[Intel-gfx] [PATCH v5 06/11] drm/i915: Enable i915 perf stream for Haswell OA unit

Robert Bragg robert at sixbynine.org
Wed Sep 14 14:19:44 UTC 2016


Gen graphics hardware can be set up to periodically write snapshots of
performance counters into a circular buffer via its Observation
Architecture and this patch exposes that capability to userspace via the
i915 perf interface.

Cc: Chris Wilson <chris at chris-wilson.co.uk>
Signed-off-by: Robert Bragg <robert at sixbynine.org>
Signed-off-by: Zhenyu Wang <zhenyuw at linux.intel.com>
---
 drivers/gpu/drm/i915/i915_drv.h         |  72 ++-
 drivers/gpu/drm/i915/i915_gem_context.c |  22 +-
 drivers/gpu/drm/i915/i915_perf.c        | 998 +++++++++++++++++++++++++++++++-
 drivers/gpu/drm/i915/i915_reg.h         | 338 +++++++++++
 drivers/gpu/drm/i915/intel_ringbuffer.c |  10 +-
 include/uapi/drm/i915_drm.h             |  70 ++-
 6 files changed, 1477 insertions(+), 33 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 5fad018..551f078 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1740,6 +1740,11 @@ struct intel_wm_config {
 	bool sprites_scaled;
 };
 
+struct i915_oa_format {
+	u32 format;
+	int size;
+};
+
 struct i915_oa_reg {
 	i915_reg_t addr;
 	u32 value;
@@ -1760,11 +1765,6 @@ struct i915_perf_stream_ops {
 	 */
 	void (*disable)(struct i915_perf_stream *stream);
 
-	/* Return: true if any i915 perf records are ready to read()
-	 * for this stream.
-	 */
-	bool (*can_read)(struct i915_perf_stream *stream);
-
 	/* Call poll_wait, passing a wait queue that will be woken
 	 * once there is something ready to read() for the stream
 	 */
@@ -1774,9 +1774,7 @@ struct i915_perf_stream_ops {
 
 	/* For handling a blocking read, wait until there is something
 	 * to ready to read() for the stream. E.g. wait on the same
-	 * wait queue that would be passed to poll_wait() until
-	 * ->can_read() returns true (if its safe to call ->can_read()
-	 * without the i915 perf lock held).
+	 * wait queue that would be passed to poll_wait().
 	 */
 	int (*wait_unlocked)(struct i915_perf_stream *stream);
 
@@ -1816,11 +1814,28 @@ struct i915_perf_stream {
 	struct list_head link;
 
 	u32 sample_flags;
+	int sample_size;
 
 	struct i915_gem_context *ctx;
 	bool enabled;
 
-	struct i915_perf_stream_ops *ops;
+	const struct i915_perf_stream_ops *ops;
+};
+
+struct i915_oa_ops {
+	void (*init_oa_buffer)(struct drm_i915_private *dev_priv);
+	int (*enable_metric_set)(struct drm_i915_private *dev_priv);
+	void (*disable_metric_set)(struct drm_i915_private *dev_priv);
+	void (*oa_enable)(struct drm_i915_private *dev_priv);
+	void (*oa_disable)(struct drm_i915_private *dev_priv);
+	void (*update_oacontrol)(struct drm_i915_private *dev_priv);
+	void (*update_hw_ctx_id_locked)(struct drm_i915_private *dev_priv,
+					u32 ctx_id);
+	int (*read)(struct i915_perf_stream *stream,
+		    char __user *buf,
+		    size_t count,
+		    size_t *offset);
+	bool (*oa_buffer_is_empty)(struct drm_i915_private *dev_priv);
 };
 
 struct drm_i915_private {
@@ -2125,16 +2140,48 @@ struct drm_i915_private {
 
 	struct {
 		bool initialized;
+
 		struct mutex lock;
 		struct list_head streams;
 
+		spinlock_t hook_lock;
+
 		struct {
-			u32 metrics_set;
+			struct i915_perf_stream *exclusive_stream;
+
+			u32 specific_ctx_id;
+
+			struct hrtimer poll_check_timer;
+			wait_queue_head_t poll_wq;
+			atomic_t pollin;
+
+			bool periodic;
+			int period_exponent;
+			int timestamp_frequency;
+
+			int tail_margin;
+
+			int metrics_set;
 
 			const struct i915_oa_reg *mux_regs;
 			int mux_regs_len;
 			const struct i915_oa_reg *b_counter_regs;
 			int b_counter_regs_len;
+
+			struct {
+				struct drm_i915_gem_object *obj;
+				struct i915_vma *vma;
+				u32 gtt_offset;
+				u8 *addr;
+				int format;
+				int format_size;
+			} oa_buffer;
+
+			u32 gen7_latched_oastatus1;
+
+			struct i915_oa_ops ops;
+			const struct i915_oa_format *oa_formats;
+			int n_builtin_sets;
 		} oa;
 	} perf;
 
@@ -3499,6 +3546,9 @@ struct drm_i915_gem_object *
 i915_gem_alloc_context_obj(struct drm_device *dev, size_t size);
 struct i915_gem_context *
 i915_gem_context_create_gvt(struct drm_device *dev);
+int i915_gem_context_pin_legacy_rcs_state(struct drm_i915_private *dev_priv,
+					  struct i915_gem_context *ctx,
+					  u64 flags);
 
 static inline struct i915_gem_context *
 i915_gem_context_lookup(struct drm_i915_file_private *file_priv, u32 id)
@@ -3545,6 +3595,8 @@ int i915_gem_context_reset_stats_ioctl(struct drm_device *dev, void *data,
 
 int i915_perf_open_ioctl(struct drm_device *dev, void *data,
 			 struct drm_file *file);
+void i915_oa_legacy_context_pin_notify(struct drm_i915_private *dev_priv,
+				       struct i915_gem_context *ctx);
 
 /* i915_gem_evict.c */
 int __must_check i915_gem_evict_something(struct i915_address_space *vm,
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index df10f4e9..62966e9 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -134,6 +134,25 @@ static int get_context_size(struct drm_i915_private *dev_priv)
 	return ret;
 }
 
+int i915_gem_context_pin_legacy_rcs_state(struct drm_i915_private *dev_priv,
+					  struct i915_gem_context *ctx,
+					  u64 flags)
+{
+	int ret;
+
+	lockdep_assert_held(&ctx->i915->drm.struct_mutex);
+
+	ret = i915_vma_pin(ctx->engine[RCS].state, 0,
+			   ctx->ggtt_alignment,
+			   flags);
+	if (ret)
+		return ret;
+
+	i915_oa_legacy_context_pin_notify(dev_priv, ctx);
+
+	return 0;
+}
+
 void i915_gem_context_free(struct kref *ctx_ref)
 {
 	struct i915_gem_context *ctx = container_of(ctx_ref, typeof(*ctx), ref);
@@ -770,7 +789,8 @@ static int do_rcs_switch(struct drm_i915_gem_request *req)
 	}
 
 	/* Trying to pin first makes error handling easier. */
-	ret = i915_vma_pin(vma, 0, to->ggtt_alignment, PIN_GLOBAL);
+	ret = i915_gem_context_pin_legacy_rcs_state(engine->i915, to,
+						    PIN_GLOBAL);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
index 87530f5..5305982 100644
--- a/drivers/gpu/drm/i915/i915_perf.c
+++ b/drivers/gpu/drm/i915/i915_perf.c
@@ -28,14 +28,860 @@
 #include <linux/sizes.h>
 
 #include "i915_drv.h"
+#include "intel_ringbuffer.h"
+#include "intel_lrc.h"
+#include "i915_oa_hsw.h"
+
+/* Must be a power of two */
+#define OA_BUFFER_SIZE		SZ_16M
+#define OA_TAKEN(tail, head)	((tail - head) & (OA_BUFFER_SIZE - 1))
+
+/* There's a HW race condition between OA unit tail pointer register updates and
+ * writes to memory whereby the tail pointer can sometimes get ahead of what's
+ * been written out to the OA buffer so far.
+ *
+ * Although this can be observed explicitly by checking for a zeroed report-id
+ * field in tail reports, it seems preferable to account for this earlier e.g.
+ * as part of the _oa_buffer_is_empty checks to minimize -EAGAIN polling cycles
+ * in this situation.
+ *
+ * To give time for the most recent reports to land before they may be copied to
+ * userspace, the driver operates as if the tail pointer effectively lags behind
+ * the HW tail pointer by 'tail_margin' bytes. The margin in bytes is calculated
+ * based on this constant in nanoseconds, the current OA sampling exponent
+ * and current report size.
+ *
+ * There is also a fallback check while reading to simply skip over reports with
+ * a zeroed report-id.
+ */
+#define OA_TAIL_MARGIN_NSEC	100000ULL
+
+/* frequency for checking whether the OA unit has written new reports to the
+ * circular OA buffer...
+ */
+#define POLL_FREQUENCY 200
+#define POLL_PERIOD (NSEC_PER_SEC / POLL_FREQUENCY)
+
+/* The maximum exponent the hardware accepts is 63 (essentially it selects one
+ * of the 64bit timestamp bits to trigger reports from) but there's currently
+ * no known use case for sampling as infrequently as once per 47 thousand years.
+ *
+ * Since the timestamps included in OA reports are only 32bits it seems
+ * reasonable to limit the OA exponent where it's still possible to account for
+ * overflow in OA report timestamps.
+ */
+#define OA_EXPONENT_MAX 31
+
+/* XXX: beware if future OA HW adds new report formats that the current
+ * code assumes all reports have a power-of-two size and ~(size - 1) can
+ * be used as a mask to align the OA tail pointer.
+ */
+static struct i915_oa_format hsw_oa_formats[I915_OA_FORMAT_MAX] = {
+	[I915_OA_FORMAT_A13]	    = { 0, 64 },
+	[I915_OA_FORMAT_A29]	    = { 1, 128 },
+	[I915_OA_FORMAT_A13_B8_C8]  = { 2, 128 },
+	/* A29_B8_C8 Disallowed as 192 bytes doesn't factor into buffer size */
+	[I915_OA_FORMAT_B4_C8]	    = { 4, 64 },
+	[I915_OA_FORMAT_A45_B8_C8]  = { 5, 256 },
+	[I915_OA_FORMAT_B4_C8_A16]  = { 6, 128 },
+	[I915_OA_FORMAT_C4_B8]	    = { 7, 64 },
+};
+
+#define SAMPLE_OA_REPORT      (1<<0)
 
 struct perf_open_properties {
 	u32 sample_flags;
 
 	u64 single_context:1;
 	u64 ctx_handle;
+
+	/* OA sampling state */
+	int metrics_set;
+	int oa_format;
+	bool oa_periodic;
+	int oa_period_exponent;
 };
 
