* [PATCH 1/4] drm/i915/chv: Use timeout mode for RC6 on chv
@ 2014-12-12 8:48 deepak.s
2014-12-12 8:48 ` [PATCH 2/4] drm/i915/chv: Populate total EU count on Cherryview deepak.s
` (3 more replies)
0 siblings, 4 replies; 30+ messages in thread
From: deepak.s @ 2014-12-12 8:48 UTC (permalink / raw)
To: intel-gfx; +Cc: daniel.vetter, Rodrigo Vivi
From: Deepak S <deepak.s@linux.intel.com>
Higher RC6 residency is observed using timeout mode
instead of EI mode. It's Recommended to use TO Method for RC6.
Signed-off-by: Deepak S <deepak.s@linux.intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
drivers/gpu/drm/i915/intel_pm.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 2316d23..2acb3de 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -4689,7 +4689,8 @@ static void cherryview_enable_rps(struct drm_device *dev)
I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10);
I915_WRITE(GEN6_RC_SLEEP, 0);
- I915_WRITE(GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */
+ /* TO threshold set to 1750 us ( 0x557 * 1.28 us) */
+ I915_WRITE(GEN6_RC6_THRESHOLD, 0x557);
/* allows RC6 residency counter to work */
I915_WRITE(VLV_COUNTER_CONTROL,
@@ -4703,7 +4704,7 @@ static void cherryview_enable_rps(struct drm_device *dev)
/* 3: Enable RC6 */
if ((intel_enable_rc6(dev) & INTEL_RC6_ENABLE) &&
(pcbr >> VLV_PCBR_ADDR_SHIFT))
- rc6_mode = GEN6_RC_CTL_EI_MODE(1);
+ rc6_mode = GEN7_RC_CTL_TO_MODE;
I915_WRITE(GEN6_RC_CONTROL, rc6_mode);
--
1.9.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 30+ messages in thread* [PATCH 2/4] drm/i915/chv: Populate total EU count on Cherryview 2014-12-12 8:48 [PATCH 1/4] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s @ 2014-12-12 8:48 ` deepak.s 2014-12-11 12:09 ` Jani Nikula 2014-12-12 8:48 ` [PATCH 3/4] drm/i915: New offset for reading frequencies on CHV deepak.s ` (2 subsequent siblings) 3 siblings, 1 reply; 30+ messages in thread From: deepak.s @ 2014-12-12 8:48 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> Starting with Cherryview, devices may have a varying number of EU for a given ID due to creative fusing. Punit support different frequency for different fuse data. We use this patch to help get total eu enabled and read the right offset to get RP0 Based upon a patch from Jeff, but reworked to only store eu_total and avoid sending info to userspace Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> --- drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ 3 files changed, 23 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index 887d88f..2bd36b6 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -598,6 +598,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) info->num_pipes = 0; } } + + if (IS_CHERRYVIEW(dev)) { + u32 fuse, mask_eu; + + fuse = I915_READ(CHV_FUSE_GT); + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | + CHV_FGT_EU_DIS_SS0_R1_MASK | + CHV_FGT_EU_DIS_SS1_R0_MASK | + CHV_FGT_EU_DIS_SS1_R1_MASK); + info->eu_total = 16 - hweight32(mask_eu); + } } /** diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 9381504..b58bad4 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -624,6 +624,7 @@ struct intel_device_info { int trans_offsets[I915_MAX_TRANSCODERS]; int palette_offsets[I915_MAX_PIPES]; int cursor_offsets[I915_MAX_PIPES]; + unsigned int eu_total; }; #undef DEFINE_FLAG diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 93fdad8..b57cba3 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -1466,6 +1466,17 @@ enum punit_power_well { #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) +/* Fuse readout registers for GT */ +#define CHV_FUSE_GT 0x182168 +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf<<CHV_FGT_EU_DIS_SS0_R0_SHIFT) +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf<<CHV_FGT_EU_DIS_SS0_R1_SHIFT) +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf<<CHV_FGT_EU_DIS_SS1_R0_SHIFT) +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf<<CHV_FGT_EU_DIS_SS1_R1_SHIFT) + #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* Re: [PATCH 2/4] drm/i915/chv: Populate total EU count on Cherryview 2014-12-12 8:48 ` [PATCH 2/4] drm/i915/chv: Populate total EU count on Cherryview deepak.s @ 2014-12-11 12:09 ` Jani Nikula 2014-12-12 12:10 ` Deepak S 0 siblings, 1 reply; 30+ messages in thread From: Jani Nikula @ 2014-12-11 12:09 UTC (permalink / raw) To: deepak.s, intel-gfx; +Cc: daniel.vetter On Fri, 12 Dec 2014, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Starting with Cherryview, devices may have a varying number of EU for > a given ID due to creative fusing. Punit support different frequency for > different fuse data. We use this patch to help get total eu enabled and > read the right offset to get RP0 > > Based upon a patch from Jeff, but reworked to only store eu_total and > avoid sending info to userspace > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> > --- > drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ > drivers/gpu/drm/i915/i915_drv.h | 1 + > drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ > 3 files changed, 23 insertions(+) > > diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c > index 887d88f..2bd36b6 100644 > --- a/drivers/gpu/drm/i915/i915_dma.c > +++ b/drivers/gpu/drm/i915/i915_dma.c > @@ -598,6 +598,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) > info->num_pipes = 0; > } > } > + > + if (IS_CHERRYVIEW(dev)) { > + u32 fuse, mask_eu; > + > + fuse = I915_READ(CHV_FUSE_GT); > + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | > + CHV_FGT_EU_DIS_SS0_R1_MASK | > + CHV_FGT_EU_DIS_SS1_R0_MASK | > + CHV_FGT_EU_DIS_SS1_R1_MASK); > + info->eu_total = 16 - hweight32(mask_eu); > + } > } > > /** > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index 9381504..b58bad4 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -624,6 +624,7 @@ struct intel_device_info { > int trans_offsets[I915_MAX_TRANSCODERS]; > int palette_offsets[I915_MAX_PIPES]; > int cursor_offsets[I915_MAX_PIPES]; > + unsigned int eu_total; > }; > > #undef DEFINE_FLAG > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h > index 93fdad8..b57cba3 100644 > --- a/drivers/gpu/drm/i915/i915_reg.h > +++ b/drivers/gpu/drm/i915/i915_reg.h > @@ -1466,6 +1466,17 @@ enum punit_power_well { > #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) > #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) > > +/* Fuse readout registers for GT */ > +#define CHV_FUSE_GT 0x182168 Should this be (VLV_DISPLAY_BASE + 0x2168)? > +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 > +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf<<CHV_FGT_EU_DIS_SS0_R0_SHIFT) > +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 > +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf<<CHV_FGT_EU_DIS_SS0_R1_SHIFT) > +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 > +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf<<CHV_FGT_EU_DIS_SS1_R0_SHIFT) > +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 > +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf<<CHV_FGT_EU_DIS_SS1_R1_SHIFT) Please add spaces both sides of "<<" if you end up sending a v2. > + > #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 > #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) > #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) > -- > 1.9.1 > > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/intel-gfx -- Jani Nikula, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH 2/4] drm/i915/chv: Populate total EU count on Cherryview 2014-12-11 12:09 ` Jani Nikula @ 2014-12-12 12:10 ` Deepak S 2014-12-12 12:33 ` [PATCH v2] " deepak.s 0 siblings, 1 reply; 30+ messages in thread From: Deepak S @ 2014-12-12 12:10 UTC (permalink / raw) To: Jani Nikula, intel-gfx; +Cc: daniel.vetter On Thursday 11 December 2014 05:39 PM, Jani Nikula wrote: > On Fri, 12 Dec 2014, deepak.s@linux.intel.com wrote: >> From: Deepak S <deepak.s@linux.intel.com> >> >> Starting with Cherryview, devices may have a varying number of EU for >> a given ID due to creative fusing. Punit support different frequency for >> different fuse data. We use this patch to help get total eu enabled and >> read the right offset to get RP0 >> >> Based upon a patch from Jeff, but reworked to only store eu_total and >> avoid sending info to userspace >> >> Signed-off-by: Deepak S <deepak.s@linux.intel.com> >> Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> >> --- >> drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ >> drivers/gpu/drm/i915/i915_drv.h | 1 + >> drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ >> 3 files changed, 23 insertions(+) >> >> diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c >> index 887d88f..2bd36b6 100644 >> --- a/drivers/gpu/drm/i915/i915_dma.c >> +++ b/drivers/gpu/drm/i915/i915_dma.c >> @@ -598,6 +598,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) >> info->num_pipes = 0; >> } >> } >> + >> + if (IS_CHERRYVIEW(dev)) { >> + u32 fuse, mask_eu; >> + >> + fuse = I915_READ(CHV_FUSE_GT); >> + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | >> + CHV_FGT_EU_DIS_SS0_R1_MASK | >> + CHV_FGT_EU_DIS_SS1_R0_MASK | >> + CHV_FGT_EU_DIS_SS1_R1_MASK); >> + info->eu_total = 16 - hweight32(mask_eu); >> + } >> } >> >> /** >> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h >> index 9381504..b58bad4 100644 >> --- a/drivers/gpu/drm/i915/i915_drv.h >> +++ b/drivers/gpu/drm/i915/i915_drv.h >> @@ -624,6 +624,7 @@ struct intel_device_info { >> int trans_offsets[I915_MAX_TRANSCODERS]; >> int palette_offsets[I915_MAX_PIPES]; >> int cursor_offsets[I915_MAX_PIPES]; >> + unsigned int eu_total; >> }; >> >> #undef DEFINE_FLAG >> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h >> index 93fdad8..b57cba3 100644 >> --- a/drivers/gpu/drm/i915/i915_reg.h >> +++ b/drivers/gpu/drm/i915/i915_reg.h >> @@ -1466,6 +1466,17 @@ enum punit_power_well { >> #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) >> #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) >> >> +/* Fuse readout registers for GT */ >> +#define CHV_FUSE_GT 0x182168 > Should this be (VLV_DISPLAY_BASE + 0x2168)? > >> +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 >> +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf<<CHV_FGT_EU_DIS_SS0_R0_SHIFT) >> +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 >> +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf<<CHV_FGT_EU_DIS_SS0_R1_SHIFT) >> +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 >> +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf<<CHV_FGT_EU_DIS_SS1_R0_SHIFT) >> +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 >> +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf<<CHV_FGT_EU_DIS_SS1_R1_SHIFT) > Please add spaces both sides of "<<" if you end up sending a v2. Thanks for review. I will address both the comment and send v2 >> + >> #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 >> #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) >> #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) >> -- >> 1.9.1 >> >> _______________________________________________ >> Intel-gfx mailing list >> Intel-gfx@lists.freedesktop.org >> http://lists.freedesktop.org/mailman/listinfo/intel-gfx _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* [PATCH v2] drm/i915/chv: Populate total EU count on Cherryview 2014-12-12 12:10 ` Deepak S @ 2014-12-12 12:33 ` deepak.s 0 siblings, 0 replies; 30+ messages in thread From: deepak.s @ 2014-12-12 12:33 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> Starting with Cherryview, devices may have a varying number of EU for a given ID due to creative fusing. Punit support different frequency for different fuse data. We use this patch to help get total eu enabled and read the right offset to get RP0 Based upon a patch from Jeff, but reworked to only store eu_total and avoid sending info to userspace v2: Format register definitions (Jani) Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> --- drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ 3 files changed, 23 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index 887d88f..2bd36b6 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -598,6 +598,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) info->num_pipes = 0; } } + + if (IS_CHERRYVIEW(dev)) { + u32 fuse, mask_eu; + + fuse = I915_READ(CHV_FUSE_GT); + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | + CHV_FGT_EU_DIS_SS0_R1_MASK | + CHV_FGT_EU_DIS_SS1_R0_MASK | + CHV_FGT_EU_DIS_SS1_R1_MASK); + info->eu_total = 16 - hweight32(mask_eu); + } } /** diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 9381504..b58bad4 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -624,6 +624,7 @@ struct intel_device_info { int trans_offsets[I915_MAX_TRANSCODERS]; int palette_offsets[I915_MAX_PIPES]; int cursor_offsets[I915_MAX_PIPES]; + unsigned int eu_total; }; #undef DEFINE_FLAG diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 93fdad8..09442c3 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -1466,6 +1466,17 @@ enum punit_power_well { #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) +/* Fuse readout registers for GT */ +#define CHV_FUSE_GT (VLV_DISPLAY_BASE + 0x2168) +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf << CHV_FGT_EU_DIS_SS0_R0_SHIFT) +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf << CHV_FGT_EU_DIS_SS0_R1_SHIFT) +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf << CHV_FGT_EU_DIS_SS1_R0_SHIFT) +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf << CHV_FGT_EU_DIS_SS1_R1_SHIFT) + #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* [PATCH 3/4] drm/i915: New offset for reading frequencies on CHV. 2014-12-12 8:48 [PATCH 1/4] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s 2014-12-12 8:48 ` [PATCH 2/4] drm/i915/chv: Populate total EU count on Cherryview deepak.s @ 2014-12-12 8:48 ` deepak.s 2014-12-12 19:09 ` Ville Syrjälä 2014-12-12 8:48 ` [PATCH 4/4] drm/i915: Skip gunit save/restore for cherryview deepak.s 2014-12-13 6:13 ` [PATCH v2] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s 3 siblings, 1 reply; 30+ messages in thread From: deepak.s @ 2014-12-12 8:48 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> Use new Sideband offset to read max/min/gaur freq based on the SKU it is running on. Based on the Number of EU, we read different bits to identify the max frequencies at which system can run. Signed-off-by: Deepak S <deepak.s@linux.intel.com> --- drivers/gpu/drm/i915/i915_drv.h | 4 +-- drivers/gpu/drm/i915/i915_reg.h | 12 ++++++++ drivers/gpu/drm/i915/intel_pm.c | 52 ++++++++++++++++++++++++++++++----- drivers/gpu/drm/i915/intel_sideband.c | 4 +-- 4 files changed, 61 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index b58bad4..0690dff 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -3016,8 +3016,8 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val); /* intel_sideband.c */ -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr); -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val); +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr); +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val); u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr); u32 vlv_gpio_nc_read(struct drm_i915_private *dev_priv, u32 reg); void vlv_gpio_nc_write(struct drm_i915_private *dev_priv, u32 reg, u32 val); diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index b57cba3..f41290c 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -602,6 +602,18 @@ enum punit_power_well { #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 +#define FB_GFX_FMAX_AT_VMAX_FUSE_MASK 0xff +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 + +#define FB_GFX_GUAR_FREQ_FUSE_MASK 0xff + +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 +#define FB_GFX_FMIN_AT_VMIN_FUSE_MASK 0xff +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 + #define PUNIT_GPU_STATUS_REG 0xdb #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 2acb3de..71b8e2f 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -4346,11 +4346,29 @@ void gen6_update_ring_freq(struct drm_device *dev) static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; + struct intel_device_info *info; u32 val, rp0; - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; - + info = (struct intel_device_info *)&dev_priv->info; + + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); + + if (info->eu_total == 8) /* (2 * 4) config */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); + else if (info->eu_total == 12) /* (2 * 6) config */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); + else if (info->eu_total == 16) /* (2 * 8) config */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); + else /* Setting (2 * 8) Min RP0 for any other combination */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); + rp0 = (rp0 & FB_GFX_FMAX_AT_VMAX_FUSE_MASK); + } else { /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & + PUNIT_GPU_STATUS_MAX_FREQ_MASK; + } return rp0; } @@ -4366,20 +4384,40 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; + struct intel_device_info *info; u32 val, rp1; - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; + info = (struct intel_device_info *)&dev_priv->info; + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); + rp1 = (val & FB_GFX_GUAR_FREQ_FUSE_MASK); + } else { /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & + PUNIT_GPU_STATUS_MAX_FREQ_MASK); + } return rp1; } static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; + struct intel_device_info *info; u32 val, rpn; + info = (struct intel_device_info *)&dev_priv->info; + + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & + FB_GFX_FMIN_AT_VMIN_FUSE_MASK); + } else { /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); + } - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; return rpn; } diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c index 01d841e..3c42eef 100644 --- a/drivers/gpu/drm/i915/intel_sideband.c +++ b/drivers/gpu/drm/i915/intel_sideband.c @@ -75,7 +75,7 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn, return 0; } -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr) { u32 val = 0; @@ -89,7 +89,7 @@ u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) return val; } -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val) +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val) { WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* Re: [PATCH 3/4] drm/i915: New offset for reading frequencies on CHV. 2014-12-12 8:48 ` [PATCH 3/4] drm/i915: New offset for reading frequencies on CHV deepak.s @ 2014-12-12 19:09 ` Ville Syrjälä 2014-12-15 6:51 ` Jani Nikula 0 siblings, 1 reply; 30+ messages in thread From: Ville Syrjälä @ 2014-12-12 19:09 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx On Fri, Dec 12, 2014 at 02:18:15PM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Use new Sideband offset to read max/min/gaur freq based on the SKU it > is running on. Based on the Number of EU, we read different bits to > identify the max frequencies at which system can run. > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > --- > drivers/gpu/drm/i915/i915_drv.h | 4 +-- > drivers/gpu/drm/i915/i915_reg.h | 12 ++++++++ > drivers/gpu/drm/i915/intel_pm.c | 52 ++++++++++++++++++++++++++++++----- > drivers/gpu/drm/i915/intel_sideband.c | 4 +-- > 4 files changed, 61 insertions(+), 11 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index b58bad4..0690dff 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -3016,8 +3016,8 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val > int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val); > > /* intel_sideband.c */ > -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr); > -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val); > +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr); > +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val); > u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr); > u32 vlv_gpio_nc_read(struct drm_i915_private *dev_priv, u32 reg); > void vlv_gpio_nc_write(struct drm_i915_private *dev_priv, u32 reg, u32 val); > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h > index b57cba3..f41290c 100644 > --- a/drivers/gpu/drm/i915/i915_reg.h > +++ b/drivers/gpu/drm/i915/i915_reg.h > @@ -602,6 +602,18 @@ enum punit_power_well { > #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ > #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ > > +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 > +#define FB_GFX_FMAX_AT_VMAX_FUSE_MASK 0xff > +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 > +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 > +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 > + This blank line makes me think FB_GFX_GUAR_FREQ_FUSE_MASK isn't part of this register. So best not leave such blank line here. I have 0x3c3c3c28 in this register, which matches what I get using the old method. > +#define FB_GFX_GUAR_FREQ_FUSE_MASK 0xff > + > +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 > +#define FB_GFX_FMIN_AT_VMIN_FUSE_MASK 0xff > +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 I have 0x69841428 here. The low 8 bits look like another freq value. What is it? I have no docs for this stuff so can't really review apart from looking at what my hardware reports. Actually, since all the values are 8 bits maybe it would be neater to just #define FB_GFX_FREQ_FUSE_MASK 0xff and use that everywhere instead of having three different definitions for the same 0xff value. > + > #define PUNIT_GPU_STATUS_REG 0xdb > #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 > #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c > index 2acb3de..71b8e2f 100644 > --- a/drivers/gpu/drm/i915/intel_pm.c > +++ b/drivers/gpu/drm/i915/intel_pm.c > @@ -4346,11 +4346,29 @@ void gen6_update_ring_freq(struct drm_device *dev) > > static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > + struct intel_device_info *info; > u32 val, rp0; > > - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; > - > + info = (struct intel_device_info *)&dev_priv->info; Pointless cast. Also the assignment could be done when declaring info, and we usually use INTEL_INFO() to get at it. > + > + if (dev->pdev->revision >= 0x20) { Do we really need this check? I would think it would be up to the Punit firmware version rather the stepping. My BSW has PCI rev 0x15, but with the latest BIOS both fuse registers already contain correct looking information. > + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); > + > + if (info->eu_total == 8) /* (2 * 4) config */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); > + else if (info->eu_total == 12) /* (2 * 6) config */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); > + else if (info->eu_total == 16) /* (2 * 8) config */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); > + else /* Setting (2 * 8) Min RP0 for any other combination */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); 'switch (INTEL_INFO(dev)->eu_total)' perhaps? > + rp0 = (rp0 & FB_GFX_FMAX_AT_VMAX_FUSE_MASK); > + } else { /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_MAX_FREQ_MASK; > + } > return rp0; > } > > @@ -4366,20 +4384,40 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) > > static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > + struct intel_device_info *info; > u32 val, rp1; > > - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); > - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; > + info = (struct intel_device_info *)&dev_priv->info; Unused here. > > + if (dev->pdev->revision >= 0x20) { > + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); > + rp1 = (val & FB_GFX_GUAR_FREQ_FUSE_MASK); > + } else { /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); > + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_MAX_FREQ_MASK); > + } > return rp1; > } > > static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > + struct intel_device_info *info; > u32 val, rpn; > + info = (struct intel_device_info *)&dev_priv->info; Unused. > + > + if (dev->pdev->revision >= 0x20) { > + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); > + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & > + FB_GFX_FMIN_AT_VMIN_FUSE_MASK); > + } else { /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); > + } > > - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; > return rpn; > } > > diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c > index 01d841e..3c42eef 100644 > --- a/drivers/gpu/drm/i915/intel_sideband.c > +++ b/drivers/gpu/drm/i915/intel_sideband.c > @@ -75,7 +75,7 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn, > return 0; > } > > -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) > +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr) Good thing you had to pass a constant. Otherwise this could have caused quite a bit of head scratching for you ;) > { > u32 val = 0; > > @@ -89,7 +89,7 @@ u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) > return