From patchwork Thu Jun 26 17:24:24 2014 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: John Harrison X-Patchwork-Id: 4429281 Return-Path: X-Original-To: patchwork-intel-gfx@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.19.201]) by patchwork1.web.kernel.org (Postfix) with ESMTP id 470A29FBBD for ; Thu, 26 Jun 2014 17:26:05 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 795CD20304 for ; Thu, 26 Jun 2014 17:26:01 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) by mail.kernel.org (Postfix) with ESMTP id 8E5C520381 for ; Thu, 26 Jun 2014 17:25:57 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 016F16E2D1; Thu, 26 Jun 2014 10:25:53 -0700 (PDT) X-Original-To: Intel-GFX@lists.freedesktop.org Delivered-To: Intel-GFX@lists.freedesktop.org Received: from mga11.intel.com (mga11.intel.com [192.55.52.93]) by gabe.freedesktop.org (Postfix) with ESMTP id 6E2906E24E for ; Thu, 26 Jun 2014 10:25:48 -0700 (PDT) Received: from fmsmga002.fm.intel.com ([10.253.24.26]) by fmsmga102.fm.intel.com with ESMTP; 26 Jun 2014 10:25:48 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.01,554,1400050800"; d="scan'208";a="561434919" Received: from johnharr-linux.iwi.intel.com ([172.28.253.52]) by fmsmga002.fm.intel.com with ESMTP; 26 Jun 2014 10:25:38 -0700 From: John.C.Harrison@Intel.com To: Intel-GFX@lists.freedesktop.org Date: Thu, 26 Jun 2014 18:24:24 +0100 Message-Id: <1403803475-16337-34-git-send-email-John.C.Harrison@Intel.com> X-Mailer: git-send-email 1.7.9.5 In-Reply-To: <1403803475-16337-1-git-send-email-John.C.Harrison@Intel.com> References: <1403803475-16337-1-git-send-email-John.C.Harrison@Intel.com> Organization: Intel Corporation (UK) Ltd. - Co. Reg. #1134945 - Pipers Way, Swindon SN3 1RJ Subject: [Intel-gfx] [RFC 33/44] drm/i915: Added trace points to scheduler X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" X-Spam-Status: No, score=-4.2 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_MED, T_RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: John Harrison Added trace points to the scheduler to track all the various events, node state transitions and other interesting things that occur. --- drivers/gpu/drm/i915/i915_gem_execbuffer.c | 2 + drivers/gpu/drm/i915/i915_scheduler.c | 31 ++++- drivers/gpu/drm/i915/i915_trace.h | 194 ++++++++++++++++++++++++++++ 3 files changed, 226 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 98cc95e..bf19e02 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -1413,6 +1413,8 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, ring->outstanding_lazy_seqno = 0; ring->preallocated_lazy_request = NULL; + trace_i915_gem_ring_queue(ring, &qe); + ret = i915_scheduler_queue_execbuffer(&qe); if (ret) goto err; diff --git a/drivers/gpu/drm/i915/i915_scheduler.c b/drivers/gpu/drm/i915/i915_scheduler.c index 71d8db4..6d0f4cb 100644 --- a/drivers/gpu/drm/i915/i915_scheduler.c +++ b/drivers/gpu/drm/i915/i915_scheduler.c @@ -87,6 +87,8 @@ int i915_scheduler_queue_execbuffer(struct i915_scheduler_queue_entry *qe) qe->params.scheduler_index = scheduler->index++; + trace_i915_scheduler_queue(qe->params.ring, qe); + scheduler->flags[qe->params.ring->id] |= i915_sf_submitting; ret = i915_gem_do_execbuffer_final(&qe->params); scheduler->flags[qe->params.ring->id] &= ~i915_sf_submitting; @@ -215,6 +217,9 @@ int i915_scheduler_queue_execbuffer(struct i915_scheduler_queue_entry *qe) not_flying = i915_scheduler_count_flying(scheduler, ring) < scheduler->min_flying; + trace_i915_scheduler_queue(ring, node); + trace_i915_scheduler_node_state_change(ring, node); + spin_unlock_irqrestore(&scheduler->lock, flags); if (not_flying) @@ -253,6 +258,8 @@ int i915_scheduler_fly_seqno(struct intel_engine_cs *ring, uint32_t seqno) node->stamp = stamp; node->status = i915_sqs_none; + trace_i915_scheduler_node_state_change(ring, node); + spin_lock_irqsave(&scheduler->lock, flags); ret = i915_scheduler_fly_node(node); spin_unlock_irqrestore(&scheduler->lock, flags); @@ -279,6 +286,9 @@ int i915_scheduler_fly_node(struct i915_scheduler_queue_entry *node) node->status = i915_sqs_flying; + trace_i915_scheduler_fly(ring, node); + trace_i915_scheduler_node_state_change(ring, node); + if (!(scheduler->flags[ring->id] & i915_sf_interrupts_enabled)) { bool success = true; @@ -343,6 +353,8 @@ static void i915_scheduler_node_requeue(struct i915_scheduler_queue_entry *node) BUG_ON(!I915_SQS_IS_FLYING(node)); node->status = i915_sqs_queued; + trace_i915_scheduler_unfly(node->params.ring, node); + trace_i915_scheduler_node_state_change(node->params.ring, node); } /* Give up on a popped node completely. For example, because it is causing the @@ -353,6 +365,8 @@ static void i915_scheduler_node_kill(struct i915_scheduler_queue_entry *node) BUG_ON(!I915_SQS_IS_FLYING(node)); node->status = i915_sqs_complete; + trace_i915_scheduler_unfly(node->params.ring, node); + trace_i915_scheduler_node_state_change(node->params.ring, node); } /* @@ -377,13 +391,17 @@ static int i915_scheduler_seqno_complete(struct intel_engine_cs *ring, uint32_t * if a completed entry is found then there is no need to scan further. */ list_for_each_entry(node, &scheduler->node_queue[ring->id], link) { - if (I915_SQS_IS_COMPLETE(node)) + if (I915_SQS_IS_COMPLETE(node)) { + trace_i915_scheduler_landing(ring, seqno, node); goto done; + } if (seqno == node->params.seqno) break; } + trace_i915_scheduler_landing(ring, seqno, node); + /* * NB: Lots of extra seqnos get added to the ring to track things * like cache flushes and page flips. So don't complain about if @@ -405,6 +423,7 @@ static int i915_scheduler_seqno_complete(struct intel_engine_cs *ring, uint32_t /* Node was in flight so mark it as complete. */ node->status = i915_sqs_complete; + trace_i915_scheduler_node_state_change(ring, node); } /* Should submit new work here if flight list is empty but the DRM @@ -425,6 +444,8 @@ int i915_scheduler_handle_IRQ(struct intel_engine_cs *ring) seqno = ring->get_seqno(ring, false); + trace_i915_scheduler_irq(ring, seqno); + if (i915.scheduler_override & i915_so_direct_submit) return 0; @@ -526,6 +547,8 @@ int i915_scheduler_remove(struct intel_engine_cs *ring) /* Launch more packets now? */ do_submit = (queued > 0) && (flying < scheduler->min_flying); + trace_i915_scheduler_remove(ring, min_seqno, do_submit); + spin_unlock_irqrestore(&scheduler->lock, flags); if (do_submit) @@ -535,6 +558,8 @@ int i915_scheduler_remove(struct intel_engine_cs *ring) node = list_first_entry(&remove, typeof(*node), link); list_del(&node->link); + trace_i915_scheduler_destroy(ring, node); + /* Release the locked buffers: */ for (i = 0; i < node->num_objs; i++) { drm_gem_object_unreference( @@ -793,6 +818,8 @@ static int i915_scheduler_pop_from_queue_locked(struct intel_engine_cs *ring, INIT_LIST_HEAD(&best->link); best->status = i915_sqs_none; + trace_i915_scheduler_node_state_change(ring, best); + ret = 0; } else { /* Can only get here if: @@ -812,6 +839,8 @@ static int i915_scheduler_pop_from_queue_locked(struct intel_engine_cs *ring, /* i915_scheduler_dump_queue_pop(ring, best); */ + trace_i915_scheduler_pop_from_queue(ring, best); + *pop_node = best; return ret; } diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h index f5aa006..bea2a49 100644 --- a/drivers/gpu/drm/i915/i915_trace.h +++ b/drivers/gpu/drm/i915/i915_trace.h @@ -9,6 +9,7 @@ #include "i915_drv.h" #include "intel_drv.h" #include "intel_ringbuffer.h" +#include "i915_scheduler.h" #undef TRACE_SYSTEM #define TRACE_SYSTEM i915 @@ -587,6 +588,199 @@ TRACE_EVENT(intel_gpu_freq_change, TP_printk("new_freq=%u", __entry->freq) ); +TRACE_EVENT(i915_scheduler_queue, + TP_PROTO(struct intel_engine_cs *ring, + struct i915_scheduler_queue_entry *node), + TP_ARGS(ring, node), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = node ? node->params.seqno : 0; + ), + + TP_printk("ring=%d, seqno=%d", + __entry->ring, __entry->seqno) +); + +TRACE_EVENT(i915_scheduler_fly, + TP_PROTO(struct intel_engine_cs *ring, + struct i915_scheduler_queue_entry *node), + TP_ARGS(ring, node), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = node ? node->params.seqno : 0; + ), + + TP_printk("ring=%d, seqno=%d", + __entry->ring, __entry->seqno) +); + +TRACE_EVENT(i915_scheduler_unfly, + TP_PROTO(struct intel_engine_cs *ring, + struct i915_scheduler_queue_entry *node), + TP_ARGS(ring, node), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = node ? node->params.seqno : 0; + ), + + TP_printk("ring=%d, seqno=%d", + __entry->ring, __entry->seqno) +); + +TRACE_EVENT(i915_scheduler_landing, + TP_PROTO(struct intel_engine_cs *ring, u32 seqno, + struct i915_scheduler_queue_entry *node), + TP_ARGS(ring, seqno, node), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + __field(u32, status) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = seqno; + __entry->status = node ? node->status : ~0U; + ), + + TP_printk("ring=%d, seqno=%d, status=%d", + __entry->ring, __entry->seqno, __entry->status) +); + +TRACE_EVENT(i915_scheduler_remove, + TP_PROTO(struct intel_engine_cs *ring, + u32 min_seqno, bool do_submit), + TP_ARGS(ring, min_seqno, do_submit), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, min_seqno) + __field(bool, do_submit) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->min_seqno = min_seqno; + __entry->do_submit = do_submit; + ), + + TP_printk("ring=%d, min_seqno = %d, do_submit=%d", + __entry->ring, __entry->min_seqno, __entry->do_submit) +); + +TRACE_EVENT(i915_scheduler_destroy, + TP_PROTO(struct intel_engine_cs *ring, + struct i915_scheduler_queue_entry *node), + TP_ARGS(ring, node), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = node ? node->params.seqno : 0; + ), + + TP_printk("ring=%d, seqno=%d", + __entry->ring, __entry->seqno) +); + +TRACE_EVENT(i915_scheduler_pop_from_queue, + TP_PROTO(struct intel_engine_cs *ring, + struct i915_scheduler_queue_entry *node), + TP_ARGS(ring, node), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = node ? node->params.seqno : 0; + ), + + TP_printk("ring=%d, seqno=%d", + __entry->ring, __entry->seqno) +); + +TRACE_EVENT(i915_scheduler_node_state_change, + TP_PROTO(struct intel_engine_cs *ring, + struct i915_scheduler_queue_entry *node), + TP_ARGS(ring, node), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + __field(u32, status) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = node->params.seqno; + __entry->status = node->status; + ), + + TP_printk("ring=%d, seqno=%d, status=%d", + __entry->ring, __entry->seqno, __entry->status) +); + +TRACE_EVENT(i915_scheduler_irq, + TP_PROTO(struct intel_engine_cs *ring, uint32_t seqno), + TP_ARGS(ring, seqno), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = seqno; + ), + + TP_printk("ring=%d, seqno=%d", __entry->ring, __entry->seqno) +); + +TRACE_EVENT(i915_gem_ring_queue, + TP_PROTO(struct intel_engine_cs *ring, + struct i915_scheduler_queue_entry *node), + TP_ARGS(ring, node), + + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, seqno) + ), + + TP_fast_assign( + __entry->ring = ring->id; + __entry->seqno = node->params.seqno; + ), + + TP_printk("ring=%d, seqno=%d", __entry->ring, __entry->seqno) +); + #endif /* _I915_TRACE_H_ */ /* This part must be outside protection */