+/* NB: This is either called via fops or the poll check hrtimer (atomic ctx)
+ *
+ * It's safe to read OA config state here unlocked, assuming that this is only
+ * called while the stream is enabled, while the global OA configuration can't
+ * be modified.
+ *
+ * Note: we don't lock around the head/tail reads even though there's the slim
+ * possibility of read() fop errors forcing a re-init of the OA buffer
+ * pointers.  A race here could result in a false positive !empty status which
+ * is acceptable.
+ */
+static bool gen7_oa_buffer_is_empty_fop_unlocked(struct drm_i915_private *dev_priv)
+{
+	int report_size = dev_priv->perf.oa.oa_buffer.format_size;
+	u32 oastatus2 = I915_READ(GEN7_OASTATUS2);
+	u32 oastatus1 = I915_READ(GEN7_OASTATUS1);
+	u32 head = oastatus2 & GEN7_OASTATUS2_HEAD_MASK;
+	u32 tail = oastatus1 & GEN7_OASTATUS1_TAIL_MASK;
+
+	return OA_TAKEN(tail, head) <
+		dev_priv->perf.oa.tail_margin + report_size;
+}
+
+/**
+ * Appends a status record to a userspace read() buffer.
+ */
+static int append_oa_status(struct i915_perf_stream *stream,
+			    char __user *buf,
+			    size_t count,
+			    size_t *offset,
+			    enum drm_i915_perf_record_type type)
+{
+	struct drm_i915_perf_record_header header = { type, 0, sizeof(header) };
+
+	if ((count - *offset) < header.size)
+		return -ENOSPC;
+
+	if (copy_to_user(buf + *offset, &header, sizeof(header)))
+		return -EFAULT;
+
+	(*offset) += header.size;
+
+	return 0;
+}
+
+/**
+ * Copies single OA report into userspace read() buffer.
+ */
+static int append_oa_sample(struct i915_perf_stream *stream,
+			    char __user *buf,
+			    size_t count,
+			    size_t *offset,
+			    const u8 *report)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+	int report_size = dev_priv->perf.oa.oa_buffer.format_size;
+	struct drm_i915_perf_record_header header;
+	u32 sample_flags = stream->sample_flags;
+
+	header.type = DRM_I915_PERF_RECORD_SAMPLE;
+	header.pad = 0;
+	header.size = stream->sample_size;
+
+	if ((count - *offset) < header.size)
+		return -ENOSPC;
+
+	buf += *offset;
+	if (copy_to_user(buf, &header, sizeof(header)))
+		return -EFAULT;
+	buf += sizeof(header);
+
+	if (sample_flags & SAMPLE_OA_REPORT) {
+		if (copy_to_user(buf, report, report_size))
+			return -EFAULT;
+	}
+
+	(*offset) += header.size;
+
+	return 0;
+}
+
+/**
+ * Copies all buffered OA reports into userspace read() buffer.
+ * @head_ptr: (inout): the head pointer before and after appending
+ *
+ * Returns 0 on success, negative error code on failure.
+ *
+ * Notably any error condition resulting in a short read (-ENOSPC or
+ * -EFAULT) will be returned even though one or more records may
+ * have been successfully copied. In this case it's up to the caller
+ * to decide if the error should be squashed before returning to
+ * userspace.
+ */
+static int gen7_append_oa_reports(struct i915_perf_stream *stream,
+				  char __user *buf,
+				  size_t count,
+				  size_t *offset,
+				  u32 *head_ptr,
+				  u32 tail)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+	int report_size = dev_priv->perf.oa.oa_buffer.format_size;
+	u8 *oa_buf_base = dev_priv->perf.oa.oa_buffer.addr;
+	int tail_margin = dev_priv->perf.oa.tail_margin;
+	u32 mask = (OA_BUFFER_SIZE - 1);
+	u32 head;
+	u32 taken;
+	int ret = 0;
+
+	BUG_ON(!stream->enabled);
+
+	head = *head_ptr - dev_priv->perf.oa.oa_buffer.gtt_offset;
+	tail -= dev_priv->perf.oa.oa_buffer.gtt_offset;
+
+	/* The OA unit is expected to wrap the tail pointer according to the OA
+	 * buffer size and since we should never write a misaligned head
+	 * pointer we don't expect to read one back either...
+	 */
+	if (tail > OA_BUFFER_SIZE || head > OA_BUFFER_SIZE ||
+	    head % report_size) {
+		DRM_ERROR("Inconsistent OA buffer pointer (head = %u, tail = %u): force restart",
+			  head, tail);
+		dev_priv->perf.oa.ops.oa_disable(dev_priv);
+		dev_priv->perf.oa.ops.oa_enable(dev_priv);
+		*head_ptr = I915_READ(GEN7_OASTATUS2) &
+			GEN7_OASTATUS2_HEAD_MASK;
+		return -EIO;
+	}
+
+
+	/* The tail pointer increases in 64 byte increments, not in report_size
+	 * steps...
+	 */
+	tail &= ~(report_size - 1);
+
+	/* Move the tail pointer back by the current tail_margin to account for
+	 * the possibility that the latest reports may not have really landed
+	 * in memory yet...
+	 */
+
+	if (OA_TAKEN(tail, head) < report_size + tail_margin)
+		return -EAGAIN;
+
+	tail -= tail_margin;
+	tail &= mask;
+
+	for (/* none */;
+	     (taken = OA_TAKEN(tail, head));
+	     head = (head + report_size) & mask) {
+		u8 *report = oa_buf_base + head;
+		u32 *report32 = (void *)report;
+
+		/* All the report sizes factor neatly into the buffer
+		 * size so we never expect to see a report split
+		 * between the beginning and end of the buffer.
+		 *
+		 * Given the initial alignment check a misalignment
+		 * here would imply a driver bug that would result
+		 * in an overrun.
+		 */
+		BUG_ON((OA_BUFFER_SIZE - head) < report_size);
+
+		/* The report-ID field for periodic samples includes
+		 * some undocumented flags related to what triggered
+		 * the report and is never expected to be zero so we
+		 * can check that the report isn't invalid before
+		 * copying it to userspace...
+		 */
+		if (report32[0] == 0) {
+			DRM_ERROR("Skipping spurious, invalid OA report\n");
+			continue;
+		}
+
+		ret = append_oa_sample(stream, buf, count, offset, report);
+		if (ret)
+			break;
+
+		/* The above report-id field sanity check is based on
+		 * the assumption that the OA buffer is initially
+		 * zeroed and we reset the field after copying so the
+		 * check is still meaningful once old reports start
+		 * being overwritten.
+		 */
+		report32[0] = 0;
+	}
+
+	*head_ptr = dev_priv->perf.oa.oa_buffer.gtt_offset + head;
+
+	return ret;
+}
+
+static int gen7_oa_read(struct i915_perf_stream *stream,
+			char __user *buf,
+			size_t count,
+			size_t *offset)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+	int report_size = dev_priv->perf.oa.oa_buffer.format_size;
+	u32 oastatus2;
+	u32 oastatus1;
+	u32 head;
+	u32 tail;
+	int ret;
+
+	BUG_ON(!dev_priv->perf.oa.oa_buffer.addr);
+
+	oastatus2 = I915_READ(GEN7_OASTATUS2);
+	oastatus1 = I915_READ(GEN7_OASTATUS1);
+
+	head = oastatus2 & GEN7_OASTATUS2_HEAD_MASK;
+	tail = oastatus1 & GEN7_OASTATUS1_TAIL_MASK;
+
+	/* XXX: On Haswell we don't have a safe way to clear oastatus1
+	 * bits while the OA unit is enabled (while the tail pointer
+	 * may be updated asynchronously) so we ignore status bits
+	 * that have already been reported to userspace.
+	 */
+	oastatus1 &= ~dev_priv->perf.oa.gen7_latched_oastatus1;
+
+	/* We treat OABUFFER_OVERFLOW as a significant error:
+	 *
+	 * - The status can be interpreted to mean that the buffer is
+	 *   currently full (with a higher precedence than OA_TAKEN()
+	 *   which will start to report a near-empty buffer after an
+	 *   overflow) but it's awkward that we can't clear the status
+	 *   on Haswell, so without a reset we won't be able to catch
+	 *   the state again.
+	 *
+	 * - Since it also implies the HW has started overwriting old
+	 *   reports it may also affect our sanity checks for invalid
+	 *   reports when copying to userspace that assume new reports
+	 *   are being written to cleared memory.
+	 *
+	 * - In the future we may want to introduce a flight recorder
+	 *   mode where the driver will automatically maintain a safe
+	 *   guard band between head/tail, avoiding this overflow
+	 *   condition, but we avoid the added driver complexity for
+	 *   now.
+	 */
+	if (unlikely(oastatus1 & GEN7_OASTATUS1_OABUFFER_OVERFLOW)) {
+		ret = append_oa_status(stream, buf, count, offset,
+				       DRM_I915_PERF_RECORD_OA_BUFFER_LOST);
+		if (ret)
+			return ret;
+
+		DRM_ERROR("OA buffer overflow: force restart");
+
+		dev_priv->perf.oa.ops.oa_disable(dev_priv);
+		dev_priv->perf.oa.ops.oa_enable(dev_priv);
+
+		oastatus2 = I915_READ(GEN7_OASTATUS2);
+		oastatus1 = I915_READ(GEN7_OASTATUS1);
+
+		head = oastatus2 & GEN7_OASTATUS2_HEAD_MASK;
+		tail = oastatus1 & GEN7_OASTATUS1_TAIL_MASK;
+	}
+
