diff options
Diffstat (limited to 'drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c')
| -rw-r--r-- | drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c | 80 |
1 files changed, 67 insertions, 13 deletions
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 97eadd08181d..2c651ec024ef 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c @@ -5,15 +5,16 @@ #include <linux/circ_buf.h> #include <linux/ktime.h> -#include <linux/time64.h> #include <linux/string_helpers.h> +#include <linux/time64.h> #include <linux/timekeeping.h> #include "i915_drv.h" +#include "i915_wait_util.h" #include "intel_guc_ct.h" #include "intel_guc_print.h" -#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GUC) +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG) enum { CT_DEAD_ALIVE = 0, CT_DEAD_SETUP, @@ -96,13 +97,40 @@ struct ct_request { struct ct_incoming_msg { struct list_head link; u32 size; - u32 msg[]; + u32 msg[] __counted_by(size); }; enum { CTB_SEND = 0, CTB_RECV = 1 }; enum { CTB_OWNER_HOST = 0 }; +/* + * Some H2G commands involve a synchronous response that the driver needs + * to wait for. In such cases, a timeout is required to prevent the driver + * from waiting forever in the case of an error (either no error response + * is defined in the protocol or something has died and requires a reset). + * The specific command may be defined as having a time bound response but + * the CT is a queue and that time guarantee only starts from the point + * when the command reaches the head of the queue and is processed by GuC. + * + * Ideally there would be a helper to report the progress of a given + * command through the CT. However, that would require a significant + * amount of work in the CT layer. In the meantime, provide a reasonable + * estimation of the worst case latency it should take for the entire + * queue to drain. And therefore, how long a caller should wait before + * giving up on their request. The current estimate is based on empirical + * measurement of a test that fills the buffer with context creation and + * destruction requests as they seem to be the slowest operation. + */ +long intel_guc_ct_max_queue_time_jiffies(void) +{ + /* + * A 4KB buffer full of context destroy commands takes a little + * over a second to process so bump that to 2s to be super safe. + */ + return (CTB_H2G_BUFFER_SIZE * HZ) / SZ_2K; +} + static void ct_receive_tasklet_func(struct tasklet_struct *t); static void ct_incoming_request_worker_func(struct work_struct *w); @@ -117,7 +145,7 @@ void intel_guc_ct_init_early(struct intel_guc_ct *ct) spin_lock_init(&ct->requests.lock); INIT_LIST_HEAD(&ct->requests.pending); INIT_LIST_HEAD(&ct->requests.incoming); -#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GUC) +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG) INIT_WORK(&ct->dead_ct_worker, ct_dead_ct_worker_func); #endif INIT_WORK(&ct->requests.worker, ct_incoming_request_worker_func); @@ -238,7 +266,7 @@ int intel_guc_ct_init(struct intel_guc_ct *ct) u32 *cmds; int err; - err = i915_inject_probe_error(guc_to_gt(guc)->i915, -ENXIO); + err = i915_inject_probe_error(guc_to_i915(guc), -ENXIO); if (err) return err; @@ -346,7 +374,7 @@ int intel_guc_ct_enable(struct intel_guc_ct *ct) ct->enabled = true; ct->stall_time = KTIME_MAX; -#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GUC) +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG) ct->dead_ct_reported = false; ct->dead_ct_reason = CT_DEAD_ALIVE; #endif @@ -1049,6 +1077,15 @@ static int ct_handle_response(struct intel_guc_ct *ct, struct ct_incoming_msg *r found = true; break; } + +#ifdef CONFIG_DRM_I915_SELFTEST + if (!found && ct_to_guc(ct)->fast_response_selftest) { + CT_DEBUG(ct, "Assuming unsolicited response due to FAST_REQUEST selftest\n"); + ct_to_guc(ct)->fast_response_selftest++; + found = true; + } +#endif + if (!found) { CT_ERROR(ct, "Unsolicited response message: len %u, data %#x (fence %u, last %u)\n", len, hxg[0], fence, ct->requests.last_fence); @@ -1112,12 +1149,11 @@ static int ct_process_request(struct intel_guc_ct *ct, struct ct_incoming_msg *r ret = 0; break; case INTEL_GUC_ACTION_NOTIFY_CRASH_DUMP_POSTED: - CT_ERROR(ct, "Received GuC crash dump notification!\n"); - ret = 0; - break; case INTEL_GUC_ACTION_NOTIFY_EXCEPTION: - CT_ERROR(ct, "Received GuC exception notification!\n"); - ret = 0; + ret = intel_guc_crash_process_msg(guc, action); + break; + case INTEL_GUC_ACTION_TLB_INVALIDATION_DONE: + ret = intel_guc_tlb_invalidation_done(guc, payload, len); break; default: ret = -EOPNOTSUPP; @@ -1190,9 +1226,17 @@ static int ct_handle_event(struct intel_guc_ct *ct, struct ct_incoming_msg *requ switch (action) { case INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_DONE: case INTEL_GUC_ACTION_DEREGISTER_CONTEXT_DONE: + case INTEL_GUC_ACTION_TLB_INVALIDATION_DONE: g2h_release_space(ct, request->size); } + /* + * TLB invalidation responses must be handled immediately as processing + * of other G2H notifications may be blocked by an invalidation request. + */ + if (action == INTEL_GUC_ACTION_TLB_INVALIDATION_DONE) + return ct_process_request(ct, request); + spin_lock_irqsave(&ct->requests.lock, flags); list_add_tail(&request->link, &ct->requests.incoming); spin_unlock_irqrestore(&ct->requests.lock, flags); @@ -1281,9 +1325,16 @@ static int ct_receive(struct intel_guc_ct *ct) static void ct_try_receive_message(struct intel_guc_ct *ct) { + struct intel_guc *guc = ct_to_guc(ct); int ret; - if (GEM_WARN_ON(!ct->enabled)) + if (!ct->enabled) { + GEM_WARN_ON(!guc_to_gt(guc)->uc.reset_in_progress); + return; + } + + /* When interrupt disabled, message handling is not expected */ + if (!guc->interrupts.enabled) return; ret = ct_receive(ct); @@ -1334,7 +1385,7 @@ void intel_guc_ct_print_info(struct intel_guc_ct *ct, ct->ctbs.recv.desc->tail); } -#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GUC) +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG) static void ct_dead_ct_worker_func(struct work_struct *w) { struct intel_guc_ct *ct = container_of(w, struct intel_guc_ct, dead_ct_worker); @@ -1343,6 +1394,9 @@ static void ct_dead_ct_worker_func(struct work_struct *w) if (ct->dead_ct_reported) return; + if (i915_error_injected()) + return; + ct->dead_ct_reported = true; guc_info(guc, "CTB is dead - reason=0x%X\n", ct->dead_ct_reason); |