val; > } > > -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val) > +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val) > { > WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); > > -- > 1.9.1 -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH 3/4] drm/i915: New offset for reading frequencies on CHV. 2014-12-12 19:09 ` Ville Syrjälä @ 2014-12-15 6:51 ` Jani Nikula 2014-12-16 12:11 ` Deepak S 0 siblings, 1 reply; 30+ messages in thread From: Jani Nikula @ 2014-12-15 6:51 UTC (permalink / raw) To: Ville Syrjälä, deepak.s; +Cc: daniel.vetter, intel-gfx On Fri, 12 Dec 2014, Ville Syrjälä <ville.syrjala@linux.intel.com> wrote: > On Fri, Dec 12, 2014 at 02:18:15PM +0530, deepak.s@linux.intel.com wrote: >> From: Deepak S <deepak.s@linux.intel.com> >> >> Use new Sideband offset to read max/min/gaur freq based on the SKU it >> is running on. Based on the Number of EU, we read different bits to >> identify the max frequencies at which system can run. >> >> Signed-off-by: Deepak S <deepak.s@linux.intel.com> >> --- >> drivers/gpu/drm/i915/i915_drv.h | 4 +-- >> drivers/gpu/drm/i915/i915_reg.h | 12 ++++++++ >> drivers/gpu/drm/i915/intel_pm.c | 52 ++++++++++++++++++++++++++++++----- >> drivers/gpu/drm/i915/intel_sideband.c | 4 +-- >> 4 files changed, 61 insertions(+), 11 deletions(-) >> >> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h >> index b58bad4..0690dff 100644 >> --- a/drivers/gpu/drm/i915/i915_drv.h >> +++ b/drivers/gpu/drm/i915/i915_drv.h >> @@ -3016,8 +3016,8 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val >> int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val); >> >> /* intel_sideband.c */ >> -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr); >> -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val); >> +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr); >> +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val); >> u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr); >> u32 vlv_gpio_nc_read(struct drm_i915_private *dev_priv, u32 reg); >> void vlv_gpio_nc_write(struct drm_i915_private *dev_priv, u32 reg, u32 val); >> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h >> index b57cba3..f41290c 100644 >> --- a/drivers/gpu/drm/i915/i915_reg.h >> +++ b/drivers/gpu/drm/i915/i915_reg.h >> @@ -602,6 +602,18 @@ enum punit_power_well { >> #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ >> #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ >> >> +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 >> +#define FB_GFX_FMAX_AT_VMAX_FUSE_MASK 0xff >> +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 >> +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 >> +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 >> + > > This blank line makes me think FB_GFX_GUAR_FREQ_FUSE_MASK isn't part of > this register. So best not leave such blank line here. > > I have 0x3c3c3c28 in this register, which matches what I get using the > old method. > >> +#define FB_GFX_GUAR_FREQ_FUSE_MASK 0xff >> + >> +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 >> +#define FB_GFX_FMIN_AT_VMIN_FUSE_MASK 0xff >> +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 > > I have 0x69841428 here. The low 8 bits look like another freq value. > What is it? > > I have no docs for this stuff so can't really review apart from looking > at what my hardware reports. > > > Actually, since all the values are 8 bits maybe it would be neater to > just > #define FB_GFX_FREQ_FUSE_MASK 0xff > and use that everywhere instead of having three different definitions > for the same 0xff value. > >> + >> #define PUNIT_GPU_STATUS_REG 0xdb >> #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 >> #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff >> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c >> index 2acb3de..71b8e2f 100644 >> --- a/drivers/gpu/drm/i915/intel_pm.c >> +++ b/drivers/gpu/drm/i915/intel_pm.c >> @@ -4346,11 +4346,29 @@ void gen6_update_ring_freq(struct drm_device *dev) >> >> static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) >> { >> + struct drm_device *dev = dev_priv->dev; >> + struct intel_device_info *info; >> u32 val, rp0; >> >> - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >> - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; >> - >> + info = (struct intel_device_info *)&dev_priv->info; > > Pointless cast. Also the assignment could be done when declaring info, > and we usually use INTEL_INFO() to get at it. > >> + >> + if (dev->pdev->revision >= 0x20) { > > Do we really need this check? I would think it would be up to the > Punit firmware version rather the stepping. My BSW has PCI rev 0x15, > but with the latest BIOS both fuse registers already contain correct > looking information. > >> + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); >> + >> + if (info->eu_total == 8) /* (2 * 4) config */ >> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); >> + else if (info->eu_total == 12) /* (2 * 6) config */ >> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); >> + else if (info->eu_total == 16) /* (2 * 8) config */ >> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); >> + else /* Setting (2 * 8) Min RP0 for any other combination */ >> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); > > 'switch (INTEL_INFO(dev)->eu_total)' perhaps? > >> + rp0 = (rp0 & FB_GFX_FMAX_AT_VMAX_FUSE_MASK); >> + } else { /* For pre-production hardware */ >> + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >> + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & >> + PUNIT_GPU_STATUS_MAX_FREQ_MASK; >> + } >> return rp0; >> } >> >> @@ -4366,20 +4384,40 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) >> >> static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) >> { >> + struct drm_device *dev = dev_priv->dev; >> + struct intel_device_info *info; >> u32 val, rp1; >> >> - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); >> - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; >> + info = (struct intel_device_info *)&dev_priv->info; > > Unused here. > >> >> + if (dev->pdev->revision >= 0x20) { > >> + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); >> + rp1 = (val & FB_GFX_GUAR_FREQ_FUSE_MASK); >> + } else { /* For pre-production hardware */ >> + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); >> + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & >> + PUNIT_GPU_STATUS_MAX_FREQ_MASK); >> + } >> return rp1; >> } >> >> static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) >> { >> + struct drm_device *dev = dev_priv->dev; >> + struct intel_device_info *info; >> u32 val, rpn; >> + info = (struct intel_device_info *)&dev_priv->info; > > Unused. > >> + >> + if (dev->pdev->revision >= 0x20) { >> + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); >> + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & >> + FB_GFX_FMIN_AT_VMIN_FUSE_MASK); >> + } else { /* For pre-production hardware */ >> + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >> + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & >> + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); >> + } >> >> - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >> - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; >> return rpn; >> } >> >> diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c >> index 01d841e..3c42eef 100644 >> --- a/drivers/gpu/drm/i915/intel_sideband.c >> +++ b/drivers/gpu/drm/i915/intel_sideband.c >> @@ -75,7 +75,7 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn, >> return 0; >> } >> >> -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) >> +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr) > > Good thing you had to pass a constant. Otherwise this could have caused > quite a bit of head scratching for you ;) This is the kind of thing that really should be a separate fix. If, for any reason, we need to revert the rest, we'll still want this. BR, Jani. > >> { >> u32 val = 0; >> >> @@ -89,7 +89,7 @@ u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) >> return val; >> } >> >> -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val) >> +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val) >> { >> WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); >> >> -- >> 1.9.1 > > -- > Ville Syrjälä > Intel OTC > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/intel-gfx -- Jani Nikula, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH 3/4] drm/i915: New offset for reading frequencies on CHV. 2014-12-15 6:51 ` Jani Nikula @ 2014-12-16 12:11 ` Deepak S 2015-01-16 15:12 ` [PATCH v2 0/3] Use new turbo offset for chv production system deepak.s 0 siblings, 1 reply; 30+ messages in thread From: Deepak S @ 2014-12-16 12:11 UTC (permalink / raw) To: Jani Nikula, Ville Syrjälä; +Cc: daniel.vetter, intel-gfx On Monday 15 December 2014 12:21 PM, Jani Nikula wrote: > On Fri, 12 Dec 2014, Ville Syrjälä <ville.syrjala@linux.intel.com> wrote: >> On Fri, Dec 12, 2014 at 02:18:15PM +0530, deepak.s@linux.intel.com wrote: >>> From: Deepak S <deepak.s@linux.intel.com> >>> >>> Use new Sideband offset to read max/min/gaur freq based on the SKU it >>> is running on. Based on the Number of EU, we read different bits to >>> identify the max frequencies at which system can run. >>> >>> Signed-off-by: Deepak S <deepak.s@linux.intel.com> >>> --- >>> drivers/gpu/drm/i915/i915_drv.h | 4 +-- >>> drivers/gpu/drm/i915/i915_reg.h | 12 ++++++++ >>> drivers/gpu/drm/i915/intel_pm.c | 52 ++++++++++++++++++++++++++++++----- >>> drivers/gpu/drm/i915/intel_sideband.c | 4 +-- >>> 4 files changed, 61 insertions(+), 11 deletions(-) >>> >>> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h >>> index b58bad4..0690dff 100644 >>> --- a/drivers/gpu/drm/i915/i915_drv.h >>> +++ b/drivers/gpu/drm/i915/i915_drv.h >>> @@ -3016,8 +3016,8 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val >>> int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val); >>> >>> /* intel_sideband.c */ >>> -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr); >>> -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val); >>> +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr); >>> +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val); >>> u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr); >>> u32 vlv_gpio_nc_read(struct drm_i915_private *dev_priv, u32 reg); >>> void vlv_gpio_nc_write(struct drm_i915_private *dev_priv, u32 reg, u32 val); >>> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h >>> index b57cba3..f41290c 100644 >>> --- a/drivers/gpu/drm/i915/i915_reg.h >>> +++ b/drivers/gpu/drm/i915/i915_reg.h >>> @@ -602,6 +602,18 @@ enum punit_power_well { >>> #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ >>> #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ >>> >>> +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 >>> +#define FB_GFX_FMAX_AT_VMAX_FUSE_MASK 0xff >>> +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 >>> +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 >>> +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 >>> + >> This blank line makes me think FB_GFX_GUAR_FREQ_FUSE_MASK isn't part of >> this register. So best not leave such blank line here. >> >> I have 0x3c3c3c28 in this register, which matches what I get using the >> old method. >> >>> +#define FB_GFX_GUAR_FREQ_FUSE_MASK 0xff >>> + >>> +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 >>> +#define FB_GFX_FMIN_AT_VMIN_FUSE_MASK 0xff >>> +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 >> I have 0x69841428 here. The low 8 bits look like another freq value. >> What is it? >> >> I have no docs for this stuff so can't really review apart from looking >> at what my hardware reports. >> >> >> Actually, since all the values are 8 bits maybe it would be neater to >> just >> #define FB_GFX_FREQ_FUSE_MASK 0xff >> and use that everywhere instead of having three different definitions >> for the same 0xff value. >> >>> + >>> #define PUNIT_GPU_STATUS_REG 0xdb >>> #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 >>> #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff >>> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c >>> index 2acb3de..71b8e2f 100644 >>> --- a/drivers/gpu/drm/i915/intel_pm.c >>> +++ b/drivers/gpu/drm/i915/intel_pm.c >>> @@ -4346,11 +4346,29 @@ void gen6_update_ring_freq(struct drm_device *dev) >>> >>> static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) >>> { >>> + struct drm_device *dev = dev_priv->dev; >>> + struct