+	if (unlikely(oastatus1 & GEN7_OASTATUS1_REPORT_LOST)) {
+		ret = append_oa_status(stream, buf, count, offset,
+				       DRM_I915_PERF_RECORD_OA_REPORT_LOST);
+		if (ret)
+			return ret;
+		dev_priv->perf.oa.gen7_latched_oastatus1 |=
+			GEN7_OASTATUS1_REPORT_LOST;
+	}
+
+	ret = gen7_append_oa_reports(stream, buf, count, offset,
+				     &head, tail);
+
+	/* All the report sizes are a power of two and the
+	 * head should always be incremented by some multiple
+	 * of the report size.
+	 *
+	 * A warning here, but notably if we later read back a
+	 * misaligned pointer we will treat that as a bug since
+	 * it could lead to a buffer overrun.
+	 */
+	WARN_ONCE(head & (report_size - 1),
+		  "i915: Writing misaligned OA head pointer");
+
+	/* Note: we update the head pointer here even if an error
+	 * was returned since the error may represent a short read
+	 * where some some reports were successfully copied.
+	 */
+	I915_WRITE(GEN7_OASTATUS2,
+		   ((head & GEN7_OASTATUS2_HEAD_MASK) |
+		    OA_MEM_SELECT_GGTT));
+
+	return ret;
+}
+
+static int i915_oa_wait_unlocked(struct i915_perf_stream *stream)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+
+	/* We would wait indefinitly if periodic sampling is not enabled */
+	if (!dev_priv->perf.oa.periodic)
+		return -EIO;
+
+	/* Note: the oa_buffer_is_empty() condition is ok to run unlocked as it
+	 * just performs mmio reads of the OA buffer head + tail pointers and
+	 * it's assumed we're handling some operation that implies the stream
+	 * can't be destroyed until completion (such as a read()) that ensures
+	 * the device + OA buffer can't disappear
+	 */
+	return wait_event_interruptible(dev_priv->perf.oa.poll_wq,
+					!dev_priv->perf.oa.ops.oa_buffer_is_empty(dev_priv));
+}
+
+static void i915_oa_poll_wait(struct i915_perf_stream *stream,
+			      struct file *file,
+			      poll_table *wait)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+
+	poll_wait(file, &dev_priv->perf.oa.poll_wq, wait);
+}
+
+static int i915_oa_read(struct i915_perf_stream *stream,
+			char __user *buf,
+			size_t count,
+			size_t *offset)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+
+	return dev_priv->perf.oa.ops.read(stream, buf, count, offset);
+}
+
+static void
+free_oa_buffer(struct drm_i915_private *i915)
+{
+	mutex_lock(&i915->drm.struct_mutex);
+
+	i915_gem_object_unpin_map(i915->perf.oa.oa_buffer.obj);
+	__i915_vma_unpin(i915->perf.oa.oa_buffer.vma);
+	i915_gem_object_put(i915->perf.oa.oa_buffer.obj);
+
+	i915->perf.oa.oa_buffer.obj = NULL;
+	i915->perf.oa.oa_buffer.gtt_offset = 0;
+	i915->perf.oa.oa_buffer.addr = NULL;
+
+	mutex_unlock(&i915->drm.struct_mutex);
+}
+
+static void i915_oa_stream_destroy(struct i915_perf_stream *stream)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+
+	BUG_ON(stream != dev_priv->perf.oa.exclusive_stream);
+
+	dev_priv->perf.oa.ops.disable_metric_set(dev_priv);
+
+	free_oa_buffer(dev_priv);
+
+	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
+	intel_runtime_pm_put(dev_priv);
+
+	dev_priv->perf.oa.exclusive_stream = NULL;
+}
+
+static void gen7_init_oa_buffer(struct drm_i915_private *dev_priv)
+{
+	/* Pre-DevBDW: OABUFFER must be set with counters off,
+	 * before OASTATUS1, but after OASTATUS2
+	 */
+	I915_WRITE(GEN7_OASTATUS2, dev_priv->perf.oa.oa_buffer.gtt_offset |
+		   OA_MEM_SELECT_GGTT); /* head */
+	I915_WRITE(GEN7_OABUFFER, dev_priv->perf.oa.oa_buffer.gtt_offset);
+	I915_WRITE(GEN7_OASTATUS1, dev_priv->perf.oa.oa_buffer.gtt_offset |
+		   OABUFFER_SIZE_16M); /* tail */
+
+	/* On Haswell we have to track which OASTATUS1 flags we've
+	 * already seen since they can't be cleared while periodic
+	 * sampling is enabled.
+	 */
+	dev_priv->perf.oa.gen7_latched_oastatus1 = 0;
+
+	/* NB: although the OA buffer will initially be allocated
+	 * zeroed via shmfs (and so this memset is redundant when
+	 * first allocating), we may re-init the OA buffer, either
+	 * when re-enabling a stream or in error/reset paths.
+	 *
+	 * The reason we clear the buffer for each re-init is for the
+	 * sanity check in gen7_append_oa_reports() that looks at the
+	 * report-id field to make sure it's non-zero which relies on
+	 * the assumption that new reports are being written to zeroed
+	 * memory...
+	 */
+	memset(dev_priv->perf.oa.oa_buffer.addr, 0, SZ_16M);
+
+	/* Maybe make ->pollin per-stream state if we support multiple
+	 * concurrent streams in the future. */
+	atomic_set(&dev_priv->perf.oa.pollin, false);
+}
+
+static int alloc_oa_buffer(struct drm_i915_private *dev_priv)
+{
+	struct drm_i915_gem_object *bo;
+	enum i915_map_type map;
+	struct i915_vma *vma;
+	int ret;
+
+	BUG_ON(dev_priv->perf.oa.oa_buffer.obj);
+
+	ret = i915_mutex_lock_interruptible(&dev_priv->drm);
+	if (ret)
+		return ret;
+
+	bo = i915_gem_object_create(&dev_priv->drm, OA_BUFFER_SIZE);
+	if (IS_ERR(bo)) {
+		DRM_ERROR("Failed to allocate OA buffer\n");
+		ret = PTR_ERR(bo);
+		goto unlock;
+	}
+	dev_priv->perf.oa.oa_buffer.obj = bo;
+
+	ret = i915_gem_object_set_cache_level(bo, I915_CACHE_LLC);
+	if (ret)
+		goto err_unref;
+
+	/* PreHSW required 512K alignment, HSW requires 16M */
+	vma = i915_gem_object_ggtt_pin(bo, NULL, 0, SZ_16M, PIN_MAPPABLE);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto err_unref;
+	}
+	dev_priv->perf.oa.oa_buffer.vma = vma;
+
+	map = HAS_LLC(dev_priv) ? I915_MAP_WB : I915_MAP_WC;
+
+	dev_priv->perf.oa.oa_buffer.gtt_offset = i915_ggtt_offset(vma);
+	dev_priv->perf.oa.oa_buffer.addr = i915_gem_object_pin_map(bo, map);
+	if (dev_priv->perf.oa.oa_buffer.addr == NULL)
+		goto err_unpin;
+
+	dev_priv->perf.oa.ops.init_oa_buffer(dev_priv);
+
+	DRM_DEBUG_DRIVER("OA Buffer initialized, gtt offset = 0x%x, vaddr = %p",
+			 dev_priv->perf.oa.oa_buffer.gtt_offset,
+			 dev_priv->perf.oa.oa_buffer.addr);
+
+	goto unlock;
+
+err_unpin:
+	__i915_vma_unpin(vma);
+
+err_unref:
+	i915_gem_object_put(bo);
+
+	dev_priv->perf.oa.oa_buffer.gtt_offset = 0;
+	dev_priv->perf.oa.oa_buffer.addr = NULL;
+	dev_priv->perf.oa.oa_buffer.vma = NULL;
+	dev_priv->perf.oa.oa_buffer.obj = NULL;
+
+unlock:
+	mutex_unlock(&dev_priv->drm.struct_mutex);
+	return ret;
+}
+
+static void config_oa_regs(struct drm_i915_private *dev_priv,
+			   const struct i915_oa_reg *regs,
+			   int n_regs)
+{
+	int i;
+
+	for (i = 0; i < n_regs; i++) {
+		const struct i915_oa_reg *reg = regs + i;
+
+		I915_WRITE(reg->addr, reg->value);
+	}
+}
+
+static int hsw_enable_metric_set(struct drm_i915_private *dev_priv)
+{
+	int ret = i915_oa_select_metric_set_hsw(dev_priv);
+
+	if (ret)
+		return ret;
+
+	I915_WRITE(GDT_CHICKEN_BITS, (I915_READ(GDT_CHICKEN_BITS) |
+				      GT_NOA_ENABLE));
+
+	/* PRM:
+	 *
+	 * OA unit is using “crclk” for its functionality. When trunk
+	 * level clock gating takes place, OA clock would be gated,
+	 * unable to count the events from non-render clock domain.
+	 * Render clock gating must be disabled when OA is enabled to
+	 * count the events from non-render domain. Unit level clock
+	 * gating for RCS should also be disabled.
+	 */
+	I915_WRITE(GEN7_MISCCPCTL, (I915_READ(GEN7_MISCCPCTL) &
+				    ~GEN7_DOP_CLOCK_GATE_ENABLE));
+	I915_WRITE(GEN6_UCGCTL1, (I915_READ(GEN6_UCGCTL1) |
+				  GEN6_CSUNIT_CLOCK_GATE_DISABLE));
+
+	config_oa_regs(dev_priv, dev_priv->perf.oa.mux_regs,
+		       dev_priv->perf.oa.mux_regs_len);
+
+	/* It apparently takes a fairly long time for a new MUX
+	 * configuration to be be applied after these register writes.
+	 * This delay duration was derived empirically based on the
+	 * render_basic config but hopefully it covers the maximum
+	 * configuration latency.
+	 *
+	 * As a fallback, the checks in _append_oa_reports() to skip
+	 * invalid OA reports do also seem to work to discard reports
+	 * generated before this config has completed - albeit not
+	 * silently.
+	 *
+	 * Unfortunately this is essentially a magic number, since we
+	 * don't currently know of a reliable mechanism for predicting
+	 * how long the MUX config will take to apply and besides
+	 * seeing invalid reports we don't know of a reliable way to
