drm/i915/selftests: Disable C-states when measuring RPS frequency response

Let's isolate the impact of cpu frequency selection on determing the GPU
throughput in response to selection of RPS frequencies.

For real systems, we do have to be concerned with the impact of
integrating c-states, p-states and rp-states, but for the sake of
proving whether or not RPS works, one baby step at a time.

For the record, as one would hope, it does not seem to impact on the
measured performance, but we do it anyway to reduce the number of
variables. Later, we can extend the testing to encourage the the
cpu/pkg to try and sleep while the GPU is busy.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200421142236.8614-1-chris@chris-wilson.co.uk
Link: https://patchwork.freedesktop.org/patch/msgid/20200421142236.8614-1-chris@chris-wilson.co.uk

diff --git a/drivers/gpu/drm/i915/gt/selftest_rps.c b/drivers/gpu/drm/i915/gt/selftest_rps.c
index 395265121e43a94c7cf6fb29a7f947b3eafdb2dd..e2afc2003caab88ca2a0b54e6a887920cf169e8f 100644 (file)
--- a/drivers/gpu/drm/i915/gt/selftest_rps.c
+++ b/drivers/gpu/drm/i915/gt/selftest_rps.c
@@ -3,6 +3,7 @@
  * Copyright © 2020 Intel Corporation
  */
 
+#include <linux/pm_qos.h>
 #include <linux/sort.h>
 
 #include "intel_engine_pm.h"
@@ -14,6 +15,9 @@
 #include "selftests/igt_spinner.h"
 #include "selftests/librapl.h"
 
+/* Try to isolate the impact of cstates from determing frequency response */
+#define CPU_LATENCY 0 /* -1 to disable pm_qos, 0 to disable cstates */
+
 static void dummy_rps_work(struct work_struct *wrk)
 {
 }
@@ -406,6 +410,7 @@ int live_rps_frequency_cs(void *arg)
        struct intel_gt *gt = arg;
        struct intel_rps *rps = &gt->rps;
        struct intel_engine_cs *engine;
+       struct pm_qos_request qos;
        enum intel_engine_id id;
        int err = 0;
 
@@ -421,6 +426,9 @@ int live_rps_frequency_cs(void *arg)
        if (INTEL_GEN(gt->i915) < 8) /* for CS simplicity */
                return 0;
 
+       if (CPU_LATENCY >= 0)
+               cpu_latency_qos_add_request(&qos, CPU_LATENCY);
+
        intel_gt_pm_wait_for_idle(gt);
        saved_work = rps->work.func;
        rps->work.func = dummy_rps_work;
@@ -527,6 +535,9 @@ err_vma:
        intel_gt_pm_wait_for_idle(gt);
        rps->work.func = saved_work;
 
+       if (CPU_LATENCY >= 0)
+               cpu_latency_qos_remove_request(&qos);
+
        return err;
 }
 
@@ -536,6 +547,7 @@ int live_rps_frequency_srm(void *arg)
        struct intel_gt *gt = arg;
        struct intel_rps *rps = &gt->rps;
        struct intel_engine_cs *engine;
+       struct pm_qos_request qos;
        enum intel_engine_id id;
        int err = 0;
 
@@ -551,6 +563,9 @@ int live_rps_frequency_srm(void *arg)
        if (INTEL_GEN(gt->i915) < 8) /* for CS simplicity */
                return 0;
 
+       if (CPU_LATENCY >= 0)
+               cpu_latency_qos_add_request(&qos, CPU_LATENCY);
+
        intel_gt_pm_wait_for_idle(gt);
        saved_work = rps->work.func;
        rps->work.func = dummy_rps_work;
@@ -656,6 +671,9 @@ err_vma:
        intel_gt_pm_wait_for_idle(gt);
        rps->work.func = saved_work;
 
+       if (CPU_LATENCY >= 0)
+               cpu_latency_qos_remove_request(&qos);
+
        return err;
 }