All of lore.kernel.org
 help / color / mirror / Atom feed
From: Robert Bragg <robert@sixbynine.org>
To: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org,
	Sourab Gupta <sourab.gupta@intel.com>,
	Daniel Vetter <daniel.vetter@intel.com>,
	Robert Bragg <robert@sixbynine.org>
Subject: [PATCH v3 09/11] drm/i915: add oa_event_min_timer_exponent sysctl
Date: Mon, 15 Aug 2016 15:41:26 +0100	[thread overview]
Message-ID: <20160815144128.7847-10-robert@sixbynine.org> (raw)
In-Reply-To: <20160815144128.7847-1-robert@sixbynine.org>

The minimal sampling period is now configurable via a
dev.i915.oa_min_timer_exponent sysctl parameter.

Following the precedent set by perf, the default is the minimum that
won't (on its own) exceed the default kernel.perf_event_max_sample_rate
default of 100000 samples/s.

Signed-off-by: Robert Bragg <robert@sixbynine.org>
---
 drivers/gpu/drm/i915/i915_perf.c | 42 ++++++++++++++++++++++++++++------------
 1 file changed, 30 insertions(+), 12 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
index c9e7104..18cb651 100644
--- a/drivers/gpu/drm/i915/i915_perf.c
+++ b/drivers/gpu/drm/i915/i915_perf.c
@@ -74,6 +74,23 @@ static u32 i915_perf_stream_paranoid = true;
  */
 #define OA_EXPONENT_MAX 31
 
+/* for sysctl proc_dointvec_minmax of i915_oa_min_timer_exponent */
+static int zero;
+static int oa_exponent_max = OA_EXPONENT_MAX;
+
+/* Theoretically we can program the OA unit to sample every 160ns but don't
+ * allow that by default unless root...
+ *
+ * The period is derived from the exponent as:
+ *
+ *   period = 80ns * 2^(exponent + 1)
+ *
+ * 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
+ */
+static u32 i915_oa_min_timer_exponent = 6;
+
 /* 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.
@@ -1266,21 +1283,13 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv,
 				return -EINVAL;
 			}
 
-			/* NB: The exponent represents a period as follows:
-			 *
-			 *   80ns * 2^(period_exponent + 1)
-			 *
-			 * Theoretically we can program the OA unit to sample
+			/* 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");
+			if (value < i915_oa_min_timer_exponent &&
+			    !capable(CAP_SYS_ADMIN)) {
+				DRM_ERROR("OA timer exponent too low without root privileges\n");
 				return -EACCES;
 			}
 
@@ -1342,6 +1351,15 @@ static struct ctl_table oa_table[] = {
 	 .mode = 0644,
 	 .proc_handler = proc_dointvec,
 	 },
+	{
+	 .procname = "oa_min_timer_exponent",
+	 .data = &i915_oa_min_timer_exponent,
+	 .maxlen = sizeof(i915_oa_min_timer_exponent),
+	 .mode = 0644,
+	 .proc_handler = proc_dointvec_minmax,
+	 .extra1 = &zero,
+	 .extra2 = &oa_exponent_max,
+	 },
 	{}
 };
 
-- 
2.9.2

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

  parent reply	other threads:[~2016-08-15 14:41 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-08-15 14:41 [PATCH v3 00/11] Enable Gen 7 Observation Architecture Robert Bragg
2016-08-15 14:41 ` [PATCH v3 01/11] drm/i915: Add i915 perf infrastructure Robert Bragg
2016-08-15 14:57   ` Chris Wilson
2016-08-16 14:59     ` Robert Bragg
2016-08-16 15:08       ` Chris Wilson
2016-08-15 14:41 ` [PATCH v3 02/11] drm/i915: rename OACONTROL GEN7_OACONTROL Robert Bragg
2016-08-15 14:41 ` [PATCH v3 03/11] drm/i915: return EACCES for check_cmd() failures Robert Bragg
2016-08-15 15:04   ` Chris Wilson
2016-08-18 21:18     ` Robert Bragg
2016-08-15 14:41 ` [PATCH v3 04/11] drm/i915: don't whitelist oacontrol in cmd parser Robert Bragg
2016-08-15 14:41 ` [PATCH v3 05/11] drm/i915: Add 'render basic' Haswell OA unit config Robert Bragg
2016-08-15 14:41 ` [PATCH v3 06/11] drm/i915: Enable i915 perf stream for Haswell OA unit Robert Bragg
2016-08-15 15:47   ` Chris Wilson
2016-08-15 14:41 ` [PATCH v3 07/11] drm/i915: advertise available metrics via sysfs Robert Bragg
2016-08-15 14:41 ` [PATCH v3 08/11] drm/i915: Add dev.i915.perf_event_paranoid sysctl option Robert Bragg
2016-08-15 14:41 ` Robert Bragg [this message]
2016-08-15 14:41 ` [PATCH v3 10/11] drm/i915: Add more Haswell OA metric sets Robert Bragg
2016-08-15 14:41 ` [PATCH v3 11/11] drm/i915: Add a kerneldoc summary for i915_perf.c Robert Bragg
2016-08-15 15:13 ` ✗ Ro.CI.BAT: failure for Enable Gen 7 Observation Architecture (rev5) Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160815144128.7847-10-robert@sixbynine.org \
    --to=robert@sixbynine.org \
    --cc=daniel.vetter@intel.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=sourab.gupta@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.