rq = NULL; + + /* Wait for last request */ + ret = i915_request_wait(last, 0, HZ * 30); + i915_request_put(last); + if (ret < 0) { + pr_err("Last request failed to complete: %d\n", ret); + goto err_spin_rq; + } + + /* Try to steal guc_id */ + rq = nop_user_request(ce[context_index], NULL); + if (IS_ERR(rq)) { + ret = PTR_ERR(rq); + pr_err("Failed to steal guc_id, %d: %d\n", context_index, ret); + goto err_spin_rq; + } + + /* Wait for request with stolen guc_id */ + ret = i915_request_wait(rq, 0, HZ); + i915_request_put(rq); + if (ret < 0) { + pr_err("Request with stolen guc_id failed to complete: %d\n", + ret); + goto err_spin_rq; + } + + /* Wait for idle */ + ret = intel_gt_wait_for_idle(gt, HZ * 30); + if (ret < 0) { + pr_err("GT failed to idle: %d\n", ret); + goto err_spin_rq; + } + + /* Verify a guc_id was stolen */ + if (guc->number_guc_id_stolen == number_guc_id_stolen) { + pr_err("No guc_id was stolen"); + ret = -EINVAL; + } else { + ret = 0; + } + +err_spin_rq: + if (spin_rq) { + igt_spinner_end(&spin); + intel_selftest_wait_for_rq(spin_rq); + i915_request_put(spin_rq); + igt_spinner_fini(&spin); + intel_gt_wait_for_idle(gt, HZ * 30); + } +err_contexts: + for (; context_index >= 0 && ce[context_index]; --context_index) + intel_context_put(ce[context_index]); +err_wakeref: + intel_runtime_pm_put(gt->uncore->rpm, wakeref); + kfree(ce); + guc->submission_state.num_guc_ids = sv; + + return ret; +} + int intel_guc_live_selftests(struct drm_i915_private *i915) { static const struct i915_subtest tests[] = { SUBTEST(intel_guc_scrub_ctbs), + SUBTEST(intel_guc_steal_guc_ids), }; - struct intel_gt *gt = &i915->gt; + struct intel_gt *gt = to_gt(i915); if (intel_gt_is_wedged(gt)) return 0; diff --git a/drivers/gpu/drm/i915/gt/uc/selftest_guc_multi_lrc.c b/drivers/gpu/drm/i915/gt/uc/selftest_guc_multi_lrc.c index 50953c8e8b53..1297ddbf7f88 100644 --- a/drivers/gpu/drm/i915/gt/uc/selftest_guc_multi_lrc.c +++ b/drivers/gpu/drm/i915/gt/uc/selftest_guc_multi_lrc.c @@ -167,7 +167,7 @@ int intel_guc_multi_lrc_live_selftests(struct drm_i915_private *i915) static const struct i915_subtest tests[] = { SUBTEST(intel_guc_multi_lrc_basic), }; - struct intel_gt *gt = &i915->gt; + struct intel_gt *gt = to_gt(i915); if (intel_gt_is_wedged(gt)) return 0; |