Skip to contentSkip to conversation

clk: qcom: gdsc: Use the default transition delay for GDSCs

PATCH In review Taniya Das 4 participants 13 messages Feb 9, 2022 – Feb 22, 2022

Subsystems (3)
lore
Share & tools
lore

Taniya Das

author
#0 lore raw

PATCH 1/2 clk: qcom: gdsc: Use the default transition delay for GDSCs

Do not update the transition delay and use the default reset values.

Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
 drivers/clk/qcom/gdsc.c | 6 +++++-
 drivers/clk/qcom/gdsc.h | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +5 -1
index 7e1dd8ccfa38..e7b213450640 100644
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
 	 */
 	mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
 	       EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-	val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+	regmap_read(sc->regmap, sc->gdscr, &val);
+
+	if (!(sc->flags & DEFAULT_TRANSITION_DELAY))
+		val |= EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
 	ret = regmap_update_bits(sc->regmap, sc->gdscr, mask, val);
 	if (ret)
 		return ret;
drivers/clk/qcom/gdsc.h +1 -0
index d7cc4c21a9d4..1bd3ecdd0b0a 100644
@@ -53,6 +53,7 @@ struct gdsc {
 #define ALWAYS_ON	BIT(6)
 #define RETAIN_FF_ENABLE	BIT(7)
 #define NO_RET_PERIPH	BIT(8)
+#define DEFAULT_TRANSITION_DELAY	BIT(9)
 	struct reset_controller_dev	*rcdev;
 	unsigned int			*resets;
 	unsigned int			reset_count;

--
Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member of the Code Aurora Forum, hosted by the Linux Foundation.

author
#1 lore raw

PATCH 2/2 clk: qcom: dispcc: Update gdsc flag for display GDSC

Update the mdss gdsc flag for SC7180/SC7280/SM8150/SM8250 to not program the transition delay.

Fixes dd3d06622138 ("clk: qcom: Add display clock controller driver for SC7180")
Fixes 1a00c962f9cd ("clk: qcom: Add display clock controller driver for SC7280")
Fixes 80a18f4a8567 ("clk: qcom: Add display clock controller driver for SM8150 and SM8250")
---
 drivers/clk/qcom/dispcc-sc7180.c | 2 +-
 drivers/clk/qcom/dispcc-sc7280.c | 2 +-
 drivers/clk/qcom/dispcc-sm8250.c | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
drivers/clk/qcom/dispcc-sc7180.c +1 -1
index 538e4963c915..c203888ab2ca 100644
@@ -629,7 +629,7 @@ static struct gdsc mdss_gdsc = {
 		.name = "mdss_gdsc",
 	},
 	.pwrsts = PWRSTS_OFF_ON,
-	.flags = HW_CTRL,
+	.flags = HW_CTRL |  DEFAULT_TRANSITION_DELAY,
 };
 
 static struct gdsc *disp_cc_sc7180_gdscs[] = {
drivers/clk/qcom/dispcc-sc7280.c +1 -1
index 4ef4ae231794..8e1a2d4a8120 100644
@@ -791,7 +791,7 @@ static struct gdsc disp_cc_mdss_core_gdsc = {
 		.name = "disp_cc_mdss_core_gdsc",
 	},
 	.pwrsts = PWRSTS_OFF_ON,
-	.flags = HW_CTRL | RETAIN_FF_ENABLE,
+	.flags = HW_CTRL | RETAIN_FF_ENABLE | DEFAULT_TRANSITION_DELAY,
 };
 
 static struct clk_regmap *disp_cc_sc7280_clocks[] = {
drivers/clk/qcom/dispcc-sm8250.c +1 -1
index 566fdfa0a15b..4d64d499a285 100644
@@ -1130,7 +1130,7 @@ static struct gdsc mdss_gdsc = {
 		.name = "mdss_gdsc",
 	},
 	.pwrsts = PWRSTS_OFF_ON,
-	.flags = HW_CTRL,
+	.flags = HW_CTRL | DEFAULT_TRANSITION_DELAY,
 	.supply = "mmcx",
 };
 

--
Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member of the Code Aurora Forum, hosted by the Linux Foundation.