intel_device_info *info; >>> u32 val, rp0; >>> >>> - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >>> - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; >>> - >>> + info = (struct intel_device_info *)&dev_priv->info; >> Pointless cast. Also the assignment could be done when declaring info, >> and we usually use INTEL_INFO() to get at it. >> >>> + >>> + if (dev->pdev->revision >= 0x20) { >> Do we really need this check? I would think it would be up to the >> Punit firmware version rather the stepping. My BSW has PCI rev 0x15, >> but with the latest BIOS both fuse registers already contain correct >> looking information. >> >>> + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); >>> + >>> + if (info->eu_total == 8) /* (2 * 4) config */ >>> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); >>> + else if (info->eu_total == 12) /* (2 * 6) config */ >>> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); >>> + else if (info->eu_total == 16) /* (2 * 8) config */ >>> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); >>> + else /* Setting (2 * 8) Min RP0 for any other combination */ >>> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); >> 'switch (INTEL_INFO(dev)->eu_total)' perhaps? >> >>> + rp0 = (rp0 & FB_GFX_FMAX_AT_VMAX_FUSE_MASK); >>> + } else { /* For pre-production hardware */ >>> + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >>> + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & >>> + PUNIT_GPU_STATUS_MAX_FREQ_MASK; >>> + } >>> return rp0; >>> } >>> >>> @@ -4366,20 +4384,40 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) >>> >>> static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) >>> { >>> + struct drm_device *dev = dev_priv->dev; >>> + struct intel_device_info *info; >>> u32 val, rp1; >>> >>> - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); >>> - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; >>> + info = (struct intel_device_info *)&dev_priv->info; >> Unused here. >> >>> >>> + if (dev->pdev->revision >= 0x20) { >>> + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); >>> + rp1 = (val & FB_GFX_GUAR_FREQ_FUSE_MASK); >>> + } else { /* For pre-production hardware */ >>> + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); >>> + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & >>> + PUNIT_GPU_STATUS_MAX_FREQ_MASK); >>> + } >>> return rp1; >>> } >>> >>> static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) >>> { >>> + struct drm_device *dev = dev_priv->dev; >>> + struct intel_device_info *info; >>> u32 val, rpn; >>> + info = (struct intel_device_info *)&dev_priv->info; >> Unused. >> >>> + >>> + if (dev->pdev->revision >= 0x20) { >>> + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); >>> + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & >>> + FB_GFX_FMIN_AT_VMIN_FUSE_MASK); >>> + } else { /* For pre-production hardware */ >>> + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >>> + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & >>> + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); >>> + } >>> >>> - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >>> - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; >>> return rpn; >>> } >>> >>> diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c >>> index 01d841e..3c42eef 100644 >>> --- a/drivers/gpu/drm/i915/intel_sideband.c >>> +++ b/drivers/gpu/drm/i915/intel_sideband.c >>> @@ -75,7 +75,7 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn, >>> return 0; >>> } >>> >>> -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) >>> +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr) >> Good thing you had to pass a constant. Otherwise this could have caused >> quite a bit of head scratching for you ;) > This is the kind of thing that really should be a separate fix. If, for > any reason, we need to revert the rest, we'll still want this. > > BR, > Jani. Agreed, Better to have a separate patch :) >>> { >>> u32 val = 0; >>> >>> @@ -89,7 +89,7 @@ u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) >>> return val; >>> } >>> >>> -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val) >>> +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val) >>> { >>> WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); >>> >>> -- >>> 1.9.1 >> -- >> Ville Syrjälä >> Intel OTC >> _______________________________________________ >> Intel-gfx mailing list >> Intel-gfx@lists.freedesktop.org >> http://lists.freedesktop.org/mailman/listinfo/intel-gfx _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* [PATCH v2 0/3] Use new turbo offset for chv production system 2014-12-16 12:11 ` Deepak S @ 2015-01-16 15:12 ` deepak.s 2015-01-16 15:12 ` [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview deepak.s ` (2 more replies) 0 siblings, 3 replies; 30+ messages in thread From: deepak.s @ 2015-01-16 15:12 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> CHV/BSW production system has new turbo offset to read different freq. This series adds the support. Deepak S (3): drm/i915/chv: Populate total EU count on Cherryview drm/i915: Increase the range of sideband address. drm/i915: New offset for reading frequencies on CHV. drivers/gpu/drm/i915/i915_dma.c | 11 ++++++++ drivers/gpu/drm/i915/i915_drv.h | 5 ++-- drivers/gpu/drm/i915/i915_reg.h | 20 +++++++++++++ drivers/gpu/drm/i915/intel_pm.c | 53 +++++++++++++++++++++++++++++------ drivers/gpu/drm/i915/intel_sideband.c | 4 +-- 5 files changed, 81 insertions(+), 12 deletions(-) -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview 2015-01-16 15:12 ` [PATCH v2 0/3] Use new turbo offset for chv production system deepak.s @ 2015-01-16 15:12 ` deepak.s 2015-01-16 16:06 ` Jeff McGee ` (2 more replies) 2015-01-16 15:12 ` [PATCH v2 2/3] drm/i915: Increase the range of sideband address deepak.s 2015-01-16 15:12 ` [PATCH v2 3/3] drm/i915: New offset for reading frequencies on CHV deepak.s 2 siblings, 3 replies; 30+ messages in thread From: deepak.s @ 2015-01-16 15:12 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> Starting with Cherryview, devices may have a varying number of EU for a given ID due to creative fusing. Punit support different frequency for different fuse data. We use this patch to help get total eu enabled and read the right offset to get RP0 Based upon a patch from Jeff, but reworked to only store eu_total and avoid sending info to userspace v2: Format register definitions (Jani) Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> --- drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ 3 files changed, 23 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index 2447de3..b868e9d 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -601,6 +601,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) info->num_pipes = 0; } } + + if (IS_CHERRYVIEW(dev)) { + u32 fuse, mask_eu; + + fuse = I915_READ(CHV_FUSE_GT); + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | + CHV_FGT_EU_DIS_SS0_R1_MASK | + CHV_FGT_EU_DIS_SS1_R0_MASK | + CHV_FGT_EU_DIS_SS1_R1_MASK); + info->eu_total = 16 - hweight32(mask_eu); + } } /** diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 66f0c60..ab1fa9e 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -653,6 +653,7 @@ struct intel_device_info { int trans_offsets[I915_MAX_TRANSCODERS]; int palette_offsets[I915_MAX_PIPES]; int cursor_offsets[I915_MAX_PIPES]; + unsigned int eu_total; }; #undef DEFINE_FLAG diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index a39bb03..d9692f9 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -1471,6 +1471,17 @@ enum punit_power_well { #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) +/* Fuse readout registers for GT */ +#define CHV_FUSE_GT (VLV_DISPLAY_BASE + 0x2168) +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf << CHV_FGT_EU_DIS_SS0_R0_SHIFT) +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf << CHV_FGT_EU_DIS_SS0_R1_SHIFT) +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf << CHV_FGT_EU_DIS_SS1_R0_SHIFT) +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf << CHV_FGT_EU_DIS_SS1_R1_SHIFT) + #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* Re: [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview 2015-01-16 15:12 ` [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview deepak.s @ 2015-01-16 16:06 ` Jeff McGee 2015-01-16 17:11 ` Ville Syrjälä 2015-01-19 9:44 ` Daniel Vetter 2 siblings, 0 replies; 30+ messages in thread From: Jeff McGee @ 2015-01-16 16:06 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx On Fri, Jan 16, 2015 at 08:42:16PM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Starting with Cherryview, devices may have a varying number of EU for > a given ID due to creative fusing. Punit support different frequency for > different fuse data. We use this patch to help get total eu enabled and > read the right offset to get RP0 > > Based upon a patch from Jeff, but reworked to only store eu_total and > avoid sending info to userspace > > v2: Format register definitions (Jani) > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> > --- > drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ > drivers/gpu/drm/i915/i915_drv.h | 1 + > drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ > 3 files changed, 23 insertions(+) > > diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c > index 2447de3..b868e9d 100644 > --- a/drivers/gpu/drm/i915/i915_dma.c > +++ b/drivers/gpu/drm/i915/i915_dma.c > @@ -601,6 +601,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) > info->num_pipes = 0; > } > } > + > + if (IS_CHERRYVIEW(dev)) { > + u32 fuse, mask_eu; > + > + fuse = I915_READ(CHV_FUSE_GT); > + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | > + CHV_FGT_EU_DIS_SS0_R1_MASK | > + CHV_FGT_EU_DIS_SS1_R0_MASK | > + CHV_FGT_EU_DIS_SS1_R1_MASK); > + info->eu_total = 16 - hweight32(mask_eu); > + } > } > > /** > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index 66f0c60..ab1fa9e 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -653,6 +653,7 @@ struct intel_device_info { > int trans_offsets[I915_MAX_TRANSCODERS]; > int palette_offsets[I915_MAX_PIPES]; > int cursor_offsets[I915_MAX_PIPES]; > + unsigned int eu_total; > }; > > #undef DEFINE_FLAG > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h > index a39bb03..d9692f9 100644 > --- a/drivers/gpu/drm/i915/i915_reg.h > +++ b/drivers/gpu/drm/i915/i915_reg.h > @@ -1471,6 +1471,17 @@ enum punit_power_well { > #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) > #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) > > +/* Fuse readout registers for GT */ > +#define CHV_FUSE_GT (VLV_DISPLAY_BASE + 0x2168) > +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 > +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf << CHV_FGT_EU_DIS_SS0_R0_SHIFT) > +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 > +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf << CHV_FGT_EU_DIS_SS0_R1_SHIFT) > +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 > +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf << CHV_FGT_EU_DIS_SS1_R0_SHIFT) > +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 > +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf << CHV_FGT_EU_DIS_SS1_R1_SHIFT) > + > #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 > #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) > #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) > -- > 1.9.1 > Looks good to me. I'm trying to get the full version of this upstreamed, but this version is likely to be accepted first. I'll rebase the user space export portions on it if needed. -Jeff _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview 2015-01-16 15:12 ` [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview deepak.s 2015-01-16 16:06 ` Jeff McGee @ 2015-01-16 17:11 ` Ville Syrjälä 2015-01-19 9:44 ` Daniel Vetter 2 siblings, 0 replies; 30+ messages in thread From: Ville Syrjälä @ 2015-01-16 17:11 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx On Fri, Jan 16, 2015 at 08:42:16PM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Starting with Cherryview, devices may have a varying number of EU for > a given ID due to creative fusing. Punit support different frequency for > different fuse data. We use this patch to help get total eu enabled and > read the right offset to get RP0 > > Based upon a patch from Jeff, but reworked to only store eu_total and > avoid sending info to userspace > > v2: Format register definitions (Jani) > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > --- > drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ > drivers/gpu/drm/i915/i915_drv.h | 1 + > drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ > 3 files changed, 23 insertions(+) > > diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c > index 2447de3..b868e9d 100644 > --- a/drivers/gpu/drm/i915/i915_dma.c > +++ b/drivers/gpu/drm/i915/i915_dma.c > @@ -601,6 +601,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) > info->num_pipes = 0; > } > } > + > + if (IS_CHERRYVIEW(dev)) { > + u32 fuse, mask_eu; > + > + fuse = I915_READ(CHV_FUSE_GT); > + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | > + CHV_FGT_EU_DIS_SS0_R1_MASK | > + CHV_FGT_EU_DIS_SS1_R0_MASK | > + CHV_FGT_EU_DIS_SS1_R1_MASK); > + info->eu_total = 16 - hweight32(mask_eu); > + } > } > > /** > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index 66f0c60..ab1fa9e 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -653,6 +653,7 @@ struct intel_device_info { > int trans_offsets[I915_MAX_TRANSCODERS]; > int palette_offsets[I915_MAX_PIPES]; > int cursor_offsets[I915_MAX_PIPES]; > + unsigned int eu_total; > }; > > #undef DEFINE_FLAG > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h > index a39bb03..d9692f9 100644 > --- a/drivers/gpu/drm/i915/i915_reg.h > +++ b/drivers/gpu/drm/i915/i915_reg.h > @@ -1471,6 +1471,17 @@ enum punit_power_well { > #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) > #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) > > +/* Fuse readout registers for GT */ > +#define CHV_FUSE_GT (VLV_DISPLAY_BASE + 0x2168) > +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 > +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf << CHV_FGT_EU_DIS_SS0_R0_SHIFT) > +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 > +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf << CHV_FGT_EU_DIS_SS0_R1_SHIFT) > +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 > +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf << CHV_FGT_EU_DIS_SS1_R0_SHIFT) > +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 > +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf << CHV_FGT_EU_DIS_SS1_R1_SHIFT) > + > #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 > #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) > #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) > -- > 1.9.1 -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview 2015-01-16 15:12 ` [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview deepak.s 2015-01-16 16:06 ` Jeff McGee 2015-01-16 17:11 ` Ville Syrjälä @ 2015-01-19 9:44 ` Daniel Vetter 2015-01-20 3:07 ` Deepak S 2 siblings, 1 reply; 30+ messages in thread From: Daniel Vetter @ 2015-01-19 9:44 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx On Fri, Jan 16, 2015 at 08:42:16PM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Starting with Cherryview, devices may have a varying number of EU for > a given ID due to creative fusing. Punit support different frequency for > different fuse data. We use this patch to help get total eu enabled and > read the right offset to get RP0 > > Based upon a patch from Jeff, but reworked to only store eu_total and > avoid sending info to userspace > > v2: Format register definitions (Jani) > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> This isn't really how sob works. The last line _must_ be the person who last touched the patch, otherwise you're doing it wrong. Since this patch doesn't seem to be from Jeff I've dropped his sobline and converted to an ack. Really if you just want to acknowledge people who have contributed to a patch (e.g. you've based your patch on some version of theirs) do that in the commit message + Cc: -Daniel > --- > drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ > drivers/gpu/drm/i915/i915_drv.h | 1 + > drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ > 3 files changed, 23 insertions(+) > > diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c > index 2447de3..b868e9d 100644 > --- a/drivers/gpu/drm/i915/i915_dma.c > +++ b/drivers/gpu/drm/i915/i915_dma.c > @@ -601,6 +601,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) > info->num_pipes = 0; > } > } > + > + if (IS_CHERRYVIEW(dev)) { > + u32 fuse, mask_eu; > + > + fuse = I915_READ(CHV_FUSE_GT); > + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | > + CHV_FGT_EU_DIS_SS0_R1_MASK | > + CHV_FGT_EU_DIS_SS1_R0_MASK | > + CHV_FGT_EU_DIS_SS1_R1_MASK); > + info->eu_total = 16 - hweight32(mask_eu); > + } > } > > /** > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index 66f0c60..ab1fa9e 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -653,6 +653,7 @@ struct intel_device_info { > int trans_offsets[I915_MAX_TRANSCODERS]; > int palette_offsets[I915_MAX_PIPES]; > int cursor_offsets[I915_MAX_PIPES]; > + unsigned int eu_total; > }; > > #undef DEFINE_FLAG > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h > index a39bb03..d9692f9 100644 > --- a/drivers/gpu/drm/i915/i915_reg.h > +++ b/drivers/gpu/drm/i915/i915_reg.h > @@ -1471,6 +1471,17 @@ enum punit_power_well { > #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) > #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) > > +/* Fuse readout registers for GT */ > +#define CHV_FUSE_GT (VLV_DISPLAY_BASE + 0x2168) > +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 > +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf << CHV_FGT_EU_DIS_SS0_R0_SHIFT) > +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 > +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf << CHV_FGT_EU_DIS_SS0_R1_SHIFT) > +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 > +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf << CHV_FGT_EU_DIS_SS1_R0_SHIFT) > +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 > +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf << CHV_FGT_EU_DIS_SS1_R1_SHIFT) > + > #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 > #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) > #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) > -- > 1.9.1 > -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview 2015-01-19 9:44 ` Daniel Vetter @ 2015-01-20 3:07 ` Deepak S 2015-01-20 9:06 ` Daniel Vetter 0 siblings, 1 reply; 30+ messages in thread From: Deepak S @ 2015-01-20 3:07 UTC (permalink / raw) To: Daniel Vetter; +Cc: daniel.vetter, intel-gfx On Monday 19 January 2015 03:14 PM, Daniel Vetter wrote: > On Fri, Jan 16, 2015 at 08:42:16PM +0530, deepak.s@linux.intel.com wrote: >> From: Deepak S <deepak.s@linux.intel.com> >> >> Starting with Cherryview, devices may have a varying number of EU for >> a given ID due to creative fusing. Punit support different frequency for >> different fuse data. We use this patch to help get total eu enabled and >> read the right offset to get RP0 >> >> Based upon a patch from Jeff, but reworked to only store eu_total and >> avoid sending info to userspace >> >> v2: Format register definitions (Jani) >> >> Signed-off-by: Deepak S <deepak.s@linux.intel.com> >> Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> > This isn't really how sob works. The last line _must_ be the person who > last touched the patch, otherwise you're doing it wrong. Since this patch > doesn't seem to be from Jeff I've dropped his sobline and converted to an > ack. > > Really if you just want to acknowledge people who have contributed to a > patch (e.g. you've based your patch on some version of theirs) do that in > the commit message + Cc: > -Daniel Oh Sorry. I was not aware of this. I will follow the guidelines in future patch submission. >> --- >> drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++ >> drivers/gpu/drm/i915/i915_drv.h | 1 + >> drivers/gpu/drm/i915/i915_reg.h | 11 +++++++++++ >> 3 files changed, 23 insertions(+) >> >> diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c >> index 2447de3..b868e9d 100644 >> --- a/drivers/gpu/drm/i915/i915_dma.c >> +++ b/drivers/gpu/drm/i915/i915_dma.c >> @@ -601,6 +601,17 @@ static void intel_device_info_runtime_init(struct drm_device *dev) >> info->num_pipes = 0; >> } >> } >> + >> + if (IS_CHERRYVIEW(dev)) { >> + u32 fuse, mask_eu; >> + >> + fuse = I915_READ(CHV_FUSE_GT); >> + mask_eu = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK | >> + CHV_FGT_EU_DIS_SS0_R1_MASK | >> + CHV_FGT_EU_DIS_SS1_R0_MASK | >> + CHV_FGT_EU_DIS_SS1_R1_MASK); >> + info->eu_total = 16 - hweight32(mask_eu); >> + } >> } >> >> /** >> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h >> index 66f0c60..ab1fa9e 100644 >> --- a/drivers/gpu/drm/i915/i915_drv.h >> +++ b/drivers/gpu/drm/i915/i915_drv.h >> @@ -653,6 +653,7 @@ struct intel_device_info { >> int trans_offsets[I915_MAX_TRANSCODERS]; >> int palette_offsets[I915_MAX_PIPES]; >> int cursor_offsets[I915_MAX_PIPES]; >> + unsigned int eu_total; >> }; >> >> #undef DEFINE_FLAG >> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h >> index a39bb03..d9692f9 100644 >> --- a/drivers/gpu/drm/i915/i915_reg.h >> +++ b/drivers/gpu/drm/i915/i915_reg.h >> @@ -1471,6 +1471,17 @@ enum punit_power_well { >> #define GEN8_RC_SEMA_IDLE_MSG_DISABLE (1 << 12) >> #define GEN8_FF_DOP_CLOCK_GATE_DISABLE (1<<10) >> >> +/* Fuse readout registers for GT */ >> +#define CHV_FUSE_GT (VLV_DISPLAY_BASE + 0x2168) >> +#define CHV_FGT_EU_DIS_SS0_R0_SHIFT 16 >> +#define CHV_FGT_EU_DIS_SS0_R0_MASK (0xf << CHV_FGT_EU_DIS_SS0_R0_SHIFT) >> +#define CHV_FGT_EU_DIS_SS0_R1_SHIFT 20 >> +#define CHV_FGT_EU_DIS_SS0_R1_MASK (0xf << CHV_FGT_EU_DIS_SS0_R1_SHIFT) >> +#define CHV_FGT_EU_DIS_SS1_R0_SHIFT 24 >> +#define CHV_FGT_EU_DIS_SS1_R0_MASK (0xf << CHV_FGT_EU_DIS_SS1_R0_SHIFT) >> +#define CHV_FGT_EU_DIS_SS1_R1_SHIFT 28 >> +#define CHV_FGT_EU_DIS_SS1_R1_MASK (0xf << CHV_FGT_EU_DIS_SS1_R1_SHIFT) >> + >> #define GEN6_BSD_SLEEP_PSMI_CONTROL 0x12050 >> #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) >> #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) >> -- >> 1.9.1 >> _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview 2015-01-20 3:07 ` Deepak S @ 2015-01-20 9:06 ` Daniel Vetter 0 siblings, 0 replies; 30+ messages in thread From: Daniel Vetter @ 2015-01-20 9:06 UTC (permalink / raw) To: Deepak S; +Cc: daniel.vetter, intel-gfx On Tue, Jan 20, 2015 at 08:37:21AM +0530, Deepak S wrote: > > On Monday 19 January 2015 03:14 PM, Daniel Vetter wrote: > >On Fri, Jan 16, 2015 at 08:42:16PM +0530, deepak.s@linux.intel.com wrote: > >>From: Deepak S <deepak.s@linux.intel.com> > >> > >>Starting with Cherryview, devices may have a varying number of EU for > >>a given ID due to creative fusing. Punit support different frequency for > >>different fuse data. We use this patch to help get total eu enabled and > >>read the right offset to get RP0 > >> > >>Based upon a patch from Jeff, but reworked to only store eu_total and > >>avoid sending info to userspace > >> > >>v2: Format register definitions (Jani) > >> > >>Signed-off-by: Deepak S <deepak.s@linux.intel.com> > >>Signed-off-by: Jeff McGee <jeff.mcgee@intel.com> > >This isn't really how sob works. The last line _must_ be the person who > >last touched the patch, otherwise you're doing it wrong. Since this patch > >doesn't seem to be from Jeff I've dropped his sobline and converted to an > >ack. > > > >Really if you just want to acknowledge people who have contributed to a > >patch (e.g. you've based your patch on some version of theirs) do that in > >the commit message + Cc: > >-Daniel > > Oh Sorry. I was not aware of this. I will follow the guidelines in future patch submission. Yeah I think most examples of sob abuse come from vpg display, so updating your guidelines would be good. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* [PATCH v2 2/3] drm/i915: Increase the range of sideband address. 2015-01-16 15:12 ` [PATCH v2 0/3] Use new turbo offset for chv production system deepak.s 2015-01-16 15:12 ` [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview deepak.s @ 2015-01-16 15:12 ` deepak.s 2015-01-16 17:11 ` Ville Syrjälä 2015-01-16 15:12 ` [PATCH v2 3/3] drm/i915: New offset for reading frequencies on CHV deepak.s 2 siblings, 1 reply; 30+ messages in thread From: deepak.s @ 2015-01-16 15:12 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> Looks like latest BSW/CHV production system has sideband address > 128. Use u32 data types to cover new offset/address range :) Signed-off-by: Deepak S <deepak.s@linux.intel.com> --- drivers/gpu/drm/i915/i915_drv.h | 4 ++-- drivers/gpu/drm/i915/intel_sideband.