+	 * explicitly check that the MUX config has landed.
+	 *
+	 * It's even possible we've miss characterized the underlying
+	 * problem - it just seems like the simplest explanation why
+	 * a delay at this location would mitigate any invalid reports.
+	 */
+	usleep_range(15000, 20000);
+
+	config_oa_regs(dev_priv, dev_priv->perf.oa.b_counter_regs,
+		       dev_priv->perf.oa.b_counter_regs_len);
+
+	return 0;
+}
+
+static void hsw_disable_metric_set(struct drm_i915_private *dev_priv)
+{
+	I915_WRITE(GEN6_UCGCTL1, (I915_READ(GEN6_UCGCTL1) &
+				  ~GEN6_CSUNIT_CLOCK_GATE_DISABLE));
+	I915_WRITE(GEN7_MISCCPCTL, (I915_READ(GEN7_MISCCPCTL) |
+				    GEN7_DOP_CLOCK_GATE_ENABLE));
+
+	I915_WRITE(GDT_CHICKEN_BITS, (I915_READ(GDT_CHICKEN_BITS) &
+				      ~GT_NOA_ENABLE));
+}
+
+static void gen7_update_oacontrol_locked(struct drm_i915_private *dev_priv)
+{
+	assert_spin_locked(&dev_priv->perf.hook_lock);
+
+	if (dev_priv->perf.oa.exclusive_stream->enabled) {
+		unsigned long ctx_id = 0;
+
+		if (dev_priv->perf.oa.exclusive_stream->ctx)
+			ctx_id = dev_priv->perf.oa.specific_ctx_id;
+
+		if (dev_priv->perf.oa.exclusive_stream->ctx == NULL || ctx_id) {
+			bool periodic = dev_priv->perf.oa.periodic;
+			u32 period_exponent = dev_priv->perf.oa.period_exponent;
+			u32 report_format = dev_priv->perf.oa.oa_buffer.format;
+
+			I915_WRITE(GEN7_OACONTROL,
+				   (ctx_id & GEN7_OACONTROL_CTX_MASK) |
+				   (period_exponent <<
+				    GEN7_OACONTROL_TIMER_PERIOD_SHIFT) |
+				   (periodic ?
+				    GEN7_OACONTROL_TIMER_ENABLE : 0) |
+				   (report_format <<
+				    GEN7_OACONTROL_FORMAT_SHIFT) |
+				   (ctx_id ?
+				    GEN7_OACONTROL_PER_CTX_ENABLE : 0) |
+				   GEN7_OACONTROL_ENABLE);
+			return;
+		}
+	}
+
+	I915_WRITE(GEN7_OACONTROL, 0);
+}
+
+static void gen7_oa_enable(struct drm_i915_private *dev_priv)
+{
+	unsigned long flags;
+
+	/* Reset buf pointers so we don't forward reports from before now.
+	 *
+	 * Think carefully if considering trying to avoid this, since it
+	 * also ensures status flags and the buffer itself are cleared
+	 * in error paths, and we have checks for invalid reports based
+	 * on the assumption that certain fields are written to zeroed
+	 * memory which this helps maintains.
+	 */
+	gen7_init_oa_buffer(dev_priv);
+
+	spin_lock_irqsave(&dev_priv->perf.hook_lock, flags);
+	gen7_update_oacontrol_locked(dev_priv);
+	spin_unlock_irqrestore(&dev_priv->perf.hook_lock, flags);
+}
+
+static void i915_oa_stream_enable(struct i915_perf_stream *stream)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+
+	dev_priv->perf.oa.ops.oa_enable(dev_priv);
+
+	if (dev_priv->perf.oa.periodic)
+		hrtimer_start(&dev_priv->perf.oa.poll_check_timer,
+			      ns_to_ktime(POLL_PERIOD),
+			      HRTIMER_MODE_REL_PINNED);
+}
+
+static void gen7_oa_disable(struct drm_i915_private *dev_priv)
+{
+	I915_WRITE(GEN7_OACONTROL, 0);
+}
+
+static void i915_oa_stream_disable(struct i915_perf_stream *stream)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+
+	dev_priv->perf.oa.ops.oa_disable(dev_priv);
+
+	if (dev_priv->perf.oa.periodic)
+		hrtimer_cancel(&dev_priv->perf.oa.poll_check_timer);
+}
+
+static u64 oa_exponent_to_ns(struct drm_i915_private *dev_priv, int exponent)
+{
+	return 1000000000ULL * (2ULL << exponent) /
+		dev_priv->perf.oa.timestamp_frequency;
+}
+
+static const struct i915_perf_stream_ops i915_oa_stream_ops = {
+	.destroy = i915_oa_stream_destroy,
+	.enable = i915_oa_stream_enable,
+	.disable = i915_oa_stream_disable,
+	.wait_unlocked = i915_oa_wait_unlocked,
+	.poll_wait = i915_oa_poll_wait,
+	.read = i915_oa_read,
+};
+
+static int i915_oa_stream_init(struct i915_perf_stream *stream,
+			       struct drm_i915_perf_open_param *param,
+			       struct perf_open_properties *props)
+{
+	struct drm_i915_private *dev_priv = stream->dev_priv;
+	int format_size;
+	int ret;
+
+	if (!(props->sample_flags & SAMPLE_OA_REPORT)) {
+		DRM_ERROR("Only OA report sampling supported\n");
+		return -EINVAL;
+	}
+
+	if (!dev_priv->perf.oa.ops.init_oa_buffer) {
+		DRM_ERROR("OA unit not supported\n");
+		return -ENODEV;
+	}
+
+	/* To avoid the complexity of having to accurately filter
+	 * counter reports and marshal to the appropriate client
+	 * we currently only allow exclusive access
+	 */
+	if (dev_priv->perf.oa.exclusive_stream) {
+		DRM_ERROR("OA unit already in use\n");
+		return -EBUSY;
+	}
+
+	if (!props->metrics_set) {
+		DRM_ERROR("OA metric set not specified\n");
+		return -EINVAL;
+	}
+
+	if (!props->oa_format) {
+		DRM_ERROR("OA report format not specified\n");
+		return -EINVAL;
+	}
+
+	stream->sample_size = sizeof(struct drm_i915_perf_record_header);
+
+	format_size = dev_priv->perf.oa.oa_formats[props->oa_format].size;
+
+	stream->sample_flags |= SAMPLE_OA_REPORT;
+	stream->sample_size += format_size;
+
+	dev_priv->perf.oa.oa_buffer.format_size = format_size;
+	BUG_ON(dev_priv->perf.oa.oa_buffer.format_size == 0);
+
+	dev_priv->perf.oa.oa_buffer.format =
+		dev_priv->perf.oa.oa_formats[props->oa_format].format;
+
+	dev_priv->perf.oa.metrics_set = props->metrics_set;
+
+	dev_priv->perf.oa.periodic = props->oa_periodic;
+	if (dev_priv->perf.oa.periodic) {
+		u64 period_ns = oa_exponent_to_ns(dev_priv,
+						  props->oa_period_exponent);
+
+		dev_priv->perf.oa.period_exponent = props->oa_period_exponent;
+
+		/* See comment for OA_TAIL_MARGIN_NSEC for details
+		 * about this tail_margin...
+		 */
+		dev_priv->perf.oa.tail_margin =
+			((OA_TAIL_MARGIN_NSEC / period_ns) + 1) * format_size;
+	}
+
+	ret = alloc_oa_buffer(dev_priv);
+	if (ret)
+		return ret;
+
+	/* PRM - observability performance counters:
+	 *
+	 *   OACONTROL, performance counter enable, note:
+	 *
+	 *   "When this bit is set, in order to have coherent counts,
+	 *   RC6 power state and trunk clock gating must be disabled.
+	 *   This can be achieved by programming MMIO registers as
+	 *   0xA094=0 and 0xA090[31]=1"
+	 *
+	 *   In our case we are expecting that taking pm + FORCEWAKE
+	 *   references will effectively disable RC6.
+	 */
+	intel_runtime_pm_get(dev_priv);
+	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);
+
+	ret = dev_priv->perf.oa.ops.enable_metric_set(dev_priv);
+	if (ret) {
+		intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
+		intel_runtime_pm_put(dev_priv);
+		free_oa_buffer(dev_priv);
+		return ret;
+	}
+
+	stream->ops = &i915_oa_stream_ops;
+
+	/* On Haswell we have to track which OASTATUS1 flags we've already
+	 * seen since they can't be cleared while periodic sampling is enabled.
+	 */
+	dev_priv->perf.oa.gen7_latched_oastatus1 = 0;
+
+	dev_priv->perf.oa.exclusive_stream = stream;
+
+	return 0;
+}
+
+static void gen7_update_hw_ctx_id_locked(struct drm_i915_private *dev_priv,
+					 u32 ctx_id)
+{
+	assert_spin_locked(&dev_priv->perf.hook_lock);
+
+	dev_priv->perf.oa.specific_ctx_id = ctx_id;
+	gen7_update_oacontrol_locked(dev_priv);
+}
+
+static void
+i915_oa_legacy_context_pin_notify_locked(struct drm_i915_private *dev_priv,
+					 struct i915_gem_context *ctx)
+{
+	assert_spin_locked(&dev_priv->perf.hook_lock);
+
+	BUG_ON(i915.enable_execlists);
+
+	if (dev_priv->perf.oa.ops.update_hw_ctx_id_locked == NULL)
+		return;
+
+	if (dev_priv->perf.oa.exclusive_stream &&
+	    dev_priv->perf.oa.exclusive_stream->ctx == ctx) {
+		struct i915_vma *vma = ctx->engine[RCS].state;
+		u32 ctx_id = i915_ggtt_offset(vma);
+
+		dev_priv->perf.oa.ops.update_hw_ctx_id_locked(dev_priv, ctx_id);
+	}
+}
+
+void i915_oa_legacy_context_pin_notify(struct drm_i915_private *dev_priv,
+				       struct i915_gem_context *ctx)
+{
+	unsigned long flags;
+
+	if (!dev_priv->perf.initialized)
+		return;
+
+	spin_lock_irqsave(&dev_priv->perf.hook_lock, flags);
+	i915_oa_legacy_context_pin_notify_locked(dev_priv, ctx);
+	spin_unlock_irqrestore(&dev_priv->perf.hook_lock, flags);
+}
+
 static ssize_t i915_perf_read_locked(struct i915_perf_stream *stream,
 				     struct file *file,
 				     char __user *buf,
@@ -79,8 +925,20 @@ static ssize_t i915_perf_read(struct file *file,
 	struct drm_i915_private *dev_priv = stream->dev_priv;
 	ssize_t ret;
 
+	/* To ensure it's handled consistently we simply treat all reads of a
+	 * disabled stream as an error. In particular it might otherwise lead
+	 * to a deadlock for blocking file descriptors...
+	 */
+	if (!stream->enabled)
+		return -EIO;
+
 	if (!(file->f_flags & O_NONBLOCK)) {
-		/* Allow false positives from stream->ops->wait_unlocked.
+		/* There's the small chance of false positives from
+		 * stream->ops->wait_unlocked.
+		 *
+		 * E.g. with single context filtering since we only wait until
+		 * oabuffer has >= 1 report we don't immediately know whether
+		 * any reports really belong to the current context
 		 */
 		do {
 			ret = stream->ops->wait_unlocked(stream);
@@ -98,21 +956,50 @@ static ssize_t i915_perf_read(struct file *file,
 		mutex_unlock(&dev_priv->perf.lock);
 	}
 
+	if (ret >= 0) {
+		/* Maybe make ->pollin per-stream state if we support multiple
+		 * concurrent streams in the future. */
+		atomic_set(&dev_priv->perf.oa.pollin, false);
+	}
+
 	return ret;
 }
 
-static unsigned int i915_perf_poll_locked(struct i915_perf_stream *stream,
+static enum hrtimer_restart oa_poll_check_timer_cb(struct hrtimer *hrtimer)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(hrtimer, typeof(*dev_priv),
+			     perf.oa.poll_check_timer);
+
+	if (!dev_priv->perf.oa.ops.oa_buffer_is_empty(dev_priv)) {
+		atomic_set(&dev_priv->perf.oa.pollin, true);
+		wake_up(&dev_priv->perf.oa.poll_wq);
+	}
+
+	hrtimer_forward_now(hrtimer, ns_to_ktime(POLL_PERIOD));
+
+	return HRTIMER_RESTART;
+}
+
+static unsigned int i915_perf_poll_locked(struct drm_i915_private *dev_priv,
+					  struct i915_perf_stream *stream,
 					  struct file *file,
 					  poll_table *wait)
 {
-	unsigned int streams = 0;
+	unsigned int events = 0;
 
 	stream->ops->poll_wait(stream, file, wait);
 
-	if (stream->ops->can_read(stream))
-		streams |= POLLIN;
+	/* Note: we don't explicitly check whether there's something to read
+	 * here since this path may be very hot depending on what else
+	 * userspace is polling, or on the timeout in use. We rely solely on
+	 * the hrtimer/oa_poll_check_timer_cb to notify us when there are
+	 * samples to read.
+	 */
+	if (atomic_read(&dev_priv->perf.oa.pollin))
+		events |= POLLIN;
 
-	return streams;
+	return events;
 }
 
 static unsigned int i915_perf_poll(struct file *file, poll_table *wait)
@@ -122,7 +1009,7 @@ static unsigned int i915_perf_poll(struct file *file, poll_table *wait)
 	int ret;
 
 	mutex_lock(&dev_priv->perf.lock);
-	ret = i915_perf_poll_locked(stream, file, wait);
+	ret = i915_perf_poll_locked(dev_priv, stream, file, wait);
 	mutex_unlock(&dev_priv->perf.lock);
 
 	return ret;
@@ -286,18 +1173,18 @@ int i915_perf_open_ioctl_locked(struct drm_device *dev,
 		goto err_ctx;
 	}
 
-	stream->sample_flags = props->sample_flags;
 	stream->dev_priv = dev_priv;
 	stream->ctx = specific_ctx;
 
-	/*
-	 * TODO: support sampling something
-	 *
-	 * For now this is as far as we can go.
+	ret = i915_oa_stream_init(stream, param, props);
+	if (ret)
+		goto err_alloc;
+
+	/* we avoid simply assigning stream->sample_flags = props->sample_flags
+	 * to have _stream_init check the combination of sample flags more
+	 * thoroughly, but still this is the expected result at this point.
 	 */
-	DRM_ERROR("Unsupported i915 perf stream configuration\n");
-	ret = -EINVAL;
-	goto err_alloc;
+	BUG_ON(stream->sample_flags != props->sample_flags);
 
 	list_add(&stream->link, &dev_priv->perf.streams);
 
@@ -382,7 +1269,56 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv,
 			props->single_context = 1;
 			props->ctx_handle = value;
 			break;
-
+		case DRM_I915_PERF_PROP_SAMPLE_OA:
+			props->sample_flags |= SAMPLE_OA_REPORT;
+			break;
+		case DRM_I915_PERF_PROP_OA_METRICS_SET:
+			if (value == 0 ||
+			    value > dev_priv->perf.oa.n_builtin_sets) {
+				DRM_ERROR("Unknown OA metric set ID");
+				return -EINVAL;
+			}
+			props->metrics_set = value;
+			break;
+		case DRM_I915_PERF_PROP_OA_FORMAT:
+			if (value == 0 || value >= I915_OA_FORMAT_MAX) {
+				DRM_ERROR("Invalid OA report format\n");
+				return -EINVAL;
+			}
+			if (!dev_priv->perf.oa.oa_formats[value].size) {
+				DRM_ERROR("Invalid OA report format\n");
+				return -EINVAL;
+			}
+			props->oa_format = value;
+			break;
+		case DRM_I915_PERF_PROP_OA_EXPONENT:
+			if (value > OA_EXPONENT_MAX) {
+				DRM_ERROR("OA timer exponent too high (> %u)\n",
+					  OA_EXPONENT_MAX);
+				return -EINVAL;
+			}
+
+			/* NB: The exponent represents a period as follows:
+			 *
+			 *   80ns * 2^(period_exponent + 1)
+			 *
+			 * Theoretically we can program the OA unit to sample
+			 * every 160ns but don't allow that by default unless
+			 * root.
+			 *
+			 * Referring to perf's
+			 * kernel.perf_event_max_sample_rate for a precedent
+			 * (100000 by default); with an OA exponent of 6 we get
+			 * a period of 10.240 microseconds -just under 100000Hz
+			 */
+			if (value < 6 && !capable(CAP_SYS_ADMIN)) {
+				DRM_ERROR("Sampling period too high without root privileges\n");
+				return -EACCES;
+			}
+
+			props->oa_periodic = true;
+			props->oa_period_exponent = value;
+			break;
 		case DRM_I915_PERF_PROP_MAX:
 			BUG();
 		}
@@ -431,8 +1367,35 @@ int i915_perf_open_ioctl(struct drm_device *dev, void *data,
 
 void i915_perf_init(struct drm_i915_private *dev_priv)
 {
+	if (!IS_HASWELL(dev_priv))
+		return;
+
+	hrtimer_init(&dev_priv->perf.oa.poll_check_timer,
+		     CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	dev_priv->perf.oa.poll_check_timer.function = oa_poll_check_timer_cb;
+	init_waitqueue_head(&dev_priv->perf.oa.poll_wq);
+
 	INIT_LIST_HEAD(&dev_priv->perf.streams);
 	mutex_init(&dev_priv->perf.lock);
+	spin_lock_init(&dev_priv->perf.hook_lock);
+
+	dev_priv->perf.oa.ops.init_oa_buffer = gen7_init_oa_buffer;
+	dev_priv->perf.oa.ops.enable_metric_set = hsw_enable_metric_set;
+	dev_priv->perf.oa.ops.disable_metric_set = hsw_disable_metric_set;
+	dev_priv->perf.oa.ops.oa_enable = gen7_oa_enable;
+	dev_priv->perf.oa.ops.oa_disable = gen7_oa_disable;
+	dev_priv->perf.oa.ops.update_hw_ctx_id_locked =
+		gen7_update_hw_ctx_id_locked;
+	dev_priv->perf.oa.ops.read = gen7_oa_read;
+	dev_priv->perf.oa.ops.oa_buffer_is_empty =
+		gen7_oa_buffer_is_empty_fop_unlocked;
+
+	dev_priv->perf.oa.timestamp_frequency = 12500000;
+
+	dev_priv->perf.oa.oa_formats = hsw_oa_formats;
+
+	dev_priv->perf.oa.n_builtin_sets =
+		i915_oa_n_builtin_metric_sets_hsw;
 
 	dev_priv->perf.initialized = true;
 }
@@ -442,7 +1405,6 @@ void i915_perf_fini(struct drm_i915_private *dev_priv)
 	if (!dev_priv->perf.initialized)
 		return;
 