#2 lore raw

Re: [v1 2/2] clk: qcom: dispcc: Update gdsc flag for display GDSC

Hi Taniya,

Quoted 18 lines · Taniya Das

Update the mdss gdsc flag for SC7180/SC7280/SM8150/SM8250 to not program the transition delay.

Fixes dd3d06622138 ("clk: qcom: Add display clock controller driver for SC7180")
Fixes 1a00c962f9cd ("clk: qcom: Add display clock controller driver for SC7280")
Fixes 80a18f4a8567 ("clk: qcom: Add display clock controller driver for SM8150 and SM8250")
---
  drivers/clk/qcom/dispcc-sc7180.c | 2 +-
  drivers/clk/qcom/dispcc-sc7280.c | 2 +-
  drivers/clk/qcom/dispcc-sm8250.c | 2 +-
  3 files changed, 3 insertions(+), 3 deletions(-)
drivers/clk/qcom/dispcc-sc7180.c +0 -0
index 538e4963c915..c203888ab2ca 100644

drivers/clk/qcom/dispcc-sc7180.c static struct gdsc mdss_gdsc = { · line 633

Quoted 6 lines
drivers/clk/qcom/dispcc-sc7180.c +1 -1
@@ -629,7 +629,7 @@ static struct gdsc mdss_gdsc = {
  		.name = "mdss_gdsc",
  	},
  	.pwrsts = PWRSTS_OFF_ON,
-	.flags = HW_CTRL,
+	.flags = HW_CTRL |  DEFAULT_TRANSITION_DELAY,
  };

There's an extra space after the |

Quoted 28 lines
drivers/clk/qcom/dispcc-sc7180.c +0 -0
  static struct gdsc *disp_cc_sc7180_gdscs[] = {
drivers/clk/qcom/dispcc-sc7280.c +1 -1
index 4ef4ae231794..8e1a2d4a8120 100644
@@ -791,7 +791,7 @@ static struct gdsc disp_cc_mdss_core_gdsc = {
  		.name = "disp_cc_mdss_core_gdsc",
  	},
  	.pwrsts = PWRSTS_OFF_ON,
-	.flags = HW_CTRL | RETAIN_FF_ENABLE,
+	.flags = HW_CTRL | RETAIN_FF_ENABLE | DEFAULT_TRANSITION_DELAY,
  };
 
  static struct clk_regmap *disp_cc_sc7280_clocks[] = {
drivers/clk/qcom/dispcc-sm8250.c +1 -2
index 566fdfa0a15b..4d64d499a285 100644
@@ -1130,7 +1130,7 @@ static struct gdsc mdss_gdsc = {
  		.name = "mdss_gdsc",
  	},
  	.pwrsts = PWRSTS_OFF_ON,
-	.flags = HW_CTRL,
+	.flags = HW_CTRL | DEFAULT_TRANSITION_DELAY,
  	.supply = "mmcx",
  };
 
--
Quoted 3 lines

Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member of the Code Aurora Forum, hosted by the Linux Foundation.

-- steev

#3 lore raw
In reply to Taniya Das
Quoted 2 lines · Taniya Das

Do not update the transition delay and use the default reset values.

Can you help me understand what "the transition delay" is? In particular I would like to know which other platforms, past, present and future should have this flag set - or not set.

Quoted 12 lines
Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
 drivers/clk/qcom/gdsc.c | 6 +++++-
 drivers/clk/qcom/gdsc.h | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +0 -0
index 7e1dd8ccfa38..e7b213450640 100644

drivers/clk/qcom/gdsc.c static int gdsc_init(struct gdsc *sc) · line 386

Quoted 8 lines
drivers/clk/qcom/gdsc.c +4 -1
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
 	 */
 	mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
 	       EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-	val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+	regmap_read(sc->regmap, sc->gdscr, &val);
+
+	if (!(sc->flags & DEFAULT_TRANSITION_DELAY))

Why is it a good idea to not have a transition delay other than the default one on all the other gdscs?

