[PATCH 1/3] drm/i915: Make ptr_unpack_bits() more function-like

[PATCH 1/3] drm/i915: Make ptr_unpack_bits() more function-like

[Chris Wilson]

ptr_unpack_bits() is a function-like macro, as such it is meant to be
replaceable by a function. In this case, we should be passing in the
out-param as a pointer.

Bizarrely this does affect code generation:

function                                     old     new   delta
i915_gem_object_pin_map                      409     389     -20

An improvement(?) in this case, but one can't help wonder what
strict-aliasing optimisations we are preventing.

The generated code looks identical in using ptr_unpack_bits (no extra
motions to stack, the pointer and bits appear to be kept in registers),
the difference appears to be code ordering and with a reorder it is able
to use smaller forward jumps.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_gem.c   | 2 +-
 drivers/gpu/drm/i915/i915_utils.h | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index ae4764b0c361..f49aca95553b 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2560,7 +2560,7 @@ void *i915_gem_object_pin_map(struct drm_i915_gem_object *obj,
 	}
 	GEM_BUG_ON(!obj->mm.pages);
 
-	ptr = ptr_unpack_bits(obj->mm.mapping, has_type);
+	ptr = ptr_unpack_bits(obj->mm.mapping, &has_type);
 	if (ptr && has_type != type) {
 		if (pinned) {
 			ret = -EBUSY;
diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h
index c5455d36b617..aca11aad5da7 100644
--- a/drivers/gpu/drm/i915/i915_utils.h
+++ b/drivers/gpu/drm/i915/i915_utils.h
@@ -77,7 +77,7 @@
 
 #define ptr_unpack_bits(ptr, bits) ({					\
 	unsigned long __v = (unsigned long)(ptr);			\
-	(bits) = __v & ~PAGE_MASK;					\
+	*(bits) = __v & ~PAGE_MASK;					\
 	(typeof(ptr))(__v & PAGE_MASK);					\
 })
 
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 2/3] drm/i915: Redefine ptr_pack_bits() and friends

[Chris Wilson]

Rebrand the current (pointer | bits) pack/unpack utility macros as
explicit bit twiddling for PAGE_SIZE so that we can use the more
flexible underlying macros for different bits.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_cmd_parser.c |  2 +-
 drivers/gpu/drm/i915/i915_gem.c        |  6 +++---
 drivers/gpu/drm/i915/i915_utils.h      | 19 +++++++++++++------
 3 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c
index 7af100f84410..1b69a1f9a1ea 100644
--- a/drivers/gpu/drm/i915/i915_cmd_parser.c
+++ b/drivers/gpu/drm/i915/i915_cmd_parser.c
@@ -1284,7 +1284,7 @@ int intel_engine_cmd_parser(struct intel_engine_cs *engine,
 
 		if (*cmd == MI_BATCH_BUFFER_END) {
 			if (needs_clflush_after) {
-				void *ptr = ptr_mask_bits(shadow_batch_obj->mm.mapping);
+				void *ptr = page_mask_bits(shadow_batch_obj->mm.mapping);
 				drm_clflush_virt_range(ptr,
 						       (void *)(cmd + 1) - ptr);
 			}
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index f49aca95553b..178f01226c6e 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2228,7 +2228,7 @@ void __i915_gem_object_put_pages(struct drm_i915_gem_object *obj,
 	if (obj->mm.mapping) {
 		void *ptr;
 
-		ptr = ptr_mask_bits(obj->mm.mapping);
+		ptr = page_mask_bits(obj->mm.mapping);
 		if (is_vmalloc_addr(ptr))
 			vunmap(ptr);
 		else
@@ -2560,7 +2560,7 @@ void *i915_gem_object_pin_map(struct drm_i915_gem_object *obj,
 	}
 	GEM_BUG_ON(!obj->mm.pages);
 
-	ptr = ptr_unpack_bits(obj->mm.mapping, &has_type);
+	ptr = page_unpack_bits(obj->mm.mapping, &has_type);
 	if (ptr && has_type != type) {
 		if (pinned) {
 			ret = -EBUSY;
@@ -2582,7 +2582,7 @@ void *i915_gem_object_pin_map(struct drm_i915_gem_object *obj,
 			goto err_unpin;
 		}
 