-	/* Currently nothing to clean up */
-
+	memset(&dev_priv->perf.oa.ops, 0, sizeof(dev_priv->perf.oa.ops));
 	dev_priv->perf.initialized = false;
 }
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 90756b2..a0ccc09 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -617,6 +617,343 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg)
 #define HSW_CS_GPR_UDW(n)               _MMIO(0x2600 + (n) * 8 + 4)
 
 #define GEN7_OACONTROL _MMIO(0x2360)
+#define  GEN7_OACONTROL_CTX_MASK	    0xFFFFF000
+#define  GEN7_OACONTROL_TIMER_PERIOD_MASK   0x3F
+#define  GEN7_OACONTROL_TIMER_PERIOD_SHIFT  6
+#define  GEN7_OACONTROL_TIMER_ENABLE	    (1<<5)
+#define  GEN7_OACONTROL_FORMAT_A13	    (0<<2)
+#define  GEN7_OACONTROL_FORMAT_A29	    (1<<2)
+#define  GEN7_OACONTROL_FORMAT_A13_B8_C8    (2<<2)
+#define  GEN7_OACONTROL_FORMAT_A29_B8_C8    (3<<2)
+#define  GEN7_OACONTROL_FORMAT_B4_C8	    (4<<2)
+#define  GEN7_OACONTROL_FORMAT_A45_B8_C8    (5<<2)
+#define  GEN7_OACONTROL_FORMAT_B4_C8_A16    (6<<2)
+#define  GEN7_OACONTROL_FORMAT_C4_B8	    (7<<2)
+#define  GEN7_OACONTROL_FORMAT_SHIFT	    2
+#define  GEN7_OACONTROL_PER_CTX_ENABLE	    (1<<1)
+#define  GEN7_OACONTROL_ENABLE		    (1<<0)
+
+#define GEN8_OACTXID _MMIO(0x2364)
+
+#define GEN8_OACONTROL _MMIO(0x2B00)
+#define  GEN8_OA_REPORT_FORMAT_A12	    (0<<2)
+#define  GEN8_OA_REPORT_FORMAT_A12_B8_C8    (2<<2)
+#define  GEN8_OA_REPORT_FORMAT_A36_B8_C8    (5<<2)
+#define  GEN8_OA_REPORT_FORMAT_C4_B8	    (7<<2)
+#define  GEN8_OA_REPORT_FORMAT_SHIFT	    2
+#define  GEN8_OA_SPECIFIC_CONTEXT_ENABLE    (1<<1)
+#define  GEN8_OA_COUNTER_ENABLE             (1<<0)
+
+#define GEN8_OACTXCONTROL _MMIO(0x2360)
+#define  GEN8_OA_TIMER_PERIOD_MASK	    0x3F
+#define  GEN8_OA_TIMER_PERIOD_SHIFT	    2
+#define  GEN8_OA_TIMER_ENABLE		    (1<<1)
+#define  GEN8_OA_COUNTER_RESUME		    (1<<0)
+
+#define GEN7_OABUFFER _MMIO(0x23B0) /* R/W */
+#define  GEN7_OABUFFER_OVERRUN_DISABLE	    (1<<3)
+#define  GEN7_OABUFFER_EDGE_TRIGGER	    (1<<2)
+#define  GEN7_OABUFFER_STOP_RESUME_ENABLE   (1<<1)
+#define  GEN7_OABUFFER_RESUME		    (1<<0)
+
+#define GEN8_OABUFFER _MMIO(0x2b14)
+
+#define GEN7_OASTATUS1 _MMIO(0x2364)
+#define  GEN7_OASTATUS1_TAIL_MASK	    0xffffffc0
+#define  GEN7_OASTATUS1_COUNTER_OVERFLOW    (1<<2)
+#define  GEN7_OASTATUS1_OABUFFER_OVERFLOW   (1<<1)
+#define  GEN7_OASTATUS1_REPORT_LOST	    (1<<0)
+
+#define GEN7_OASTATUS2 _MMIO(0x2368)
+#define GEN7_OASTATUS2_HEAD_MASK    0xffffffc0
+
+#define GEN8_OASTATUS _MMIO(0x2b08)
+#define  GEN8_OASTATUS_OVERRUN_STATUS	    (1<<3)
+#define  GEN8_OASTATUS_COUNTER_OVERFLOW     (1<<2)
+#define  GEN8_OASTATUS_OABUFFER_OVERFLOW    (1<<1)
+#define  GEN8_OASTATUS_REPORT_LOST	    (1<<0)
+
+#define GEN8_OAHEADPTR _MMIO(0x2B0C)
+#define GEN8_OATAILPTR _MMIO(0x2B10)
+
+#define OABUFFER_SIZE_128K  (0<<3)
+#define OABUFFER_SIZE_256K  (1<<3)
+#define OABUFFER_SIZE_512K  (2<<3)
+#define OABUFFER_SIZE_1M    (3<<3)
+#define OABUFFER_SIZE_2M    (4<<3)
+#define OABUFFER_SIZE_4M    (5<<3)
+#define OABUFFER_SIZE_8M    (6<<3)
+#define OABUFFER_SIZE_16M   (7<<3)
+
+#define OA_MEM_SELECT_GGTT  (1<<0)
+
+#define EU_PERF_CNTL0	    _MMIO(0xe458)
+
+#define GDT_CHICKEN_BITS    _MMIO(0x9840)
+#define GT_NOA_ENABLE	    0x00000080
+
+/*
+ * OA Boolean state
+ */
+
+#define OAREPORTTRIG1 _MMIO(0x2740)
+#define OAREPORTTRIG1_THRESHOLD_MASK 0xffff
+#define OAREPORTTRIG1_EDGE_LEVEL_TRIGER_SELECT_MASK 0xffff0000 /* 0=level */
+
+#define OAREPORTTRIG2 _MMIO(0x2744)
+#define OAREPORTTRIG2_INVERT_A_0  (1<<0)
+#define OAREPORTTRIG2_INVERT_A_1  (1<<1)
+#define OAREPORTTRIG2_INVERT_A_2  (1<<2)
+#define OAREPORTTRIG2_INVERT_A_3  (1<<3)
+#define OAREPORTTRIG2_INVERT_A_4  (1<<4)
+#define OAREPORTTRIG2_INVERT_A_5  (1<<5)
+#define OAREPORTTRIG2_INVERT_A_6  (1<<6)
+#define OAREPORTTRIG2_INVERT_A_7  (1<<7)
+#define OAREPORTTRIG2_INVERT_A_8  (1<<8)
+#define OAREPORTTRIG2_INVERT_A_9  (1<<9)
+#define OAREPORTTRIG2_INVERT_A_10 (1<<10)
+#define OAREPORTTRIG2_INVERT_A_11 (1<<11)
+#define OAREPORTTRIG2_INVERT_A_12 (1<<12)
+#define OAREPORTTRIG2_INVERT_A_13 (1<<13)
+#define OAREPORTTRIG2_INVERT_A_14 (1<<14)
+#define OAREPORTTRIG2_INVERT_A_15 (1<<15)
+#define OAREPORTTRIG2_INVERT_B_0  (1<<16)
+#define OAREPORTTRIG2_INVERT_B_1  (1<<17)
+#define OAREPORTTRIG2_INVERT_B_2  (1<<18)
+#define OAREPORTTRIG2_INVERT_B_3  (1<<19)
+#define OAREPORTTRIG2_INVERT_C_0  (1<<20)
+#define OAREPORTTRIG2_INVERT_C_1  (1<<21)
+#define OAREPORTTRIG2_INVERT_D_0  (1<<22)
+#define OAREPORTTRIG2_THRESHOLD_ENABLE	    (1<<23)
+#define OAREPORTTRIG2_REPORT_TRIGGER_ENABLE (1<<31)
+
+#define OAREPORTTRIG3 _MMIO(0x2748)
+#define OAREPORTTRIG3_NOA_SELECT_MASK	    0xf
+#define OAREPORTTRIG3_NOA_SELECT_8_SHIFT    0
+#define OAREPORTTRIG3_NOA_SELECT_9_SHIFT    4
+#define OAREPORTTRIG3_NOA_SELECT_10_SHIFT   8
+#define OAREPORTTRIG3_NOA_SELECT_11_SHIFT   12
+#define OAREPORTTRIG3_NOA_SELECT_12_SHIFT   16
+#define OAREPORTTRIG3_NOA_SELECT_13_SHIFT   20
+#define OAREPORTTRIG3_NOA_SELECT_14_SHIFT   24
+#define OAREPORTTRIG3_NOA_SELECT_15_SHIFT   28
+
+#define OAREPORTTRIG4 _MMIO(0x274c)
+#define OAREPORTTRIG4_NOA_SELECT_MASK	    0xf
+#define OAREPORTTRIG4_NOA_SELECT_0_SHIFT    0
+#define OAREPORTTRIG4_NOA_SELECT_1_SHIFT    4
+#define OAREPORTTRIG4_NOA_SELECT_2_SHIFT    8
+#define OAREPORTTRIG4_NOA_SELECT_3_SHIFT    12
+#define OAREPORTTRIG4_NOA_SELECT_4_SHIFT    16
+#define OAREPORTTRIG4_NOA_SELECT_5_SHIFT    20
+#define OAREPORTTRIG4_NOA_SELECT_6_SHIFT    24
+#define OAREPORTTRIG4_NOA_SELECT_7_SHIFT    28
+
+#define OAREPORTTRIG5 _MMIO(0x2750)
+#define OAREPORTTRIG5_THRESHOLD_MASK 0xffff
+#define OAREPORTTRIG5_EDGE_LEVEL_TRIGER_SELECT_MASK 0xffff0000 /* 0=level */
+
+#define OAREPORTTRIG6 _MMIO(0x2754)
+#define OAREPORTTRIG6_INVERT_A_0  (1<<0)
+#define OAREPORTTRIG6_INVERT_A_1  (1<<1)
+#define OAREPORTTRIG6_INVERT_A_2  (1<<2)
+#define OAREPORTTRIG6_INVERT_A_3  (1<<3)
+#define OAREPORTTRIG6_INVERT_A_4  (1<<4)
+#define OAREPORTTRIG6_INVERT_A_5  (1<<5)
+#define OAREPORTTRIG6_INVERT_A_6  (1<<6)
+#define OAREPORTTRIG6_INVERT_A_7  (1<<7)
+#define OAREPORTTRIG6_INVERT_A_8  (1<<8)
+#define OAREPORTTRIG6_INVERT_A_9  (1<<9)
+#define OAREPORTTRIG6_INVERT_A_10 (1<<10)
+#define OAREPORTTRIG6_INVERT_A_11 (1<<11)
+#define OAREPORTTRIG6_INVERT_A_12 (1<<12)
+#define OAREPORTTRIG6_INVERT_A_13 (1<<13)
+#define OAREPORTTRIG6_INVERT_A_14 (1<<14)
+#define OAREPORTTRIG6_INVERT_A_15 (1<<15)
+#define OAREPORTTRIG6_INVERT_B_0  (1<<16)
+#define OAREPORTTRIG6_INVERT_B_1  (1<<17)
+#define OAREPORTTRIG6_INVERT_B_2  (1<<18)
+#define OAREPORTTRIG6_INVERT_B_3  (1<<19)
+#define OAREPORTTRIG6_INVERT_C_0  (1<<20)
+#define OAREPORTTRIG6_INVERT_C_1  (1<<21)
+#define OAREPORTTRIG6_INVERT_D_0  (1<<22)
+#define OAREPORTTRIG6_THRESHOLD_ENABLE	    (1<<23)
+#define OAREPORTTRIG6_REPORT_TRIGGER_ENABLE (1<<31)
+
+#define OAREPORTTRIG7 _MMIO(0x2758)
+#define OAREPORTTRIG7_NOA_SELECT_MASK	    0xf
+#define OAREPORTTRIG7_NOA_SELECT_8_SHIFT    0
+#define OAREPORTTRIG7_NOA_SELECT_9_SHIFT    4
+#define OAREPORTTRIG7_NOA_SELECT_10_SHIFT   8
+#define OAREPORTTRIG7_NOA_SELECT_11_SHIFT   12
+#define OAREPORTTRIG7_NOA_SELECT_12_SHIFT   16
+#define OAREPORTTRIG7_NOA_SELECT_13_SHIFT   20
+#define OAREPORTTRIG7_NOA_SELECT_14_SHIFT   24
+#define OAREPORTTRIG7_NOA_SELECT_15_SHIFT   28
+
+#define OAREPORTTRIG8 _MMIO(0x275c)
+#define OAREPORTTRIG8_NOA_SELECT_MASK	    0xf
+#define OAREPORTTRIG8_NOA_SELECT_0_SHIFT    0
+#define OAREPORTTRIG8_NOA_SELECT_1_SHIFT    4
+#define OAREPORTTRIG8_NOA_SELECT_2_SHIFT    8
+#define OAREPORTTRIG8_NOA_SELECT_3_SHIFT    12
+#define OAREPORTTRIG8_NOA_SELECT_4_SHIFT    16
+#define OAREPORTTRIG8_NOA_SELECT_5_SHIFT    20
+#define OAREPORTTRIG8_NOA_SELECT_6_SHIFT    24
+#define OAREPORTTRIG8_NOA_SELECT_7_SHIFT    28
+
+#define OASTARTTRIG1 _MMIO(0x2710)
+#define OASTARTTRIG1_THRESHOLD_COUNT_MASK_MBZ 0xffff0000
+#define OASTARTTRIG1_THRESHOLD_MASK	      0xffff
+
+#define OASTARTTRIG2 _MMIO(0x2714)
+#define OASTARTTRIG2_INVERT_A_0 (1<<0)
+#define OASTARTTRIG2_INVERT_A_1 (1<<1)
+#define OASTARTTRIG2_INVERT_A_2 (1<<2)
+#define OASTARTTRIG2_INVERT_A_3 (1<<3)
+#define OASTARTTRIG2_INVERT_A_4 (1<<4)
+#define OASTARTTRIG2_INVERT_A_5 (1<<5)
+#define OASTARTTRIG2_INVERT_A_6 (1<<6)
+#define OASTARTTRIG2_INVERT_A_7 (1<<7)
+#define OASTARTTRIG2_INVERT_A_8 (1<<8)
+#define OASTARTTRIG2_INVERT_A_9 (1<<9)
+#define OASTARTTRIG2_INVERT_A_10 (1<<10)
+#define OASTARTTRIG2_INVERT_A_11 (1<<11)
+#define OASTARTTRIG2_INVERT_A_12 (1<<12)
+#define OASTARTTRIG2_INVERT_A_13 (1<<13)
+#define OASTARTTRIG2_INVERT_A_14 (1<<14)
+#define OASTARTTRIG2_INVERT_A_15 (1<<15)
+#define OASTARTTRIG2_INVERT_B_0 (1<<16)
+#define OASTARTTRIG2_INVERT_B_1 (1<<17)
+#define OASTARTTRIG2_INVERT_B_2 (1<<18)
+#define OASTARTTRIG2_INVERT_B_3 (1<<19)
+#define OASTARTTRIG2_INVERT_C_0 (1<<20)
+#define OASTARTTRIG2_INVERT_C_1 (1<<21)
+#define OASTARTTRIG2_INVERT_D_0 (1<<22)
+#define OASTARTTRIG2_THRESHOLD_ENABLE	    (1<<23)
+#define OASTARTTRIG2_START_TRIG_FLAG_MBZ    (1<<24)
+#define OASTARTTRIG2_EVENT_SELECT_0  (1<<28)
+#define OASTARTTRIG2_EVENT_SELECT_1  (1<<29)
+#define OASTARTTRIG2_EVENT_SELECT_2  (1<<30)
+#define OASTARTTRIG2_EVENT_SELECT_3  (1<<31)
+
+#define OASTARTTRIG3 _MMIO(0x2718)
+#define OASTARTTRIG3_NOA_SELECT_MASK	   0xf
+#define OASTARTTRIG3_NOA_SELECT_8_SHIFT    0
+#define OASTARTTRIG3_NOA_SELECT_9_SHIFT    4
+#define OASTARTTRIG3_NOA_SELECT_10_SHIFT   8
+#define OASTARTTRIG3_NOA_SELECT_11_SHIFT   12
+#define OASTARTTRIG3_NOA_SELECT_12_SHIFT   16
+#define OASTARTTRIG3_NOA_SELECT_13_SHIFT   20
+#define OASTARTTRIG3_NOA_SELECT_14_SHIFT   24
+#define OASTARTTRIG3_NOA_SELECT_15_SHIFT   28
+
+#define OASTARTTRIG4 _MMIO(0x271c)
+#define OASTARTTRIG4_NOA_SELECT_MASK	    0xf
+#define OASTARTTRIG4_NOA_SELECT_0_SHIFT    0
+#define OASTARTTRIG4_NOA_SELECT_1_SHIFT    4
+#define OASTARTTRIG4_NOA_SELECT_2_SHIFT    8
+#define OASTARTTRIG4_NOA_SELECT_3_SHIFT    12
+#define OASTARTTRIG4_NOA_SELECT_4_SHIFT    16
+#define OASTARTTRIG4_NOA_SELECT_5_SHIFT    20
+#define OASTARTTRIG4_NOA_SELECT_6_SHIFT    24
+#define OASTARTTRIG4_NOA_SELECT_7_SHIFT    28
+
+#define OASTARTTRIG5 _MMIO(0x2720)
+#define OASTARTTRIG5_THRESHOLD_COUNT_MASK_MBZ 0xffff0000
+#define OASTARTTRIG5_THRESHOLD_MASK	      0xffff
+
+#define OASTARTTRIG6 _MMIO(0x2724)
+#define OASTARTTRIG6_INVERT_A_0 (1<<0)
+#define OASTARTTRIG6_INVERT_A_1 (1<<1)
+#define OASTARTTRIG6_INVERT_A_2 (1<<2)
+#define OASTARTTRIG6_INVERT_A_3 (1<<3)
+#define OASTARTTRIG6_INVERT_A_4 (1<<4)
+#define OASTARTTRIG6_INVERT_A_5 (1<<5)
+#define OASTARTTRIG6_INVERT_A_6 (1<<6)
+#define OASTARTTRIG6_INVERT_A_7 (1<<7)
+#define OASTARTTRIG6_INVERT_A_8 (1<<8)
+#define OASTARTTRIG6_INVERT_A_9 (1<<9)
+#define OASTARTTRIG6_INVERT_A_10 (1<<10)
+#define OASTARTTRIG6_INVERT_A_11 (1<<11)
+#define OASTARTTRIG6_INVERT_A_12 (1<<12)
+#define OASTARTTRIG6_INVERT_A_13 (1<<13)
+#define OASTARTTRIG6_INVERT_A_14 (1<<14)
+#define OASTARTTRIG6_INVERT_A_15 (1<<15)
+#define OASTARTTRIG6_INVERT_B_0 (1<<16)
+#define OASTARTTRIG6_INVERT_B_1 (1<<17)
+#define OASTARTTRIG6_INVERT_B_2 (1<<18)
+#define OASTARTTRIG6_INVERT_B_3 (1<<19)
+#define OASTARTTRIG6_INVERT_C_0 (1<<20)
+#define OASTARTTRIG6_INVERT_C_1 (1<<21)
+#define OASTARTTRIG6_INVERT_D_0 (1<<22)
+#define OASTARTTRIG6_THRESHOLD_ENABLE	    (1<<23)
+#define OASTARTTRIG6_START_TRIG_FLAG_MBZ    (1<<24)
+#define OASTARTTRIG6_EVENT_SELECT_4  (1<<28)
+#define OASTARTTRIG6_EVENT_SELECT_5  (1<<29)
+#define OASTARTTRIG6_EVENT_SELECT_6  (1<<30)
+#define OASTARTTRIG6_EVENT_SELECT_7  (1<<31)
+
+#define OASTARTTRIG7 _MMIO(0x2728)
+#define OASTARTTRIG7_NOA_SELECT_MASK	   0xf
+#define OASTARTTRIG7_NOA_SELECT_8_SHIFT    0
+#define OASTARTTRIG7_NOA_SELECT_9_SHIFT    4
+#define OASTARTTRIG7_NOA_SELECT_10_SHIFT   8
+#define OASTARTTRIG7_NOA_SELECT_11_SHIFT   12
+#define OASTARTTRIG7_NOA_SELECT_12_SHIFT   16
+#define OASTARTTRIG7_NOA_SELECT_13_SHIFT   20
+#define OASTARTTRIG7_NOA_SELECT_14_SHIFT   24
+#define OASTARTTRIG7_NOA_SELECT_15_SHIFT   28
+
+#define OASTARTTRIG8 _MMIO(0x272c)
+#define OASTARTTRIG8_NOA_SELECT_MASK	   0xf
+#define OASTARTTRIG8_NOA_SELECT_0_SHIFT    0
+#define OASTARTTRIG8_NOA_SELECT_1_SHIFT    4
+#define OASTARTTRIG8_NOA_SELECT_2_SHIFT    8
+#define OASTARTTRIG8_NOA_SELECT_3_SHIFT    12
+#define OASTARTTRIG8_NOA_SELECT_4_SHIFT    16
+#define OASTARTTRIG8_NOA_SELECT_5_SHIFT    20
+#define OASTARTTRIG8_NOA_SELECT_6_SHIFT    24
+#define OASTARTTRIG8_NOA_SELECT_7_SHIFT    28
+
+/* CECX_0 */
+#define OACEC_COMPARE_LESS_OR_EQUAL	6
+#define OACEC_COMPARE_NOT_EQUAL		5
+#define OACEC_COMPARE_LESS_THAN		4
+#define OACEC_COMPARE_GREATER_OR_EQUAL	3
+#define OACEC_COMPARE_EQUAL		2
+#define OACEC_COMPARE_GREATER_THAN	1
+#define OACEC_COMPARE_ANY_EQUAL		0
+
+#define OACEC_COMPARE_VALUE_MASK    0xffff
+#define OACEC_COMPARE_VALUE_SHIFT   3
+
+#define OACEC_SELECT_NOA	(0<<19)
+#define OACEC_SELECT_PREV	(1<<19)
+#define OACEC_SELECT_BOOLEAN	(2<<19)
+
+/* CECX_1 */
+#define OACEC_MASK_MASK		    0xffff
+#define OACEC_CONSIDERATIONS_MASK   0xffff
+#define OACEC_CONSIDERATIONS_SHIFT  16
+
+#define OACEC0_0 _MMIO(0x2770)
+#define OACEC0_1 _MMIO(0x2774)
+#define OACEC1_0 _MMIO(0x2778)
+#define OACEC1_1 _MMIO(0x277c)
+#define OACEC2_0 _MMIO(0x2780)
+#define OACEC2_1 _MMIO(0x2784)
+#define OACEC3_0 _MMIO(0x2788)
+#define OACEC3_1 _MMIO(0x278c)
+#define OACEC4_0 _MMIO(0x2790)
+#define OACEC4_1 _MMIO(0x2794)
+#define OACEC5_0 _MMIO(0x2798)
+#define OACEC5_1 _MMIO(0x279c)
+#define OACEC6_0 _MMIO(0x27a0)
+#define OACEC6_1 _MMIO(0x27a4)
+#define OACEC7_0 _MMIO(0x27a8)
+#define OACEC7_1 _MMIO(0x27ac)
+
 
 #define _GEN7_PIPEA_DE_LOAD_SL	0x70068
 #define _GEN7_PIPEB_DE_LOAD_SL	0x71068
@@ -6977,6 +7314,7 @@ enum {
 # define GEN6_RCCUNIT_CLOCK_GATE_DISABLE		(1 << 11)
 
 #define GEN6_UCGCTL3				_MMIO(0x9408)
+# define GEN6_OACSUNIT_CLOCK_GATE_DISABLE		(1 << 20)
 
 #define GEN7_UCGCTL4				_MMIO(0x940c)
 #define  GEN7_L3BANK2X_CLOCK_GATE_DISABLE	(1<<25)
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 7a74750..22d5ff1 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -2043,8 +2043,14 @@ static int intel_ring_context_pin(struct i915_gem_context *ctx,
 		if (ret)
 			goto error;
 
-		ret = i915_vma_pin(ce->state, 0, ctx->ggtt_alignment,
-				   PIN_GLOBAL | PIN_HIGH);
+		if (engine->id == RCS) {
+			u64 vma_flags = PIN_GLOBAL | PIN_HIGH;
+			ret = i915_gem_context_pin_legacy_rcs_state(engine->i915,
+								    ctx,
+								    vma_flags);
+		} else
+			ret = i915_vma_pin(ce->state, 0, ctx->ggtt_alignment,
+					   PIN_GLOBAL | PIN_HIGH);
 		if (ret)
 			goto error;
 	}
diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h
index 7777678..5ad3747 100644
--- a/include/uapi/drm/i915_drm.h
+++ b/include/uapi/drm/i915_drm.h
@@ -1224,6 +1224,18 @@ struct drm_i915_gem_context_param {
 	__u64 value;
 };
 
+enum drm_i915_oa_format {
+	I915_OA_FORMAT_A13 = 1,
+	I915_OA_FORMAT_A29,
+	I915_OA_FORMAT_A13_B8_C8,
+	I915_OA_FORMAT_B4_C8,
+	I915_OA_FORMAT_A45_B8_C8,
+	I915_OA_FORMAT_B4_C8_A16,
+	I915_OA_FORMAT_C4_B8,
+
+	I915_OA_FORMAT_MAX	    /* non-ABI */
+};
+
 enum drm_i915_perf_property_id {
 	/**
 	 * Open the stream for a specific context handle (as used with
@@ -1232,6 +1244,32 @@ enum drm_i915_perf_property_id {
 	 */
 	DRM_I915_PERF_PROP_CTX_HANDLE = 1,
 
+	/**
+	 * A value of 1 requests the inclusion of raw OA unit reports as
+	 * part of stream samples.
+	 */
+	DRM_I915_PERF_PROP_SAMPLE_OA,
+
+	/**
+	 * The value specifies which set of OA unit metrics should be
+	 * be configured, defining the contents of any OA unit reports.
+	 */
+	DRM_I915_PERF_PROP_OA_METRICS_SET,
+
+	/**
+	 * The value specifies the size and layout of OA unit reports.
+	 */
+	DRM_I915_PERF_PROP_OA_FORMAT,
+
+	/**
+	 * Specifying this property implicitly requests periodic OA unit
+	 * sampling and (at least on Haswell) the sampling frequency is derived
+	 * from this exponent as follows:
+	 *
+	 *   80ns * 2^(period_exponent + 1)
+	 */
+	DRM_I915_PERF_PROP_OA_EXPONENT,
+
 	DRM_I915_PERF_PROP_MAX /* non-ABI */
 };
 
@@ -1251,7 +1289,22 @@ struct drm_i915_perf_open_param {
 	__u64 __user properties_ptr;
 };
 
+/**
+ * Enable data capture for a stream that was either opened in a disabled state
+ * via I915_PERF_FLAG_DIABLED or was later disabled via I915_PERF_IOCTL_DISABLE.
+ *
+ * It is intended to be cheaper to disable and enable a stream than it may be
+ * to close and re-open a stream with the same configuration.
+ *
+ * It's undefined whether any pending data for the stream will be lost.
+ */
 #define I915_PERF_IOCTL_ENABLE	_IO('i', 0x0)
+
+/**
+ * Disable data capture for a stream.
+ *
+ * It is an error to try and read a stream that is disabled.
+ */
 #define I915_PERF_IOCTL_DISABLE	_IO('i', 0x1)
 
 /**
@@ -1275,17 +1328,30 @@ enum drm_i915_perf_record_type {
 	 * every sample.
 	 *
 	 * The order of these sample properties given by userspace has no
-	 * affect on the ordering of data within a sample. The order will be
+	 * affect on the ordering of data within a sample. The order is
 	 * documented here.
 	 *
 	 * struct {
 	 *     struct drm_i915_perf_record_header header;
 	 *
-	 *     TODO: itemize extensible sample data here
+	 *     { u32 oa_report[]; } && DRM_I915_PERF_PROP_SAMPLE_OA
 	 * };
 	 */
 	DRM_I915_PERF_RECORD_SAMPLE = 1,
 
+	/*
+	 * Indicates that one or more OA reports were not written by the
+	 * hardware. This can happen for example if an MI_REPORT_PERF_COUNT
+	 * command collides with periodic sampling - which would be more likely
+	 * at higher sampling frequencies.
+	 */
+	DRM_I915_PERF_RECORD_OA_REPORT_LOST = 2,
+
+	/**
+	 * An error occurred that resulted in all pending OA reports being lost.
+	 */
+	DRM_I915_PERF_RECORD_OA_BUFFER_LOST = 3,
+
 	DRM_I915_PERF_RECORD_MAX /* non-ABI */
 };
 
-- 
2.9.2



More information about the Intel-gfx mailing list