Regards,
Bjorn

Quoted 20 lines
drivers/clk/qcom/gdsc.c +1 -0
+		val |= EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
 	ret = regmap_update_bits(sc->regmap, sc->gdscr, mask, val);
 	if (ret)
 		return ret;
drivers/clk/qcom/gdsc.h +1 -1
index d7cc4c21a9d4..1bd3ecdd0b0a 100644
@@ -53,6 +53,7 @@ struct gdsc {
 #define ALWAYS_ON	BIT(6)
 #define RETAIN_FF_ENABLE	BIT(7)
 #define NO_RET_PERIPH	BIT(8)
+#define DEFAULT_TRANSITION_DELAY	BIT(9)
 	struct reset_controller_dev	*rcdev;
 	unsigned int			*resets;
 	unsigned int			reset_count;
--

Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member of the Code Aurora Forum, hosted by the Linux Foundation.

#4 lore raw
In reply to Taniya Das
Quoted 14 lines · Taniya Das

Do not update the transition delay and use the default reset values.

Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
 drivers/clk/qcom/gdsc.c | 6 +++++-
 drivers/clk/qcom/gdsc.h | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +0 -0
index 7e1dd8ccfa38..e7b213450640 100644

drivers/clk/qcom/gdsc.c static int gdsc_init(struct gdsc *sc) · line 386

Quoted 8 lines
drivers/clk/qcom/gdsc.c +4 -1
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
 	 */
 	mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
 	       EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-	val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+	regmap_read(sc->regmap, sc->gdscr, &val);
+
+	if (!(sc->flags & DEFAULT_TRANSITION_DELAY))

I dug a little bit more into this and noticed that on various platforms CLK_DIS_WAIT_VAL for the GPU_CX GDSC is supposed to be 8 (whereas both hw default and CLK_DIS_WAIT_VAL is 2).

I'm not able to find anything helpful in the git log describing what the value does, but it seems that a "just use hw default" flag won't cut it for this scenario.

Regards,
Bjorn

Quoted 20 lines
drivers/clk/qcom/gdsc.c +1 -0
+		val |= EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
 	ret = regmap_update_bits(sc->regmap, sc->gdscr, mask, val);
 	if (ret)
 		return ret;
drivers/clk/qcom/gdsc.h +1 -1
index d7cc4c21a9d4..1bd3ecdd0b0a 100644
@@ -53,6 +53,7 @@ struct gdsc {
 #define ALWAYS_ON	BIT(6)
 #define RETAIN_FF_ENABLE	BIT(7)
 #define NO_RET_PERIPH	BIT(8)
+#define DEFAULT_TRANSITION_DELAY	BIT(9)
 	struct reset_controller_dev	*rcdev;
 	unsigned int			*resets;
 	unsigned int			reset_count;
--

Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member of the Code Aurora Forum, hosted by the Linux Foundation.

maintainer
#5 lore raw

Quoting Bjorn Andersson (2022-02-09 14:35:08)

Quoted 29 lines · Taniya Das
Quoted 15 lines

Do not update the transition delay and use the default reset values.

Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
 drivers/clk/qcom/gdsc.c | 6 +++++-
 drivers/clk/qcom/gdsc.h | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +0 -0
index 7e1dd8ccfa38..e7b213450640 100644

drivers/clk/qcom/gdsc.c static int gdsc_init(struct gdsc *sc) · line 386

Quoted 8 lines
drivers/clk/qcom/gdsc.c +4 -1
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
       */
      mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
             EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-     val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+     regmap_read(sc->regmap, sc->gdscr, &val);
+
+     if (!(sc->flags & DEFAULT_TRANSITION_DELAY))

I dug a little bit more into this and noticed that on various platforms CLK_DIS_WAIT_VAL for the GPU_CX GDSC is supposed to be 8 (whereas both hw default and CLK_DIS_WAIT_VAL is 2).

Quoted 4 lines

I'm not able to find anything helpful in the git log describing what the value does, but it seems that a "just use hw default" flag won't cut it for this scenario.