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index ab1fa9e..272088c 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -3130,8 +3130,8 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val); /* intel_sideband.c */ -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr); -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val); +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr); +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val); u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr); u32 vlv_gpio_nc_read(struct drm_i915_private *dev_priv, u32 reg); void vlv_gpio_nc_write(struct drm_i915_private *dev_priv, u32 reg, u32 val); diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c index 01d841e..3c42eef 100644 --- a/drivers/gpu/drm/i915/intel_sideband.c +++ b/drivers/gpu/drm/i915/intel_sideband.c @@ -75,7 +75,7 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn, return 0; } -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr) { u32 val = 0; @@ -89,7 +89,7 @@ u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) return val; } -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val) +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val) { WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* Re: [PATCH v2 2/3] drm/i915: Increase the range of sideband address. 2015-01-16 15:12 ` [PATCH v2 2/3] drm/i915: Increase the range of sideband address deepak.s @ 2015-01-16 17:11 ` Ville Syrjälä 0 siblings, 0 replies; 30+ messages in thread From: Ville Syrjälä @ 2015-01-16 17:11 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx On Fri, Jan 16, 2015 at 08:42:17PM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Looks like latest BSW/CHV production system has sideband address > 128. > Use u32 data types to cover new offset/address range :) > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > --- > drivers/gpu/drm/i915/i915_drv.h | 4 ++-- > drivers/gpu/drm/i915/intel_sideband.c | 4 ++-- > 2 files changed, 4 insertions(+), 4 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index ab1fa9e..272088c 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -3130,8 +3130,8 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val > int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val); > > /* intel_sideband.c */ > -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr); > -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val); > +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr); > +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val); > u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr); > u32 vlv_gpio_nc_read(struct drm_i915_private *dev_priv, u32 reg); > void vlv_gpio_nc_write(struct drm_i915_private *dev_priv, u32 reg, u32 val); > diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c > index 01d841e..3c42eef 100644 > --- a/drivers/gpu/drm/i915/intel_sideband.c > +++ b/drivers/gpu/drm/i915/intel_sideband.c > @@ -75,7 +75,7 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn, > return 0; > } > > -u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) > +u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr) > { > u32 val = 0; > > @@ -89,7 +89,7 @@ u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr) > return val; > } > > -void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val) > +void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val) > { > WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); > > -- > 1.9.1 -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* [PATCH v2 3/3] drm/i915: New offset for reading frequencies on CHV. 2015-01-16 15:12 ` [PATCH v2 0/3] Use new turbo offset for chv production system deepak.s 2015-01-16 15:12 ` [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview deepak.s 2015-01-16 15:12 ` [PATCH v2 2/3] drm/i915: Increase the range of sideband address deepak.s @ 2015-01-16 15:12 ` deepak.s 2015-01-16 17:09 ` Ville Syrjälä 2 siblings, 1 reply; 30+ messages in thread From: deepak.s @ 2015-01-16 15:12 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> Use new Sideband offset to read max/min/gaur freq based on the SKU it is running on. Based on the Number of EU, we read different bits to identify the max frequencies at which system can run. v2: reuse mask definitions & INTEL_INFO() to get device info (Ville) Signed-off-by: Deepak S <deepak.s@linux.intel.com> --- drivers/gpu/drm/i915/i915_reg.h | 9 +++++++ drivers/gpu/drm/i915/intel_pm.c | 53 ++++++++++++++++++++++++++++++++++------- 2 files changed, 54 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index d9692f9..2dcb1b3 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -605,6 +605,15 @@ enum punit_power_well { #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 +#define FB_GFX_FREQ_FUSE_MASK 0xff +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 + +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 + #define PUNIT_GPU_STATUS_REG 0xdb #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 03fc7f2..c010d5c 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -4338,11 +4338,32 @@ void gen6_update_ring_freq(struct drm_device *dev) static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; u32 val, rp0; - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; - + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); + + switch (INTEL_INFO(dev)->eu_total) { + case 8: + /* (2 * 4) config */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); + case 12: + /* (2 * 6) config */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); + case 16: + /* (2 * 8) config */ + default: + /* Setting (2 * 8) Min RP0 for any other combination */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); + } + rp0 = (rp0 & FB_GFX_FREQ_FUSE_MASK); + } else { + /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & + PUNIT_GPU_STATUS_MAX_FREQ_MASK; + } return rp0; } @@ -4358,20 +4379,36 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; u32 val, rp1; - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; - + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); + rp1 = (val & FB_GFX_FREQ_FUSE_MASK); + } else { + /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & + PUNIT_GPU_STATUS_MAX_FREQ_MASK); + } return rp1; } static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; u32 val, rpn; - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & + FB_GFX_FREQ_FUSE_MASK); + } else { /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); + } + return rpn; } -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* Re: [PATCH v2 3/3] drm/i915: New offset for reading frequencies on CHV. 2015-01-16 15:12 ` [PATCH v2 3/3] drm/i915: New offset for reading frequencies on CHV deepak.s @ 2015-01-16 17:09 ` Ville Syrjälä 2015-01-17 5:34 ` Deepak S 0 siblings, 1 reply; 30+ messages in thread From: Ville Syrjälä @ 2015-01-16 17:09 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx On Fri, Jan 16, 2015 at 08:42:18PM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Use new Sideband offset to read max/min/gaur freq based on the SKU it > is running on. Based on the Number of EU, we read different bits to > identify the max frequencies at which system can run. > > v2: reuse mask definitions & INTEL_INFO() to get device info (Ville) > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > --- > drivers/gpu/drm/i915/i915_reg.h | 9 +++++++ > drivers/gpu/drm/i915/intel_pm.c | 53 ++++++++++++++++++++++++++++++++++------- > 2 files changed, 54 insertions(+), 8 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h > index d9692f9..2dcb1b3 100644 > --- a/drivers/gpu/drm/i915/i915_reg.h > +++ b/drivers/gpu/drm/i915/i915_reg.h > @@ -605,6 +605,15 @@ enum punit_power_well { > #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ > #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ > > +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 > +#define FB_GFX_FREQ_FUSE_MASK 0xff > +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 > +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 > +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 > + > +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 > +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 > + > #define PUNIT_GPU_STATUS_REG 0xdb > #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 > #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c > index 03fc7f2..c010d5c 100644 > --- a/drivers/gpu/drm/i915/intel_pm.c > +++ b/drivers/gpu/drm/i915/intel_pm.c > @@ -4338,11 +4338,32 @@ void gen6_update_ring_freq(struct drm_device *dev) > > static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > u32 val, rp0; > > - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; > - > + if (dev->pdev->revision >= 0x20) { > + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); > + > + switch (INTEL_INFO(dev)->eu_total) { > + case 8: > + /* (2 * 4) config */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); break; > + case 12: > + /* (2 * 6) config */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); break; > + case 16: > + /* (2 * 8) config */ > + default: > + /* Setting (2 * 8) Min RP0 for any other combination */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); Maybe break; here too if you feel like it :) Hmm. Now that I started to think about it, might we be expecting EUs to be fused off in some other configurations? In that case the switch statement might be not be the best idea, or we'd need to use the gnu case range extension. But we can maybe worry about that later since it might require more investigative work, or I might be totally off base here anyway, and we should get this patch in ASAP. So if you add the missing break statements this patch can have: Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > + } > + rp0 = (rp0 & FB_GFX_FREQ_FUSE_MASK); > + } else { > + /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_MAX_FREQ_MASK; > + } > return rp0; > } > > @@ -4358,20 +4379,36 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) > > static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > u32 val, rp1; > > - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); > - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; > - > + if (dev->pdev->revision >= 0x20) { > + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); > + rp1 = (val & FB_GFX_FREQ_FUSE_MASK); > + } else { > + /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); > + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_MAX_FREQ_MASK); > + } > return rp1; > } > > static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > u32 val, rpn; > > - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; > + if (dev->pdev->revision >= 0x20) { > + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); > + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & > + FB_GFX_FREQ_FUSE_MASK); > + } else { /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); > + } > + > return rpn; > } > > -- > 1.9.1 -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH v2 3/3] drm/i915: New offset for reading frequencies on CHV. 2015-01-16 17:09 ` Ville Syrjälä @ 2015-01-17 5:34 ` Deepak S 2015-01-17 5:35 ` [PATCH v3] " deepak.s 0 siblings, 1 reply; 30+ messages in thread From: Deepak S @ 2015-01-17 5:34 UTC (permalink / raw) To: Ville Syrjälä; +Cc: daniel.vetter, intel-gfx On Friday 16 January 2015 10:39 PM, Ville Syrjälä wrote: > On Fri, Jan 16, 2015 at 08:42:18PM +0530, deepak.s@linux.intel.com wrote: >> From: Deepak S <deepak.s@linux.intel.com> >> >> Use new Sideband offset to read max/min/gaur freq based on the SKU it >> is running on. Based on the Number of EU, we read different bits to >> identify the max frequencies at which system can run. >> >> v2: reuse mask definitions & INTEL_INFO() to get device info (Ville) >> >> Signed-off-by: Deepak S <deepak.s@linux.intel.com> >> --- >> drivers/gpu/drm/i915/i915_reg.h | 9 +++++++ >> drivers/gpu/drm/i915/intel_pm.c | 53 ++++++++++++++++++++++++++++++++++------- >> 2 files changed, 54 insertions(+), 8 deletions(-) >> >> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h >> index d9692f9..2dcb1b3 100644 >> --- a/drivers/gpu/drm/i915/i915_reg.h >> +++ b/drivers/gpu/drm/i915/i915_reg.h >> @@ -605,6 +605,15 @@ enum punit_power_well { >> #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ >> #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ >> >> +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 >> +#define FB_GFX_FREQ_FUSE_MASK 0xff >> +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 >> +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 >> +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 >> + >> +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 >> +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 >> + >> #define PUNIT_GPU_STATUS_REG 0xdb >> #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 >> #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff >> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c >> index 03fc7f2..c010d5c 100644 >> --- a/drivers/gpu/drm/i915/intel_pm.c >> +++ b/drivers/gpu/drm/i915/intel_pm.c >> @@ -4338,11 +4338,32 @@ void gen6_update_ring_freq(struct drm_device *dev) >> >> static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) >> { >> + struct drm_device *dev = dev_priv->dev; >> u32 val, rp0; >> >> - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >> - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; >> - >> + if (dev->pdev->revision >= 0x20) { >> + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); >> + >> + switch (INTEL_INFO(dev)->eu_total) { >> + case 8: >> + /* (2 * 4) config */ >> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); > break; > >> + case 12: >> + /* (2 * 6) config */ >> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); > break; > >> + case 16: >> + /* (2 * 8) config */ >> + default: >> + /* Setting (2 * 8) Min RP0 for any other combination */ >> + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); > Maybe break; here too if you feel like it :) :) Thanks for the review > Hmm. Now that I started to think about it, might we be expecting EUs to > be fused off in some other configurations? In that case the switch > statement might be not be the best idea, or we'd need to use the gnu > case range extension. But we can maybe worry about that later since it > might require more investigative work, or I might be totally off base here > anyway, and we should get this patch in ASAP. > > So if you add the missing break statements this patch can have: > Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > >> + } >> + rp0 = (rp0 & FB_GFX_FREQ_FUSE_MASK); >> + } else { >> + /* For pre-production hardware */ >> + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >> + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & >> + PUNIT_GPU_STATUS_MAX_FREQ_MASK; >> + } >> return rp0; >> } >> >> @@ -4358,20 +4379,36 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) >> >> static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) >> { >> + struct drm_device *dev = dev_priv->dev; >> u32 val, rp1; >> >> - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); >> - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; >> - >> + if (dev->pdev->revision >= 0x20) { >> + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); >> + rp1 = (val & FB_GFX_FREQ_FUSE_MASK); >> + } else { >> + /* For pre-production hardware */ >> + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); >> + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & >> + PUNIT_GPU_STATUS_MAX_FREQ_MASK); >> + } >> return rp1; >> } >> >> static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) >> { >> + struct drm_device *dev = dev_priv->dev; >> u32 val, rpn; >> >> - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >> - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; >> + if (dev->pdev->revision >= 0x20) { >> + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); >> + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & >> + FB_GFX_FREQ_FUSE_MASK); >> + } else { /* For pre-production hardware */ >> + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); >> + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & >> + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); >> + } >> + >> return rpn; >> } >> >> -- >> 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* [PATCH v3] drm/i915: New offset for reading frequencies on CHV. 2015-01-17 5:34 ` Deepak S @ 2015-01-17 5:35 ` deepak.s 2015-01-19 9:47 ` Daniel Vetter 0 siblings, 1 reply; 30+ messages in thread From: deepak.s @ 2015-01-17 5:35 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> Use new Sideband offset to read max/min/gaur freq based on the SKU it is running on. Based on the Number of EU, we read different bits to identify the max frequencies at which system can run. v2: reuse mask definitions & INTEL_INFO() to get device info (Ville) v3: add break in switch conditions (Ville) Signed-off-by: Deepak S <deepak.s@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala at linux.intel.com> --- drivers/gpu/drm/i915/i915_reg.h | 9 +++++++ drivers/gpu/drm/i915/intel_pm.c | 54 +++++++++++++++++++++++++++++++++++------ 2 files changed, 56 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index d9692f9..2dcb1b3 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -605,6 +605,15 @@ enum punit_power_well { #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 +#define FB_GFX_FREQ_FUSE_MASK 0xff +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 + +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 + #define PUNIT_GPU_STATUS_REG 0xdb #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 03fc7f2..b73d601 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -4338,11 +4338,35 @@ void gen6_update_ring_freq(struct drm_device *dev) static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; u32 val, rp0; - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); + switch (INTEL_INFO(dev)->eu_total) { + case 8: + /* (2 * 4) config */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); + break; + case 12: + /* (2 * 6) config */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); + break; + case 16: + /* (2 * 8) config */ + default: + /* Setting (2 * 8) Min RP0 for any other combination */ + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); + break; + } + rp0 = (rp0 & FB_GFX_FREQ_FUSE_MASK); + } else { + /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & + PUNIT_GPU_STATUS_MAX_FREQ_MASK; + } return rp0; } @@ -4358,20 +4382,36 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; u32 val, rp1; - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; - + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); + rp1 = (val & FB_GFX_FREQ_FUSE_MASK); + } else { + /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & + PUNIT_GPU_STATUS_MAX_FREQ_MASK); + } return rp1; } static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) { + struct drm_device *dev = dev_priv->dev; u32 val, rpn; - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; + if (dev->pdev->revision >= 0x20) { + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & + FB_GFX_FREQ_FUSE_MASK); + } else { /* For pre-production hardware */ + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); + } + return rpn; } -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* Re: [PATCH v3] drm/i915: New offset for reading frequencies on CHV. 2015-01-17 5:35 ` [PATCH v3] " deepak.s @ 2015-01-19 9:47 ` Daniel Vetter 0 siblings, 0 replies; 30+ messages in thread From: Daniel Vetter @ 2015-01-19 9:47 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx On Sat, Jan 17, 2015 at 11:05:59AM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Use new Sideband offset to read max/min/gaur freq based on the SKU it > is running on. Based on the Number of EU, we read different bits to > identify the max frequencies at which system can run. > > v2: reuse mask definitions & INTEL_INFO() to get device info (Ville) > > v3: add break in switch conditions (Ville) > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > Reviewed-by: Ville Syrjälä <ville.syrjala at linux.intel.com> All 3 patches merged, thanks. -Daniel > --- > drivers/gpu/drm/i915/i915_reg.h | 9 +++++++ > drivers/gpu/drm/i915/intel_pm.c | 54 +++++++++++++++++++++++++++++++++++------ > 2 files changed, 56 insertions(+), 7 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h > index d9692f9..2dcb1b3 100644 > --- a/drivers/gpu/drm/i915/i915_reg.h > +++ b/drivers/gpu/drm/i915/i915_reg.h > @@ -605,6 +605,15 @@ enum punit_power_well { > #define PUNIT_FUSE_BUS2 0xf6 /* bits 47:40 */ > #define PUNIT_FUSE_BUS1 0xf5 /* bits 55:48 */ > > +#define FB_GFX_FMAX_AT_VMAX_FUSE 0x136 > +#define FB_GFX_FREQ_FUSE_MASK 0xff > +#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT 24 > +#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT 16 > +#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT 8 > + > +#define FB_GFX_FMIN_AT_VMIN_FUSE 0x137 > +#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT 8 > + > #define PUNIT_GPU_STATUS_REG 0xdb > #define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT 16 > #define PUNIT_GPU_STATUS_MAX_FREQ_MASK 0xff > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c > index 03fc7f2..b73d601 100644 > --- a/drivers/gpu/drm/i915/intel_pm.c > +++ b/drivers/gpu/drm/i915/intel_pm.c > @@ -4338,11 +4338,35 @@ void gen6_update_ring_freq(struct drm_device *dev) > > static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > u32 val, rp0; > > - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > - rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; > + if (dev->pdev->revision >= 0x20) { > + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); > > + switch (INTEL_INFO(dev)->eu_total) { > + case 8: > + /* (2 * 4) config */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); > + break; > + case 12: > + /* (2 * 6) config */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT); > + break; > + case 16: > + /* (2 * 8) config */ > + default: > + /* Setting (2 * 8) Min RP0 for any other combination */ > + rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT); > + break; > + } > + rp0 = (rp0 & FB_GFX_FREQ_FUSE_MASK); > + } else { > + /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > + rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_MAX_FREQ_MASK; > + } > return rp0; > } > > @@ -4358,20 +4382,36 @@ static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv) > > static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > u32 val, rp1; > > - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); > - rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK; > - > + if (dev->pdev->revision >= 0x20) { > + val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); > + rp1 = (val & FB_GFX_FREQ_FUSE_MASK); > + } else { > + /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); > + rp1 = ((val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_MAX_FREQ_MASK); > + } > return rp1; > } > > static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv) > { > + struct drm_device *dev = dev_priv->dev; > u32 val, rpn; > > - val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > - rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK; > + if (dev->pdev->revision >= 0x20) { > + val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE); > + rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) & > + FB_GFX_FREQ_FUSE_MASK); > + } else { /* For pre-production hardware */ > + val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG); > + rpn = ((val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & > + PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK); > + } > + > return rpn; > } > > -- > 1.9.1 > -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* [PATCH 4/4] drm/i915: Skip gunit save/restore for cherryview 2014-12-12 8:48 [PATCH 1/4] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s 2014-12-12 8:48 ` [PATCH 2/4] drm/i915/chv: Populate total EU count on Cherryview deepak.s 2014-12-12 8:48 ` [PATCH 3/4] drm/i915: New offset for reading frequencies on CHV deepak.s @ 2014-12-12 8:48 ` deepak.s 2014-12-11 12:02 ` Ville Syrjälä 2014-12-13 6:13 ` [PATCH v2] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s 3 siblings, 1 reply; 30+ messages in thread From: deepak.s @ 2014-12-12 8:48 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter From: Deepak S <deepak.s@linux.intel.com> With cherryview onwards, Gunit hardware itself save and restore all the Gunit registers. Skipping the "vlv_save_gunit_s0ix_state" & "vlv_restore_gunit_s0ix_state" for cherryview in S3/S0ix sequence. Signed-off-by: Deepak S <deepak.s@linux.intel.com> --- drivers/gpu/drm/i915/i915_drv.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 6c09bf8..6257f66 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1299,7 +1299,9 @@ static int vlv_suspend_complete(struct drm_i915_private *dev_priv) err = vlv_allow_gt_wake(dev_priv, false); if (err) goto err2; - vlv_save_gunit_s0ix_state(dev_priv); + + if (!IS_CHERRYVIEW(dev_priv->dev)) + vlv_save_gunit_s0ix_state(dev_priv); err = vlv_force_gfx_clock(dev_priv, false); if (err) @@ -1330,7 +1332,8 @@ static int vlv_resume_prepare(struct drm_i915_private *dev_priv, */ ret = vlv_force_gfx_clock(dev_priv, true); - vlv_restore_gunit_s0ix_state(dev_priv); + if (!IS_CHERRYVIEW(dev_priv->dev)) + vlv_restore_gunit_s0ix_state(dev_priv); err = vlv_allow_gt_wake(dev_priv, true); if (!ret) -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* Re: [PATCH 4/4] drm/i915: Skip gunit save/restore for cherryview 2014-12-12 8:48 ` [PATCH 4/4] drm/i915: Skip gunit save/restore for cherryview deepak.s @ 2014-12-11 12:02 ` Ville Syrjälä 2014-12-15 15:16 ` Daniel Vetter 0 siblings, 1 reply; 30+ messages in thread From: Ville Syrjälä @ 2014-12-11 12:02 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx On Fri, Dec 12, 2014 at 02:18:16PM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > With cherryview onwards, Gunit hardware itself save and restore all the > Gunit registers. Skipping the "vlv_save_gunit_s0ix_state" & > "vlv_restore_gunit_s0ix_state" for cherryview in S3/S0ix sequence. > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> I had the same patch tucked away somewhere, but didn't bother to send it since I've not yet looked into s0ix. In any case the docs agree with this so: Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > --- > drivers/gpu/drm/i915/i915_drv.c | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c > index 6c09bf8..6257f66 100644 > --- a/drivers/gpu/drm/i915/i915_drv.c > +++ b/drivers/gpu/drm/i915/i915_drv.c > @@ -1299,7 +1299,9 @@ static int vlv_suspend_complete(struct drm_i915_private *dev_priv) > err = vlv_allow_gt_wake(dev_priv, false); > if (err) > goto err2; > - vlv_save_gunit_s0ix_state(dev_priv); > + > + if (!IS_CHERRYVIEW(dev_priv->dev)) > + vlv_save_gunit_s0ix_state(dev_priv); > > err = vlv_force_gfx_clock(dev_priv, false); > if (err) > @@ -1330,7 +1332,8 @@ static int vlv_resume_prepare(struct drm_i915_private *dev_priv, > */ > ret = vlv_force_gfx_clock(dev_priv, true); > > - vlv_restore_gunit_s0ix_state(dev_priv); > + if (!IS_CHERRYVIEW(dev_priv->dev)) > + vlv_restore_gunit_s0ix_state(dev_priv); > > err = vlv_allow_gt_wake(dev_priv, true); > if (!ret) > -- > 1.9.1 -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH 4/4] drm/i915: Skip gunit save/restore for cherryview 2014-12-11 12:02 ` Ville Syrjälä @ 2014-12-15 15:16 ` Daniel Vetter 0 siblings, 0 replies; 30+ messages in thread From: Daniel Vetter @ 2014-12-15 15:16 UTC (permalink / raw) To: Ville Syrjälä; +Cc: daniel.vetter, intel-gfx On Thu, Dec 11, 2014 at 02:02:52PM +0200, Ville Syrjälä wrote: > On Fri, Dec 12, 2014 at 02:18:16PM +0530, deepak.s@linux.intel.com wrote: > > From: Deepak S <deepak.s@linux.intel.com> > > > > With cherryview onwards, Gunit hardware itself save and restore all the > > Gunit registers. Skipping the "vlv_save_gunit_s0ix_state" & > > "vlv_restore_gunit_s0ix_state" for cherryview in S3/S0ix sequence. > > > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > > I had the same patch tucked away somewhere, but didn't bother to send it > since I've not yet looked into s0ix. In any case the docs agree with > this so: > > Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Merged patch 1&4 from this series, thanks. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* [PATCH v2] drm/i915/chv: Use timeout mode for RC6 on chv 2014-12-12 8:48 [PATCH 1/4] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s ` (2 preceding siblings ...) 2014-12-12 8:48 ` [PATCH 4/4] drm/i915: Skip gunit save/restore for cherryview deepak.s @ 2014-12-13 6:13 ` deepak.s 2014-12-12 16:34 ` Ville Syrjälä 3 siblings, 1 reply; 30+ messages in thread From: deepak.s @ 2014-12-13 6:13 UTC (permalink / raw) To: intel-gfx; +Cc: daniel.vetter, Rodrigo Vivi From: Deepak S <deepak.s@linux.intel.com> Higher RC6 residency is observed using timeout mode instead of EI mode. It's Recommended to use TO Method for RC6. v2: Add comment about timeout threshold. (Tom) Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> --- drivers/gpu/drm/i915/intel_pm.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 2316d23..2acb3de 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -4689,7 +4689,8 @@ static void cherryview_enable_rps(struct drm_device *dev) I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10); I915_WRITE(GEN6_RC_SLEEP, 0); - I915_WRITE(GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */ + /* TO threshold set to 1750 us ( 0x557 * 1.28 us) */ + I915_WRITE(GEN6_RC6_THRESHOLD, 0x557); /* allows RC6 residency counter to work */ I915_WRITE(VLV_COUNTER_CONTROL, @@ -4703,7 +4704,7 @@ static void cherryview_enable_rps(struct drm_device *dev) /* 3: Enable RC6 */ if ((intel_enable_rc6(dev) & INTEL_RC6_ENABLE) && (pcbr >> VLV_PCBR_ADDR_SHIFT)) - rc6_mode = GEN6_RC_CTL_EI_MODE(1); + rc6_mode = GEN7_RC_CTL_TO_MODE; I915_WRITE(GEN6_RC_CONTROL, rc6_mode); -- 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply related [flat|nested] 30+ messages in thread
* Re: [PATCH v2] drm/i915/chv: Use timeout mode for RC6 on chv 2014-12-13 6:13 ` [PATCH v2] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s @ 2014-12-12 16:34 ` Ville Syrjälä 2014-12-16 12:09 ` Deepak S 0 siblings, 1 reply; 30+ messages in thread From: Ville Syrjälä @ 2014-12-12 16:34 UTC (permalink / raw) To: deepak.s; +Cc: daniel.vetter, intel-gfx, Rodrigo Vivi On Sat, Dec 13, 2014 at 11:43:27AM +0530, deepak.s@linux.intel.com wrote: > From: Deepak S <deepak.s@linux.intel.com> > > Higher RC6 residency is observed using timeout mode > instead of EI mode. It's Recommended to use TO Method for RC6. > > v2: Add comment about timeout threshold. (Tom) Yeah if TO is better let's just use it. The 1750us value is what the BIOS spec recommends, so: Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > > Signed-off-by: Deepak S <deepak.s@linux.intel.com> > Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Why is Rodrigo's sob here? > --- > drivers/gpu/drm/i915/intel_pm.c | 5 +++-- > 1 file changed, 3 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c > index 2316d23..2acb3de 100644 > --- a/drivers/gpu/drm/i915/intel_pm.c > +++ b/drivers/gpu/drm/i915/intel_pm.c > @@ -4689,7 +4689,8 @@ static void cherryview_enable_rps(struct drm_device *dev) > I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10); > I915_WRITE(GEN6_RC_SLEEP, 0); > > - I915_WRITE(GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */ > + /* TO threshold set to 1750 us ( 0x557 * 1.28 us) */ > + I915_WRITE(GEN6_RC6_THRESHOLD, 0x557); > > /* allows RC6 residency counter to work */ > I915_WRITE(VLV_COUNTER_CONTROL, > @@ -4703,7 +4704,7 @@ static void cherryview_enable_rps(struct drm_device *dev) > /* 3: Enable RC6 */ > if ((intel_enable_rc6(dev) & INTEL_RC6_ENABLE) && > (pcbr >> VLV_PCBR_ADDR_SHIFT)) > - rc6_mode = GEN6_RC_CTL_EI_MODE(1); > + rc6_mode = GEN7_RC_CTL_TO_MODE; > > I915_WRITE(GEN6_RC_CONTROL, rc6_mode); > > -- > 1.9.1 -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH v2] drm/i915/chv: Use timeout mode for RC6 on chv 2014-12-12 16:34 ` Ville Syrjälä @ 2014-12-16 12:09 ` Deepak S 2014-12-15 15:12 ` Daniel Vetter 0 siblings, 1 reply; 30+ messages in thread From: Deepak S @ 2014-12-16 12:09 UTC (permalink / raw) To: Ville Syrjälä; +Cc: daniel.vetter, intel-gfx, Rodrigo Vivi On Friday 12 December 2014 10:04 PM, Ville Syrjälä wrote: > On Sat, Dec 13, 2014 at 11:43:27AM +0530, deepak.s@linux.intel.com wrote: >> From: Deepak S <deepak.s@linux.intel.com> >> >> Higher RC6 residency is observed using timeout mode >> instead of EI mode. It's Recommended to use TO Method for RC6. >> >> v2: Add comment about timeout threshold. (Tom) > Yeah if TO is better let's just use it. The 1750us value is what > the BIOS spec recommends, so: > > Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > >> Signed-off-by: Deepak S <deepak.s@linux.intel.com> >> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> > Why is Rodrigo's sob here? Rodrigo had resubmitted my patch with his sob. just retained :) >> --- >> drivers/gpu/drm/i915/intel_pm.c | 5 +++-- >> 1 file changed, 3 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c >> index 2316d23..2acb3de 100644 >> --- a/drivers/gpu/drm/i915/intel_pm.c >> +++ b/drivers/gpu/drm/i915/intel_pm.c >> @@ -4689,7 +4689,8 @@ static void cherryview_enable_rps(struct drm_device *dev) >> I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10); >> I915_WRITE(GEN6_RC_SLEEP, 0); >> >> - I915_WRITE(GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */ >> + /* TO threshold set to 1750 us ( 0x557 * 1.28 us) */ >> + I915_WRITE(GEN6_RC6_THRESHOLD, 0x557); >> >> /* allows RC6 residency counter to work */ >> I915_WRITE(VLV_COUNTER_CONTROL, >> @@ -4703,7 +4704,7 @@ static void cherryview_enable_rps(struct drm_device *dev) >> /* 3: Enable RC6 */ >> if ((intel_enable_rc6(dev) & INTEL_RC6_ENABLE) && >> (pcbr >> VLV_PCBR_ADDR_SHIFT)) >> - rc6_mode = GEN6_RC_CTL_EI_MODE(1); >> + rc6_mode = GEN7_RC_CTL_TO_MODE; >> >> I915_WRITE(GEN6_RC_CONTROL, rc6_mode); >> >> -- >> 1.9.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
* Re: [PATCH v2] drm/i915/chv: Use timeout mode for RC6 on chv 2014-12-16 12:09 ` Deepak S @ 2014-12-15 15:12 ` Daniel Vetter 0 siblings, 0 replies; 30+ messages in thread From: Daniel Vetter @ 2014-12-15 15:12 UTC (permalink / raw) To: Deepak S; +Cc: daniel.vetter, intel-gfx, Rodrigo Vivi On Tue, Dec 16, 2014 at 05:39:19PM +0530, Deepak S wrote: > > On Friday 12 December 2014 10:04 PM, Ville Syrjälä wrote: > >On Sat, Dec 13, 2014 at 11:43:27AM +0530, deepak.s@linux.intel.com wrote: > >>From: Deepak S <deepak.s@linux.intel.com> > >> > >>Higher RC6 residency is observed using timeout mode > >>instead of EI mode. It's Recommended to use TO Method for RC6. > >> > >>v2: Add comment about timeout threshold. (Tom) > >Yeah if TO is better let's just use it. The 1750us value is what > >the BIOS spec recommends, so: > > > >Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > > > >>Signed-off-by: Deepak S <deepak.s@linux.intel.com> > >>Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> > >Why is Rodrigo's sob here? > > Rodrigo had resubmitted my patch with his sob. just retained :) Sob should document the path a patch took, not every possible person who every touch a patch. If you want to acknowledge indirect contributions just mention them in the commit message (E.g. "based on a patch by" or "squash in fixup from $person"). The bangalore team like to pile up sob lines especially, which isn't really how it's supposed to work. A few things to check: - First sob should be the original author. - Last sob should be the submitter of the patch. Anyway, back to merging patches for me ;-) -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx ^ permalink raw reply [flat|nested] 30+ messages in thread
end of thread, other threads:[~2015-01-20 9:06 UTC | newest] Thread overview: 30+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2014-12-12 8:48 [PATCH 1/4] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s 2014-12-12 8:48 ` [PATCH 2/4] drm/i915/chv: Populate total EU count on Cherryview deepak.s 2014-12-11 12:09 ` Jani Nikula 2014-12-12 12:10 ` Deepak S 2014-12-12 12:33 ` [PATCH v2] " deepak.s 2014-12-12 8:48 ` [PATCH 3/4] drm/i915: New offset for reading frequencies on CHV deepak.s 2014-12-12 19:09 ` Ville Syrjälä 2014-12-15 6:51 ` Jani Nikula 2014-12-16 12:11 ` Deepak S 2015-01-16 15:12 ` [PATCH v2 0/3] Use new turbo offset for chv production system deepak.s 2015-01-16 15:12 ` [PATCH v2 1/3] drm/i915/chv: Populate total EU count on Cherryview deepak.s 2015-01-16 16:06 ` Jeff McGee 2015-01-16 17:11 ` Ville Syrjälä 2015-01-19 9:44 ` Daniel Vetter 2015-01-20 3:07 ` Deepak S 2015-01-20 9:06 ` Daniel Vetter 2015-01-16 15:12 ` [PATCH v2 2/3] drm/i915: Increase the range of sideband address deepak.s 2015-01-16 17:11 ` Ville Syrjälä 2015-01-16 15:12 ` [PATCH v2 3/3] drm/i915: New offset for reading frequencies on CHV deepak.s 2015-01-16 17:09 ` Ville Syrjälä 2015-01-17 5:34 ` Deepak S 2015-01-17 5:35 ` [PATCH v3] " deepak.s 2015-01-19 9:47 ` Daniel Vetter 2014-12-12 8:48 ` [PATCH 4/4] drm/i915: Skip gunit save/restore for cherryview deepak.s 2014-12-11 12:02 ` Ville Syrjälä 2014-12-15 15:16 ` Daniel Vetter 2014-12-13 6:13 ` [PATCH v2] drm/i915/chv: Use timeout mode for RC6 on chv deepak.s 2014-12-12 16:34 ` Ville Syrjälä 2014-12-16 12:09 ` Deepak S 2014-12-15 15:12 ` Daniel Vetter
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox