* [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management
@ 2017-10-06 12:13 Sagar Arun Kamble
2017-10-06 12:13 ` [PATCH v2 01/11] drm/i915: Separate RPS and RC6 handling for gen6+ Sagar Arun Kamble
` (12 more replies)
0 siblings, 13 replies; 22+ messages in thread
From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw)
To: intel-gfx; +Cc: Beuchat, Marc
With GuC based SLPC, frequency control will be moved to GuC and Host will
continue to control RC6 and LLC ring frequency setup. This needs separate
handling of RPS, RC6 and LLC ring frequencies in i915 flows. We still
continue use the *gt_powersave routines with separate status variables
for RPS, RC6, ring frequency as pm.rps.enabled, pm.rc6.enabled and
pm.llc_pstate.configured respectively in dev_priv.
Post this, with SLPC changes integrated we can just skip the Host RPS path
in i915 PM flows.
v2: Added new patch 2. Addressed review comments. Pending review for last
3 patches and patch 2 currently.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Beuchat, Marc <marc.beuchat@intel.com>
Sagar Arun Kamble (11):
drm/i915: Separate RPS and RC6 handling for gen6+
drm/i915: Remove superfluous IS_BDW checks and non-BDW changes from
gen8_enable_rps
drm/i915: Separate RPS and RC6 handling for BDW
drm/i915: Separate RPS and RC6 handling for VLV
drm/i915: Separate RPS and RC6 handling for CHV
drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm"
drm/i915: Name structure in dev_priv that contains RPS/RC6 state as
"pm"
drm/i915: Rename intel_enable_rc6 to intel_rc6_enabled
drm/i915: Create generic function to setup LLC ring frequency table
drm/i915: Create generic functions to control RC6, RPS
drm/i915: Introduce separate status variable for RC6 and LLC ring
frequency setup
drivers/gpu/drm/i915/i915_debugfs.c | 158 +++----
drivers/gpu/drm/i915/i915_drv.c | 10 +-
drivers/gpu/drm/i915/i915_drv.h | 24 +-
drivers/gpu/drm/i915/i915_gem_request.c | 2 +-
drivers/gpu/drm/i915/i915_gpu_error.c | 4 +-
drivers/gpu/drm/i915/i915_guc_submission.c | 8 +-
drivers/gpu/drm/i915/i915_irq.c | 94 +++--
drivers/gpu/drm/i915/i915_sysfs.c | 72 ++--
drivers/gpu/drm/i915/intel_cdclk.c | 40 +-
drivers/gpu/drm/i915/intel_display.c | 12 +-
drivers/gpu/drm/i915/intel_drv.h | 14 +-
drivers/gpu/drm/i915/intel_guc.c | 3 +-
drivers/gpu/drm/i915/intel_pm.c | 637 +++++++++++++++++------------
drivers/gpu/drm/i915/intel_runtime_pm.c | 26 +-
drivers/gpu/drm/i915/intel_sideband.c | 6 +-
15 files changed, 626 insertions(+), 484 deletions(-)
--
1.9.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 22+ messages in thread* [PATCH v2 01/11] drm/i915: Separate RPS and RC6 handling for gen6+ 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 02/11] drm/i915: Remove superfluous IS_BDW checks and non-BDW changes from gen8_enable_rps Sagar Arun Kamble ` (11 subsequent siblings) 12 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx This patch separates enable/disable of RC6 and RPS for gen6+ platforms prior to VLV. Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/i915_debugfs.c | 17 +++++++------- drivers/gpu/drm/i915/intel_pm.c | 44 +++++++++++++++++++++++++------------ 2 files changed, 38 insertions(+), 23 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 44aae25..ca65600 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -1137,6 +1137,13 @@ static int i915_frequency_info(struct seq_file *m, void *unused) pm_iir = I915_READ(GEN8_GT_IIR(2)); pm_mask = I915_READ(GEN6_PMINTRMSK); } + seq_printf(m, "Video Turbo Mode: %s\n", + yesno(rpmodectl & GEN6_RP_MEDIA_TURBO)); + seq_printf(m, "HW control enabled: %s\n", + yesno(rpmodectl & GEN6_RP_ENABLE)); + seq_printf(m, "SW control enabled: %s\n", + yesno((rpmodectl & GEN6_RP_MEDIA_MODE_MASK) == + GEN6_RP_MEDIA_SW_MODE)); seq_printf(m, "PM IER=0x%08x IMR=0x%08x ISR=0x%08x IIR=0x%08x, MASK=0x%08x\n", pm_ier, pm_imr, pm_isr, pm_iir, pm_mask); seq_printf(m, "pm_intrmsk_mbz: 0x%08x\n", @@ -1480,7 +1487,7 @@ static int vlv_drpc_info(struct seq_file *m) static int gen6_drpc_info(struct seq_file *m) { struct drm_i915_private *dev_priv = node_to_i915(m->private); - u32 rpmodectl1, gt_core_status, rcctl1, rc6vids = 0; + u32 gt_core_status, rcctl1, rc6vids = 0; u32 gen9_powergate_enable = 0, gen9_powergate_status = 0; unsigned forcewake_count; int count = 0; @@ -1499,7 +1506,6 @@ static int gen6_drpc_info(struct seq_file *m) gt_core_status = I915_READ_FW(GEN6_GT_CORE_STATUS); trace_i915_reg_rw(false, GEN6_GT_CORE_STATUS, gt_core_status, 4, true); - rpmodectl1 = I915_READ(GEN6_RP_CONTROL); rcctl1 = I915_READ(GEN6_RC_CONTROL); if (INTEL_GEN(dev_priv) >= 9) { gen9_powergate_enable = I915_READ(GEN9_PG_ENABLE); @@ -1510,13 +1516,6 @@ static int gen6_drpc_info(struct seq_file *m) sandybridge_pcode_read(dev_priv, GEN6_PCODE_READ_RC6VIDS, &rc6vids); mutex_unlock(&dev_priv->rps.hw_lock); - seq_printf(m, "Video Turbo Mode: %s\n", - yesno(rpmodectl1 & GEN6_RP_MEDIA_TURBO)); - seq_printf(m, "HW control enabled: %s\n", - yesno(rpmodectl1 & GEN6_RP_ENABLE)); - seq_printf(m, "SW control enabled: %s\n", - yesno((rpmodectl1 & GEN6_RP_MEDIA_MODE_MASK) == - GEN6_RP_MEDIA_SW_MODE)); seq_printf(m, "RC1e Enabled: %s\n", yesno(rcctl1 & GEN6_RC_CTL_RC1e_ENABLE)); seq_printf(m, "RC6 Enabled: %s\n", diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 171b21f..f118764 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6332,9 +6332,13 @@ static void gen9_disable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RP_CONTROL, 0); } -static void gen6_disable_rps(struct drm_i915_private *dev_priv) +static void gen6_disable_rc6(struct drm_i915_private *dev_priv) { I915_WRITE(GEN6_RC_CONTROL, 0); +} + +static void gen6_disable_rps(struct drm_i915_private *dev_priv) +{ I915_WRITE(GEN6_RPNSWREQ, 1 << 31); I915_WRITE(GEN6_RP_CONTROL, 0); } @@ -6692,7 +6696,7 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } -static void gen6_enable_rps(struct drm_i915_private *dev_priv) +static void gen6_enable_rc6(struct drm_i915_private *dev_priv) { struct intel_engine_cs *engine; enum intel_engine_id id; @@ -6703,12 +6707,6 @@ static void gen6_enable_rps(struct drm_i915_private *dev_priv) WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); - /* Here begins a magic sequence of register writes to enable - * auto-downclocking. - * - * Perhaps there might be some value in exposing these to - * userspace... - */ I915_WRITE(GEN6_RC_STATE, 0); /* Clear the DBG now so we don't confuse earlier errors */ @@ -6762,12 +6760,6 @@ static void gen6_enable_rps(struct drm_i915_private *dev_priv) GEN6_RC_CTL_EI_MODE(1) | GEN6_RC_CTL_HW_ENABLE); - /* Power down if completely idle for over 50ms */ - I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 50000); - I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10); - - reset_rps(dev_priv, gen6_set_rps); - rc6vids = 0; ret = sandybridge_pcode_read(dev_priv, GEN6_PCODE_READ_RC6VIDS, &rc6vids); if (IS_GEN6(dev_priv) && ret) { @@ -6785,6 +6777,28 @@ static void gen6_enable_rps(struct drm_i915_private *dev_priv) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } + +static void gen6_enable_rps(struct drm_i915_private *dev_priv) +{ + WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + + /* Here begins a magic sequence of register writes to enable + * auto-downclocking. + * + * Perhaps there might be some value in exposing these to + * userspace... + */ + intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); + + /* Power down if completely idle for over 50ms */ + I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 50000); + I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10); + + reset_rps(dev_priv, gen6_set_rps); + + intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); +} + static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) { int min_freq = 15; @@ -7934,6 +7948,7 @@ void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) } else if (IS_VALLEYVIEW(dev_priv)) { valleyview_disable_rps(dev_priv); } else if (INTEL_GEN(dev_priv) >= 6) { + gen6_disable_rc6(dev_priv); gen6_disable_rps(dev_priv); } else if (IS_IRONLAKE_M(dev_priv)) { ironlake_disable_drps(dev_priv); @@ -7970,6 +7985,7 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) gen8_enable_rps(dev_priv); gen6_update_ring_freq(dev_priv); } else if (INTEL_GEN(dev_priv) >= 6) { + gen6_enable_rc6(dev_priv); gen6_enable_rps(dev_priv); gen6_update_ring_freq(dev_priv); } else if (IS_IRONLAKE_M(dev_priv)) { -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v2 02/11] drm/i915: Remove superfluous IS_BDW checks and non-BDW changes from gen8_enable_rps 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 01/11] drm/i915: Separate RPS and RC6 handling for gen6+ Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:24 ` Chris Wilson 2017-10-06 12:13 ` [PATCH v2 03/11] drm/i915: Separate RPS and RC6 handling for BDW Sagar Arun Kamble ` (10 subsequent siblings) 12 siblings, 1 reply; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx This patch removes all IS_BROADWELL checks and non-BDW changes from gen8_enable_rps as it is called only for BROADWELL. Suggested-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> --- drivers/gpu/drm/i915/intel_pm.c | 17 +++++------------ 1 file changed, 5 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index f118764..ffd0224 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6642,23 +6642,16 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) for_each_engine(engine, dev_priv, id) I915_WRITE(RING_MAX_IDLE(engine->mmio_base), 10); I915_WRITE(GEN6_RC_SLEEP, 0); - if (IS_BROADWELL(dev_priv)) - I915_WRITE(GEN6_RC6_THRESHOLD, 625); /* 800us/1.28 for TO */ - else - I915_WRITE(GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */ + I915_WRITE(GEN6_RC6_THRESHOLD, 625); /* 800us/1.28 for TO */ /* 3: Enable RC6 */ if (intel_enable_rc6() & INTEL_RC6_ENABLE) rc6_mask = GEN6_RC_CTL_RC6_ENABLE; intel_print_rc6_info(dev_priv, rc6_mask); - if (IS_BROADWELL(dev_priv)) - I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE | - GEN7_RC_CTL_TO_MODE | - rc6_mask); - else - I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE | - GEN6_RC_CTL_EI_MODE(1) | - rc6_mask); + + I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE | + GEN7_RC_CTL_TO_MODE | + rc6_mask); /* 4 Program defaults and thresholds for RPS*/ I915_WRITE(GEN6_RPNSWREQ, -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH v2 02/11] drm/i915: Remove superfluous IS_BDW checks and non-BDW changes from gen8_enable_rps 2017-10-06 12:13 ` [PATCH v2 02/11] drm/i915: Remove superfluous IS_BDW checks and non-BDW changes from gen8_enable_rps Sagar Arun Kamble @ 2017-10-06 12:24 ` Chris Wilson 0 siblings, 0 replies; 22+ messages in thread From: Chris Wilson @ 2017-10-06 12:24 UTC (permalink / raw) To: Sagar Arun Kamble, intel-gfx Quoting Sagar Arun Kamble (2017-10-06 13:13:31) > This patch removes all IS_BROADWELL checks and non-BDW changes from > gen8_enable_rps as it is called only for BROADWELL. > > Suggested-by: Chris Wilson <chris@chris-wilson.co.uk> > Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> > Cc: Imre Deak <imre.deak@intel.com> > Cc: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> > Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v2 03/11] drm/i915: Separate RPS and RC6 handling for BDW 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 01/11] drm/i915: Separate RPS and RC6 handling for gen6+ Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 02/11] drm/i915: Remove superfluous IS_BDW checks and non-BDW changes from gen8_enable_rps Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 04/11] drm/i915: Separate RPS and RC6 handling for VLV Sagar Arun Kamble ` (9 subsequent siblings) 12 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx This patch separates RC6 and RPS enabling for BDW. RC6/RPS Disabling are handled through gen6 functions. PM Programming guide recommends a sequence within forcewakes to configure RC6, RPS and ring frequencies in sequence. With this patch the order is still maintained. v2: Update sequence numbers in RC6 programming and comment about intent of reset_rps during gen8_enable_rps. (Radoslaw) v3: Rebase. Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/intel_pm.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index ffd0224..e26e9db 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6619,7 +6619,7 @@ static void gen9_enable_rc6(struct drm_i915_private *dev_priv) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } -static void gen8_enable_rps(struct drm_i915_private *dev_priv) +static void gen8_enable_rc6(struct drm_i915_private *dev_priv) { struct intel_engine_cs *engine; enum intel_engine_id id; @@ -6628,7 +6628,7 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) /* 1a: Software RC state - RC0 */ I915_WRITE(GEN6_RC_STATE, 0); - /* 1c & 1d: Get forcewake during program sequence. Although the driver + /* 1b: Get forcewake during program sequence. Although the driver * hasn't enabled a state yet where we need forcewake, BIOS may have.*/ intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); @@ -6653,7 +6653,14 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) GEN7_RC_CTL_TO_MODE | rc6_mask); - /* 4 Program defaults and thresholds for RPS*/ + intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); +} + +static void gen8_enable_rps(struct drm_i915_private *dev_priv) +{ + intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); + + /* 1 Program defaults and thresholds for RPS*/ I915_WRITE(GEN6_RPNSWREQ, HSW_FREQUENCY(dev_priv->rps.rp1_freq)); I915_WRITE(GEN6_RC_VIDEO_FREQ, @@ -6673,7 +6680,7 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10); - /* 5: Enable RPS */ + /* 2: Enable RPS */ I915_WRITE(GEN6_RP_CONTROL, GEN6_RP_MEDIA_TURBO | GEN6_RP_MEDIA_HW_NORMAL_MODE | @@ -6682,8 +6689,6 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) GEN6_RP_UP_BUSY_AVG | GEN6_RP_DOWN_IDLE_AVG); - /* 6: Ring frequency + overclocking (our driver does this later */ - reset_rps(dev_priv, gen6_set_rps); intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); @@ -7975,6 +7980,7 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) gen6_update_ring_freq(dev_priv); } else if (IS_BROADWELL(dev_priv)) { + gen8_enable_rc6(dev_priv); gen8_enable_rps(dev_priv); gen6_update_ring_freq(dev_priv); } else if (INTEL_GEN(dev_priv) >= 6) { -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v2 04/11] drm/i915: Separate RPS and RC6 handling for VLV 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (2 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 03/11] drm/i915: Separate RPS and RC6 handling for BDW Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 05/11] drm/i915: Separate RPS and RC6 handling for CHV Sagar Arun Kamble ` (8 subsequent siblings) 12 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx This patch separates enable/disable of RC6 and RPS for VLV. v2: Removed unnecessary comments about forcewakes while enabling RC6/RPS. Added changes to output turbo control status for VLV in i915_frequency_info. Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/i915_debugfs.c | 24 +++++++-------- drivers/gpu/drm/i915/intel_pm.c | 61 ++++++++++++++++++++++++------------- 2 files changed, 51 insertions(+), 34 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index ca65600..142474a 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -1042,9 +1042,19 @@ static int i915_frequency_info(struct seq_file *m, void *unused) seq_printf(m, "Current P-state: %d\n", (rgvstat & MEMSTAT_PSTATE_MASK) >> MEMSTAT_PSTATE_SHIFT); } else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { - u32 freq_sts; + u32 rpmodectl, freq_sts; mutex_lock(&dev_priv->rps.hw_lock); + + rpmodectl = I915_READ(GEN6_RP_CONTROL); + seq_printf(m, "Video Turbo Mode: %s\n", + yesno(rpmodectl & GEN6_RP_MEDIA_TURBO)); + seq_printf(m, "HW control enabled: %s\n", + yesno(rpmodectl & GEN6_RP_ENABLE)); + seq_printf(m, "SW control enabled: %s\n", + yesno((rpmodectl & GEN6_RP_MEDIA_MODE_MASK) == + GEN6_RP_MEDIA_SW_MODE)); + freq_sts = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); seq_printf(m, "PUNIT_REG_GPU_FREQ_STS: 0x%08x\n", freq_sts); seq_printf(m, "DDR freq: %d MHz\n", dev_priv->mem_freq); @@ -1455,21 +1465,11 @@ static void print_rc6_res(struct seq_file *m, static int vlv_drpc_info(struct seq_file *m) { struct drm_i915_private *dev_priv = node_to_i915(m->private); - u32 rpmodectl1, rcctl1, pw_status; + u32 rcctl1, pw_status; pw_status = I915_READ(VLV_GTLC_PW_STATUS); - rpmodectl1 = I915_READ(GEN6_RP_CONTROL); rcctl1 = I915_READ(GEN6_RC_CONTROL); - seq_printf(m, "Video Turbo Mode: %s\n", - yesno(rpmodectl1 & GEN6_RP_MEDIA_TURBO)); - seq_printf(m, "Turbo enabled: %s\n", - yesno(rpmodectl1 & GEN6_RP_ENABLE)); - seq_printf(m, "HW control enabled: %s\n", - yesno(rpmodectl1 & GEN6_RP_ENABLE)); - seq_printf(m, "SW control enabled: %s\n", - yesno((rpmodectl1 & GEN6_RP_MEDIA_MODE_MASK) == - GEN6_RP_MEDIA_SW_MODE)); seq_printf(m, "RC6 Enabled: %s\n", yesno(rcctl1 & (GEN7_RC_CTL_TO_MODE | GEN6_RC_CTL_EI_MODE(1)))); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index e26e9db..a55cee4 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6348,9 +6348,9 @@ static void cherryview_disable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RC_CONTROL, 0); } -static void valleyview_disable_rps(struct drm_i915_private *dev_priv) +static void valleyview_disable_rc6(struct drm_i915_private *dev_priv) { - /* we're doing forcewake before Disabling RC6, + /* We're doing forcewake before Disabling RC6, * This what the BIOS expects when going into suspend */ intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); @@ -6359,6 +6359,11 @@ static void valleyview_disable_rps(struct drm_i915_private *dev_priv) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } +static void valleyview_disable_rps(struct drm_i915_private *dev_priv) +{ + I915_WRITE(GEN6_RP_CONTROL, 0); +} + static void intel_print_rc6_info(struct drm_i915_private *dev_priv, u32 mode) { if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { @@ -7282,11 +7287,11 @@ static void cherryview_enable_rps(struct drm_i915_private *dev_priv) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } -static void valleyview_enable_rps(struct drm_i915_private *dev_priv) +static void valleyview_enable_rc6(struct drm_i915_private *dev_priv) { struct intel_engine_cs *engine; enum intel_engine_id id; - u32 gtfifodbg, val, rc6_mode = 0; + u32 gtfifodbg, rc6_mode = 0; WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); @@ -7299,28 +7304,11 @@ static void valleyview_enable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GTFIFODBG, gtfifodbg); } - /* If VLV, Forcewake all wells, else re-direct to regular path */ intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); /* Disable RC states. */ I915_WRITE(GEN6_RC_CONTROL, 0); - I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 1000000); - I915_WRITE(GEN6_RP_UP_THRESHOLD, 59400); - I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 245000); - I915_WRITE(GEN6_RP_UP_EI, 66000); - I915_WRITE(GEN6_RP_DOWN_EI, 350000); - - I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10); - - I915_WRITE(GEN6_RP_CONTROL, - GEN6_RP_MEDIA_TURBO | - GEN6_RP_MEDIA_HW_NORMAL_MODE | - GEN6_RP_MEDIA_IS_GFX | - GEN6_RP_ENABLE | - GEN6_RP_UP_BUSY_AVG | - GEN6_RP_DOWN_IDLE_CONT); - I915_WRITE(GEN6_RC6_WAKE_RATE_LIMIT, 0x00280000); I915_WRITE(GEN6_RC_EVALUATION_INTERVAL, 125000); I915_WRITE(GEN6_RC_IDLE_HYSTERSIS, 25); @@ -7330,7 +7318,7 @@ static void valleyview_enable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RC6_THRESHOLD, 0x557); - /* allows RC6 residency counter to work */ + /* Allows RC6 residency counter to work */ I915_WRITE(VLV_COUNTER_CONTROL, _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH | VLV_MEDIA_RC0_COUNT_EN | @@ -7345,6 +7333,33 @@ static void valleyview_enable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RC_CONTROL, rc6_mode); + intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); +} + +static void valleyview_enable_rps(struct drm_i915_private *dev_priv) +{ + u32 val; + + WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + + intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); + + I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 1000000); + I915_WRITE(GEN6_RP_UP_THRESHOLD, 59400); + I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 245000); + I915_WRITE(GEN6_RP_UP_EI, 66000); + I915_WRITE(GEN6_RP_DOWN_EI, 350000); + + I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10); + + I915_WRITE(GEN6_RP_CONTROL, + GEN6_RP_MEDIA_TURBO | + GEN6_RP_MEDIA_HW_NORMAL_MODE | + GEN6_RP_MEDIA_IS_GFX | + GEN6_RP_ENABLE | + GEN6_RP_UP_BUSY_AVG | + GEN6_RP_DOWN_IDLE_CONT); + /* Setting Fixed Bias */ val = VLV_OVERRIDE_EN | VLV_SOC_TDP_EN | @@ -7944,6 +7959,7 @@ void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) } else if (IS_CHERRYVIEW(dev_priv)) { cherryview_disable_rps(dev_priv); } else if (IS_VALLEYVIEW(dev_priv)) { + valleyview_disable_rc6(dev_priv); valleyview_disable_rps(dev_priv); } else if (INTEL_GEN(dev_priv) >= 6) { gen6_disable_rc6(dev_priv); @@ -7973,6 +7989,7 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) if (IS_CHERRYVIEW(dev_priv)) { cherryview_enable_rps(dev_priv); } else if (IS_VALLEYVIEW(dev_priv)) { + valleyview_enable_rc6(dev_priv); valleyview_enable_rps(dev_priv); } else if (INTEL_GEN(dev_priv) >= 9) { gen9_enable_rc6(dev_priv); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v2 05/11] drm/i915: Separate RPS and RC6 handling for CHV 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (3 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 04/11] drm/i915: Separate RPS and RC6 handling for VLV Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 06/11] drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm" Sagar Arun Kamble ` (7 subsequent siblings) 12 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx This patch separates enable/disable of RC6 and RPS for CHV. v2: Fixed comment. Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++------ 1 file changed, 24 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index a55cee4..843fdc7 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6343,11 +6343,16 @@ static void gen6_disable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RP_CONTROL, 0); } -static void cherryview_disable_rps(struct drm_i915_private *dev_priv) +static void cherryview_disable_rc6(struct drm_i915_private *dev_priv) { I915_WRITE(GEN6_RC_CONTROL, 0); } +static void cherryview_disable_rps(struct drm_i915_private *dev_priv) +{ + I915_WRITE(GEN6_RP_CONTROL, 0); +} + static void valleyview_disable_rc6(struct drm_i915_private *dev_priv) { /* We're doing forcewake before Disabling RC6, @@ -7198,11 +7203,11 @@ static void valleyview_cleanup_gt_powersave(struct drm_i915_private *dev_priv) valleyview_cleanup_pctx(dev_priv); } -static void cherryview_enable_rps(struct drm_i915_private *dev_priv) +static void cherryview_enable_rc6(struct drm_i915_private *dev_priv) { struct intel_engine_cs *engine; enum intel_engine_id id; - u32 gtfifodbg, val, rc6_mode = 0, pcbr; + u32 gtfifodbg, rc6_mode = 0, pcbr; WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); @@ -7235,7 +7240,7 @@ static void cherryview_enable_rps(struct drm_i915_private *dev_priv) /* TO threshold set to 500 us ( 0x186 * 1.28 us) */ I915_WRITE(GEN6_RC6_THRESHOLD, 0x186); - /* allows RC6 residency counter to work */ + /* Allows RC6 residency counter to work */ I915_WRITE(VLV_COUNTER_CONTROL, _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH | VLV_MEDIA_RC6_COUNT_EN | @@ -7251,7 +7256,18 @@ static void cherryview_enable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RC_CONTROL, rc6_mode); - /* 4 Program defaults and thresholds for RPS*/ + intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); +} + +static void cherryview_enable_rps(struct drm_i915_private *dev_priv) +{ + u32 val; + + WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + + intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); + + /* 1: Program defaults and thresholds for RPS*/ I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 1000000); I915_WRITE(GEN6_RP_UP_THRESHOLD, 59400); I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 245000); @@ -7260,7 +7276,7 @@ static void cherryview_enable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10); - /* 5: Enable RPS */ + /* 2: Enable RPS */ I915_WRITE(GEN6_RP_CONTROL, GEN6_RP_MEDIA_HW_NORMAL_MODE | GEN6_RP_MEDIA_IS_GFX | @@ -7957,6 +7973,7 @@ void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) gen9_disable_rc6(dev_priv); gen9_disable_rps(dev_priv); } else if (IS_CHERRYVIEW(dev_priv)) { + cherryview_disable_rc6(dev_priv); cherryview_disable_rps(dev_priv); } else if (IS_VALLEYVIEW(dev_priv)) { valleyview_disable_rc6(dev_priv); @@ -7987,6 +8004,7 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) mutex_lock(&dev_priv->rps.hw_lock); if (IS_CHERRYVIEW(dev_priv)) { + cherryview_enable_rc6(dev_priv); cherryview_enable_rps(dev_priv); } else if (IS_VALLEYVIEW(dev_priv)) { valleyview_enable_rc6(dev_priv); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v2 06/11] drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm" 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (4 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 05/11] drm/i915: Separate RPS and RC6 handling for CHV Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:40 ` Chris Wilson 2017-10-06 12:13 ` [PATCH v2 07/11] drm/i915: Name structure in dev_priv that contains RPS/RC6 state as "pm" Sagar Arun Kamble ` (6 subsequent siblings) 12 siblings, 1 reply; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx We were using dev_priv->pm for runtime power management related state. This patch renames it to "rpm" which looks more apt. Will be using pm for state containing RPS/RC6 state in the next patch. Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/i915_drv.c | 8 ++++---- drivers/gpu/drm/i915/i915_drv.h | 2 +- drivers/gpu/drm/i915/i915_gpu_error.c | 4 ++-- drivers/gpu/drm/i915/i915_irq.c | 8 ++++---- drivers/gpu/drm/i915/intel_drv.h | 10 +++++----- drivers/gpu/drm/i915/intel_pm.c | 4 ++-- drivers/gpu/drm/i915/intel_runtime_pm.c | 10 +++++----- 7 files changed, 23 insertions(+), 23 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 66fc156..a365860 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -2544,12 +2544,12 @@ static int intel_runtime_suspend(struct device *kdev) intel_uncore_suspend(dev_priv); enable_rpm_wakeref_asserts(dev_priv); - WARN_ON_ONCE(atomic_read(&dev_priv->pm.wakeref_count)); + WARN_ON_ONCE(atomic_read(&dev_priv->rpm.wakeref_count)); if (intel_uncore_arm_unclaimed_mmio_detection(dev_priv)) DRM_ERROR("Unclaimed access detected prior to suspending\n"); - dev_priv->pm.suspended = true; + dev_priv->rpm.suspended = true; /* * FIXME: We really should find a document that references the arguments @@ -2595,11 +2595,11 @@ static int intel_runtime_resume(struct device *kdev) DRM_DEBUG_KMS("Resuming device\n"); - WARN_ON_ONCE(atomic_read(&dev_priv->pm.wakeref_count)); + WARN_ON_ONCE(atomic_read(&dev_priv->rpm.wakeref_count)); disable_rpm_wakeref_asserts(dev_priv); intel_opregion_notify_adapter(dev_priv, PCI_D0); - dev_priv->pm.suspended = false; + dev_priv->rpm.suspended = false; if (intel_uncore_unclaimed_mmio(dev_priv)) DRM_DEBUG_DRIVER("Unclaimed access during suspend, bios?\n"); diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 1fc7080..7313312 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -2526,7 +2526,7 @@ struct drm_i915_private { bool distrust_bios_wm; } wm; - struct i915_runtime_pm pm; + struct i915_runtime_pm rpm; struct { bool initialized; diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index dc91b32..635002a 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -1674,8 +1674,8 @@ static void i915_capture_gen_state(struct drm_i915_private *dev_priv, struct i915_gpu_state *error) { error->awake = dev_priv->gt.awake; - error->wakelock = atomic_read(&dev_priv->pm.wakeref_count); - error->suspended = dev_priv->pm.suspended; + error->wakelock = atomic_read(&dev_priv->rpm.wakeref_count); + error->suspended = dev_priv->rpm.suspended; error->iommu = -1; #ifdef CONFIG_INTEL_IOMMU diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index de77713..d5953a4 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -4178,7 +4178,7 @@ int intel_irq_install(struct drm_i915_private *dev_priv) * interrupts as enabled _before_ actually enabling them to avoid * special cases in our ordering checks. */ - dev_priv->pm.irqs_enabled = true; + dev_priv->rpm.irqs_enabled = true; return drm_irq_install(&dev_priv->drm, dev_priv->drm.pdev->irq); } @@ -4194,7 +4194,7 @@ void intel_irq_uninstall(struct drm_i915_private *dev_priv) { drm_irq_uninstall(&dev_priv->drm); intel_hpd_cancel_work(dev_priv); - dev_priv->pm.irqs_enabled = false; + dev_priv->rpm.irqs_enabled = false; } /** @@ -4207,7 +4207,7 @@ void intel_irq_uninstall(struct drm_i915_private *dev_priv) void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv) { dev_priv->drm.driver->irq_uninstall(&dev_priv->drm); - dev_priv->pm.irqs_enabled = false; + dev_priv->rpm.irqs_enabled = false; synchronize_irq(dev_priv->drm.irq); } @@ -4220,7 +4220,7 @@ void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv) */ void intel_runtime_pm_enable_interrupts(struct drm_i915_private *dev_priv) { - dev_priv->pm.irqs_enabled = true; + dev_priv->rpm.irqs_enabled = true; dev_priv->drm.driver->irq_preinstall(&dev_priv->drm); dev_priv->drm.driver->irq_postinstall(&dev_priv->drm); } diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 0cab667..b5bfddd 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1254,7 +1254,7 @@ static inline bool intel_irqs_enabled(struct drm_i915_private *dev_priv) * We only use drm_irq_uninstall() at unload and VT switch, so * this is the only thing we need to check. */ - return dev_priv->pm.irqs_enabled; + return dev_priv->rpm.irqs_enabled; } int intel_get_crtc_scanline(struct intel_crtc *crtc); @@ -1790,7 +1790,7 @@ void intel_display_power_put(struct drm_i915_private *dev_priv, static inline void assert_rpm_device_not_suspended(struct drm_i915_private *dev_priv) { - WARN_ONCE(dev_priv->pm.suspended, + WARN_ONCE(dev_priv->rpm.suspended, "Device suspended during HW access\n"); } @@ -1798,7 +1798,7 @@ void intel_display_power_put(struct drm_i915_private *dev_priv, assert_rpm_wakelock_held(struct drm_i915_private *dev_priv) { assert_rpm_device_not_suspended(dev_priv); - WARN_ONCE(!atomic_read(&dev_priv->pm.wakeref_count), + WARN_ONCE(!atomic_read(&dev_priv->rpm.wakeref_count), "RPM wakelock ref not held during HW access"); } @@ -1823,7 +1823,7 @@ void intel_display_power_put(struct drm_i915_private *dev_priv, static inline void disable_rpm_wakeref_asserts(struct drm_i915_private *dev_priv) { - atomic_inc(&dev_priv->pm.wakeref_count); + atomic_inc(&dev_priv->rpm.wakeref_count); } /** @@ -1840,7 +1840,7 @@ void intel_display_power_put(struct drm_i915_private *dev_priv, static inline void enable_rpm_wakeref_asserts(struct drm_i915_private *dev_priv) { - atomic_dec(&dev_priv->pm.wakeref_count); + atomic_dec(&dev_priv->rpm.wakeref_count); } void intel_runtime_pm_get(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 843fdc7..2ad6ebf 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -9350,8 +9350,8 @@ void intel_pm_setup(struct drm_i915_private *dev_priv) __intel_autoenable_gt_powersave); atomic_set(&dev_priv->rps.num_waiters, 0); - dev_priv->pm.suspended = false; - atomic_set(&dev_priv->pm.wakeref_count, 0); + dev_priv->rpm.suspended = false; + atomic_set(&dev_priv->rpm.wakeref_count, 0); } static u64 vlv_residency_raw(struct drm_i915_private *dev_priv, diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index 3791c3f..a37ad0f 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -187,7 +187,7 @@ bool __intel_display_power_is_enabled(struct drm_i915_private *dev_priv, struct i915_power_well *power_well; bool is_enabled; - if (dev_priv->pm.suspended) + if (dev_priv->rpm.suspended) return false; is_enabled = true; @@ -3128,7 +3128,7 @@ void intel_runtime_pm_get(struct drm_i915_private *dev_priv) ret = pm_runtime_get_sync(kdev); WARN_ONCE(ret < 0, "pm_runtime_get_sync() failed: %d\n", ret); - atomic_inc(&dev_priv->pm.wakeref_count); + atomic_inc(&dev_priv->rpm.wakeref_count); assert_rpm_wakelock_held(dev_priv); } @@ -3162,7 +3162,7 @@ bool intel_runtime_pm_get_if_in_use(struct drm_i915_private *dev_priv) return false; } - atomic_inc(&dev_priv->pm.wakeref_count); + atomic_inc(&dev_priv->rpm.wakeref_count); assert_rpm_wakelock_held(dev_priv); return true; @@ -3193,7 +3193,7 @@ void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv) assert_rpm_wakelock_held(dev_priv); pm_runtime_get_noresume(kdev); - atomic_inc(&dev_priv->pm.wakeref_count); + atomic_inc(&dev_priv->rpm.wakeref_count); } /** @@ -3210,7 +3210,7 @@ void intel_runtime_pm_put(struct drm_i915_private *dev_priv) struct device *kdev = &pdev->dev; assert_rpm_wakelock_held(dev_priv); - atomic_dec(&dev_priv->pm.wakeref_count); + atomic_dec(&dev_priv->rpm.wakeref_count); pm_runtime_mark_last_busy(kdev); pm_runtime_put_autosuspend(kdev); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH v2 06/11] drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm" 2017-10-06 12:13 ` [PATCH v2 06/11] drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm" Sagar Arun Kamble @ 2017-10-06 12:40 ` Chris Wilson 2017-10-06 15:00 ` Sagar Arun Kamble 0 siblings, 1 reply; 22+ messages in thread From: Chris Wilson @ 2017-10-06 12:40 UTC (permalink / raw) To: Sagar Arun Kamble, intel-gfx Quoting Sagar Arun Kamble (2017-10-06 13:13:35) > We were using dev_priv->pm for runtime power management related state. > This patch renames it to "rpm" which looks more apt. Will be using pm > for state containing RPS/RC6 state in the next patch. > > Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> > Cc: Imre Deak <imre.deak@intel.com> > Cc: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> > Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> > Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Thinking about this again, rpm, pm are very close. How about if we used i915->runtime_pm and i915->gt_pm (or i915->gt.pm)? Imre, any thoughts? -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v2 06/11] drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm" 2017-10-06 12:40 ` Chris Wilson @ 2017-10-06 15:00 ` Sagar Arun Kamble 0 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 15:00 UTC (permalink / raw) To: Chris Wilson, intel-gfx On 10/6/2017 6:10 PM, Chris Wilson wrote: > Quoting Sagar Arun Kamble (2017-10-06 13:13:35) >> We were using dev_priv->pm for runtime power management related state. >> This patch renames it to "rpm" which looks more apt. Will be using pm >> for state containing RPS/RC6 state in the next patch. >> >> Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> >> Cc: Imre Deak <imre.deak@intel.com> >> Cc: Chris Wilson <chris@chris-wilson.co.uk> >> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> >> Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> >> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> > Thinking about this again, rpm, pm are very close. How about if we used > i915->runtime_pm and i915->gt_pm (or i915->gt.pm)? Imre, any thoughts? rps.hw_lock/pcu_lock is used by display too, so I just kept it pm. should we pull rps.hw_lock/pcu_lock out into drm_i915_private and then gt_pm would be good. > -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v2 07/11] drm/i915: Name structure in dev_priv that contains RPS/RC6 state as "pm" 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (5 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 06/11] drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm" Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 08/11] drm/i915: Rename intel_enable_rc6 to intel_rc6_enabled Sagar Arun Kamble ` (5 subsequent siblings) 12 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx Prepared substructure rps for RPS related state. autoenable_work and pcu_lock are used for RC6 hence they are defined outside rps structure. Renamed the RPS lock as pcu_lock. v2: Rebase. Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Imre Deak <imre.deak@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/i915_debugfs.c | 117 +++++----- drivers/gpu/drm/i915/i915_drv.c | 2 +- drivers/gpu/drm/i915/i915_drv.h | 12 +- drivers/gpu/drm/i915/i915_gem_request.c | 2 +- drivers/gpu/drm/i915/i915_guc_submission.c | 8 +- drivers/gpu/drm/i915/i915_irq.c | 86 +++---- drivers/gpu/drm/i915/i915_sysfs.c | 70 +++--- drivers/gpu/drm/i915/intel_cdclk.c | 40 ++-- drivers/gpu/drm/i915/intel_display.c | 12 +- drivers/gpu/drm/i915/intel_drv.h | 2 +- drivers/gpu/drm/i915/intel_pm.c | 363 +++++++++++++++-------------- drivers/gpu/drm/i915/intel_runtime_pm.c | 16 +- drivers/gpu/drm/i915/intel_sideband.c | 6 +- 13 files changed, 379 insertions(+), 357 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 142474a..7e7c195 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -1027,6 +1027,7 @@ static int i915_error_state_open(struct inode *inode, struct file *file) static int i915_frequency_info(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); + struct intel_rps *rps = &dev_priv->pm.rps; int ret = 0; intel_runtime_pm_get(dev_priv); @@ -1044,7 +1045,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused) } else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { u32 rpmodectl, freq_sts; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); rpmodectl = I915_READ(GEN6_RP_CONTROL); seq_printf(m, "Video Turbo Mode: %s\n", @@ -1063,21 +1064,21 @@ static int i915_frequency_info(struct seq_file *m, void *unused) intel_gpu_freq(dev_priv, (freq_sts >> 8) & 0xff)); seq_printf(m, "current GPU freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.cur_freq)); + intel_gpu_freq(dev_priv, rps->cur_freq)); seq_printf(m, "max GPU freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.max_freq)); + intel_gpu_freq(dev_priv, rps->max_freq)); seq_printf(m, "min GPU freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.min_freq)); + intel_gpu_freq(dev_priv, rps->min_freq)); seq_printf(m, "idle GPU freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.idle_freq)); + intel_gpu_freq(dev_priv, rps->idle_freq)); seq_printf(m, "efficient (RPe) frequency: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.efficient_freq)); - mutex_unlock(&dev_priv->rps.hw_lock); + intel_gpu_freq(dev_priv, rps->efficient_freq)); + mutex_unlock(&dev_priv->pm.pcu_lock); } else if (INTEL_GEN(dev_priv) >= 6) { u32 rp_state_limits; u32 gt_perf_status; @@ -1157,7 +1158,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused) seq_printf(m, "PM IER=0x%08x IMR=0x%08x ISR=0x%08x IIR=0x%08x, MASK=0x%08x\n", pm_ier, pm_imr, pm_isr, pm_iir, pm_mask); seq_printf(m, "pm_intrmsk_mbz: 0x%08x\n", - dev_priv->rps.pm_intrmsk_mbz); + rps->pm_intrmsk_mbz); seq_printf(m, "GT_PERF_STATUS: 0x%08x\n", gt_perf_status); seq_printf(m, "Render p-state ratio: %d\n", (gt_perf_status & (INTEL_GEN(dev_priv) >= 9 ? 0x1ff00 : 0xff00)) >> 8); @@ -1177,8 +1178,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused) rpcurup, GT_PM_INTERVAL_TO_US(dev_priv, rpcurup)); seq_printf(m, "RP PREV UP: %d (%dus)\n", rpprevup, GT_PM_INTERVAL_TO_US(dev_priv, rpprevup)); - seq_printf(m, "Up threshold: %d%%\n", - dev_priv->rps.up_threshold); + seq_printf(m, "Up threshold: %d%%\n", rps->up_threshold); seq_printf(m, "RP CUR DOWN EI: %d (%dus)\n", rpdownei, GT_PM_INTERVAL_TO_US(dev_priv, rpdownei)); @@ -1186,8 +1186,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused) rpcurdown, GT_PM_INTERVAL_TO_US(dev_priv, rpcurdown)); seq_printf(m, "RP PREV DOWN: %d (%dus)\n", rpprevdown, GT_PM_INTERVAL_TO_US(dev_priv, rpprevdown)); - seq_printf(m, "Down threshold: %d%%\n", - dev_priv->rps.down_threshold); + seq_printf(m, "Down threshold: %d%%\n", rps->down_threshold); max_freq = (IS_GEN9_LP(dev_priv) ? rp_state_cap >> 0 : rp_state_cap >> 16) & 0xff; @@ -1209,22 +1208,22 @@ static int i915_frequency_info(struct seq_file *m, void *unused) seq_printf(m, "Max non-overclocked (RP0) frequency: %dMHz\n", intel_gpu_freq(dev_priv, max_freq)); seq_printf(m, "Max overclocked frequency: %dMHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.max_freq)); + intel_gpu_freq(dev_priv, rps->max_freq)); seq_printf(m, "Current freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.cur_freq)); + intel_gpu_freq(dev_priv, rps->cur_freq)); seq_printf(m, "Actual freq: %d MHz\n", cagf); seq_printf(m, "Idle freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.idle_freq)); + intel_gpu_freq(dev_priv, rps->idle_freq)); seq_printf(m, "Min freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.min_freq)); + intel_gpu_freq(dev_priv, rps->min_freq)); seq_printf(m, "Boost freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.boost_freq)); + intel_gpu_freq(dev_priv, rps->boost_freq)); seq_printf(m, "Max freq: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.max_freq)); + intel_gpu_freq(dev_priv, rps->max_freq)); seq_printf(m, "efficient (RPe) frequency: %d MHz\n", - intel_gpu_freq(dev_priv, dev_priv->rps.efficient_freq)); + intel_gpu_freq(dev_priv, rps->efficient_freq)); } else { seq_puts(m, "no P-state info available\n"); } @@ -1512,9 +1511,9 @@ static int gen6_drpc_info(struct seq_file *m) gen9_powergate_status = I915_READ(GEN9_PWRGT_DOMAIN_STATUS); } - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); sandybridge_pcode_read(dev_priv, GEN6_PCODE_READ_RC6VIDS, &rc6vids); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); seq_printf(m, "RC1e Enabled: %s\n", yesno(rcctl1 & GEN6_RC_CTL_RC1e_ENABLE)); @@ -1789,19 +1788,19 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) intel_runtime_pm_get(dev_priv); - ret = mutex_lock_interruptible(&dev_priv->rps.hw_lock); + ret = mutex_lock_interruptible(&dev_priv->pm.pcu_lock); if (ret) goto out; if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { /* Convert GT frequency to 50 HZ units */ min_gpu_freq = - dev_priv->rps.min_freq_softlimit / GEN9_FREQ_SCALER; + dev_priv->pm.rps.min_freq_softlimit / GEN9_FREQ_SCALER; max_gpu_freq = - dev_priv->rps.max_freq_softlimit / GEN9_FREQ_SCALER; + dev_priv->pm.rps.max_freq_softlimit / GEN9_FREQ_SCALER; } else { - min_gpu_freq = dev_priv->rps.min_freq_softlimit; - max_gpu_freq = dev_priv->rps.max_freq_softlimit; + min_gpu_freq = dev_priv->pm.rps.min_freq_softlimit; + max_gpu_freq = dev_priv->pm.rps.max_freq_softlimit; } seq_puts(m, "GPU freq (MHz)\tEffective CPU freq (MHz)\tEffective Ring freq (MHz)\n"); @@ -1820,7 +1819,7 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) ((ia_freq >> 8) & 0xff) * 100); } - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); out: intel_runtime_pm_put(dev_priv); @@ -2254,25 +2253,26 @@ static int i915_rps_boost_info(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); struct drm_device *dev = &dev_priv->drm; + struct intel_rps *rps = &dev_priv->pm.rps; struct drm_file *file; - seq_printf(m, "RPS enabled? %d\n", dev_priv->rps.enabled); + seq_printf(m, "RPS enabled? %d\n", rps->enabled); seq_printf(m, "GPU busy? %s [%d requests]\n", yesno(dev_priv->gt.awake), dev_priv->gt.active_requests); seq_printf(m, "CPU waiting? %d\n", count_irq_waiters(dev_priv)); seq_printf(m, "Boosts outstanding? %d\n", - atomic_read(&dev_priv->rps.num_waiters)); + atomic_read(&rps->num_waiters)); seq_printf(m, "Frequency requested %d\n", - intel_gpu_freq(dev_priv, dev_priv->rps.cur_freq)); + intel_gpu_freq(dev_priv, rps->cur_freq)); seq_printf(m, " min hard:%d, soft:%d; max soft:%d, hard:%d\n", - intel_gpu_freq(dev_priv, dev_priv->rps.min_freq), - intel_gpu_freq(dev_priv, dev_priv->rps.min_freq_softlimit), - intel_gpu_freq(dev_priv, dev_priv->rps.max_freq_softlimit), - intel_gpu_freq(dev_priv, dev_priv->rps.max_freq)); + intel_gpu_freq(dev_priv, rps->min_freq), + intel_gpu_freq(dev_priv, rps->min_freq_softlimit), + intel_gpu_freq(dev_priv, rps->max_freq_softlimit), + intel_gpu_freq(dev_priv, rps->max_freq)); seq_printf(m, " idle:%d, efficient:%d, boost:%d\n", - intel_gpu_freq(dev_priv, dev_priv->rps.idle_freq), - intel_gpu_freq(dev_priv, dev_priv->rps.efficient_freq), - intel_gpu_freq(dev_priv, dev_priv->rps.boost_freq)); + intel_gpu_freq(dev_priv, rps->idle_freq), + intel_gpu_freq(dev_priv, rps->efficient_freq), + intel_gpu_freq(dev_priv, rps->boost_freq)); mutex_lock(&dev->filelist_mutex); list_for_each_entry_reverse(file, &dev->filelist, lhead) { @@ -2288,11 +2288,11 @@ static int i915_rps_boost_info(struct seq_file *m, void *data) rcu_read_unlock(); } seq_printf(m, "Kernel (anonymous) boosts: %d\n", - atomic_read(&dev_priv->rps.boosts)); + atomic_read(&rps->boosts)); mutex_unlock(&dev->filelist_mutex); if (INTEL_GEN(dev_priv) >= 6 && - dev_priv->rps.enabled && + rps->enabled && dev_priv->gt.active_requests) { u32 rpup, rpupei; u32 rpdown, rpdownei; @@ -2305,13 +2305,13 @@ static int i915_rps_boost_info(struct seq_file *m, void *data) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); seq_printf(m, "\nRPS Autotuning (current \"%s\" window):\n", - rps_power_to_str(dev_priv->rps.power)); + rps_power_to_str(rps->power)); seq_printf(m, " Avg. up: %d%% [above threshold? %d%%]\n", rpup && rpupei ? 100 * rpup / rpupei : 0, - dev_priv->rps.up_threshold); + rps->up_threshold); seq_printf(m, " Avg. down: %d%% [below threshold? %d%%]\n", rpdown && rpdownei ? 100 * rpdown / rpdownei : 0, - dev_priv->rps.down_threshold); + rps->down_threshold); } else { seq_puts(m, "\nRPS Autotuning inactive\n"); } @@ -4392,7 +4392,7 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf, if (INTEL_GEN(dev_priv) < 6) return -ENODEV; - *val = intel_gpu_freq(dev_priv, dev_priv->rps.max_freq_softlimit); + *val = intel_gpu_freq(dev_priv, dev_priv->pm.rps.max_freq_softlimit); return 0; } @@ -4400,6 +4400,7 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf, i915_max_freq_set(void *data, u64 val) { struct drm_i915_private *dev_priv = data; + struct intel_rps *rps = &dev_priv->pm.rps; u32 hw_max, hw_min; int ret; @@ -4408,7 +4409,7 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf, DRM_DEBUG_DRIVER("Manually setting max freq to %llu\n", val); - ret = mutex_lock_interruptible(&dev_priv->rps.hw_lock); + ret = mutex_lock_interruptible(&dev_priv->pm.pcu_lock); if (ret) return ret; @@ -4417,20 +4418,20 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf, */ val = intel_freq_opcode(dev_priv, val); - hw_max = dev_priv->rps.max_freq; - hw_min = dev_priv->rps.min_freq; + hw_max = rps->max_freq; + hw_min = rps->min_freq; - if (val < hw_min || val > hw_max || val < dev_priv->rps.min_freq_softlimit) { - mutex_unlock(&dev_priv->rps.hw_lock); + if (val < hw_min || val > hw_max || val < rps->min_freq_softlimit) { + mutex_unlock(&dev_priv->pm.pcu_lock); return -EINVAL; } - dev_priv->rps.max_freq_softlimit = val; + rps->max_freq_softlimit = val; if (intel_set_rps(dev_priv, val)) DRM_DEBUG_DRIVER("failed to update RPS to new softlimit\n"); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); return 0; } @@ -4447,7 +4448,7 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf, if (INTEL_GEN(dev_priv) < 6) return -ENODEV; - *val = intel_gpu_freq(dev_priv, dev_priv->rps.min_freq_softlimit); + *val = intel_gpu_freq(dev_priv, dev_priv->pm.rps.min_freq_softlimit); return 0; } @@ -4463,7 +4464,7 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf, DRM_DEBUG_DRIVER("Manually setting min freq to %llu\n", val); - ret = mutex_lock_interruptible(&dev_priv->rps.hw_lock); + ret = mutex_lock_interruptible(&dev_priv->pm.pcu_lock); if (ret) return ret; @@ -4472,21 +4473,21 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf, */ val = intel_freq_opcode(dev_priv, val); - hw_max = dev_priv->rps.max_freq; - hw_min = dev_priv->rps.min_freq; + hw_max = dev_priv->pm.rps.max_freq; + hw_min = dev_priv->pm.rps.min_freq; if (val < hw_min || - val > hw_max || val > dev_priv->rps.max_freq_softlimit) { - mutex_unlock(&dev_priv->rps.hw_lock); + val > hw_max || val > dev_priv->pm.rps.max_freq_softlimit) { + mutex_unlock(&dev_priv->pm.pcu_lock); return -EINVAL; } - dev_priv->rps.min_freq_softlimit = val; + dev_priv->pm.rps.min_freq_softlimit = val; if (intel_set_rps(dev_priv, val)) DRM_DEBUG_DRIVER("failed to update RPS to new softlimit\n"); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); return 0; } diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index a365860..a6af7d5 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -2502,7 +2502,7 @@ static int intel_runtime_suspend(struct device *kdev) struct drm_i915_private *dev_priv = to_i915(dev); int ret; - if (WARN_ON_ONCE(!(dev_priv->rps.enabled && intel_enable_rc6()))) + if (WARN_ON_ONCE(!(dev_priv->pm.rps.enabled && intel_enable_rc6()))) return -ENODEV; if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev_priv))) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 7313312..45944a8 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1315,7 +1315,7 @@ struct intel_rps_ei { u32 media_c0; }; -struct intel_gen6_power_mgmt { +struct intel_rps { /* * work, interrupts_enabled and pm_iir are protected by * dev_priv->irq_lock @@ -1356,12 +1356,16 @@ struct intel_gen6_power_mgmt { enum { LOW_POWER, BETWEEN, HIGH_POWER } power; bool enabled; - struct delayed_work autoenable_work; atomic_t num_waiters; atomic_t boosts; /* manual wa residency calculations */ struct intel_rps_ei ei; +}; + +struct intel_gen6_power_mgmt { + struct intel_rps rps; + struct delayed_work autoenable_work; /* * Protects RPS/RC6 register access and PCU communication. @@ -1369,7 +1373,7 @@ struct intel_gen6_power_mgmt { * this lock may be held for long periods of time when * talking to hw - so only take it when talking to hw! */ - struct mutex hw_lock; + struct mutex pcu_lock; }; /* defined intel_pm.c */ @@ -2415,7 +2419,7 @@ struct drm_i915_private { u32 edram_cap; /* gen6+ rps state */ - struct intel_gen6_power_mgmt rps; + struct intel_gen6_power_mgmt pm; /* ilk-only ips/rps state. Everything in here is protected by the global * mchdev_lock in intel_pm.c */ diff --git a/drivers/gpu/drm/i915/i915_gem_request.c b/drivers/gpu/drm/i915/i915_gem_request.c index b100b38..be1005d 100644 --- a/drivers/gpu/drm/i915/i915_gem_request.c +++ b/drivers/gpu/drm/i915/i915_gem_request.c @@ -416,7 +416,7 @@ static void i915_gem_request_retire(struct drm_i915_gem_request *request) spin_lock_irq(&request->lock); if (request->waitboost) - atomic_dec(&request->i915->rps.num_waiters); + atomic_dec(&request->i915->pm.rps.num_waiters); dma_fence_signal_locked(&request->fence); spin_unlock_irq(&request->lock); diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c index 7460ab4..0cc9122 100644 --- a/drivers/gpu/drm/i915/i915_guc_submission.c +++ b/drivers/gpu/drm/i915/i915_guc_submission.c @@ -1064,8 +1064,8 @@ static void guc_interrupts_capture(struct drm_i915_private *dev_priv) * Here we CLEAR REDIRECT_TO_GUC bit in pm_intrmsk_mbz, which will * result in the register bit being left SET! */ - dev_priv->rps.pm_intrmsk_mbz |= ARAT_EXPIRED_INTRMSK; - dev_priv->rps.pm_intrmsk_mbz &= ~GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC; + dev_priv->pm.rps.pm_intrmsk_mbz |= ARAT_EXPIRED_INTRMSK; + dev_priv->pm.rps.pm_intrmsk_mbz &= ~GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC; } static void guc_interrupts_release(struct drm_i915_private *dev_priv) @@ -1088,8 +1088,8 @@ static void guc_interrupts_release(struct drm_i915_private *dev_priv) I915_WRITE(GUC_VCS2_VCS1_IER, 0); I915_WRITE(GUC_WD_VECS_IER, 0); - dev_priv->rps.pm_intrmsk_mbz |= GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC; - dev_priv->rps.pm_intrmsk_mbz &= ~ARAT_EXPIRED_INTRMSK; + dev_priv->pm.rps.pm_intrmsk_mbz |= GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC; + dev_priv->pm.rps.pm_intrmsk_mbz &= ~ARAT_EXPIRED_INTRMSK; } int i915_guc_submission_enable(struct drm_i915_private *dev_priv) diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index d5953a4..3ad4df0 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -404,19 +404,19 @@ void gen6_reset_rps_interrupts(struct drm_i915_private *dev_priv) { spin_lock_irq(&dev_priv->irq_lock); gen6_reset_pm_iir(dev_priv, dev_priv->pm_rps_events); - dev_priv->rps.pm_iir = 0; + dev_priv->pm.rps.pm_iir = 0; spin_unlock_irq(&dev_priv->irq_lock); } void gen6_enable_rps_interrupts(struct drm_i915_private *dev_priv) { - if (READ_ONCE(dev_priv->rps.interrupts_enabled)) + if (READ_ONCE(dev_priv->pm.rps.interrupts_enabled)) return; spin_lock_irq(&dev_priv->irq_lock); - WARN_ON_ONCE(dev_priv->rps.pm_iir); + WARN_ON_ONCE(dev_priv->pm.rps.pm_iir); WARN_ON_ONCE(I915_READ(gen6_pm_iir(dev_priv)) & dev_priv->pm_rps_events); - dev_priv->rps.interrupts_enabled = true; + dev_priv->pm.rps.interrupts_enabled = true; gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); spin_unlock_irq(&dev_priv->irq_lock); @@ -424,11 +424,11 @@ void gen6_enable_rps_interrupts(struct drm_i915_private *dev_priv) void gen6_disable_rps_interrupts(struct drm_i915_private *dev_priv) { - if (!READ_ONCE(dev_priv->rps.interrupts_enabled)) + if (!READ_ONCE(dev_priv->pm.rps.interrupts_enabled)) return; spin_lock_irq(&dev_priv->irq_lock); - dev_priv->rps.interrupts_enabled = false; + dev_priv->pm.rps.interrupts_enabled = false; I915_WRITE(GEN6_PMINTRMSK, gen6_sanitize_rps_pm_mask(dev_priv, ~0u)); @@ -442,7 +442,7 @@ void gen6_disable_rps_interrupts(struct drm_i915_private *dev_priv) * we will reset the GPU to minimum frequencies, so the current * state of the worker can be discarded. */ - cancel_work_sync(&dev_priv->rps.work); + cancel_work_sync(&dev_priv->pm.rps.work); gen6_reset_rps_interrupts(dev_priv); } @@ -1119,12 +1119,12 @@ static void vlv_c0_read(struct drm_i915_private *dev_priv, void gen6_rps_reset_ei(struct drm_i915_private *dev_priv) { - memset(&dev_priv->rps.ei, 0, sizeof(dev_priv->rps.ei)); + memset(&dev_priv->pm.rps.ei, 0, sizeof(dev_priv->pm.rps.ei)); } static u32 vlv_wa_c0_ei(struct drm_i915_private *dev_priv, u32 pm_iir) { - const struct intel_rps_ei *prev = &dev_priv->rps.ei; + const struct intel_rps_ei *prev = &dev_priv->pm.rps.ei; struct intel_rps_ei now; u32 events = 0; @@ -1151,28 +1151,29 @@ static u32 vlv_wa_c0_ei(struct drm_i915_private *dev_priv, u32 pm_iir) c0 = max(render, media); c0 *= 1000 * 100 << 8; /* to usecs and scale to threshold% */ - if (c0 > time * dev_priv->rps.up_threshold) + if (c0 > time * dev_priv->pm.rps.up_threshold) events = GEN6_PM_RP_UP_THRESHOLD; - else if (c0 < time * dev_priv->rps.down_threshold) + else if (c0 < time * dev_priv->pm.rps.down_threshold) events = GEN6_PM_RP_DOWN_THRESHOLD; } - dev_priv->rps.ei = now; + dev_priv->pm.rps.ei = now; return events; } static void gen6_pm_rps_work(struct work_struct *work) { struct drm_i915_private *dev_priv = - container_of(work, struct drm_i915_private, rps.work); + container_of(work, struct drm_i915_private, pm.rps.work); + struct intel_rps *rps = &dev_priv->pm.rps; bool client_boost = false; int new_delay, adj, min, max; u32 pm_iir = 0; spin_lock_irq(&dev_priv->irq_lock); - if (dev_priv->rps.interrupts_enabled) { - pm_iir = fetch_and_zero(&dev_priv->rps.pm_iir); - client_boost = atomic_read(&dev_priv->rps.num_waiters); + if (rps->interrupts_enabled) { + pm_iir = fetch_and_zero(&rps->pm_iir); + client_boost = atomic_read(&rps->num_waiters); } spin_unlock_irq(&dev_priv->irq_lock); @@ -1181,18 +1182,18 @@ static void gen6_pm_rps_work(struct work_struct *work) if ((pm_iir & dev_priv->pm_rps_events) == 0 && !client_boost) goto out; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); pm_iir |= vlv_wa_c0_ei(dev_priv, pm_iir); - adj = dev_priv->rps.last_adj; - new_delay = dev_priv->rps.cur_freq; - min = dev_priv->rps.min_freq_softlimit; - max = dev_priv->rps.max_freq_softlimit; + adj = rps->last_adj; + new_delay = rps->cur_freq; + min = rps->min_freq_softlimit; + max = rps->max_freq_softlimit; if (client_boost) - max = dev_priv->rps.max_freq; - if (client_boost && new_delay < dev_priv->rps.boost_freq) { - new_delay = dev_priv->rps.boost_freq; + max = rps->max_freq; + if (client_boost && new_delay < rps->boost_freq) { + new_delay = rps->boost_freq; adj = 0; } else if (pm_iir & GEN6_PM_RP_UP_THRESHOLD) { if (adj > 0) @@ -1200,15 +1201,15 @@ static void gen6_pm_rps_work(struct work_struct *work) else /* CHV needs even encode values */ adj = IS_CHERRYVIEW(dev_priv) ? 2 : 1; - if (new_delay >= dev_priv->rps.max_freq_softlimit) + if (new_delay >= rps->max_freq_softlimit) adj = 0; } else if (client_boost) { adj = 0; } else if (pm_iir & GEN6_PM_RP_DOWN_TIMEOUT) { - if (dev_priv->rps.cur_freq > dev_priv->rps.efficient_freq) - new_delay = dev_priv->rps.efficient_freq; - else if (dev_priv->rps.cur_freq > dev_priv->rps.min_freq_softlimit) - new_delay = dev_priv->rps.min_freq_softlimit; + if (rps->cur_freq > rps->efficient_freq) + new_delay = rps->efficient_freq; + else if (rps->cur_freq > rps->min_freq_softlimit) + new_delay = rps->min_freq_softlimit; adj = 0; } else if (pm_iir & GEN6_PM_RP_DOWN_THRESHOLD) { if (adj < 0) @@ -1216,13 +1217,13 @@ static void gen6_pm_rps_work(struct work_struct *work) else /* CHV needs even encode values */ adj = IS_CHERRYVIEW(dev_priv) ? -2 : -1; - if (new_delay <= dev_priv->rps.min_freq_softlimit) + if (new_delay <= rps->min_freq_softlimit) adj = 0; } else { /* unknown event */ adj = 0; } - dev_priv->rps.last_adj = adj; + rps->last_adj = adj; /* sysfs frequency interfaces may have snuck in while servicing the * interrupt @@ -1232,15 +1233,15 @@ static void gen6_pm_rps_work(struct work_struct *work) if (intel_set_rps(dev_priv, new_delay)) { DRM_DEBUG_DRIVER("Failed to set new GPU frequency\n"); - dev_priv->rps.last_adj = 0; + rps->last_adj = 0; } - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); out: /* Make sure not to corrupt PMIMR state used by ringbuffer on GEN6 */ spin_lock_irq(&dev_priv->irq_lock); - if (dev_priv->rps.interrupts_enabled) + if (rps->interrupts_enabled) gen6_unmask_pm_irq(dev_priv, dev_priv->pm_rps_events); spin_unlock_irq(&dev_priv->irq_lock); } @@ -1721,12 +1722,14 @@ static void i9xx_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, * the work queue. */ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir) { + struct intel_rps *rps = &dev_priv->pm.rps; + if (pm_iir & dev_priv->pm_rps_events) { spin_lock(&dev_priv->irq_lock); gen6_mask_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events); - if (dev_priv->rps.interrupts_enabled) { - dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events; - schedule_work(&dev_priv->rps.work); + if (rps->interrupts_enabled) { + rps->pm_iir |= pm_iir & dev_priv->pm_rps_events; + schedule_work(&rps->work); } spin_unlock(&dev_priv->irq_lock); } @@ -4012,11 +4015,12 @@ static irqreturn_t i965_irq_handler(int irq, void *arg) void intel_irq_init(struct drm_i915_private *dev_priv) { struct drm_device *dev = &dev_priv->drm; + struct intel_rps *rps = &dev_priv->pm.rps; int i; intel_hpd_init_work(dev_priv); - INIT_WORK(&dev_priv->rps.work, gen6_pm_rps_work); + INIT_WORK(&rps->work, gen6_pm_rps_work); INIT_WORK(&dev_priv->l3_parity.error_work, ivybridge_parity_work); for (i = 0; i < MAX_L3_SLICES; ++i) @@ -4032,7 +4036,7 @@ void intel_irq_init(struct drm_i915_private *dev_priv) else dev_priv->pm_rps_events = GEN6_PM_RPS_EVENTS; - dev_priv->rps.pm_intrmsk_mbz = 0; + rps->pm_intrmsk_mbz = 0; /* * SNB,IVB,HSW can while VLV,CHV may hard hang on looping batchbuffer @@ -4041,10 +4045,10 @@ void intel_irq_init(struct drm_i915_private *dev_priv) * TODO: verify if this can be reproduced on VLV,CHV. */ if (INTEL_GEN(dev_priv) <= 7) - dev_priv->rps.pm_intrmsk_mbz |= GEN6_PM_RP_UP_EI_EXPIRED; + rps->pm_intrmsk_mbz |= GEN6_PM_RP_UP_EI_EXPIRED; if (INTEL_GEN(dev_priv) >= 8) - dev_priv->rps.pm_intrmsk_mbz |= GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC; + rps->pm_intrmsk_mbz |= GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC; if (IS_GEN2(dev_priv)) { /* Gen2 doesn't have a hardware frame counter */ diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c index d61c872..c16e907 100644 --- a/drivers/gpu/drm/i915/i915_sysfs.c +++ b/drivers/gpu/drm/i915/i915_sysfs.c @@ -246,7 +246,7 @@ static ssize_t gt_act_freq_mhz_show(struct device *kdev, intel_runtime_pm_get(dev_priv); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { u32 freq; freq = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); @@ -261,7 +261,7 @@ static ssize_t gt_act_freq_mhz_show(struct device *kdev, ret = (rpstat & GEN6_CAGF_MASK) >> GEN6_CAGF_SHIFT; ret = intel_gpu_freq(dev_priv, ret); } - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); intel_runtime_pm_put(dev_priv); @@ -275,7 +275,7 @@ static ssize_t gt_cur_freq_mhz_show(struct device *kdev, return snprintf(buf, PAGE_SIZE, "%d\n", intel_gpu_freq(dev_priv, - dev_priv->rps.cur_freq)); + dev_priv->pm.rps.cur_freq)); } static ssize_t gt_boost_freq_mhz_show(struct device *kdev, struct device_attribute *attr, char *buf) @@ -284,7 +284,7 @@ static ssize_t gt_boost_freq_mhz_show(struct device *kdev, struct device_attribu return snprintf(buf, PAGE_SIZE, "%d\n", intel_gpu_freq(dev_priv, - dev_priv->rps.boost_freq)); + dev_priv->pm.rps.boost_freq)); } static ssize_t gt_boost_freq_mhz_store(struct device *kdev, @@ -301,12 +301,12 @@ static ssize_t gt_boost_freq_mhz_store(struct device *kdev, /* Validate against (static) hardware limits */ val = intel_freq_opcode(dev_priv, val); - if (val < dev_priv->rps.min_freq || val > dev_priv->rps.max_freq) + if (val < dev_priv->pm.rps.min_freq || val > dev_priv->pm.rps.max_freq) return -EINVAL; - mutex_lock(&dev_priv->rps.hw_lock); - dev_priv->rps.boost_freq = val; - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); + dev_priv->pm.rps.boost_freq = val; + mutex_unlock(&dev_priv->pm.pcu_lock); return count; } @@ -318,7 +318,7 @@ static ssize_t vlv_rpe_freq_mhz_show(struct device *kdev, return snprintf(buf, PAGE_SIZE, "%d\n", intel_gpu_freq(dev_priv, - dev_priv->rps.efficient_freq)); + dev_priv->pm.rps.efficient_freq)); } static ssize_t gt_max_freq_mhz_show(struct device *kdev, struct device_attribute *attr, char *buf) @@ -327,7 +327,7 @@ static ssize_t gt_max_freq_mhz_show(struct device *kdev, struct device_attribute return snprintf(buf, PAGE_SIZE, "%d\n", intel_gpu_freq(dev_priv, - dev_priv->rps.max_freq_softlimit)); + dev_priv->pm.rps.max_freq_softlimit)); } static ssize_t gt_max_freq_mhz_store(struct device *kdev, @@ -344,34 +344,34 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev, intel_runtime_pm_get(dev_priv); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); val = intel_freq_opcode(dev_priv, val); - if (val < dev_priv->rps.min_freq || - val > dev_priv->rps.max_freq || - val < dev_priv->rps.min_freq_softlimit) { - mutex_unlock(&dev_priv->rps.hw_lock); + if (val < dev_priv->pm.rps.min_freq || + val > dev_priv->pm.rps.max_freq || + val < dev_priv->pm.rps.min_freq_softlimit) { + mutex_unlock(&dev_priv->pm.pcu_lock); intel_runtime_pm_put(dev_priv); return -EINVAL; } - if (val > dev_priv->rps.rp0_freq) + if (val > dev_priv->pm.rps.rp0_freq) DRM_DEBUG("User requested overclocking to %d\n", intel_gpu_freq(dev_priv, val)); - dev_priv->rps.max_freq_softlimit = val; + dev_priv->pm.rps.max_freq_softlimit = val; - val = clamp_t(int, dev_priv->rps.cur_freq, - dev_priv->rps.min_freq_softlimit, - dev_priv->rps.max_freq_softlimit); + val = clamp_t(int, dev_priv->pm.rps.cur_freq, + dev_priv->pm.rps.min_freq_softlimit, + dev_priv->pm.rps.max_freq_softlimit); /* We still need *_set_rps to process the new max_delay and * update the interrupt limits and PMINTRMSK even though * frequency request may be unchanged. */ ret = intel_set_rps(dev_priv, val); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); intel_runtime_pm_put(dev_priv); @@ -384,7 +384,7 @@ static ssize_t gt_min_freq_mhz_show(struct device *kdev, struct device_attribute return snprintf(buf, PAGE_SIZE, "%d\n", intel_gpu_freq(dev_priv, - dev_priv->rps.min_freq_softlimit)); + dev_priv->pm.rps.min_freq_softlimit)); } static ssize_t gt_min_freq_mhz_store(struct device *kdev, @@ -401,30 +401,30 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev, intel_runtime_pm_get(dev_priv); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); val = intel_freq_opcode(dev_priv, val); - if (val < dev_priv->rps.min_freq || - val > dev_priv->rps.max_freq || - val > dev_priv->rps.max_freq_softlimit) { - mutex_unlock(&dev_priv->rps.hw_lock); + if (val < dev_priv->pm.rps.min_freq || + val > dev_priv->pm.rps.max_freq || + val > dev_priv->pm.rps.max_freq_softlimit) { + mutex_unlock(&dev_priv->pm.pcu_lock); intel_runtime_pm_put(dev_priv); return -EINVAL; } - dev_priv->rps.min_freq_softlimit = val; + dev_priv->pm.rps.min_freq_softlimit = val; - val = clamp_t(int, dev_priv->rps.cur_freq, - dev_priv->rps.min_freq_softlimit, - dev_priv->rps.max_freq_softlimit); + val = clamp_t(int, dev_priv->pm.rps.cur_freq, + dev_priv->pm.rps.min_freq_softlimit, + dev_priv->pm.rps.max_freq_softlimit); /* We still need *_set_rps to process the new min_delay and * update the interrupt limits and PMINTRMSK even though * frequency request may be unchanged. */ ret = intel_set_rps(dev_priv, val); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); intel_runtime_pm_put(dev_priv); @@ -451,11 +451,11 @@ static ssize_t gt_rp_mhz_show(struct device *kdev, struct device_attribute *attr u32 val; if (attr == &dev_attr_gt_RP0_freq_mhz) - val = intel_gpu_freq(dev_priv, dev_priv->rps.rp0_freq); + val = intel_gpu_freq(dev_priv, dev_priv->pm.rps.rp0_freq); else if (attr == &dev_attr_gt_RP1_freq_mhz) - val = intel_gpu_freq(dev_priv, dev_priv->rps.rp1_freq); + val = intel_gpu_freq(dev_priv, dev_priv->pm.rps.rp1_freq); else if (attr == &dev_attr_gt_RPn_freq_mhz) - val = intel_gpu_freq(dev_priv, dev_priv->rps.min_freq); + val = intel_gpu_freq(dev_priv, dev_priv->pm.rps.min_freq); else BUG(); diff --git a/drivers/gpu/drm/i915/intel_cdclk.c b/drivers/gpu/drm/i915/intel_cdclk.c index 87fc42b..6cf7eef 100644 --- a/drivers/gpu/drm/i915/intel_cdclk.c +++ b/drivers/gpu/drm/i915/intel_cdclk.c @@ -503,7 +503,7 @@ static void vlv_set_cdclk(struct drm_i915_private *dev_priv, else cmd = 0; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ); val &= ~DSPFREQGUAR_MASK; val |= (cmd << DSPFREQGUAR_SHIFT); @@ -513,7 +513,7 @@ static void vlv_set_cdclk(struct drm_i915_private *dev_priv, 50)) { DRM_ERROR("timed out waiting for CDclk change\n"); } - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); mutex_lock(&dev_priv->sb_lock); @@ -590,7 +590,7 @@ static void chv_set_cdclk(struct drm_i915_private *dev_priv, */ cmd = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, cdclk) - 1; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ); val &= ~DSPFREQGUAR_MASK_CHV; val |= (cmd << DSPFREQGUAR_SHIFT_CHV); @@ -600,7 +600,7 @@ static void chv_set_cdclk(struct drm_i915_private *dev_priv, 50)) { DRM_ERROR("timed out waiting for CDclk change\n"); } - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); intel_update_cdclk(dev_priv); @@ -656,10 +656,10 @@ static void bdw_set_cdclk(struct drm_i915_private *dev_priv, "trying to change cdclk frequency with cdclk not enabled\n")) return; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); ret = sandybridge_pcode_write(dev_priv, BDW_PCODE_DISPLAY_FREQ_CHANGE_REQ, 0x0); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); if (ret) { DRM_ERROR("failed to inform pcode about cdclk change\n"); return; @@ -712,9 +712,9 @@ static void bdw_set_cdclk(struct drm_i915_private *dev_priv, LCPLL_CD_SOURCE_FCLK_DONE) == 0, 1)) DRM_ERROR("Switching back to LCPLL failed\n"); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ, data); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); I915_WRITE(CDCLK_FREQ, DIV_ROUND_CLOSEST(cdclk, 1000) - 1); @@ -928,12 +928,12 @@ static void skl_set_cdclk(struct drm_i915_private *dev_priv, WARN_ON((cdclk == 24000) != (vco == 0)); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); ret = skl_pcode_request(dev_priv, SKL_PCODE_CDCLK_CONTROL, SKL_CDCLK_PREPARE_FOR_CHANGE, SKL_CDCLK_READY_FOR_CHANGE, SKL_CDCLK_READY_FOR_CHANGE, 3); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); if (ret) { DRM_ERROR("Failed to inform PCU about cdclk change (%d)\n", ret); @@ -975,9 +975,9 @@ static void skl_set_cdclk(struct drm_i915_private *dev_priv, POSTING_READ(CDCLK_CTL); /* inform PCU of the change */ - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); sandybridge_pcode_write(dev_priv, SKL_PCODE_CDCLK_CONTROL, pcu_ack); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); intel_update_cdclk(dev_priv); } @@ -1268,10 +1268,10 @@ static void bxt_set_cdclk(struct drm_i915_private *dev_priv, } /* Inform power controller of upcoming frequency change */ - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); ret = sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ, 0x80000000); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); if (ret) { DRM_ERROR("PCode CDCLK freq change notify failed (err %d, freq %d)\n", @@ -1300,10 +1300,10 @@ static void bxt_set_cdclk(struct drm_i915_private *dev_priv, val |= BXT_CDCLK_SSA_PRECHARGE_ENABLE; I915_WRITE(CDCLK_CTL, val); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); ret = sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ, DIV_ROUND_UP(cdclk, 25000)); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); if (ret) { DRM_ERROR("PCode CDCLK freq set failed, (err %d, freq %d)\n", @@ -1518,12 +1518,12 @@ static void cnl_set_cdclk(struct drm_i915_private *dev_priv, u32 val, divider, pcu_ack; int ret; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); ret = skl_pcode_request(dev_priv, SKL_PCODE_CDCLK_CONTROL, SKL_CDCLK_PREPARE_FOR_CHANGE, SKL_CDCLK_READY_FOR_CHANGE, SKL_CDCLK_READY_FOR_CHANGE, 3); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); if (ret) { DRM_ERROR("Failed to inform PCU about cdclk change (%d)\n", ret); @@ -1575,9 +1575,9 @@ static void cnl_set_cdclk(struct drm_i915_private *dev_priv, I915_WRITE(CDCLK_CTL, val); /* inform PCU of the change */ - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); sandybridge_pcode_write(dev_priv, SKL_PCODE_CDCLK_CONTROL, pcu_ack); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); intel_update_cdclk(dev_priv); } diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index cdb2e25..41b21e4 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -4955,10 +4955,10 @@ void hsw_enable_ips(struct intel_crtc *crtc) assert_plane_enabled(dev_priv, crtc->plane); if (IS_BROADWELL(dev_priv)) { - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); WARN_ON(sandybridge_pcode_write(dev_priv, DISPLAY_IPS_CONTROL, IPS_ENABLE | IPS_PCODE_CONTROL)); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); /* Quoting Art Runyan: "its not safe to expect any particular * value in IPS_CTL bit 31 after enabling IPS through the * mailbox." Moreover, the mailbox may return a bogus state, @@ -4988,9 +4988,9 @@ void hsw_disable_ips(struct intel_crtc *crtc) assert_plane_enabled(dev_priv, crtc->plane); if (IS_BROADWELL(dev_priv)) { - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); WARN_ON(sandybridge_pcode_write(dev_priv, DISPLAY_IPS_CONTROL, 0)); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); /* wait for pcode to finish disabling IPS, which may take up to 42ms */ if (intel_wait_for_register(dev_priv, IPS_CTL, IPS_ENABLE, 0, @@ -8872,11 +8872,11 @@ static uint32_t hsw_read_dcomp(struct drm_i915_private *dev_priv) static void hsw_write_dcomp(struct drm_i915_private *dev_priv, uint32_t val) { if (IS_HASWELL(dev_priv)) { - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); if (sandybridge_pcode_write(dev_priv, GEN6_PCODE_WRITE_D_COMP, val)) DRM_DEBUG_KMS("Failed to write to D_COMP\n"); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); } else { I915_WRITE(D_COMP_BDW, val); POSTING_READ(D_COMP_BDW); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index b5bfddd..c28e690 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1243,7 +1243,7 @@ void intel_pch_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv, static inline u32 gen6_sanitize_rps_pm_mask(const struct drm_i915_private *i915, u32 mask) { - return mask & ~i915->rps.pm_intrmsk_mbz; + return mask & ~i915->pm.rps.pm_intrmsk_mbz; } void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 2ad6ebf..bf3349e 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -322,7 +322,7 @@ static void chv_set_memory_dvfs(struct drm_i915_private *dev_priv, bool enable) { u32 val; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); val = vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2); if (enable) @@ -337,14 +337,14 @@ static void chv_set_memory_dvfs(struct drm_i915_private *dev_priv, bool enable) FORCE_DDR_FREQ_REQ_ACK) == 0, 3)) DRM_ERROR("timed out waiting for Punit DDR DVFS request\n"); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); } static void chv_set_memory_pm5(struct drm_i915_private *dev_priv, bool enable) { u32 val; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ); if (enable) @@ -353,7 +353,7 @@ static void chv_set_memory_pm5(struct drm_i915_private *dev_priv, bool enable) val &= ~DSP_MAXFIFO_PM5_ENABLE; vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); } #define FW_WM(value, plane) \ @@ -2790,11 +2790,11 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv, /* read the first set of memory latencies[0:3] */ val = 0; /* data0 to be programmed to 0 for first set */ - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); ret = sandybridge_pcode_read(dev_priv, GEN9_PCODE_READ_MEM_LATENCY, &val); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); if (ret) { DRM_ERROR("SKL Mailbox read error = %d\n", ret); @@ -2811,11 +2811,11 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv, /* read the second set of memory latencies[4:7] */ val = 1; /* data0 to be programmed to 1 for second set */ - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); ret = sandybridge_pcode_read(dev_priv, GEN9_PCODE_READ_MEM_LATENCY, &val); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); if (ret) { DRM_ERROR("SKL Mailbox read error = %d\n", ret); return; @@ -3608,13 +3608,13 @@ static bool skl_needs_memory_bw_wa(struct intel_atomic_state *state) return 0; DRM_DEBUG_KMS("Enabling the SAGV\n"); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); ret = sandybridge_pcode_write(dev_priv, GEN9_PCODE_SAGV_CONTROL, GEN9_SAGV_ENABLE); /* We don't need to wait for the SAGV when enabling */ - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); /* * Some skl systems, pre-release machines in particular, @@ -3645,14 +3645,14 @@ static bool skl_needs_memory_bw_wa(struct intel_atomic_state *state) return 0; DRM_DEBUG_KMS("Disabling the SAGV\n"); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); /* bspec says to keep retrying for at least 1 ms */ ret = skl_pcode_request(dev_priv, GEN9_PCODE_SAGV_CONTROL, GEN9_SAGV_DISABLE, GEN9_SAGV_IS_DISABLED, GEN9_SAGV_IS_DISABLED, 1); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); /* * Some skl systems, pre-release machines in particular, @@ -5619,7 +5619,7 @@ void vlv_wm_get_hw_state(struct drm_device *dev) wm->level = VLV_WM_LEVEL_PM2; if (IS_CHERRYVIEW(dev_priv)) { - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ); if (val & DSP_MAXFIFO_PM5_ENABLE) @@ -5649,7 +5649,7 @@ void vlv_wm_get_hw_state(struct drm_device *dev) wm->level = VLV_WM_LEVEL_DDR_DVFS; } - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); } for_each_intel_crtc(dev, crtc) { @@ -5995,13 +5995,13 @@ static u32 intel_rps_limits(struct drm_i915_private *dev_priv, u8 val) * frequency, if the down threshold expires in that window we will not * receive a down interrupt. */ if (INTEL_GEN(dev_priv) >= 9) { - limits = (dev_priv->rps.max_freq_softlimit) << 23; - if (val <= dev_priv->rps.min_freq_softlimit) - limits |= (dev_priv->rps.min_freq_softlimit) << 14; + limits = (dev_priv->pm.rps.max_freq_softlimit) << 23; + if (val <= dev_priv->pm.rps.min_freq_softlimit) + limits |= (dev_priv->pm.rps.min_freq_softlimit) << 14; } else { - limits = dev_priv->rps.max_freq_softlimit << 24; - if (val <= dev_priv->rps.min_freq_softlimit) - limits |= dev_priv->rps.min_freq_softlimit << 16; + limits = dev_priv->pm.rps.max_freq_softlimit << 24; + if (val <= dev_priv->pm.rps.min_freq_softlimit) + limits |= dev_priv->pm.rps.min_freq_softlimit << 16; } return limits; @@ -6009,39 +6009,40 @@ static u32 intel_rps_limits(struct drm_i915_private *dev_priv, u8 val) static void gen6_set_rps_thresholds(struct drm_i915_private *dev_priv, u8 val) { + struct intel_rps *rps = &dev_priv->pm.rps; int new_power; u32 threshold_up = 0, threshold_down = 0; /* in % */ u32 ei_up = 0, ei_down = 0; - new_power = dev_priv->rps.power; - switch (dev_priv->rps.power) { + new_power = rps->power; + switch (rps->power) { case LOW_POWER: - if (val > dev_priv->rps.efficient_freq + 1 && - val > dev_priv->rps.cur_freq) + if (val > rps->efficient_freq + 1 && + val > rps->cur_freq) new_power = BETWEEN; break; case BETWEEN: - if (val <= dev_priv->rps.efficient_freq && - val < dev_priv->rps.cur_freq) + if (val <= rps->efficient_freq && + val < rps->cur_freq) new_power = LOW_POWER; - else if (val >= dev_priv->rps.rp0_freq && - val > dev_priv->rps.cur_freq) + else if (val >= rps->rp0_freq && + val > rps->cur_freq) new_power = HIGH_POWER; break; case HIGH_POWER: - if (val < (dev_priv->rps.rp1_freq + dev_priv->rps.rp0_freq) >> 1 && - val < dev_priv->rps.cur_freq) + if (val < (rps->rp1_freq + rps->rp0_freq) >> 1 && + val < rps->cur_freq) new_power = BETWEEN; break; } /* Max/min bins are special */ - if (val <= dev_priv->rps.min_freq_softlimit) + if (val <= rps->min_freq_softlimit) new_power = LOW_POWER; - if (val >= dev_priv->rps.max_freq_softlimit) + if (val >= rps->max_freq_softlimit) new_power = HIGH_POWER; - if (new_power == dev_priv->rps.power) + if (new_power == rps->power) return; /* Note the units here are not exactly 1us, but 1280ns. */ @@ -6104,10 +6105,10 @@ static void gen6_set_rps_thresholds(struct drm_i915_private *dev_priv, u8 val) GEN6_RP_DOWN_IDLE_AVG); skip_hw_write: - dev_priv->rps.power = new_power; - dev_priv->rps.up_threshold = threshold_up; - dev_priv->rps.down_threshold = threshold_down; - dev_priv->rps.last_adj = 0; + rps->power = new_power; + rps->up_threshold = threshold_up; + rps->down_threshold = threshold_down; + rps->last_adj = 0; } static u32 gen6_rps_pm_mask(struct drm_i915_private *dev_priv, u8 val) @@ -6115,9 +6116,9 @@ static u32 gen6_rps_pm_mask(struct drm_i915_private *dev_priv, u8 val) u32 mask = 0; /* We use UP_EI_EXPIRED interupts for both up/down in manual mode */ - if (val > dev_priv->rps.min_freq_softlimit) + if (val > dev_priv->pm.rps.min_freq_softlimit) mask |= GEN6_PM_RP_UP_EI_EXPIRED | GEN6_PM_RP_DOWN_THRESHOLD | GEN6_PM_RP_DOWN_TIMEOUT; - if (val < dev_priv->rps.max_freq_softlimit) + if (val < dev_priv->pm.rps.max_freq_softlimit) mask |= GEN6_PM_RP_UP_EI_EXPIRED | GEN6_PM_RP_UP_THRESHOLD; mask &= dev_priv->pm_rps_events; @@ -6133,7 +6134,7 @@ static int gen6_set_rps(struct drm_i915_private *dev_priv, u8 val) /* min/max delay may still have been modified so be sure to * write the limits value. */ - if (val != dev_priv->rps.cur_freq) { + if (val != dev_priv->pm.rps.cur_freq) { gen6_set_rps_thresholds(dev_priv, val); if (INTEL_GEN(dev_priv) >= 9) @@ -6155,7 +6156,7 @@ static int gen6_set_rps(struct drm_i915_private *dev_priv, u8 val) I915_WRITE(GEN6_RP_INTERRUPT_LIMITS, intel_rps_limits(dev_priv, val)); I915_WRITE(GEN6_PMINTRMSK, gen6_rps_pm_mask(dev_priv, val)); - dev_priv->rps.cur_freq = val; + dev_priv->pm.rps.cur_freq = val; trace_intel_gpu_freq_change(intel_gpu_freq(dev_priv, val)); return 0; @@ -6171,7 +6172,7 @@ static int valleyview_set_rps(struct drm_i915_private *dev_priv, u8 val) I915_WRITE(GEN6_PMINTRMSK, gen6_rps_pm_mask(dev_priv, val)); - if (val != dev_priv->rps.cur_freq) { + if (val != dev_priv->pm.rps.cur_freq) { err = vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, val); if (err) return err; @@ -6179,7 +6180,7 @@ static int valleyview_set_rps(struct drm_i915_private *dev_priv, u8 val) gen6_set_rps_thresholds(dev_priv, val); } - dev_priv->rps.cur_freq = val; + dev_priv->pm.rps.cur_freq = val; trace_intel_gpu_freq_change(intel_gpu_freq(dev_priv, val)); return 0; @@ -6194,10 +6195,10 @@ static int valleyview_set_rps(struct drm_i915_private *dev_priv, u8 val) */ static void vlv_set_rps_idle(struct drm_i915_private *dev_priv) { - u32 val = dev_priv->rps.idle_freq; + u32 val = dev_priv->pm.rps.idle_freq; int err; - if (dev_priv->rps.cur_freq <= val) + if (dev_priv->pm.rps.cur_freq <= val) return; /* The punit delays the write of the frequency and voltage until it @@ -6222,30 +6223,32 @@ static void vlv_set_rps_idle(struct drm_i915_private *dev_priv) void gen6_rps_busy(struct drm_i915_private *dev_priv) { - mutex_lock(&dev_priv->rps.hw_lock); - if (dev_priv->rps.enabled) { + struct intel_rps *rps = &dev_priv->pm.rps; + + mutex_lock(&dev_priv->pm.pcu_lock); + if (rps->enabled) { u8 freq; if (dev_priv->pm_rps_events & GEN6_PM_RP_UP_EI_EXPIRED) gen6_rps_reset_ei(dev_priv); I915_WRITE(GEN6_PMINTRMSK, - gen6_rps_pm_mask(dev_priv, dev_priv->rps.cur_freq)); + gen6_rps_pm_mask(dev_priv, rps->cur_freq)); gen6_enable_rps_interrupts(dev_priv); /* Use the user's desired frequency as a guide, but for better * performance, jump directly to RPe as our starting frequency. */ - freq = max(dev_priv->rps.cur_freq, - dev_priv->rps.efficient_freq); + freq = max(rps->cur_freq, + rps->efficient_freq); if (intel_set_rps(dev_priv, clamp(freq, - dev_priv->rps.min_freq_softlimit, - dev_priv->rps.max_freq_softlimit))) + rps->min_freq_softlimit, + rps->max_freq_softlimit))) DRM_DEBUG_DRIVER("Failed to set idle frequency\n"); } - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); } void gen6_rps_idle(struct drm_i915_private *dev_priv) @@ -6257,17 +6260,17 @@ void gen6_rps_idle(struct drm_i915_private *dev_priv) */ gen6_disable_rps_interrupts(dev_priv); - mutex_lock(&dev_priv->rps.hw_lock); - if (dev_priv->rps.enabled) { + mutex_lock(&dev_priv->pm.pcu_lock); + if (dev_priv->pm.rps.enabled) { if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) vlv_set_rps_idle(dev_priv); else - gen6_set_rps(dev_priv, dev_priv->rps.idle_freq); - dev_priv->rps.last_adj = 0; + gen6_set_rps(dev_priv, dev_priv->pm.rps.idle_freq); + dev_priv->pm.rps.last_adj = 0; I915_WRITE(GEN6_PMINTRMSK, gen6_sanitize_rps_pm_mask(dev_priv, ~0)); } - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); } void gen6_rps_boost(struct drm_i915_gem_request *rq, @@ -6280,13 +6283,13 @@ void gen6_rps_boost(struct drm_i915_gem_request *rq, /* This is intentionally racy! We peek at the state here, then * validate inside the RPS worker. */ - if (!i915->rps.enabled) + if (!i915->pm.rps.enabled) return; boost = false; spin_lock_irqsave(&rq->lock, flags); if (!rq->waitboost && !i915_gem_request_completed(rq)) { - atomic_inc(&i915->rps.num_waiters); + atomic_inc(&i915->pm.rps.num_waiters); rq->waitboost = true; boost = true; } @@ -6294,22 +6297,22 @@ void gen6_rps_boost(struct drm_i915_gem_request *rq, if (!boost) return; - if (READ_ONCE(i915->rps.cur_freq) < i915->rps.boost_freq) - schedule_work(&i915->rps.work); + if (READ_ONCE(i915->pm.rps.cur_freq) < i915->pm.rps.boost_freq) + schedule_work(&i915->pm.rps.work); - atomic_inc(rps ? &rps->boosts : &i915->rps.boosts); + atomic_inc(rps ? &rps->boosts : &i915->pm.rps.boosts); } int intel_set_rps(struct drm_i915_private *dev_priv, u8 val) { int err; - lockdep_assert_held(&dev_priv->rps.hw_lock); - GEM_BUG_ON(val > dev_priv->rps.max_freq); - GEM_BUG_ON(val < dev_priv->rps.min_freq); + lockdep_assert_held(&dev_priv->pm.pcu_lock); + GEM_BUG_ON(val > dev_priv->pm.rps.max_freq); + GEM_BUG_ON(val < dev_priv->pm.rps.min_freq); - if (!dev_priv->rps.enabled) { - dev_priv->rps.cur_freq = val; + if (!dev_priv->pm.rps.enabled) { + dev_priv->pm.rps.cur_freq = val; return 0; } @@ -6496,19 +6499,19 @@ static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv) /* static values from HW: RP0 > RP1 > RPn (min_freq) */ if (IS_GEN9_LP(dev_priv)) { u32 rp_state_cap = I915_READ(BXT_RP_STATE_CAP); - dev_priv->rps.rp0_freq = (rp_state_cap >> 16) & 0xff; - dev_priv->rps.rp1_freq = (rp_state_cap >> 8) & 0xff; - dev_priv->rps.min_freq = (rp_state_cap >> 0) & 0xff; + dev_priv->pm.rps.rp0_freq = (rp_state_cap >> 16) & 0xff; + dev_priv->pm.rps.rp1_freq = (rp_state_cap >> 8) & 0xff; + dev_priv->pm.rps.min_freq = (rp_state_cap >> 0) & 0xff; } else { u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP); - dev_priv->rps.rp0_freq = (rp_state_cap >> 0) & 0xff; - dev_priv->rps.rp1_freq = (rp_state_cap >> 8) & 0xff; - dev_priv->rps.min_freq = (rp_state_cap >> 16) & 0xff; + dev_priv->pm.rps.rp0_freq = (rp_state_cap >> 0) & 0xff; + dev_priv->pm.rps.rp1_freq = (rp_state_cap >> 8) & 0xff; + dev_priv->pm.rps.min_freq = (rp_state_cap >> 16) & 0xff; } /* hw_max = RP0 until we check for overclocking */ - dev_priv->rps.max_freq = dev_priv->rps.rp0_freq; + dev_priv->pm.rps.max_freq = dev_priv->pm.rps.rp0_freq; - dev_priv->rps.efficient_freq = dev_priv->rps.rp1_freq; + dev_priv->pm.rps.efficient_freq = dev_priv->pm.rps.rp1_freq; if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv) || IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { u32 ddcc_status = 0; @@ -6516,33 +6519,33 @@ static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv) if (sandybridge_pcode_read(dev_priv, HSW_PCODE_DYNAMIC_DUTY_CYCLE_CONTROL, &ddcc_status) == 0) - dev_priv->rps.efficient_freq = + dev_priv->pm.rps.efficient_freq = clamp_t(u8, ((ddcc_status >> 8) & 0xff), - dev_priv->rps.min_freq, - dev_priv->rps.max_freq); + dev_priv->pm.rps.min_freq, + dev_priv->pm.rps.max_freq); } if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { /* Store the frequency values in 16.66 MHZ units, which is * the natural hardware unit for SKL */ - dev_priv->rps.rp0_freq *= GEN9_FREQ_SCALER; - dev_priv->rps.rp1_freq *= GEN9_FREQ_SCALER; - dev_priv->rps.min_freq *= GEN9_FREQ_SCALER; - dev_priv->rps.max_freq *= GEN9_FREQ_SCALER; - dev_priv->rps.efficient_freq *= GEN9_FREQ_SCALER; + dev_priv->pm.rps.rp0_freq *= GEN9_FREQ_SCALER; + dev_priv->pm.rps.rp1_freq *= GEN9_FREQ_SCALER; + dev_priv->pm.rps.min_freq *= GEN9_FREQ_SCALER; + dev_priv->pm.rps.max_freq *= GEN9_FREQ_SCALER; + dev_priv->pm.rps.efficient_freq *= GEN9_FREQ_SCALER; } } static void reset_rps(struct drm_i915_private *dev_priv, int (*set)(struct drm_i915_private *, u8)) { - u8 freq = dev_priv->rps.cur_freq; + u8 freq = dev_priv->pm.rps.cur_freq; /* force a reset */ - dev_priv->rps.power = -1; - dev_priv->rps.cur_freq = -1; + dev_priv->pm.rps.power = -1; + dev_priv->pm.rps.cur_freq = -1; if (set(dev_priv, freq)) DRM_ERROR("Failed to reset RPS to initial values\n"); @@ -6555,7 +6558,7 @@ static void gen9_enable_rps(struct drm_i915_private *dev_priv) /* Program defaults and thresholds for RPS*/ I915_WRITE(GEN6_RC_VIDEO_FREQ, - GEN9_FREQUENCY(dev_priv->rps.rp1_freq)); + GEN9_FREQUENCY(dev_priv->pm.rps.rp1_freq)); /* 1 second timeout*/ I915_WRITE(GEN6_RP_DOWN_TIMEOUT, @@ -6672,16 +6675,16 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) /* 1 Program defaults and thresholds for RPS*/ I915_WRITE(GEN6_RPNSWREQ, - HSW_FREQUENCY(dev_priv->rps.rp1_freq)); + HSW_FREQUENCY(dev_priv->pm.rps.rp1_freq)); I915_WRITE(GEN6_RC_VIDEO_FREQ, - HSW_FREQUENCY(dev_priv->rps.rp1_freq)); + HSW_FREQUENCY(dev_priv->pm.rps.rp1_freq)); /* NB: Docs say 1s, and 1000000 - which aren't equivalent */ I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 100000000 / 128); /* 1 second timeout */ /* Docs recommend 900MHz, and 300 MHz respectively */ I915_WRITE(GEN6_RP_INTERRUPT_LIMITS, - dev_priv->rps.max_freq_softlimit << 24 | - dev_priv->rps.min_freq_softlimit << 16); + dev_priv->pm.rps.max_freq_softlimit << 24 | + dev_priv->pm.rps.min_freq_softlimit << 16); I915_WRITE(GEN6_RP_UP_THRESHOLD, 7600000 / 128); /* 76ms busyness per EI, 90% */ I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 31300000 / 128); /* 313ms busyness per EI, 70%*/ @@ -6713,7 +6716,7 @@ static void gen6_enable_rc6(struct drm_i915_private *dev_priv) int rc6_mode; int ret; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); I915_WRITE(GEN6_RC_STATE, 0); @@ -6788,7 +6791,7 @@ static void gen6_enable_rc6(struct drm_i915_private *dev_priv) static void gen6_enable_rps(struct drm_i915_private *dev_priv) { - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); /* Here begins a magic sequence of register writes to enable * auto-downclocking. @@ -6816,7 +6819,7 @@ static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) int scaling_factor = 180; struct cpufreq_policy *policy; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); policy = cpufreq_cpu_get(0); if (policy) { @@ -6839,11 +6842,11 @@ static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { /* Convert GT frequency to 50 HZ units */ - min_gpu_freq = dev_priv->rps.min_freq / GEN9_FREQ_SCALER; - max_gpu_freq = dev_priv->rps.max_freq / GEN9_FREQ_SCALER; + min_gpu_freq = dev_priv->pm.rps.min_freq / GEN9_FREQ_SCALER; + max_gpu_freq = dev_priv->pm.rps.max_freq / GEN9_FREQ_SCALER; } else { - min_gpu_freq = dev_priv->rps.min_freq; - max_gpu_freq = dev_priv->rps.max_freq; + min_gpu_freq = dev_priv->pm.rps.min_freq; + max_gpu_freq = dev_priv->pm.rps.max_freq; } /* @@ -7094,17 +7097,18 @@ static void valleyview_cleanup_pctx(struct drm_i915_private *dev_priv) static void vlv_init_gpll_ref_freq(struct drm_i915_private *dev_priv) { - dev_priv->rps.gpll_ref_freq = + dev_priv->pm.rps.gpll_ref_freq = vlv_get_cck_clock(dev_priv, "GPLL ref", CCK_GPLL_CLOCK_CONTROL, dev_priv->czclk_freq); DRM_DEBUG_DRIVER("GPLL reference freq: %d kHz\n", - dev_priv->rps.gpll_ref_freq); + dev_priv->pm.rps.gpll_ref_freq); } static void valleyview_init_gt_powersave(struct drm_i915_private *dev_priv) { + struct intel_rps *rps = &dev_priv->pm.rps; u32 val; valleyview_setup_pctx(dev_priv); @@ -7126,30 +7130,31 @@ static void valleyview_init_gt_powersave(struct drm_i915_private *dev_priv) } DRM_DEBUG_DRIVER("DDR speed: %d MHz\n", dev_priv->mem_freq); - dev_priv->rps.max_freq = valleyview_rps_max_freq(dev_priv); - dev_priv->rps.rp0_freq = dev_priv->rps.max_freq; + rps->max_freq = valleyview_rps_max_freq(dev_priv); + rps->rp0_freq = rps->max_freq; DRM_DEBUG_DRIVER("max GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.max_freq), - dev_priv->rps.max_freq); + intel_gpu_freq(dev_priv, rps->max_freq), + rps->max_freq); - dev_priv->rps.efficient_freq = valleyview_rps_rpe_freq(dev_priv); + rps->efficient_freq = valleyview_rps_rpe_freq(dev_priv); DRM_DEBUG_DRIVER("RPe GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.efficient_freq), - dev_priv->rps.efficient_freq); + intel_gpu_freq(dev_priv, rps->efficient_freq), + rps->efficient_freq); - dev_priv->rps.rp1_freq = valleyview_rps_guar_freq(dev_priv); + rps->rp1_freq = valleyview_rps_guar_freq(dev_priv); DRM_DEBUG_DRIVER("RP1(Guar Freq) GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.rp1_freq), - dev_priv->rps.rp1_freq); + intel_gpu_freq(dev_priv, rps->rp1_freq), + rps->rp1_freq); - dev_priv->rps.min_freq = valleyview_rps_min_freq(dev_priv); + rps->min_freq = valleyview_rps_min_freq(dev_priv); DRM_DEBUG_DRIVER("min GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.min_freq), - dev_priv->rps.min_freq); + intel_gpu_freq(dev_priv, rps->min_freq), + rps->min_freq); } static void cherryview_init_gt_powersave(struct drm_i915_private *dev_priv) { + struct intel_rps *rps = &dev_priv->pm.rps; u32 val; cherryview_setup_pctx(dev_priv); @@ -7170,31 +7175,29 @@ static void cherryview_init_gt_powersave(struct drm_i915_private *dev_priv) } DRM_DEBUG_DRIVER("DDR speed: %d MHz\n", dev_priv->mem_freq); - dev_priv->rps.max_freq = cherryview_rps_max_freq(dev_priv); - dev_priv->rps.rp0_freq = dev_priv->rps.max_freq; + rps->max_freq = cherryview_rps_max_freq(dev_priv); + rps->rp0_freq = rps->max_freq; DRM_DEBUG_DRIVER("max GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.max_freq), - dev_priv->rps.max_freq); + intel_gpu_freq(dev_priv, rps->max_freq), + rps->max_freq); - dev_priv->rps.efficient_freq = cherryview_rps_rpe_freq(dev_priv); + rps->efficient_freq = cherryview_rps_rpe_freq(dev_priv); DRM_DEBUG_DRIVER("RPe GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.efficient_freq), - dev_priv->rps.efficient_freq); + intel_gpu_freq(dev_priv, rps->efficient_freq), + rps->efficient_freq); - dev_priv->rps.rp1_freq = cherryview_rps_guar_freq(dev_priv); + rps->rp1_freq = cherryview_rps_guar_freq(dev_priv); DRM_DEBUG_DRIVER("RP1(Guar) GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.rp1_freq), - dev_priv->rps.rp1_freq); + intel_gpu_freq(dev_priv, rps->rp1_freq), + rps->rp1_freq); - dev_priv->rps.min_freq = cherryview_rps_min_freq(dev_priv); + rps->min_freq = cherryview_rps_min_freq(dev_priv); DRM_DEBUG_DRIVER("min GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.min_freq), - dev_priv->rps.min_freq); + intel_gpu_freq(dev_priv, rps->min_freq), + rps->min_freq); - WARN_ONCE((dev_priv->rps.max_freq | - dev_priv->rps.efficient_freq | - dev_priv->rps.rp1_freq | - dev_priv->rps.min_freq) & 1, + WARN_ONCE((rps->max_freq | rps->efficient_freq | rps->rp1_freq | + rps->min_freq) & 1, "Odd GPU freq values\n"); } @@ -7209,7 +7212,7 @@ static void cherryview_enable_rc6(struct drm_i915_private *dev_priv) enum intel_engine_id id; u32 gtfifodbg, rc6_mode = 0, pcbr; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); gtfifodbg = I915_READ(GTFIFODBG) & ~(GT_FIFO_SBDEDICATE_FREE_ENTRY_CHV | GT_FIFO_FREE_ENTRIES_CHV); @@ -7263,7 +7266,7 @@ static void cherryview_enable_rps(struct drm_i915_private *dev_priv) { u32 val; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); @@ -7309,7 +7312,7 @@ static void valleyview_enable_rc6(struct drm_i915_private *dev_priv) enum intel_engine_id id; u32 gtfifodbg, rc6_mode = 0; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); valleyview_check_pctx(dev_priv); @@ -7356,7 +7359,7 @@ static void valleyview_enable_rps(struct drm_i915_private *dev_priv) { u32 val; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); @@ -7583,7 +7586,7 @@ static unsigned long __i915_gfx_val(struct drm_i915_private *dev_priv) lockdep_assert_held(&mchdev_lock); - pxvid = I915_READ(PXVFREQ(dev_priv->rps.cur_freq)); + pxvid = I915_READ(PXVFREQ(dev_priv->pm.rps.cur_freq)); pxvid = (pxvid >> 24) & 0x7f; ext_v = pvid_to_extvid(dev_priv, pxvid); @@ -7870,6 +7873,8 @@ static void intel_init_emon(struct drm_i915_private *dev_priv) void intel_init_gt_powersave(struct drm_i915_private *dev_priv) { + struct intel_rps *rps = &dev_priv->pm.rps; + /* * RPM depends on RC6 to save restore the GT HW context, so make RC6 a * requirement. @@ -7880,7 +7885,7 @@ void intel_init_gt_powersave(struct drm_i915_private *dev_priv) } mutex_lock(&dev_priv->drm.struct_mutex); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); /* Initialize RPS limits (for userspace) */ if (IS_CHERRYVIEW(dev_priv)) @@ -7891,16 +7896,16 @@ void intel_init_gt_powersave(struct drm_i915_private *dev_priv) gen6_init_rps_frequencies(dev_priv); /* Derive initial user preferences/limits from the hardware limits */ - dev_priv->rps.idle_freq = dev_priv->rps.min_freq; - dev_priv->rps.cur_freq = dev_priv->rps.idle_freq; + rps->idle_freq = rps->min_freq; + rps->cur_freq = rps->idle_freq; - dev_priv->rps.max_freq_softlimit = dev_priv->rps.max_freq; - dev_priv->rps.min_freq_softlimit = dev_priv->rps.min_freq; + rps->max_freq_softlimit = rps->max_freq; + rps->min_freq_softlimit = rps->min_freq; if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) - dev_priv->rps.min_freq_softlimit = + rps->min_freq_softlimit = max_t(int, - dev_priv->rps.efficient_freq, + rps->efficient_freq, intel_freq_opcode(dev_priv, 450)); /* After setting max-softlimit, find the overclock max freq */ @@ -7911,16 +7916,16 @@ void intel_init_gt_powersave(struct drm_i915_private *dev_priv) sandybridge_pcode_read(dev_priv, GEN6_READ_OC_PARAMS, ¶ms); if (params & BIT(31)) { /* OC supported */ DRM_DEBUG_DRIVER("Overclocking supported, max: %dMHz, overclock: %dMHz\n", - (dev_priv->rps.max_freq & 0xff) * 50, + (rps->max_freq & 0xff) * 50, (params & 0xff) * 50); - dev_priv->rps.max_freq = params & 0xff; + rps->max_freq = params & 0xff; } } /* Finally allow us to boost to max by default */ - dev_priv->rps.boost_freq = dev_priv->rps.max_freq; + rps->boost_freq = rps->max_freq; - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); mutex_unlock(&dev_priv->drm.struct_mutex); intel_autoenable_gt_powersave(dev_priv); @@ -7948,7 +7953,7 @@ void intel_suspend_gt_powersave(struct drm_i915_private *dev_priv) if (INTEL_GEN(dev_priv) < 6) return; - if (cancel_delayed_work_sync(&dev_priv->rps.autoenable_work)) + if (cancel_delayed_work_sync(&dev_priv->pm.autoenable_work)) intel_runtime_pm_put(dev_priv); /* gen6_rps_idle() will be called later to disable interrupts */ @@ -7956,7 +7961,7 @@ void intel_suspend_gt_powersave(struct drm_i915_private *dev_priv) void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) { - dev_priv->rps.enabled = true; /* force disabling */ + dev_priv->pm.rps.enabled = true; /* force disabling */ intel_disable_gt_powersave(dev_priv); gen6_reset_rps_interrupts(dev_priv); @@ -7964,10 +7969,10 @@ void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) { - if (!READ_ONCE(dev_priv->rps.enabled)) + if (!READ_ONCE(dev_priv->pm.rps.enabled)) return; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); if (INTEL_GEN(dev_priv) >= 9) { gen9_disable_rc6(dev_priv); @@ -7985,8 +7990,8 @@ void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) ironlake_disable_drps(dev_priv); } - dev_priv->rps.enabled = false; - mutex_unlock(&dev_priv->rps.hw_lock); + dev_priv->pm.rps.enabled = false; + mutex_unlock(&dev_priv->pm.pcu_lock); } void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) @@ -7994,14 +7999,14 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) /* We shouldn't be disabling as we submit, so this should be less * racy than it appears! */ - if (READ_ONCE(dev_priv->rps.enabled)) + if (READ_ONCE(dev_priv->pm.rps.enabled)) return; /* Powersaving is controlled by the host when inside a VM */ if (intel_vgpu_active(dev_priv)) return; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); if (IS_CHERRYVIEW(dev_priv)) { cherryview_enable_rc6(dev_priv); @@ -8027,24 +8032,24 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) intel_init_emon(dev_priv); } - WARN_ON(dev_priv->rps.max_freq < dev_priv->rps.min_freq); - WARN_ON(dev_priv->rps.idle_freq > dev_priv->rps.max_freq); + WARN_ON(dev_priv->pm.rps.max_freq < dev_priv->pm.rps.min_freq); + WARN_ON(dev_priv->pm.rps.idle_freq > dev_priv->pm.rps.max_freq); - WARN_ON(dev_priv->rps.efficient_freq < dev_priv->rps.min_freq); - WARN_ON(dev_priv->rps.efficient_freq > dev_priv->rps.max_freq); + WARN_ON(dev_priv->pm.rps.efficient_freq < dev_priv->pm.rps.min_freq); + WARN_ON(dev_priv->pm.rps.efficient_freq > dev_priv->pm.rps.max_freq); - dev_priv->rps.enabled = true; - mutex_unlock(&dev_priv->rps.hw_lock); + dev_priv->pm.rps.enabled = true; + mutex_unlock(&dev_priv->pm.pcu_lock); } static void __intel_autoenable_gt_powersave(struct work_struct *work) { struct drm_i915_private *dev_priv = - container_of(work, typeof(*dev_priv), rps.autoenable_work.work); + container_of(work, typeof(*dev_priv), pm.autoenable_work.work); struct intel_engine_cs *rcs; struct drm_i915_gem_request *req; - if (READ_ONCE(dev_priv->rps.enabled)) + if (READ_ONCE(dev_priv->pm.rps.enabled)) goto out; rcs = dev_priv->engine[RCS]; @@ -8074,7 +8079,7 @@ static void __intel_autoenable_gt_powersave(struct work_struct *work) void intel_autoenable_gt_powersave(struct drm_i915_private *dev_priv) { - if (READ_ONCE(dev_priv->rps.enabled)) + if (READ_ONCE(dev_priv->pm.rps.enabled)) return; if (IS_IRONLAKE_M(dev_priv)) { @@ -8094,7 +8099,7 @@ void intel_autoenable_gt_powersave(struct drm_i915_private *dev_priv) * runtime resume it's necessary). */ if (queue_delayed_work(dev_priv->wq, - &dev_priv->rps.autoenable_work, + &dev_priv->pm.autoenable_work, round_jiffies_up_relative(HZ))) intel_runtime_pm_get_noresume(dev_priv); } @@ -9123,7 +9128,7 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val { int status; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); /* GEN6_PCODE_* are outside of the forcewake domain, we can * use te fw I915_READ variants to reduce the amount of work @@ -9170,7 +9175,7 @@ int sandybridge_pcode_write(struct drm_i915_private *dev_priv, { int status; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); /* GEN6_PCODE_* are outside of the forcewake domain, we can * use te fw I915_READ variants to reduce the amount of work @@ -9247,7 +9252,7 @@ int skl_pcode_request(struct drm_i915_private *dev_priv, u32 mbox, u32 request, u32 status; int ret; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); #define COND skl_pcode_try_request(dev_priv, mbox, request, reply_mask, reply, \ &status) @@ -9289,31 +9294,39 @@ int skl_pcode_request(struct drm_i915_private *dev_priv, u32 mbox, u32 request, static int byt_gpu_freq(struct drm_i915_private *dev_priv, int val) { + struct intel_rps *rps = &dev_priv->pm.rps; + /* * N = val - 0xb7 * Slow = Fast = GPLL ref * N */ - return DIV_ROUND_CLOSEST(dev_priv->rps.gpll_ref_freq * (val - 0xb7), 1000); + return DIV_ROUND_CLOSEST(rps->gpll_ref_freq * (val - 0xb7), 1000); } static int byt_freq_opcode(struct drm_i915_private *dev_priv, int val) { - return DIV_ROUND_CLOSEST(1000 * val, dev_priv->rps.gpll_ref_freq) + 0xb7; + struct intel_rps *rps = &dev_priv->pm.rps; + + return DIV_ROUND_CLOSEST(1000 * val, rps->gpll_ref_freq) + 0xb7; } static int chv_gpu_freq(struct drm_i915_private *dev_priv, int val) { + struct intel_rps *rps = &dev_priv->pm.rps; + /* * N = val / 2 * CU (slow) = CU2x (fast) / 2 = GPLL ref * N / 2 */ - return DIV_ROUND_CLOSEST(dev_priv->rps.gpll_ref_freq * val, 2 * 2 * 1000); + return DIV_ROUND_CLOSEST(rps->gpll_ref_freq * val, 2 * 2 * 1000); } static int chv_freq_opcode(struct drm_i915_private *dev_priv, int val) { + struct intel_rps *rps = &dev_priv->pm.rps; + /* CHV needs even values */ - return DIV_ROUND_CLOSEST(2 * 1000 * val, dev_priv->rps.gpll_ref_freq) * 2; + return DIV_ROUND_CLOSEST(2 * 1000 * val, rps->gpll_ref_freq) * 2; } int intel_gpu_freq(struct drm_i915_private *dev_priv, int val) @@ -9344,11 +9357,11 @@ int intel_freq_opcode(struct drm_i915_private *dev_priv, int val) void intel_pm_setup(struct drm_i915_private *dev_priv) { - mutex_init(&dev_priv->rps.hw_lock); + mutex_init(&dev_priv->pm.pcu_lock); - INIT_DELAYED_WORK(&dev_priv->rps.autoenable_work, + INIT_DELAYED_WORK(&dev_priv->pm.autoenable_work, __intel_autoenable_gt_powersave); - atomic_set(&dev_priv->rps.num_waiters, 0); + atomic_set(&dev_priv->pm.rps.num_waiters, 0); dev_priv->rpm.suspended = false; atomic_set(&dev_priv->rpm.wakeref_count, 0); diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index a37ad0f..1b5d0a8 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -785,7 +785,7 @@ static void vlv_set_power_well(struct drm_i915_private *dev_priv, state = enable ? PUNIT_PWRGT_PWR_ON(power_well_id) : PUNIT_PWRGT_PWR_GATE(power_well_id); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); #define COND \ ((vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask) == state) @@ -806,7 +806,7 @@ static void vlv_set_power_well(struct drm_i915_private *dev_priv, #undef COND out: - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); } static void vlv_power_well_enable(struct drm_i915_private *dev_priv, @@ -833,7 +833,7 @@ static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv, mask = PUNIT_PWRGT_MASK(power_well_id); ctrl = PUNIT_PWRGT_PWR_ON(power_well_id); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); state = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask; /* @@ -852,7 +852,7 @@ static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv, ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL) & mask; WARN_ON(ctrl != state); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); return enabled; } @@ -1364,7 +1364,7 @@ static bool chv_pipe_power_well_enabled(struct drm_i915_private *dev_priv, bool enabled; u32 state, ctrl; - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); state = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSS_MASK(pipe); /* @@ -1381,7 +1381,7 @@ static bool chv_pipe_power_well_enabled(struct drm_i915_private *dev_priv, ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSC_MASK(pipe); WARN_ON(ctrl << 16 != state); - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); return enabled; } @@ -1396,7 +1396,7 @@ static void chv_set_pipe_power_well(struct drm_i915_private *dev_priv, state = enable ? DP_SSS_PWR_ON(pipe) : DP_SSS_PWR_GATE(pipe); - mutex_lock(&dev_priv->rps.hw_lock); + mutex_lock(&dev_priv->pm.pcu_lock); #define COND \ ((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSS_MASK(pipe)) == state) @@ -1417,7 +1417,7 @@ static void chv_set_pipe_power_well(struct drm_i915_private *dev_priv, #undef COND out: - mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->pm.pcu_lock); } static void chv_pipe_power_well_enable(struct drm_i915_private *dev_priv, diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c index 7d971cb..6a71a54 100644 --- a/drivers/gpu/drm/i915/intel_sideband.c +++ b/drivers/gpu/drm/i915/intel_sideband.c @@ -81,7 +81,7 @@ u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr) { u32 val = 0; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); mutex_lock(&dev_priv->sb_lock); vlv_sideband_rw(dev_priv, PCI_DEVFN(0, 0), IOSF_PORT_PUNIT, @@ -95,7 +95,7 @@ int vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val) { int err; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); mutex_lock(&dev_priv->sb_lock); err = vlv_sideband_rw(dev_priv, PCI_DEVFN(0, 0), IOSF_PORT_PUNIT, @@ -125,7 +125,7 @@ u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr) { u32 val = 0; - WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); + WARN_ON(!mutex_is_locked(&dev_priv->pm.pcu_lock)); mutex_lock(&dev_priv->sb_lock); vlv_sideband_rw(dev_priv, PCI_DEVFN(0, 0), IOSF_PORT_NC, -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v2 08/11] drm/i915: Rename intel_enable_rc6 to intel_rc6_enabled 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (6 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 07/11] drm/i915: Name structure in dev_priv that contains RPS/RC6 state as "pm" Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 09/11] drm/i915: Create generic function to setup LLC ring frequency table Sagar Arun Kamble ` (4 subsequent siblings) 12 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx This function gives the status of RC6, whether disabled or if enabled then which state. intel_enable_rc6 will be used for enabling RC6 in the next patch. v2: Rebase. Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Imre Deak <imre.deak@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> Reviewed-by: Ewelina Musial <ewelina.musial@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> --- drivers/gpu/drm/i915/i915_drv.c | 2 +- drivers/gpu/drm/i915/i915_sysfs.c | 2 +- drivers/gpu/drm/i915/intel_drv.h | 2 +- drivers/gpu/drm/i915/intel_guc.c | 3 ++- drivers/gpu/drm/i915/intel_pm.c | 12 ++++++------ 5 files changed, 11 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index a6af7d5..470807c 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -2502,7 +2502,7 @@ static int intel_runtime_suspend(struct device *kdev) struct drm_i915_private *dev_priv = to_i915(dev); int ret; - if (WARN_ON_ONCE(!(dev_priv->pm.rps.enabled && intel_enable_rc6()))) + if (WARN_ON_ONCE(!(dev_priv->pm.rps.enabled && intel_rc6_enabled()))) return -ENODEV; if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev_priv))) diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c index c16e907..8add849 100644 --- a/drivers/gpu/drm/i915/i915_sysfs.c +++ b/drivers/gpu/drm/i915/i915_sysfs.c @@ -49,7 +49,7 @@ static u32 calc_residency(struct drm_i915_private *dev_priv, static ssize_t show_rc6_mask(struct device *kdev, struct device_attribute *attr, char *buf) { - return snprintf(buf, PAGE_SIZE, "%x\n", intel_enable_rc6()); + return snprintf(buf, PAGE_SIZE, "%x\n", intel_rc6_enabled()); } static ssize_t diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index c28e690..02fb641 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1902,7 +1902,7 @@ int skl_check_pipe_max_pixel_rate(struct intel_crtc *intel_crtc, struct intel_crtc_state *cstate); void intel_init_ipc(struct drm_i915_private *dev_priv); void intel_enable_ipc(struct drm_i915_private *dev_priv); -static inline int intel_enable_rc6(void) +static inline int intel_rc6_enabled(void) { return i915_modparams.enable_rc6; } diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index bbe4c32..9e18c4f 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -137,7 +137,8 @@ int intel_guc_sample_forcewake(struct intel_guc *guc) action[0] = INTEL_GUC_ACTION_SAMPLE_FORCEWAKE; /* WaRsDisableCoarsePowerGating:skl,bxt */ - if (!intel_enable_rc6() || NEEDS_WaRsDisableCoarsePowerGating(dev_priv)) + if (!intel_rc6_enabled() || + NEEDS_WaRsDisableCoarsePowerGating(dev_priv)) action[1] = 0; else /* bit 0 and 1 are for Render and Media domain separately */ diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index bf3349e..f7c1d5e 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6612,7 +6612,7 @@ static void gen9_enable_rc6(struct drm_i915_private *dev_priv) I915_WRITE(GEN9_RENDER_PG_IDLE_HYSTERESIS, 25); /* 3a: Enable RC6 */ - if (intel_enable_rc6() & INTEL_RC6_ENABLE) + if (intel_rc6_enabled() & INTEL_RC6_ENABLE) rc6_mask = GEN6_RC_CTL_RC6_ENABLE; DRM_INFO("RC6 %s\n", onoff(rc6_mask & GEN6_RC_CTL_RC6_ENABLE)); I915_WRITE(GEN6_RC6_THRESHOLD, 37500); /* 37.5/125ms per EI */ @@ -6658,7 +6658,7 @@ static void gen8_enable_rc6(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RC6_THRESHOLD, 625); /* 800us/1.28 for TO */ /* 3: Enable RC6 */ - if (intel_enable_rc6() & INTEL_RC6_ENABLE) + if (intel_rc6_enabled() & INTEL_RC6_ENABLE) rc6_mask = GEN6_RC_CTL_RC6_ENABLE; intel_print_rc6_info(dev_priv, rc6_mask); @@ -6751,7 +6751,7 @@ static void gen6_enable_rc6(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RC6pp_THRESHOLD, 64000); /* unused */ /* Check if we are enabling RC6 */ - rc6_mode = intel_enable_rc6(); + rc6_mode = intel_rc6_enabled(); if (rc6_mode & INTEL_RC6_ENABLE) rc6_mask |= GEN6_RC_CTL_RC6_ENABLE; @@ -7253,7 +7253,7 @@ static void cherryview_enable_rc6(struct drm_i915_private *dev_priv) pcbr = I915_READ(VLV_PCBR); /* 3: Enable RC6 */ - if ((intel_enable_rc6() & INTEL_RC6_ENABLE) && + if ((intel_rc6_enabled() & INTEL_RC6_ENABLE) && (pcbr >> VLV_PCBR_ADDR_SHIFT)) rc6_mode = GEN7_RC_CTL_TO_MODE; @@ -7345,7 +7345,7 @@ static void valleyview_enable_rc6(struct drm_i915_private *dev_priv) VLV_MEDIA_RC6_COUNT_EN | VLV_RENDER_RC6_COUNT_EN)); - if (intel_enable_rc6() & INTEL_RC6_ENABLE) + if (intel_rc6_enabled() & INTEL_RC6_ENABLE) rc6_mode = GEN7_RC_CTL_TO_MODE | VLV_RC_CTL_CTX_RST_PARALLEL; intel_print_rc6_info(dev_priv, rc6_mode); @@ -9417,7 +9417,7 @@ u64 intel_rc6_residency_us(struct drm_i915_private *dev_priv, { u64 time_hw, units, div; - if (!intel_enable_rc6()) + if (!intel_rc6_enabled()) return 0; intel_runtime_pm_get(dev_priv); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v2 09/11] drm/i915: Create generic function to setup LLC ring frequency table 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (7 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 08/11] drm/i915: Rename intel_enable_rc6 to intel_rc6_enabled Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:43 ` Chris Wilson 2017-10-06 12:13 ` [PATCH v2 10/11] drm/i915: Create generic functions to control RC6, RPS Sagar Arun Kamble ` (3 subsequent siblings) 12 siblings, 1 reply; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx Prepared intel_update_ring_freq function to setup ring frequency for applicable platforms determined by macro HAS_LLC. v2: Replaced NEEDS_RING_FREQ_UPDATE with HAS_LLC macro. (Chris) Added check while calling from intel_enable_gt_powersave. Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> --- drivers/gpu/drm/i915/intel_pm.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index f7c1d5e..ce2dc5b 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -7871,6 +7871,11 @@ static void intel_init_emon(struct drm_i915_private *dev_priv) dev_priv->ips.corr = (lcfuse & LCFUSE_HIV_MASK); } +static inline void intel_update_ring_freq(struct drm_i915_private *i915) +{ + gen6_update_ring_freq(i915); +} + void intel_init_gt_powersave(struct drm_i915_private *dev_priv) { struct intel_rps *rps = &dev_priv->pm.rps; @@ -8017,21 +8022,20 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) } else if (INTEL_GEN(dev_priv) >= 9) { gen9_enable_rc6(dev_priv); gen9_enable_rps(dev_priv); - if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) - gen6_update_ring_freq(dev_priv); } else if (IS_BROADWELL(dev_priv)) { gen8_enable_rc6(dev_priv); gen8_enable_rps(dev_priv); - gen6_update_ring_freq(dev_priv); } else if (INTEL_GEN(dev_priv) >= 6) { gen6_enable_rc6(dev_priv); gen6_enable_rps(dev_priv); - gen6_update_ring_freq(dev_priv); } else if (IS_IRONLAKE_M(dev_priv)) { ironlake_enable_drps(dev_priv); intel_init_emon(dev_priv); } + if (HAS_LLC(dev_priv)) + intel_update_ring_freq(dev_priv); + WARN_ON(dev_priv->pm.rps.max_freq < dev_priv->pm.rps.min_freq); WARN_ON(dev_priv->pm.rps.idle_freq > dev_priv->pm.rps.max_freq); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH v2 09/11] drm/i915: Create generic function to setup LLC ring frequency table 2017-10-06 12:13 ` [PATCH v2 09/11] drm/i915: Create generic function to setup LLC ring frequency table Sagar Arun Kamble @ 2017-10-06 12:43 ` Chris Wilson 0 siblings, 0 replies; 22+ messages in thread From: Chris Wilson @ 2017-10-06 12:43 UTC (permalink / raw) To: Sagar Arun Kamble, intel-gfx Quoting Sagar Arun Kamble (2017-10-06 13:13:38) > Prepared intel_update_ring_freq function to setup ring frequency > for applicable platforms determined by macro HAS_LLC. > > v2: Replaced NEEDS_RING_FREQ_UPDATE with HAS_LLC macro. (Chris) > Added check while calling from intel_enable_gt_powersave. > > Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> > Cc: Imre Deak <imre.deak@intel.com> > Cc: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> > Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> > --- > drivers/gpu/drm/i915/intel_pm.c | 12 ++++++++---- > 1 file changed, 8 insertions(+), 4 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c > index f7c1d5e..ce2dc5b 100644 > --- a/drivers/gpu/drm/i915/intel_pm.c > +++ b/drivers/gpu/drm/i915/intel_pm.c > @@ -7871,6 +7871,11 @@ static void intel_init_emon(struct drm_i915_private *dev_priv) > dev_priv->ips.corr = (lcfuse & LCFUSE_HIV_MASK); > } > > +static inline void intel_update_ring_freq(struct drm_i915_private *i915) > +{ > + gen6_update_ring_freq(i915); > +} > + > void intel_init_gt_powersave(struct drm_i915_private *dev_priv) > { > struct intel_rps *rps = &dev_priv->pm.rps; > @@ -8017,21 +8022,20 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) > } else if (INTEL_GEN(dev_priv) >= 9) { > gen9_enable_rc6(dev_priv); > gen9_enable_rps(dev_priv); > - if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) > - gen6_update_ring_freq(dev_priv); > } else if (IS_BROADWELL(dev_priv)) { > gen8_enable_rc6(dev_priv); > gen8_enable_rps(dev_priv); > - gen6_update_ring_freq(dev_priv); > } else if (INTEL_GEN(dev_priv) >= 6) { > gen6_enable_rc6(dev_priv); > gen6_enable_rps(dev_priv); > - gen6_update_ring_freq(dev_priv); > } else if (IS_IRONLAKE_M(dev_priv)) { > ironlake_enable_drps(dev_priv); > intel_init_emon(dev_priv); > } > > + if (HAS_LLC(dev_priv)) > + intel_update_ring_freq(dev_priv); Hmm, actually makes sense at a hw architectural level. Some doubts as to whether it will hold forever, but that's a problem for tomorrow. Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v2 10/11] drm/i915: Create generic functions to control RC6, RPS 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (8 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 09/11] drm/i915: Create generic function to setup LLC ring frequency table Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:46 ` Chris Wilson 2017-10-06 12:13 ` [PATCH v2 11/11] drm/i915: Introduce separate status variable for RC6 and LLC ring frequency setup Sagar Arun Kamble ` (2 subsequent siblings) 12 siblings, 1 reply; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx Prepared generic functions intel_enable_rc6, intel_disable_rc6, intel_enable_rps and intel_disable_rps functions to setup RC6/RPS based on platforms. v2: Make intel_enable/disable_rc6/rps static. (Chris) Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> --- drivers/gpu/drm/i915/intel_pm.c | 97 ++++++++++++++++++++++++++--------------- 1 file changed, 62 insertions(+), 35 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index ce2dc5b..03264fe 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -7972,75 +7972,102 @@ void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) gen6_reset_rps_interrupts(dev_priv); } -void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) +static void intel_disable_rc6(struct drm_i915_private *dev_priv) { - if (!READ_ONCE(dev_priv->pm.rps.enabled)) - return; - - mutex_lock(&dev_priv->pm.pcu_lock); - - if (INTEL_GEN(dev_priv) >= 9) { + if (INTEL_GEN(dev_priv) >= 9) gen9_disable_rc6(dev_priv); - gen9_disable_rps(dev_priv); - } else if (IS_CHERRYVIEW(dev_priv)) { + else if (IS_CHERRYVIEW(dev_priv)) cherryview_disable_rc6(dev_priv); - cherryview_disable_rps(dev_priv); - } else if (IS_VALLEYVIEW(dev_priv)) { + else if (IS_VALLEYVIEW(dev_priv)) valleyview_disable_rc6(dev_priv); - valleyview_disable_rps(dev_priv); - } else if (INTEL_GEN(dev_priv) >= 6) { + else if (INTEL_GEN(dev_priv) >= 6) gen6_disable_rc6(dev_priv); +} + +static void intel_disable_rps(struct drm_i915_private *dev_priv) +{ + if (INTEL_GEN(dev_priv) >= 9) + gen9_disable_rps(dev_priv); + else if (IS_CHERRYVIEW(dev_priv)) + cherryview_disable_rps(dev_priv); + else if (IS_VALLEYVIEW(dev_priv)) + valleyview_disable_rps(dev_priv); + else if (INTEL_GEN(dev_priv) >= 6) gen6_disable_rps(dev_priv); - } else if (IS_IRONLAKE_M(dev_priv)) { + else if (IS_IRONLAKE_M(dev_priv)) ironlake_disable_drps(dev_priv); - } - - dev_priv->pm.rps.enabled = false; - mutex_unlock(&dev_priv->pm.pcu_lock); } -void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) +void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) { - /* We shouldn't be disabling as we submit, so this should be less - * racy than it appears! - */ - if (READ_ONCE(dev_priv->pm.rps.enabled)) - return; - - /* Powersaving is controlled by the host when inside a VM */ - if (intel_vgpu_active(dev_priv)) + if (!READ_ONCE(dev_priv->pm.rps.enabled)) return; mutex_lock(&dev_priv->pm.pcu_lock); - if (IS_CHERRYVIEW(dev_priv)) { + intel_disable_rc6(dev_priv); + intel_disable_rps(dev_priv); + + dev_priv->pm.rps.enabled = false; + mutex_unlock(&dev_priv->pm.pcu_lock); +} + +static void intel_enable_rc6(struct drm_i915_private *dev_priv) +{ + if (IS_CHERRYVIEW(dev_priv)) cherryview_enable_rc6(dev_priv); + else if (IS_VALLEYVIEW(dev_priv)) + valleyview_enable_rc6(dev_priv); + else if (INTEL_GEN(dev_priv) >= 9) + gen9_enable_rc6(dev_priv); + else if (IS_BROADWELL(dev_priv)) + gen8_enable_rc6(dev_priv); + else if (INTEL_GEN(dev_priv) >= 6) + gen6_enable_rc6(dev_priv); +} + +static void intel_enable_rps(struct drm_i915_private *dev_priv) +{ + if (IS_CHERRYVIEW(dev_priv)) { cherryview_enable_rps(dev_priv); } else if (IS_VALLEYVIEW(dev_priv)) { - valleyview_enable_rc6(dev_priv); valleyview_enable_rps(dev_priv); } else if (INTEL_GEN(dev_priv) >= 9) { - gen9_enable_rc6(dev_priv); gen9_enable_rps(dev_priv); } else if (IS_BROADWELL(dev_priv)) { - gen8_enable_rc6(dev_priv); gen8_enable_rps(dev_priv); } else if (INTEL_GEN(dev_priv) >= 6) { - gen6_enable_rc6(dev_priv); gen6_enable_rps(dev_priv); } else if (IS_IRONLAKE_M(dev_priv)) { ironlake_enable_drps(dev_priv); intel_init_emon(dev_priv); } - if (HAS_LLC(dev_priv)) - intel_update_ring_freq(dev_priv); - WARN_ON(dev_priv->pm.rps.max_freq < dev_priv->pm.rps.min_freq); WARN_ON(dev_priv->pm.rps.idle_freq > dev_priv->pm.rps.max_freq); WARN_ON(dev_priv->pm.rps.efficient_freq < dev_priv->pm.rps.min_freq); WARN_ON(dev_priv->pm.rps.efficient_freq > dev_priv->pm.rps.max_freq); +} + +void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) +{ + /* We shouldn't be disabling as we submit, so this should be less + * racy than it appears! + */ + if (READ_ONCE(dev_priv->pm.rps.enabled)) + return; + + /* Powersaving is controlled by the host when inside a VM */ + if (intel_vgpu_active(dev_priv)) + return; + + mutex_lock(&dev_priv->pm.pcu_lock); + + intel_enable_rc6(dev_priv); + intel_enable_rps(dev_priv); + if (HAS_LLC(dev_priv)) + intel_update_ring_freq(dev_priv); dev_priv->pm.rps.enabled = true; mutex_unlock(&dev_priv->pm.pcu_lock); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH v2 10/11] drm/i915: Create generic functions to control RC6, RPS 2017-10-06 12:13 ` [PATCH v2 10/11] drm/i915: Create generic functions to control RC6, RPS Sagar Arun Kamble @ 2017-10-06 12:46 ` Chris Wilson 2017-10-06 15:02 ` Sagar Arun Kamble 0 siblings, 1 reply; 22+ messages in thread From: Chris Wilson @ 2017-10-06 12:46 UTC (permalink / raw) To: Sagar Arun Kamble, intel-gfx Quoting Sagar Arun Kamble (2017-10-06 13:13:39) > Prepared generic functions intel_enable_rc6, intel_disable_rc6, > intel_enable_rps and intel_disable_rps functions to setup RC6/RPS > based on platforms. > > v2: Make intel_enable/disable_rc6/rps static. (Chris) > > Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> > Cc: Imre Deak <imre.deak@intel.com> > Cc: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> > Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> > --- > drivers/gpu/drm/i915/intel_pm.c | 97 ++++++++++++++++++++++++++--------------- > 1 file changed, 62 insertions(+), 35 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c > index ce2dc5b..03264fe 100644 > --- a/drivers/gpu/drm/i915/intel_pm.c > +++ b/drivers/gpu/drm/i915/intel_pm.c > @@ -7972,75 +7972,102 @@ void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) > gen6_reset_rps_interrupts(dev_priv); > } > > -void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) > +static void intel_disable_rc6(struct drm_i915_private *dev_priv) > { > - if (!READ_ONCE(dev_priv->pm.rps.enabled)) > - return; > - > - mutex_lock(&dev_priv->pm.pcu_lock); lockdep_assert_held(dev_priv->pm.pcu_lock); ? We often skip it for statics, unless we know we are planning on adding an interface that may not take the lock. Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v2 10/11] drm/i915: Create generic functions to control RC6, RPS 2017-10-06 12:46 ` Chris Wilson @ 2017-10-06 15:02 ` Sagar Arun Kamble 0 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 15:02 UTC (permalink / raw) To: Chris Wilson, intel-gfx On 10/6/2017 6:16 PM, Chris Wilson wrote: > Quoting Sagar Arun Kamble (2017-10-06 13:13:39) >> Prepared generic functions intel_enable_rc6, intel_disable_rc6, >> intel_enable_rps and intel_disable_rps functions to setup RC6/RPS >> based on platforms. >> >> v2: Make intel_enable/disable_rc6/rps static. (Chris) >> >> Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> >> Cc: Imre Deak <imre.deak@intel.com> >> Cc: Chris Wilson <chris@chris-wilson.co.uk> >> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> >> Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> >> --- >> drivers/gpu/drm/i915/intel_pm.c | 97 ++++++++++++++++++++++++++--------------- >> 1 file changed, 62 insertions(+), 35 deletions(-) >> >> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c >> index ce2dc5b..03264fe 100644 >> --- a/drivers/gpu/drm/i915/intel_pm.c >> +++ b/drivers/gpu/drm/i915/intel_pm.c >> @@ -7972,75 +7972,102 @@ void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) >> gen6_reset_rps_interrupts(dev_priv); >> } >> >> -void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) >> +static void intel_disable_rc6(struct drm_i915_private *dev_priv) >> { >> - if (!READ_ONCE(dev_priv->pm.rps.enabled)) >> - return; >> - >> - mutex_lock(&dev_priv->pm.pcu_lock); > lockdep_assert_held(dev_priv->pm.pcu_lock); ? > > We often skip it for statics, unless we know we are planning on adding > an interface that may not take the lock. Sure will add this. Thanks. > > Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> > -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v2 11/11] drm/i915: Introduce separate status variable for RC6 and LLC ring frequency setup 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (9 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 10/11] drm/i915: Create generic functions to control RC6, RPS Sagar Arun Kamble @ 2017-10-06 12:13 ` Sagar Arun Kamble 2017-10-06 12:55 ` Chris Wilson 2017-10-06 14:10 ` ✓ Fi.CI.BAT: success for drm/i915: Separate RC6, RPS, LLC ring Frequency management Patchwork 2017-10-06 18:06 ` ✓ Fi.CI.IGT: " Patchwork 12 siblings, 1 reply; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 12:13 UTC (permalink / raw) To: intel-gfx Defined new struct intel_rc6 to hold RC6 specific state and intel_ring_pstate to hold ring specific state. v2: s/intel_ring_pstate/intel_llc_pstate and rebase. (Chris) Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> --- drivers/gpu/drm/i915/i915_drv.c | 2 +- drivers/gpu/drm/i915/i915_drv.h | 10 ++++++++ drivers/gpu/drm/i915/intel_pm.c | 57 +++++++++++++++++++++++++++++++---------- 3 files changed, 54 insertions(+), 15 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 470807c..154f231 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -2502,7 +2502,7 @@ static int intel_runtime_suspend(struct device *kdev) struct drm_i915_private *dev_priv = to_i915(dev); int ret; - if (WARN_ON_ONCE(!(dev_priv->pm.rps.enabled && intel_rc6_enabled()))) + if (WARN_ON_ONCE(!(dev_priv->pm.rc6.enabled && intel_rc6_enabled()))) return -ENODEV; if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev_priv))) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 45944a8..a07aa71 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1363,8 +1363,18 @@ struct intel_rps { struct intel_rps_ei ei; }; +struct intel_rc6 { + bool enabled; +}; + +struct intel_llc_pstate { + bool configured; +}; + struct intel_gen6_power_mgmt { struct intel_rps rps; + struct intel_rc6 rc6; + struct intel_llc_pstate llc_pstate; struct delayed_work autoenable_work; /* diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 03264fe..df36a6f 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -7873,7 +7873,12 @@ static void intel_init_emon(struct drm_i915_private *dev_priv) static inline void intel_update_ring_freq(struct drm_i915_private *i915) { + if (READ_ONCE(i915->pm.llc_pstate.configured)) + return; + gen6_update_ring_freq(i915); + + i915->pm.llc_pstate.configured = true; } void intel_init_gt_powersave(struct drm_i915_private *dev_priv) @@ -7966,7 +7971,8 @@ void intel_suspend_gt_powersave(struct drm_i915_private *dev_priv) void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) { - dev_priv->pm.rps.enabled = true; /* force disabling */ + dev_priv->pm.rps.enabled = true; /* force RPS disabling */ + dev_priv->pm.rc6.enabled = true; /* force RC6 disabling */ intel_disable_gt_powersave(dev_priv); gen6_reset_rps_interrupts(dev_priv); @@ -7974,6 +7980,9 @@ void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) static void intel_disable_rc6(struct drm_i915_private *dev_priv) { + if (!READ_ONCE(dev_priv->pm.rc6.enabled)) + return; + if (INTEL_GEN(dev_priv) >= 9) gen9_disable_rc6(dev_priv); else if (IS_CHERRYVIEW(dev_priv)) @@ -7982,10 +7991,15 @@ static void intel_disable_rc6(struct drm_i915_private *dev_priv) valleyview_disable_rc6(dev_priv); else if (INTEL_GEN(dev_priv) >= 6) gen6_disable_rc6(dev_priv); + + dev_priv->pm.rc6.enabled = false; } static void intel_disable_rps(struct drm_i915_private *dev_priv) { + if (!READ_ONCE(dev_priv->pm.rps.enabled)) + return; + if (INTEL_GEN(dev_priv) >= 9) gen9_disable_rps(dev_priv); else if (IS_CHERRYVIEW(dev_priv)) @@ -7996,24 +8010,30 @@ static void intel_disable_rps(struct drm_i915_private *dev_priv) gen6_disable_rps(dev_priv); else if (IS_IRONLAKE_M(dev_priv)) ironlake_disable_drps(dev_priv); + + dev_priv->pm.rps.enabled = false; } void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) { - if (!READ_ONCE(dev_priv->pm.rps.enabled)) - return; - mutex_lock(&dev_priv->pm.pcu_lock); intel_disable_rc6(dev_priv); intel_disable_rps(dev_priv); + if (HAS_LLC(dev_priv)) + dev_priv->pm.llc_pstate.configured = false; - dev_priv->pm.rps.enabled = false; mutex_unlock(&dev_priv->pm.pcu_lock); } static void intel_enable_rc6(struct drm_i915_private *dev_priv) { + /* We shouldn't be disabling as we submit, so this should be less + * racy than it appears! + */ + if (READ_ONCE(dev_priv->pm.rc6.enabled)) + return; + if (IS_CHERRYVIEW(dev_priv)) cherryview_enable_rc6(dev_priv); else if (IS_VALLEYVIEW(dev_priv)) @@ -8024,10 +8044,18 @@ static void intel_enable_rc6(struct drm_i915_private *dev_priv) gen8_enable_rc6(dev_priv); else if (INTEL_GEN(dev_priv) >= 6) gen6_enable_rc6(dev_priv); + + dev_priv->pm.rc6.enabled = true; } static void intel_enable_rps(struct drm_i915_private *dev_priv) { + /* We shouldn't be disabling as we submit, so this should be less + * racy than it appears! + */ + if (READ_ONCE(dev_priv->pm.rps.enabled)) + return; + if (IS_CHERRYVIEW(dev_priv)) { cherryview_enable_rps(dev_priv); } else if (IS_VALLEYVIEW(dev_priv)) { @@ -8048,16 +8076,12 @@ static void intel_enable_rps(struct drm_i915_private *dev_priv) WARN_ON(dev_priv->pm.rps.efficient_freq < dev_priv->pm.rps.min_freq); WARN_ON(dev_priv->pm.rps.efficient_freq > dev_priv->pm.rps.max_freq); + + dev_priv->pm.rps.enabled = true; } void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) { - /* We shouldn't be disabling as we submit, so this should be less - * racy than it appears! - */ - if (READ_ONCE(dev_priv->pm.rps.enabled)) - return; - /* Powersaving is controlled by the host when inside a VM */ if (intel_vgpu_active(dev_priv)) return; @@ -8069,7 +8093,6 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) if (HAS_LLC(dev_priv)) intel_update_ring_freq(dev_priv); - dev_priv->pm.rps.enabled = true; mutex_unlock(&dev_priv->pm.pcu_lock); } @@ -8080,7 +8103,10 @@ static void __intel_autoenable_gt_powersave(struct work_struct *work) struct intel_engine_cs *rcs; struct drm_i915_gem_request *req; - if (READ_ONCE(dev_priv->pm.rps.enabled)) + if (READ_ONCE(dev_priv->pm.rps.enabled) && + READ_ONCE(dev_priv->pm.rc6.enabled) && + !(HAS_LLC(dev_priv) ^ + READ_ONCE(dev_priv->pm.llc_pstate.configured))) goto out; rcs = dev_priv->engine[RCS]; @@ -8110,7 +8136,10 @@ static void __intel_autoenable_gt_powersave(struct work_struct *work) void intel_autoenable_gt_powersave(struct drm_i915_private *dev_priv) { - if (READ_ONCE(dev_priv->pm.rps.enabled)) + if (READ_ONCE(dev_priv->pm.rps.enabled) && + READ_ONCE(dev_priv->pm.rc6.enabled) && + !(HAS_LLC(dev_priv) ^ + READ_ONCE(dev_priv->pm.llc_pstate.configured))) return; if (IS_IRONLAKE_M(dev_priv)) { -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH v2 11/11] drm/i915: Introduce separate status variable for RC6 and LLC ring frequency setup 2017-10-06 12:13 ` [PATCH v2 11/11] drm/i915: Introduce separate status variable for RC6 and LLC ring frequency setup Sagar Arun Kamble @ 2017-10-06 12:55 ` Chris Wilson 2017-10-06 15:08 ` Sagar Arun Kamble 0 siblings, 1 reply; 22+ messages in thread From: Chris Wilson @ 2017-10-06 12:55 UTC (permalink / raw) To: Sagar Arun Kamble, intel-gfx Quoting Sagar Arun Kamble (2017-10-06 13:13:40) > Defined new struct intel_rc6 to hold RC6 specific state and > intel_ring_pstate to hold ring specific state. > > v2: s/intel_ring_pstate/intel_llc_pstate and rebase. (Chris) > > Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> > Cc: Imre Deak <imre.deak@intel.com> > Cc: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> > Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> > --- > drivers/gpu/drm/i915/i915_drv.c | 2 +- > drivers/gpu/drm/i915/i915_drv.h | 10 ++++++++ > drivers/gpu/drm/i915/intel_pm.c | 57 +++++++++++++++++++++++++++++++---------- > 3 files changed, 54 insertions(+), 15 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c > index 470807c..154f231 100644 > --- a/drivers/gpu/drm/i915/i915_drv.c > +++ b/drivers/gpu/drm/i915/i915_drv.c > @@ -2502,7 +2502,7 @@ static int intel_runtime_suspend(struct device *kdev) > struct drm_i915_private *dev_priv = to_i915(dev); > int ret; > > - if (WARN_ON_ONCE(!(dev_priv->pm.rps.enabled && intel_rc6_enabled()))) > + if (WARN_ON_ONCE(!(dev_priv->pm.rc6.enabled && intel_rc6_enabled()))) > return -ENODEV; > > if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev_priv))) > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index 45944a8..a07aa71 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -1363,8 +1363,18 @@ struct intel_rps { > struct intel_rps_ei ei; > }; > > +struct intel_rc6 { > + bool enabled; > +}; > + > +struct intel_llc_pstate { > + bool configured; > +}; > + > struct intel_gen6_power_mgmt { > struct intel_rps rps; > + struct intel_rc6 rc6; > + struct intel_llc_pstate llc_pstate; > struct delayed_work autoenable_work; > > /* > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c > index 03264fe..df36a6f 100644 > --- a/drivers/gpu/drm/i915/intel_pm.c > +++ b/drivers/gpu/drm/i915/intel_pm.c > @@ -7873,7 +7873,12 @@ static void intel_init_emon(struct drm_i915_private *dev_priv) > > static inline void intel_update_ring_freq(struct drm_i915_private *i915) > { > + if (READ_ONCE(i915->pm.llc_pstate.configured)) > + return; Tell me about how you expect the locking around this function to be. The READ_ONCE() implies that we are doing a optimistic peek outside of a lock, but then we set configured without acquiring a lock, so I assume we are inside some lock. That looks true for all, we don't need READ_ONCE() anymore as we only inspect inside the mutex (and so READ_ONCE is giving the wrong impression). > + > gen6_update_ring_freq(i915); > + > + i915->pm.llc_pstate.configured = true; > } > > void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) > { > - if (!READ_ONCE(dev_priv->pm.rps.enabled)) > - return; > - > mutex_lock(&dev_priv->pm.pcu_lock); > > intel_disable_rc6(dev_priv); > intel_disable_rps(dev_priv); > + if (HAS_LLC(dev_priv)) > + dev_priv->pm.llc_pstate.configured = false; Always clear it? If no llc, it can never be configured. Hmm, better if we just made it symmetrical with s/intel_update_ring_freq/intel_enable_llc_pstate/ and intel_disable_llc_pstate here. > > - dev_priv->pm.rps.enabled = false; > mutex_unlock(&dev_priv->pm.pcu_lock); > } > @@ -8080,7 +8103,10 @@ static void __intel_autoenable_gt_powersave(struct work_struct *work) > struct intel_engine_cs *rcs; > struct drm_i915_gem_request *req; > > - if (READ_ONCE(dev_priv->pm.rps.enabled)) > + if (READ_ONCE(dev_priv->pm.rps.enabled) && > + READ_ONCE(dev_priv->pm.rc6.enabled) && > + !(HAS_LLC(dev_priv) ^ > + READ_ONCE(dev_priv->pm.llc_pstate.configured))) > goto out; This optimisation has lost its appeal :) Kill it, if we need something like it we can try again later. -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v2 11/11] drm/i915: Introduce separate status variable for RC6 and LLC ring frequency setup 2017-10-06 12:55 ` Chris Wilson @ 2017-10-06 15:08 ` Sagar Arun Kamble 0 siblings, 0 replies; 22+ messages in thread From: Sagar Arun Kamble @ 2017-10-06 15:08 UTC (permalink / raw) To: Chris Wilson, intel-gfx On 10/6/2017 6:25 PM, Chris Wilson wrote: > Quoting Sagar Arun Kamble (2017-10-06 13:13:40) >> Defined new struct intel_rc6 to hold RC6 specific state and >> intel_ring_pstate to hold ring specific state. >> >> v2: s/intel_ring_pstate/intel_llc_pstate and rebase. (Chris) >> >> Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> >> Cc: Imre Deak <imre.deak@intel.com> >> Cc: Chris Wilson <chris@chris-wilson.co.uk> >> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> >> Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com> >> --- >> drivers/gpu/drm/i915/i915_drv.c | 2 +- >> drivers/gpu/drm/i915/i915_drv.h | 10 ++++++++ >> drivers/gpu/drm/i915/intel_pm.c | 57 +++++++++++++++++++++++++++++++---------- >> 3 files changed, 54 insertions(+), 15 deletions(-) >> >> diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c >> index 470807c..154f231 100644 >> --- a/drivers/gpu/drm/i915/i915_drv.c >> +++ b/drivers/gpu/drm/i915/i915_drv.c >> @@ -2502,7 +2502,7 @@ static int intel_runtime_suspend(struct device *kdev) >> struct drm_i915_private *dev_priv = to_i915(dev); >> int ret; >> >> - if (WARN_ON_ONCE(!(dev_priv->pm.rps.enabled && intel_rc6_enabled()))) >> + if (WARN_ON_ONCE(!(dev_priv->pm.rc6.enabled && intel_rc6_enabled()))) >> return -ENODEV; >> >> if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev_priv))) >> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h >> index 45944a8..a07aa71 100644 >> --- a/drivers/gpu/drm/i915/i915_drv.h >> +++ b/drivers/gpu/drm/i915/i915_drv.h >> @@ -1363,8 +1363,18 @@ struct intel_rps { >> struct intel_rps_ei ei; >> }; >> >> +struct intel_rc6 { >> + bool enabled; >> +}; >> + >> +struct intel_llc_pstate { >> + bool configured; >> +}; >> + >> struct intel_gen6_power_mgmt { >> struct intel_rps rps; >> + struct intel_rc6 rc6; >> + struct intel_llc_pstate llc_pstate; >> struct delayed_work autoenable_work; >> >> /* >> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c >> index 03264fe..df36a6f 100644 >> --- a/drivers/gpu/drm/i915/intel_pm.c >> +++ b/drivers/gpu/drm/i915/intel_pm.c >> @@ -7873,7 +7873,12 @@ static void intel_init_emon(struct drm_i915_private *dev_priv) >> >> static inline void intel_update_ring_freq(struct drm_i915_private *i915) >> { >> + if (READ_ONCE(i915->pm.llc_pstate.configured)) >> + return; > Tell me about how you expect the locking around this function to be. > > The READ_ONCE() implies that we are doing a optimistic peek outside of a > lock, but then we set configured without acquiring a lock, so I assume > we are inside some lock. > > That looks true for all, we don't need READ_ONCE() anymore as we only > inspect inside the mutex (and so READ_ONCE is giving the wrong > impression). > >> + >> gen6_update_ring_freq(i915); >> + >> + i915->pm.llc_pstate.configured = true; >> } >> >> void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) >> { >> - if (!READ_ONCE(dev_priv->pm.rps.enabled)) >> - return; >> - >> mutex_lock(&dev_priv->pm.pcu_lock); >> >> intel_disable_rc6(dev_priv); >> intel_disable_rps(dev_priv); >> + if (HAS_LLC(dev_priv)) >> + dev_priv->pm.llc_pstate.configured = false; > Always clear it? If no llc, it can never be configured. > Hmm, better if we just made it symmetrical with > s/intel_update_ring_freq/intel_enable_llc_pstate/ and > intel_disable_llc_pstate here. Will update. > >> >> - dev_priv->pm.rps.enabled = false; >> mutex_unlock(&dev_priv->pm.pcu_lock); >> } >> @@ -8080,7 +8103,10 @@ static void __intel_autoenable_gt_powersave(struct work_struct *work) >> struct intel_engine_cs *rcs; >> struct drm_i915_gem_request *req; >> >> - if (READ_ONCE(dev_priv->pm.rps.enabled)) >> + if (READ_ONCE(dev_priv->pm.rps.enabled) && >> + READ_ONCE(dev_priv->pm.rc6.enabled) && >> + !(HAS_LLC(dev_priv) ^ >> + READ_ONCE(dev_priv->pm.llc_pstate.configured))) >> goto out; > This optimisation has lost its appeal :) > > Kill it, if we need something like it we can try again later. > -Chris Sure. Understood that using READ_ONCE inside lock was unnecessary. Will remove this triple condition. _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* ✓ Fi.CI.BAT: success for drm/i915: Separate RC6, RPS, LLC ring Frequency management 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (10 preceding siblings ...) 2017-10-06 12:13 ` [PATCH v2 11/11] drm/i915: Introduce separate status variable for RC6 and LLC ring frequency setup Sagar Arun Kamble @ 2017-10-06 14:10 ` Patchwork 2017-10-06 18:06 ` ✓ Fi.CI.IGT: " Patchwork 12 siblings, 0 replies; 22+ messages in thread From: Patchwork @ 2017-10-06 14:10 UTC (permalink / raw) To: Sagar Arun Kamble; +Cc: intel-gfx == Series Details == Series: drm/i915: Separate RC6, RPS, LLC ring Frequency management URL : https://patchwork.freedesktop.org/series/31487/ State : success == Summary == Series 31487v1 drm/i915: Separate RC6, RPS, LLC ring Frequency management https://patchwork.freedesktop.org/api/1.0/series/31487/revisions/1/mbox/ Test kms_pipe_crc_basic: Subgroup suspend-read-crc-pipe-b: dmesg-warn -> PASS (fi-byt-n2820) fdo#101705 fdo#101705 https://bugs.freedesktop.org/show_bug.cgi?id=101705 fi-bdw-5557u total:289 pass:268 dwarn:0 dfail:0 fail:0 skip:21 time:457s fi-bdw-gvtdvm total:289 pass:265 dwarn:0 dfail:0 fail:0 skip:24 time:471s fi-blb-e6850 total:289 pass:223 dwarn:1 dfail:0 fail:0 skip:65 time:394s fi-bsw-n3050 total:289 pass:243 dwarn:0 dfail:0 fail:0 skip:46 time:570s fi-bwr-2160 total:289 pass:183 dwarn:0 dfail:0 fail:0 skip:106 time:288s fi-bxt-dsi total:289 pass:259 dwarn:0 dfail:0 fail:0 skip:30 time:527s fi-bxt-j4205 total:289 pass:260 dwarn:0 dfail:0 fail:0 skip:29 time:535s fi-byt-j1900 total:289 pass:253 dwarn:1 dfail:0 fail:0 skip:35 time:541s fi-byt-n2820 total:289 pass:250 dwarn:0 dfail:0 fail:0 skip:39 time:526s fi-cfl-s total:289 pass:256 dwarn:1 dfail:0 fail:0 skip:32 time:563s fi-cnl-y total:289 pass:262 dwarn:0 dfail:0 fail:0 skip:27 time:620s fi-elk-e7500 total:289 pass:229 dwarn:0 dfail:0 fail:0 skip:60 time:441s fi-glk-1 total:289 pass:261 dwarn:0 dfail:0 fail:0 skip:28 time:595s fi-hsw-4770 total:289 pass:262 dwarn:0 dfail:0 fail:0 skip:27 time:444s fi-hsw-4770r total:289 pass:262 dwarn:0 dfail:0 fail:0 skip:27 time:419s fi-ivb-3520m total:289 pass:260 dwarn:0 dfail:0 fail:0 skip:29 time:502s fi-ivb-3770 total:289 pass:260 dwarn:0 dfail:0 fail:0 skip:29 time:480s fi-kbl-7500u total:289 pass:264 dwarn:1 dfail:0 fail:0 skip:24 time:503s fi-kbl-7560u total:289 pass:270 dwarn:0 dfail:0 fail:0 skip:19 time:582s fi-kbl-7567u total:289 pass:265 dwarn:4 dfail:0 fail:0 skip:20 time:488s fi-kbl-r total:289 pass:262 dwarn:0 dfail:0 fail:0 skip:27 time:595s fi-pnv-d510 total:289 pass:222 dwarn:1 dfail:0 fail:0 skip:66 time:657s fi-skl-6260u total:289 pass:269 dwarn:0 dfail:0 fail:0 skip:20 time:470s fi-skl-6700hq total:289 pass:263 dwarn:0 dfail:0 fail:0 skip:26 time:654s fi-skl-6700k total:289 pass:265 dwarn:0 dfail:0 fail:0 skip:24 time:537s fi-skl-6770hq total:289 pass:269 dwarn:0 dfail:0 fail:0 skip:20 time:516s fi-skl-gvtdvm total:289 pass:266 dwarn:0 dfail:0 fail:0 skip:23 time:476s fi-snb-2520m total:289 pass:250 dwarn:0 dfail:0 fail:0 skip:39 time:582s fi-snb-2600 total:289 pass:249 dwarn:0 dfail:0 fail:0 skip:40 time:432s 7dacd1f2e70cb3202e2b153d76b05b601d099082 drm-tip: 2017y-10m-06d-12h-29m-28s UTC integration manifest 60b2fc17d783 drm/i915: Introduce separate status variable for RC6 and LLC ring frequency setup 5428c7e4323e drm/i915: Create generic functions to control RC6, RPS 56752ac404db drm/i915: Create generic function to setup LLC ring frequency table a147d1b841eb drm/i915: Rename intel_enable_rc6 to intel_rc6_enabled d116f1e68bef drm/i915: Name structure in dev_priv that contains RPS/RC6 state as "pm" b876a9da937a drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm" c177de7b23d4 drm/i915: Separate RPS and RC6 handling for CHV 3d1464310118 drm/i915: Separate RPS and RC6 handling for VLV d5e6fb79e822 drm/i915: Separate RPS and RC6 handling for BDW 926add4a05ec drm/i915: Remove superfluous IS_BDW checks and non-BDW changes from gen8_enable_rps 86b76d399bbd drm/i915: Separate RPS and RC6 handling for gen6+ == Logs == For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_5930/ _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
* ✓ Fi.CI.IGT: success for drm/i915: Separate RC6, RPS, LLC ring Frequency management 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble ` (11 preceding siblings ...) 2017-10-06 14:10 ` ✓ Fi.CI.BAT: success for drm/i915: Separate RC6, RPS, LLC ring Frequency management Patchwork @ 2017-10-06 18:06 ` Patchwork 12 siblings, 0 replies; 22+ messages in thread From: Patchwork @ 2017-10-06 18:06 UTC (permalink / raw) To: Sagar Arun Kamble; +Cc: intel-gfx == Series Details == Series: drm/i915: Separate RC6, RPS, LLC ring Frequency management URL : https://patchwork.freedesktop.org/series/31487/ State : success == Summary == Test kms_setmode: Subgroup basic: fail -> PASS (shard-hsw) fdo#99912 fdo#99912 https://bugs.freedesktop.org/show_bug.cgi?id=99912 shard-hsw total:2446 pass:1329 dwarn:6 dfail:0 fail:8 skip:1103 time:10141s == Logs == For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_5930/shards.html _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 22+ messages in thread
end of thread, other threads:[~2017-10-06 18:06 UTC | newest] Thread overview: 22+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-10-06 12:13 [PATCH v2 00/11] drm/i915: Separate RC6, RPS, LLC ring Frequency management Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 01/11] drm/i915: Separate RPS and RC6 handling for gen6+ Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 02/11] drm/i915: Remove superfluous IS_BDW checks and non-BDW changes from gen8_enable_rps Sagar Arun Kamble 2017-10-06 12:24 ` Chris Wilson 2017-10-06 12:13 ` [PATCH v2 03/11] drm/i915: Separate RPS and RC6 handling for BDW Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 04/11] drm/i915: Separate RPS and RC6 handling for VLV Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 05/11] drm/i915: Separate RPS and RC6 handling for CHV Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 06/11] drm/i915: Name i915_runtime_pm structure in dev_priv as "rpm" Sagar Arun Kamble 2017-10-06 12:40 ` Chris Wilson 2017-10-06 15:00 ` Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 07/11] drm/i915: Name structure in dev_priv that contains RPS/RC6 state as "pm" Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 08/11] drm/i915: Rename intel_enable_rc6 to intel_rc6_enabled Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 09/11] drm/i915: Create generic function to setup LLC ring frequency table Sagar Arun Kamble 2017-10-06 12:43 ` Chris Wilson 2017-10-06 12:13 ` [PATCH v2 10/11] drm/i915: Create generic functions to control RC6, RPS Sagar Arun Kamble 2017-10-06 12:46 ` Chris Wilson 2017-10-06 15:02 ` Sagar Arun Kamble 2017-10-06 12:13 ` [PATCH v2 11/11] drm/i915: Introduce separate status variable for RC6 and LLC ring frequency setup Sagar Arun Kamble 2017-10-06 12:55 ` Chris Wilson 2017-10-06 15:08 ` Sagar Arun Kamble 2017-10-06 14:10 ` ✓ Fi.CI.BAT: success for drm/i915: Separate RC6, RPS, LLC ring Frequency management Patchwork 2017-10-06 18:06 ` ✓ Fi.CI.IGT: " Patchwork
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.