I'd prefer we invert the logic so that we don't need to litter this flag all over the place. I recall that the wait values were incorrect a long time ago on early gdsc using designs but hopefully they've been fixed now and we can simply use the default power on reset (POR) values.

#6 lore raw
Quoted 35 lines · Stephen Boyd

Quoting Bjorn Andersson (2022-02-09 14:35:08)

Quoted 29 lines · Taniya Das
Quoted 15 lines

Do not update the transition delay and use the default reset values.

Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
 drivers/clk/qcom/gdsc.c | 6 +++++-
 drivers/clk/qcom/gdsc.h | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +0 -0
index 7e1dd8ccfa38..e7b213450640 100644

drivers/clk/qcom/gdsc.c static int gdsc_init(struct gdsc *sc) · line 386

Quoted 8 lines
drivers/clk/qcom/gdsc.c +4 -1
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
       */
      mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
             EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-     val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+     regmap_read(sc->regmap, sc->gdscr, &val);
+
+     if (!(sc->flags & DEFAULT_TRANSITION_DELAY))

I dug a little bit more into this and noticed that on various platforms CLK_DIS_WAIT_VAL for the GPU_CX GDSC is supposed to be 8 (whereas both hw default and CLK_DIS_WAIT_VAL is 2).

Quoted 4 lines

I'm not able to find anything helpful in the git log describing what the value does, but it seems that a "just use hw default" flag won't cut it for this scenario.

Quoted 4 lines

I'd prefer we invert the logic so that we don't need to litter this flag all over the place. I recall that the wait values were incorrect a long time ago on early gdsc using designs but hopefully they've been fixed now and we can simply use the default power on reset (POR) values.

Are you suggesting that we make it the default to not update the values and then provide means to specify it where needed?

I like that suggestion.

But as mentioned in my reply yesterday, GPU_CX on several platforms needs a different CLK_DIS_WAIT_VAL - and not the same value.

Are these values ever 0? Or could we simply add the three numbers to struct gdsc and have 0 denote "use hw default"?

Regards,
Bjorn

maintainer
#7 lore raw

Quoting Bjorn Andersson (2022-02-10 11:32:03)

Quoted 9 lines · Stephen Boyd
Quoted 5 lines

I'd prefer we invert the logic so that we don't need to litter this flag all over the place. I recall that the wait values were incorrect a long time ago on early gdsc using designs but hopefully they've been fixed now and we can simply use the default power on reset (POR) values.

Are you suggesting that we make it the default to not update the values and then provide means to specify it where needed?

Sure!

Quoted 8 lines

I like that suggestion.

But as mentioned in my reply yesterday, GPU_CX on several platforms needs a different CLK_DIS_WAIT_VAL - and not the same value.

Are these values ever 0? Or could we simply add the three numbers to struct gdsc and have 0 denote "use hw default"?

I don't think they're ever set to zero. That would most likely just break things because they need a few clk cycles to wait between transitioning states. Otherwise the GDSC gets stuck and the clks don't work.

author
#8 lore raw
In reply to Stephen Boyd

Hi Stephen, Bjorn,

Thanks for your comments.

Quoted 30 lines · Stephen Boyd

Quoting Bjorn Andersson (2022-02-09 14:35:08)

Quoted 29 lines · Taniya Das
Quoted 15 lines

Do not update the transition delay and use the default reset values.

Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
  drivers/clk/qcom/gdsc.c | 6 +++++-
  drivers/clk/qcom/gdsc.h | 1 +
  2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +0 -0
index 7e1dd8ccfa38..e7b213450640 100644

drivers/clk/qcom/gdsc.c static int gdsc_init(struct gdsc *sc) · line 386

Quoted 8 lines
drivers/clk/qcom/gdsc.c +4 -1
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
        */
       mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
              EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-     val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+     regmap_read(sc->regmap, sc->gdscr, &val);
+
+     if (!(sc->flags & DEFAULT_TRANSITION_DELAY))

I dug a little bit more into this and noticed that on various platforms CLK_DIS_WAIT_VAL for the GPU_CX GDSC is supposed to be 8 (whereas both hw default and CLK_DIS_WAIT_VAL is 2).

Yes, only for certain GPU_CC these would be updated and that too in case the design team suggests. Downstream we would set the value from probe itself, or we can pick these from device tree as required and suggested.

Quoted 4 lines
Quoted 4 lines

I'm not able to find anything helpful in the git log describing what the value does, but it seems that a "just use hw default" flag won't cut it for this scenario.

This value is used for the number of clock cycles it would wait before the GDSCR ACK signals/halting the clock.

Quoted 5 lines

I'd prefer we invert the logic so that we don't need to litter this flag all over the place. I recall that the wait values were incorrect a long time ago on early gdsc using designs but hopefully they've been fixed now and we can simply use the default power on reset (POR) values.

I am okay to invert the logic, but I am not sure if they could cause any issues to the older targets. They were broken in HW design long back, but got fixed in most families of target and most GDSCs.

As mentioned if explicitly they need to be updated, it is best to do from the probe. This was done in SC7180 GPUCC driver.

        /* Configure clk_dis_wait for gpu_cx_gdsc */
        regmap_update_bits(regmap, 0x106c, CLK_DIS_WAIT_MASK,
                                        8 << CLK_DIS_WAIT_SHIFT);

Please let me know if we are okay to add the invert logic.

Signature · 4 lines
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation.

--
#9 lore raw
Quoted 36 lines · Taniya Das

Hi Stephen, Bjorn,

Thanks for your comments.

Quoted 30 lines · Stephen Boyd

Quoting Bjorn Andersson (2022-02-09 14:35:08)

Quoted 29 lines · Taniya Das
Quoted 15 lines

Do not update the transition delay and use the default reset values.

Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
  drivers/clk/qcom/gdsc.c | 6 +++++-
  drivers/clk/qcom/gdsc.h | 1 +
  2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +0 -0
index 7e1dd8ccfa38..e7b213450640 100644

drivers/clk/qcom/gdsc.c static int gdsc_init(struct gdsc *sc) · line 386

Quoted 8 lines
drivers/clk/qcom/gdsc.c +4 -1
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
        */
       mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
              EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-     val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+     regmap_read(sc->regmap, sc->gdscr, &val);
+
+     if (!(sc->flags & DEFAULT_TRANSITION_DELAY))

I dug a little bit more into this and noticed that on various platforms CLK_DIS_WAIT_VAL for the GPU_CX GDSC is supposed to be 8 (whereas both hw default and CLK_DIS_WAIT_VAL is 2).

Quoted 4 lines

Yes, only for certain GPU_CC these would be updated and that too in case the design team suggests. Downstream we would set the value from probe itself, or we can pick these from device tree as required and suggested.

I don't expect that value to be "configurable", so pushing it to DT doesn't seem like the proper solution.

Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

I'm not able to find anything helpful in the git log describing what the value does, but it seems that a "just use hw default" flag won't cut it for this scenario.

This value is used for the number of clock cycles it would wait before the GDSCR ACK signals/halting the clock.

That makes sense.

Quoted 10 lines
Quoted 5 lines

I'd prefer we invert the logic so that we don't need to litter this flag all over the place. I recall that the wait values were incorrect a long time ago on early gdsc using designs but hopefully they've been fixed now and we can simply use the default power on reset (POR) values.

I am okay to invert the logic, but I am not sure if they could cause any issues to the older targets. They were broken in HW design long back, but got fixed in most families of target and most GDSCs.

I don't fancy us having a flag with the purpose of "don't set the timings to 2, 8 and 2" and then rely on open coded writes in probe to set it to something else where needed.

So I definitely would prefer to flip this around, to make the cases where we want to write different values explicit.

But as you say, unless we make sure that all existing platforms do write 2, 8 and 2 we risk introducing regressions from the current behavior.

Quoted 6 lines

As mentioned if explicitly they need to be updated, it is best to do from the probe.
This was done in SC7180 GPUCC driver.

        /* Configure clk_dis_wait for gpu_cx_gdsc */
        regmap_update_bits(regmap, 0x106c, CLK_DIS_WAIT_MASK,
                                        8 << CLK_DIS_WAIT_SHIFT);

But we call regmap_update_bits() from probe, which sets the CLK_DIS_WAIT to 8, then we call qcom_cc_really_probe() which will end up in gdsc_init() which replaces that with a 2.

Perhaps I'm missing something?

Quoted 4 lines

Please let me know if we are okay to add the invert logic.

I'm still favoring a scheme where we add 3 integers to struct gdsc and in gdsc_init() we check if they are non-zero we write the value to the register.

Although being a big patch, we could maintain the existing behaviour by giving all existing struct gdsc definitions the values 2, 8 and 2 to avoid regressions and we (everyone) can then go through the platforms one by one and remove the unnecessary assignments - but more importantly, double check with downstream if they need a different value.

This will also have the side effect going forward that it will be more explicit and people writing gdsc definitions should double check these values.

Regards,
Bjorn

author
#10 lore raw

Hello Bjorn,

Thanks for your comments.

Quoted 85 lines · Bjorn Andersson
Quoted 36 lines · Taniya Das

Hi Stephen, Bjorn,

Thanks for your comments.

Quoted 30 lines · Stephen Boyd

Quoting Bjorn Andersson (2022-02-09 14:35:08)

Quoted 29 lines · Taniya Das
Quoted 15 lines

Do not update the transition delay and use the default reset values.

Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
   drivers/clk/qcom/gdsc.c | 6 +++++-
   drivers/clk/qcom/gdsc.h | 1 +
   2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +0 -0
index 7e1dd8ccfa38..e7b213450640 100644

drivers/clk/qcom/gdsc.c static int gdsc_init(struct gdsc *sc) · line 386

Quoted 8 lines
drivers/clk/qcom/gdsc.c +4 -1
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
         */
        mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
               EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-     val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+     regmap_read(sc->regmap, sc->gdscr, &val);
+
+     if (!(sc->flags & DEFAULT_TRANSITION_DELAY))

I dug a little bit more into this and noticed that on various platforms CLK_DIS_WAIT_VAL for the GPU_CX GDSC is supposed to be 8 (whereas both hw default and CLK_DIS_WAIT_VAL is 2).

Quoted 4 lines

Yes, only for certain GPU_CC these would be updated and that too in case the design team suggests. Downstream we would set the value from probe itself, or we can pick these from device tree as required and suggested.

I don't expect that value to be "configurable", so pushing it to DT doesn't seem like the proper solution.

Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

I'm not able to find anything helpful in the git log describing what the value does, but it seems that a "just use hw default" flag won't cut it for this scenario.

This value is used for the number of clock cycles it would wait before the GDSCR ACK signals/halting the clock.

That makes sense.

Quoted 10 lines
Quoted 5 lines

I'd prefer we invert the logic so that we don't need to litter this flag all over the place. I recall that the wait values were incorrect a long time ago on early gdsc using designs but hopefully they've been fixed now and we can simply use the default power on reset (POR) values.

I am okay to invert the logic, but I am not sure if they could cause any issues to the older targets. They were broken in HW design long back, but got fixed in most families of target and most GDSCs.

I don't fancy us having a flag with the purpose of "don't set the timings to 2, 8 and 2" and then rely on open coded writes in probe to set it to something else where needed.

So I definitely would prefer to flip this around, to make the cases where we want to write different values explicit.

But as you say, unless we make sure that all existing platforms do write 2, 8 and 2 we risk introducing regressions from the current behavior.

Quoted 6 lines

As mentioned if explicitly they need to be updated, it is best to do from the probe.
This was done in SC7180 GPUCC driver.

         /* Configure clk_dis_wait for gpu_cx_gdsc */
         regmap_update_bits(regmap, 0x106c, CLK_DIS_WAIT_MASK,
                                         8 << CLK_DIS_WAIT_SHIFT);
Quoted 6 lines

But we call regmap_update_bits() from probe, which sets the CLK_DIS_WAIT to 8, then we call qcom_cc_really_probe() which will end up in gdsc_init() which replaces that with a 2.

Perhaps I'm missing something?

It was my miss when I did a cleanup to move the DIS_WAIT_VAL before registering the clocks.

Quoted 9 lines
Quoted 4 lines

Please let me know if we are okay to add the invert logic.

I'm still favoring a scheme where we add 3 integers to struct gdsc and in gdsc_init() we check if they are non-zero we write the value to the register.

Sure, will update the gdsc_init() to default "2, 8, 2" in case of non-zero value.

Quoted 7 lines

Although being a big patch, we could maintain the existing behaviour by giving all existing struct gdsc definitions the values 2, 8 and 2 to avoid regressions and we (everyone) can then go through the platforms one by one and remove the unnecessary assignments - but more importantly, double check with downstream if they need a different value.

Sure, that would help.

Quoted 6 lines

This will also have the side effect going forward that it will be more explicit and people writing gdsc definitions should double check these values.

Regards,
Bjorn

Signature · 4 lines
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation.

--
maintainer
#11 lore raw
In reply to Taniya Das

Quoting Taniya Das (2022-02-09 09:25:12)

Quoted 1 line

Do not update the transition delay and use the default reset values.

Why? I know why because I'm CCed on the internal bug but nobody else knows what's going on here. Please clearly describe what you're fixing, i.e. that you're seeing black screens on sc7180 because the gdsc is being enabled/disabled so rapidly that the hardware is getting borked. Lots more detail please. Otherwise we're going to ignore this patch series as nonsense because we don't have enough information to realize that it is important.

#12 lore raw
In reply to Taniya Das
Quoted 14 lines · Taniya Das

Do not update the transition delay and use the default reset values.

Fixes 45dd0e55317cc ("clk: qcom: Add support for GDSCs)
---
 drivers/clk/qcom/gdsc.c | 6 +++++-
 drivers/clk/qcom/gdsc.h | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)
drivers/clk/qcom/gdsc.c +0 -0
index 7e1dd8ccfa38..e7b213450640 100644

drivers/clk/qcom/gdsc.c static int gdsc_init(struct gdsc *sc) · line 387

Quoted 9 lines
drivers/clk/qcom/gdsc.c +5 -1
@@ -380,7 +380,11 @@ static int gdsc_init(struct gdsc *sc)
 	 */
 	mask = HW_CONTROL_MASK | SW_OVERRIDE_MASK |
 	       EN_REST_WAIT_MASK | EN_FEW_WAIT_MASK | CLK_DIS_WAIT_MASK;
-	val = EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;
+
+	regmap_read(sc->regmap, sc->gdscr, &val);
+
+	if (!(sc->flags & DEFAULT_TRANSITION_DELAY))
+		val |= EN_REST_WAIT_VAL | EN_FEW_WAIT_VAL | CLK_DIS_WAIT_VAL;

Forgot to point out this yesterday, but these are multi-bit values, so you need to mask out whatever bits are already set before filling in the new ones.

Regards,
Bjorn

Quoted 19 lines
drivers/clk/qcom/gdsc.c +0 -0
 	ret = regmap_update_bits(sc->regmap, sc->gdscr, mask, val);
 	if (ret)
 		return ret;
drivers/clk/qcom/gdsc.h +1 -1
index d7cc4c21a9d4..1bd3ecdd0b0a 100644
@@ -53,6 +53,7 @@ struct gdsc {
 #define ALWAYS_ON	BIT(6)
 #define RETAIN_FF_ENABLE	BIT(7)
 #define NO_RET_PERIPH	BIT(8)
+#define DEFAULT_TRANSITION_DELAY	BIT(9)
 	struct reset_controller_dev	*rcdev;
 	unsigned int			*resets;
 	unsigned int			reset_count;
--

Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member of the Code Aurora Forum, hosted by the Linux Foundation.