-		obj->mm.mapping = ptr_pack_bits(ptr, type);
+		obj->mm.mapping = page_pack_bits(ptr, type);
 	}
 
 out_unlock:
diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h
index aca11aad5da7..f0500c65726d 100644
--- a/drivers/gpu/drm/i915/i915_utils.h
+++ b/drivers/gpu/drm/i915/i915_utils.h
@@ -70,20 +70,27 @@
 #define overflows_type(x, T) \
 	(sizeof(x) > sizeof(T) && (x) >> (sizeof(T) * BITS_PER_BYTE))
 
-#define ptr_mask_bits(ptr) ({						\
+#define ptr_mask_bits(ptr, n) ({					\
 	unsigned long __v = (unsigned long)(ptr);			\
-	(typeof(ptr))(__v & PAGE_MASK);					\
+	(typeof(ptr))(__v & -BIT(n));					\
 })
 
-#define ptr_unpack_bits(ptr, bits) ({					\
+#define ptr_unmask_bits(ptr, n) ((unsigned long)(ptr) & (BIT(n) - 1))
+
+#define ptr_unpack_bits(ptr, bits, n) ({				\
 	unsigned long __v = (unsigned long)(ptr);			\
-	*(bits) = __v & ~PAGE_MASK;					\
-	(typeof(ptr))(__v & PAGE_MASK);					\
+	*(bits) = __v & (BIT(n) - 1);					\
+	(typeof(ptr))(__v & -BIT(n));					\
 })
 
-#define ptr_pack_bits(ptr, bits)					\
+#define ptr_pack_bits(ptr, bits, n)					\
 	((typeof(ptr))((unsigned long)(ptr) | (bits)))
 
+#define page_mask_bits(ptr) ptr_mask_bits(ptr, PAGE_SHIFT)
+#define page_unmask_bits(ptr) ptr_unmask_bits(ptr, PAGE_SHIFT)
+#define page_pack_bits(ptr, bits) ptr_pack_bits(ptr, bits, PAGE_SHIFT)
+#define page_unpack_bits(ptr, bits) ptr_unpack_bits(ptr, bits, PAGE_SHIFT)
+
 #define ptr_offset(ptr, member) offsetof(typeof(*(ptr)), member)
 
 #define fetch_and_zero(ptr) ({						\
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 3/3] drm/i915/execlists: Pack the count into the low bits of the port.request

[Chris Wilson]

add/remove: 1/1 grow/shrink: 5/4 up/down: 391/-578 (-187)
function                                     old     new   delta
execlists_submit_ports                       262     471    +209
port_assign.isra                               -     136    +136
capture                                     6344    6359     +15
reset_common_ring                            438     452     +14
execlists_submit_request                     228     238     +10
gen8_init_common_ring                        334     341      +7
intel_engine_is_idle                         106     105      -1
i915_engine_info                            2314    2290     -24
__i915_gem_set_wedged_BKL                    485     411     -74
intel_lrc_irq_handler                       1789    1604    -185
execlists_update_context                     294       -    -294

The most important change there is the improve to the
intel_lrc_irq_handler and excclist_submit_ports (net improvement since
execlists_update_context is now inlined).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
---
 drivers/gpu/drm/i915/i915_debugfs.c        |  32 ++++---
 drivers/gpu/drm/i915/i915_gem.c            |   6 +-
 drivers/gpu/drm/i915/i915_gpu_error.c      |  13 ++-
 drivers/gpu/drm/i915/i915_guc_submission.c |  18 ++--
 drivers/gpu/drm/i915/intel_engine_cs.c     |   2 +-
 drivers/gpu/drm/i915/intel_lrc.c           | 133 ++++++++++++++++-------------
 drivers/gpu/drm/i915/intel_ringbuffer.h    |   8 +-
 7 files changed, 120 insertions(+), 92 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index d689e511744e..0a123779f693 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -3317,6 +3317,7 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 		if (i915.enable_execlists) {
 			u32 ptr, read, write;
 			struct rb_node *rb;
+			unsigned int idx;
 
 			seq_printf(m, "\tExeclist status: 0x%08x %08x\n",
 				   I915_READ(RING_EXECLIST_STATUS_LO(engine)),
@@ -3334,8 +3335,7 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 			if (read > write)
 				write += GEN8_CSB_ENTRIES;
 			while (read < write) {
-				unsigned int idx = ++read % GEN8_CSB_ENTRIES;
-
+				idx = ++read % GEN8_CSB_ENTRIES;
 				seq_printf(m, "\tExeclist CSB[%d]: 0x%08x, context: %d\n",
 					   idx,
 					   I915_READ(RING_CONTEXT_STATUS_BUF_LO(engine, idx)),
@@ -3343,21 +3343,19 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 			}
 
 			rcu_read_lock();
-			rq = READ_ONCE(engine->execlist_port[0].request);
-			if (rq) {
-				seq_printf(m, "\t\tELSP[0] count=%d, ",
-					   engine->execlist_port[0].count);
-				print_request(m, rq, "rq: ");
-			} else {
-				seq_printf(m, "\t\tELSP[0] idle\n");
-			}
-			rq = READ_ONCE(engine->execlist_port[1].request);
-			if (rq) {
-				seq_printf(m, "\t\tELSP[1] count=%d, ",
-					   engine->execlist_port[1].count);
-				print_request(m, rq, "rq: ");
-			} else {
-				seq_printf(m, "\t\tELSP[1] idle\n");
+			for (idx = 0; idx < ARRAY_SIZE(engine->execlist_port); idx++) {
+				unsigned int count;
+
+				rq = port_unpack(&engine->execlist_port[idx],
+						 &count);
+				if (rq) {
+					seq_printf(m, "\t\tELSP[%d] count=%d, ",
+						   idx, count);
+					print_request(m, rq, "rq: ");
+				} else {
+					seq_printf(m, "\t\tELSP[%d] idle\n",
+						   idx);
+				}
 			}
 			rcu_read_unlock();
 
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 178f01226c6e..9ad5c7b8e36b 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2967,12 +2967,14 @@ static void engine_set_wedged(struct intel_engine_cs *engine)
 	 */
 
 	if (i915.enable_execlists) {
+		struct execlist_port *port = engine->execlist_port;
 		unsigned long flags;
+		unsigned int n;
 
 		spin_lock_irqsave(&engine->timeline->lock, flags);
 
-		i915_gem_request_put(engine->execlist_port[0].request);
-		i915_gem_request_put(engine->execlist_port[1].request);
+		for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++)
+			i915_gem_request_put(port_request(&port[n]));
 		memset(engine->execlist_port, 0, sizeof(engine->execlist_port));
 		engine->execlist_queue = RB_ROOT;
 		engine->execlist_first = NULL;
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index 8effc59f5cb5..c86a3bd89468 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -1316,12 +1316,17 @@ static void engine_record_requests(struct intel_engine_cs *engine,
 static void error_record_engine_execlists(struct intel_engine_cs *engine,
 					  struct drm_i915_error_engine *ee)
 {
+	const struct execlist_port *port = engine->execlist_port;
 	unsigned int n;
 
-	for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++)
-		if (engine->execlist_port[n].request)
-			record_request(engine->execlist_port[n].request,
-				       &ee->execlist[n]);
+	for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++) {
+		struct drm_i915_gem_request *rq = port_request(&port[n]);
+
+		if (!rq)
+			break;
+
+		record_request(rq, &ee->execlist[n]);
+	}
 }
 
 static void record_context(struct drm_i915_error_context *e,
diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index 1642fff9cf13..370373c97b81 100644
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -658,7 +658,7 @@ static void nested_enable_signaling(struct drm_i915_gem_request *rq)
 static bool i915_guc_dequeue(struct intel_engine_cs *engine)
 {
 	struct execlist_port *port = engine->execlist_port;
-	struct drm_i915_gem_request *last = port[0].request;
+	struct drm_i915_gem_request *last = port[0].request_count;
 	struct rb_node *rb;
 	bool submit = false;
 
@@ -672,7 +672,7 @@ static bool i915_guc_dequeue(struct intel_engine_cs *engine)
 			if (port != engine->execlist_port)
 				break;
 
-			i915_gem_request_assign(&port->request, last);
+			i915_gem_request_assign(&port->request_count, last);
 			nested_enable_signaling(last);
 			port++;
 		}
@@ -688,7 +688,7 @@ static bool i915_guc_dequeue(struct intel_engine_cs *engine)
 		submit = true;
 	}
 	if (submit) {
-		i915_gem_request_assign(&port->request, last);
+		i915_gem_request_assign(&port->request_count, last);
 		nested_enable_signaling(last);
 		engine->execlist_first = rb;
 	}
@@ -705,17 +705,19 @@ static void i915_guc_irq_handler(unsigned long data)
 	bool submit;
 
 	do {
-		rq = port[0].request;
+		rq = port[0].request_count;
 		while (rq && i915_gem_request_completed(rq)) {
 			trace_i915_gem_request_out(rq);
 			i915_gem_request_put(rq);
-			port[0].request = port[1].request;
-			port[1].request = NULL;
-			rq = port[0].request;
+
+			port[0].request_count = port[1].request_count;
+			port[1].request_count = NULL;
+
+			rq = port[0].request_count;
 		}
 
 		submit = false;
-		if (!port[1].request)
+		if (!port[1].request_count)
 			submit = i915_guc_dequeue(engine);
 	} while (submit);
 }
diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c
index 854e8e0c836b..fca7160a29c2 100644
--- a/drivers/gpu/drm/i915/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/intel_engine_cs.c
@@ -1096,7 +1096,7 @@ bool intel_engine_is_idle(struct intel_engine_cs *engine)
 		return false;
 
 	/* Both ports drained, no more ELSP submission? */
-	if (engine->execlist_port[0].request)
+	if (port_request(&engine->execlist_port[0]))
 		return false;
 
 	/* Ring stopped? */
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index 0dc1cc4ad6e7..15bd5119b050 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -342,39 +342,32 @@ static u64 execlists_update_context(struct drm_i915_gem_request *rq)
 
 static void execlists_submit_ports(struct intel_engine_cs *engine)
 {
-	struct drm_i915_private *dev_priv = engine->i915;
 	struct execlist_port *port = engine->execlist_port;
 	u32 __iomem *elsp =
-		dev_priv->regs + i915_mmio_reg_offset(RING_ELSP(engine));
-	u64 desc[2];
-
-	GEM_BUG_ON(port[0].count > 1);
-	if (!port[0].count)
-		execlists_context_status_change(port[0].request,
-						INTEL_CONTEXT_SCHEDULE_IN);
-	desc[0] = execlists_update_context(port[0].request);
-	GEM_DEBUG_EXEC(port[0].context_id = upper_32_bits(desc[0]));
-	port[0].count++;
-
-	if (port[1].request) {
-		GEM_BUG_ON(port[1].count);
-		execlists_context_status_change(port[1].request,
-						INTEL_CONTEXT_SCHEDULE_IN);
-		desc[1] = execlists_update_context(port[1].request);
-		GEM_DEBUG_EXEC(port[1].context_id = upper_32_bits(desc[1]));
-		port[1].count = 1;
-	} else {
-		desc[1] = 0;
-	}
-	GEM_BUG_ON(desc[0] == desc[1]);
-
-	/* You must always write both descriptors in the order below. */
-	writel(upper_32_bits(desc[1]), elsp);
-	writel(lower_32_bits(desc[1]), elsp);
+		engine->i915->regs + i915_mmio_reg_offset(RING_ELSP(engine));
+	unsigned int n;
+
+	for (n = ARRAY_SIZE(engine->execlist_port); n--; ) {
+		struct drm_i915_gem_request *rq;
+		unsigned int count;
+		u64 desc;
+
+		rq = port_unpack(&port[n], &count);
+		if (rq) {
+			GEM_BUG_ON(count > !n);
+			if (!count++)
+				execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_IN);
+			port[n].request_count = port_pack(rq, count);
+			desc = execlists_update_context(rq);
+			GEM_DEBUG_EXEC(port[n].context_id = upper_32_bits(desc));
+		} else {
+			GEM_BUG_ON(!n);
+			desc = 0;
+		}
 
-	writel(upper_32_bits(desc[0]), elsp);
-	/* The context is automatically loaded after the following */
-	writel(lower_32_bits(desc[0]), elsp);
+		writel(upper_32_bits(desc), elsp);
+		writel(lower_32_bits(desc), elsp);
+	}
 }
 
 static bool ctx_single_port_submission(const struct i915_gem_context *ctx)
@@ -395,6 +388,18 @@ static bool can_merge_ctx(const struct i915_gem_context *prev,
 	return true;
 }
 
+static void port_assign(struct execlist_port *port,
+			struct drm_i915_gem_request *rq)
+{
+	GEM_BUG_ON(rq == port_request(port));
+
+	if (port->request_count)
+		i915_gem_request_put(port_request(port));
+
+	port->request_count =
+		port_pack(i915_gem_request_get(rq), port_count(port));
+}
+
 static void execlists_dequeue(struct intel_engine_cs *engine)
 {
 	struct drm_i915_gem_request *last;
@@ -402,7 +407,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 	struct rb_node *rb;
 	bool submit = false;
 
-	last = port->request;
+	last = port_request(port);
 	if (last)
 		/* WaIdleLiteRestore:bdw,skl
 		 * Apply the wa NOOPs to prevent ring:HEAD == req:TAIL
@@ -412,7 +417,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 		 */
 		last->tail = last->wa_tail;
 
-	GEM_BUG_ON(port[1].request);
+	GEM_BUG_ON(port[1].request_count);
 
 	/* Hardware submission is through 2 ports. Conceptually each port
 	 * has a (RING_START, RING_HEAD, RING_TAIL) tuple. RING_START is
@@ -469,7 +474,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 
 			GEM_BUG_ON(last->ctx == cursor->ctx);
 
-			i915_gem_request_assign(&port->request, last);
+			if (submit)
+				port_assign(port, last);
 			port++;
 		}
 
@@ -484,7 +490,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 		submit = true;
 	}
 	if (submit) {
-		i915_gem_request_assign(&port->request, last);
+		port_assign(port, last);
 		engine->execlist_first = rb;
 	}
 	spin_unlock_irq(&engine->timeline->lock);
@@ -495,14 +501,14 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 
 static bool execlists_elsp_idle(struct intel_engine_cs *engine)
 {
-	return !engine->execlist_port[0].request;
+	return !port_count(&engine->execlist_port[0]);
 }
 
 static bool execlists_elsp_ready(const struct intel_engine_cs *engine)
 {
 	const struct execlist_port *port = engine->execlist_port;
 
-	return port[0].count + port[1].count < 2;
+	return port_count(&port[0]) + port_count(&port[1]) < 2;
 }
 
 /*
@@ -543,7 +549,9 @@ static void intel_lrc_irq_handler(unsigned long data)
 		tail = GEN8_CSB_WRITE_PTR(head);
 		head = GEN8_CSB_READ_PTR(head);
 		while (head != tail) {
+			struct drm_i915_gem_request *rq;
 			unsigned int status;
+			unsigned int count;
 
 			if (++head == GEN8_CSB_ENTRIES)
 				head = 0;
@@ -573,20 +581,24 @@ static void intel_lrc_irq_handler(unsigned long data)
 			GEM_DEBUG_BUG_ON(readl(buf + 2 * head + 1) !=
 					 port[0].context_id);
 
-			GEM_BUG_ON(port[0].count == 0);
-			if (--port[0].count == 0) {
+			rq = port_unpack(&port[0], &count);
+			GEM_BUG_ON(count == 0);
+			if (--count == 0) {
 				GEM_BUG_ON(status & GEN8_CTX_STATUS_PREEMPTED);
-				GEM_BUG_ON(!i915_gem_request_completed(port[0].request));
-				execlists_context_status_change(port[0].request,
-								INTEL_CONTEXT_SCHEDULE_OUT);
+				GEM_BUG_ON(!i915_gem_request_completed(rq));
+				execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_OUT);
+
+				trace_i915_gem_request_out(rq);
+				i915_gem_request_put(rq);
 
-				trace_i915_gem_request_out(port[0].request);
-				i915_gem_request_put(port[0].request);
 				port[0] = port[1];
 				memset(&port[1], 0, sizeof(port[1]));
+			} else {
+				port[0].request_count = port_pack(rq, count);
 			}
 
-			GEM_BUG_ON(port[0].count == 0 &&
+			/* After the final element, the hw should be idle */
+			GEM_BUG_ON(port_count(&port[0]) == 0 &&
 				   !(status & GEN8_CTX_STATUS_ACTIVE_IDLE));
 		}
 
@@ -1139,11 +1151,6 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine)
 	return ret;
 }
 
-static u32 port_seqno(struct execlist_port *port)
-{
-	return port->request ? port->request->global_seqno : 0;
-}
-
 static int gen8_init_common_ring(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
@@ -1168,12 +1175,22 @@ static int gen8_init_common_ring(struct intel_engine_cs *engine)
 	/* After a GPU reset, we may have requests to replay */
 	clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted);
 	if (!i915.enable_guc_submission && !execlists_elsp_idle(engine)) {
-		DRM_DEBUG_DRIVER("Restarting %s from requests [0x%x, 0x%x]\n",
-				 engine->name,
-				 port_seqno(&engine->execlist_port[0]),
-				 port_seqno(&engine->execlist_port[1]));
-		engine->execlist_port[0].count = 0;
-		engine->execlist_port[1].count = 0;
+		struct execlist_port *port = engine->execlist_port;
+		unsigned int n;
+
+		for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++) {
+			if (!port[n].request_count)
+				break;
+
+			DRM_DEBUG_DRIVER("Restarting %s from 0x%x [%d]\n",
+					 engine->name,
+					 port_request(&port[n])->global_seqno,
+					 n);
+
+			/* Discard the current inflight count */
+			port[n].request_count = port_request(&port[n]);
+		}
+
 		execlists_submit_ports(engine);
 	}
 
@@ -1252,13 +1269,13 @@ static void reset_common_ring(struct intel_engine_cs *engine,
 	intel_ring_update_space(request->ring);
 
 	/* Catch up with any missed context-switch interrupts */
-	if (request->ctx != port[0].request->ctx) {
-		i915_gem_request_put(port[0].request);
+	if (request->ctx != port_request(&port[0])->ctx) {
+		i915_gem_request_put(port_request(&port[0]));
 		port[0] = port[1];
 		memset(&port[1], 0, sizeof(port[1]));
 	}
 
-	GEM_BUG_ON(request->ctx != port[0].request->ctx);
+	GEM_BUG_ON(request->ctx != port_request(&port[0])->ctx);
 
 	/* Reset WaIdleLiteRestore:bdw,skl as well */
 	request->tail =
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index cbe61d3f31da..97417aedf2bd 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -371,8 +371,12 @@ struct intel_engine_cs {
 	/* Execlists */
 	struct tasklet_struct irq_tasklet;
 	struct execlist_port {
-		struct drm_i915_gem_request *request;
-		unsigned int count;
+		struct drm_i915_gem_request *request_count;
+#define EXECLIST_COUNT_BITS 2
+#define port_request(p) ptr_mask_bits((p)->request_count, EXECLIST_COUNT_BITS)
+#define port_count(p) ptr_unmask_bits((p)->request_count, EXECLIST_COUNT_BITS)
+#define port_pack(rq, count) ptr_pack_bits(rq, count, EXECLIST_COUNT_BITS)
+#define port_unpack(p, count) ptr_unpack_bits((p)->request_count, count, EXECLIST_COUNT_BITS)
 		GEM_DEBUG_DECL(u32 context_id);
 	} execlist_port[2];
 	struct rb_root execlist_queue;
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

✓ Fi.CI.BAT: success for series starting with [1/3] drm/i915: Make ptr_unpack_bits() more function-like

[Patchwork]

== Series Details ==

Series: series starting with [1/3] drm/i915: Make ptr_unpack_bits() more function-like
URL   : https://patchwork.freedesktop.org/series/22432/
State : success

== Summary ==

Series 22432v1 Series without cover letter
https://patchwork.freedesktop.org/api/1.0/series/22432/revisions/1/mbox/

Test gem_exec_flush:
        Subgroup basic-batch-kernel-default-uc:
                fail       -> PASS       (fi-snb-2600) fdo#100007
Test gem_exec_suspend:
        Subgroup basic-s4-devices:
                pass       -> DMESG-WARN (fi-bxt-t5700) fdo#100125
                pass       -> DMESG-WARN (fi-kbl-7560u) fdo#100125
Test kms_pipe_crc_basic:
        Subgroup suspend-read-crc-pipe-a:
                fail       -> PASS       (fi-skl-6700k) fdo#100367
Test prime_vgem:
        Subgroup basic-fence-flip:
                dmesg-warn -> PASS       (fi-byt-n2820) fdo#100094

fdo#100007 https://bugs.freedesktop.org/show_bug.cgi?id=100007
fdo#100125 https://bugs.freedesktop.org/show_bug.cgi?id=100125
fdo#100367 https://bugs.freedesktop.org/show_bug.cgi?id=100367
fdo#100094 https://bugs.freedesktop.org/show_bug.cgi?id=100094

fi-bdw-5557u     total:278  pass:267  dwarn:0   dfail:0   fail:0   skip:11  time: 429s
fi-bdw-gvtdvm    total:278  pass:256  dwarn:8   dfail:0   fail:0   skip:14  time: 427s
fi-bsw-n3050     total:278  pass:239  dwarn:0   dfail:0   fail:0   skip:39  time: 566s
fi-bxt-j4205     total:278  pass:259  dwarn:0   dfail:0   fail:0   skip:19  time: 507s
fi-bxt-t5700     total:278  pass:257  dwarn:1   dfail:0   fail:0   skip:20  time: 556s
fi-byt-j1900     total:278  pass:251  dwarn:0   dfail:0   fail:0   skip:27  time: 481s
fi-byt-n2820     total:278  pass:247  dwarn:0   dfail:0   fail:0   skip:31  time: 476s
fi-hsw-4770      total:278  pass:262  dwarn:0   dfail:0   fail:0   skip:16  time: 407s
fi-hsw-4770r     total:278  pass:262  dwarn:0   dfail:0   fail:0   skip:16  time: 407s
fi-ilk-650       total:278  pass:228  dwarn:0   dfail:0   fail:0   skip:50  time: 435s
fi-ivb-3520m     total:278  pass:260  dwarn:0   dfail:0   fail:0   skip:18  time: 496s
fi-ivb-3770      total:278  pass:260  dwarn:0   dfail:0   fail:0   skip:18  time: 485s
fi-kbl-7500u     total:278  pass:260  dwarn:0   dfail:0   fail:0   skip:18  time: 454s
fi-kbl-7560u     total:278  pass:267  dwarn:1   dfail:0   fail:0   skip:10  time: 576s
fi-skl-6260u     total:278  pass:268  dwarn:0   dfail:0   fail:0   skip:10  time: 452s
fi-skl-6700hq    total:278  pass:261  dwarn:0   dfail:0   fail:0   skip:17  time: 571s
fi-skl-6700k     total:278  pass:256  dwarn:4   dfail:0   fail:0   skip:18  time: 463s
fi-skl-6770hq    total:278  pass:268  dwarn:0   dfail:0   fail:0   skip:10  time: 489s
fi-skl-gvtdvm    total:278  pass:265  dwarn:0   dfail:0   fail:0   skip:13  time: 430s
fi-snb-2520m     total:278  pass:250  dwarn:0   dfail:0   fail:0   skip:28  time: 534s
fi-snb-2600      total:278  pass:249  dwarn:0   dfail:0   fail:0   skip:29  time: 408s

38a5db0d3643f519d86f42029b7a93e624807a71 drm-tip: 2017y-04m-04d-13h-24m-06s UTC integration manifest
a6d54fa drm/i915/execlists: Pack the count into the low bits of the port.request
c096e1d drm/i915: Redefine ptr_pack_bits() and friends
c6ec960 drm/i915: Make ptr_unpack_bits() more function-like

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_4393/
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx