Skip to contentSkip to conversation

Minor improvements for Intel pinctrl

PATCH Superseded Raag Jadav 3 participants 14 messages Jun 13, 2023 – Jun 19, 2023

Subsystems (3)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Raag Jadav

author
#0 lore raw

v3 cover Minor improvements for Intel pinctrl

This series implements minor improvements for Intel pinctrl driver.

The optimizations are as tested with gcc 7.5.0 with default -O2.

Changes since v2:
- Drop redundant patches.
- Update commit message.

Changes since v1:
- Update commit message.
- Specify compiler options used to measure optimizations.
- Drop redundant comments.

Raag Jadav (3):
pinctrl: intel: refine ->set_mux() hook
pinctrl: intel: refine ->irq_set_type() hook
pinctrl: intel: simplify exit path of ->gpio_request_enable() hook

 drivers/pinctrl/intel/pinctrl-intel.c | 48 +++++++++++++--------------
 1 file changed, 24 insertions(+), 24 deletions(-)
Signature · 1 line
2.17.1
author
#1 lore raw

PATCH v3 1/3 pinctrl: intel: refine ->set_mux() hook

Utilize a temporary variable for common shift operation in ->set_mux() hook and improve readability while saving a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-3 (-3)

Function                                     old     new   delta
intel_pinmux_set_mux                         245     242      -3

Total: Before=10472, After=10469, chg -0.03%

---
 drivers/pinctrl/intel/pinctrl-intel.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/pinctrl/intel/pinctrl-intel.c +5 -4
index c7a71c49df0a..e8adf2580321 100644
@@ -411,18 +411,19 @@ static int intel_pinmux_set_mux(struct pinctrl_dev *pctldev,
 	/* Now enable the mux setting for each pin in the group */
 	for (i = 0; i < grp->grp.npins; i++) {
 		void __iomem *padcfg0;
-		u32 value;
+		u32 value, pmode;
 
 		padcfg0 = intel_get_padcfg(pctrl, grp->grp.pins[i], PADCFG0);
-		value = readl(padcfg0);
 
+		value = readl(padcfg0);
 		value &= ~PADCFG0_PMODE_MASK;
 
 		if (grp->modes)
-			value |= grp->modes[i] << PADCFG0_PMODE_SHIFT;
+			pmode = grp->modes[i];
 		else
-			value |= grp->mode << PADCFG0_PMODE_SHIFT;
+			pmode = grp->mode;
 
+		value |= pmode << PADCFG0_PMODE_SHIFT;
 		writel(value, padcfg0);
 	}
 
Signature · 1 line
2.17.1
#2 lore raw

Re: [PATCH v3 1/3] pinctrl: intel: refine ->set_mux() hook

Quoted 7 lines · Raag Jadav

Utilize a temporary variable for common shift operation in ->set_mux() hook and improve readability while saving a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-3 (-3)

Function                                     old     new   delta
intel_pinmux_set_mux                         245     242      -3

Total: Before=10472, After=10469, chg -0.03%

Pushed to my review and testing queue, thanks!

Signature · 2 lines
With Best Regards,
Andy Shevchenko
author
#3 lore raw

PATCH v3 2/3 pinctrl: intel: refine ->irq_set_type() hook

In reply to Raag Jadav

Utilize a temporary variable for common shift operation in ->irq_set_type() hook and improve readability. While at it, simplify if-else-if chain and save a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-16 (-16)

Function                                     old     new   delta
intel_gpio_irq_type                          317     301     -16

Total: Before=10469, After=10453, chg -0.15%

---
 drivers/pinctrl/intel/pinctrl-intel.c | 19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)
drivers/pinctrl/intel/pinctrl-intel.c +10 -9
index e8adf2580321..3f78066b1837 100644
@@ -1128,8 +1128,8 @@ static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)
 	struct intel_pinctrl *pctrl = gpiochip_get_data(gc);
 	unsigned int pin = intel_gpio_to_pin(pctrl, irqd_to_hwirq(d), NULL, NULL);
 	unsigned long flags;
+	u32 value, rxevcfg;
 	void __iomem *reg;
-	u32 value;
 
 	reg = intel_get_padcfg(pctrl, pin, PADCFG0);
 	if (!reg)
@@ -1150,23 +1150,24 @@ static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)
 	intel_gpio_set_gpio_mode(reg);
 
 	value = readl(reg);
-
 	value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
 
 	if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
-		value |= PADCFG0_RXEVCFG_EDGE_BOTH << PADCFG0_RXEVCFG_SHIFT;
+		rxevcfg = PADCFG0_RXEVCFG_EDGE_BOTH;
 	} else if (type & IRQ_TYPE_EDGE_FALLING) {
-		value |= PADCFG0_RXEVCFG_EDGE << PADCFG0_RXEVCFG_SHIFT;
-		value |= PADCFG0_RXINV;
+		rxevcfg = PADCFG0_RXEVCFG_EDGE;
 	} else if (type & IRQ_TYPE_EDGE_RISING) {
-		value |= PADCFG0_RXEVCFG_EDGE << PADCFG0_RXEVCFG_SHIFT;
+		rxevcfg = PADCFG0_RXEVCFG_EDGE;
 	} else if (type & IRQ_TYPE_LEVEL_MASK) {
-		if (type & IRQ_TYPE_LEVEL_LOW)
-			value |= PADCFG0_RXINV;
+		rxevcfg = PADCFG0_RXEVCFG_LEVEL;
 	} else {
-		value |= PADCFG0_RXEVCFG_DISABLED << PADCFG0_RXEVCFG_SHIFT;
+		rxevcfg = PADCFG0_RXEVCFG_DISABLED;
 	}
 
+	if (type == IRQ_TYPE_EDGE_FALLING || type == IRQ_TYPE_LEVEL_LOW)
+		value |= PADCFG0_RXINV;
+
+	value |= rxevcfg << PADCFG0_RXEVCFG_SHIFT;
 	writel(value, reg);
 
 	if (type & IRQ_TYPE_EDGE_BOTH)
Signature · 1 line
2.17.1
#4 lore raw

Re: [PATCH v3 2/3] pinctrl: intel: refine ->irq_set_type() hook

Quoted 8 lines · Raag Jadav

Utilize a temporary variable for common shift operation in ->irq_set_type() hook and improve readability. While at it, simplify if-else-if chain and save a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-16 (-16)

Function                                     old     new   delta
intel_gpio_irq_type                          317     301     -16

Total: Before=10469, After=10453, chg -0.15%

...

Quoted 18 lines
drivers/pinctrl/intel/pinctrl-intel.c +4 -7
 	value = readl(reg);
-
 	value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
 
 	if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
-		value |= PADCFG0_RXEVCFG_EDGE_BOTH << PADCFG0_RXEVCFG_SHIFT;
+		rxevcfg = PADCFG0_RXEVCFG_EDGE_BOTH;
 	} else if (type & IRQ_TYPE_EDGE_FALLING) {
-		value |= PADCFG0_RXEVCFG_EDGE << PADCFG0_RXEVCFG_SHIFT;
-		value |= PADCFG0_RXINV;
+		rxevcfg = PADCFG0_RXEVCFG_EDGE;
 	} else if (type & IRQ_TYPE_EDGE_RISING) {
-		value |= PADCFG0_RXEVCFG_EDGE << PADCFG0_RXEVCFG_SHIFT;
+		rxevcfg = PADCFG0_RXEVCFG_EDGE;
 	} else if (type & IRQ_TYPE_LEVEL_MASK) {
-		if (type & IRQ_TYPE_LEVEL_LOW)
-			value |= PADCFG0_RXINV;
+		rxevcfg = PADCFG0_RXEVCFG_LEVEL;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type) · line 1171

Quoted 10 lines
drivers/pinctrl/intel/pinctrl-intel.c +5 -1
 	} else {
-		value |= PADCFG0_RXEVCFG_DISABLED << PADCFG0_RXEVCFG_SHIFT;
+		rxevcfg = PADCFG0_RXEVCFG_DISABLED;
 	}
 
+	if (type == IRQ_TYPE_EDGE_FALLING || type == IRQ_TYPE_LEVEL_LOW)
+		value |= PADCFG0_RXINV;
+
+	value |= rxevcfg << PADCFG0_RXEVCFG_SHIFT;
 	writel(value, reg);

Looking at this I realized that entire temporary variable assignments can be done outside of spin lock. You probably would need another one for keeping rxinv value.

Will it give us any memory reduction in comparison to the current code?

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#5 lore raw

RE: [PATCH v3 2/3] pinctrl: intel: refine ->irq_set_type() hook

Quoted 41 lines · Raag Jadav
Quoted 8 lines

Utilize a temporary variable for common shift operation in ->irq_set_type() hook and improve readability. While at it, simplify if-else-if chain and save a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-16 (-16)

Function                                     old     new   delta
intel_gpio_irq_type                          317     301     -16

Total: Before=10469, After=10453, chg -0.15%

...

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 6 lines
drivers/pinctrl/intel/pinctrl-intel.c +0 -2
 	value = readl(reg);
-
 	value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
 
 	if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
-		value |= PADCFG0_RXEVCFG_EDGE_BOTH <<

PADCFG0_RXEVCFG_SHIFT;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 3 lines
drivers/pinctrl/intel/pinctrl-intel.c +1 -1
+		rxevcfg = PADCFG0_RXEVCFG_EDGE_BOTH;
 	} else if (type & IRQ_TYPE_EDGE_FALLING) {
-		value |= PADCFG0_RXEVCFG_EDGE <<

PADCFG0_RXEVCFG_SHIFT;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 4 lines
drivers/pinctrl/intel/pinctrl-intel.c +1 -2
-		value |= PADCFG0_RXINV;
+		rxevcfg = PADCFG0_RXEVCFG_EDGE;
 	} else if (type & IRQ_TYPE_EDGE_RISING) {
-		value |= PADCFG0_RXEVCFG_EDGE <<

PADCFG0_RXEVCFG_SHIFT;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 7 lines
drivers/pinctrl/intel/pinctrl-intel.c +2 -3
+		rxevcfg = PADCFG0_RXEVCFG_EDGE;
 	} else if (type & IRQ_TYPE_LEVEL_MASK) {
-		if (type & IRQ_TYPE_LEVEL_LOW)
-			value |= PADCFG0_RXINV;
+		rxevcfg = PADCFG0_RXEVCFG_LEVEL;
 	} else {
-		value |= PADCFG0_RXEVCFG_DISABLED <<

PADCFG0_RXEVCFG_SHIFT;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 4 lines
drivers/pinctrl/intel/pinctrl-intel.c +2 -0
+		rxevcfg = PADCFG0_RXEVCFG_DISABLED;
 	}
 
+	if (type == IRQ_TYPE_EDGE_FALLING || type ==

IRQ_TYPE_LEVEL_LOW)

Quoted 8 lines

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type) · line 1171

Quoted 4 lines
drivers/pinctrl/intel/pinctrl-intel.c +3 -0
+		value |= PADCFG0_RXINV;
+
+	value |= rxevcfg << PADCFG0_RXEVCFG_SHIFT;
 	writel(value, reg);

Looking at this I realized that entire temporary variable assignments can be done outside of spin lock. You probably would need another one for keeping rxinv value.

Something like this?

        u32 value, rxevcfg;
        u32 rxinv = 0;

        if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE_BOTH;
        } else if (type & IRQ_TYPE_EDGE_FALLING) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE;
        } else if (type & IRQ_TYPE_EDGE_RISING) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE;
        } else if (type & IRQ_TYPE_LEVEL_MASK) {
                rxevcfg = PADCFG0_RXEVCFG_LEVEL;
        } else {
                rxevcfg = PADCFG0_RXEVCFG_DISABLED;
        }

        if (type == IRQ_TYPE_EDGE_FALLING || type == IRQ_TYPE_LEVEL_LOW)
                rxinv = PADCFG0_RXINV;

        raw_spin_lock_irqsave(&pctrl->lock, flags);

        intel_gpio_set_gpio_mode(reg);

        value = readl(reg);

        value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
        value |= rxinv;
        value |= rxevcfg << PADCFG0_RXEVCFG_SHIFT;

        writel(value, reg);
Quoted 1 line

Will it give us any memory reduction in comparison to the current code?

add/remove: 0/0 grow/shrink: 1/0 up/down: 4/0 (4)

Function                                     old     new   delta
intel_gpio_irq_type                          317     321      +4

Total: Before=10469, After=10473, chg +0.04%

Unfortunately gcc doesn't seem to consider this as best of the sequence, and I'm not entirely sure why.

#6 lore raw

Re: [PATCH v3 2/3] pinctrl: intel: refine ->irq_set_type() hook

Quoted 74 lines · Raag
Quoted 41 lines · Raag Jadav
Quoted 8 lines

Utilize a temporary variable for common shift operation in ->irq_set_type() hook and improve readability. While at it, simplify if-else-if chain and save a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-16 (-16)

Function                                     old     new   delta
intel_gpio_irq_type                          317     301     -16

Total: Before=10469, After=10453, chg -0.15%

...

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 6 lines
drivers/pinctrl/intel/pinctrl-intel.c +0 -2
 	value = readl(reg);
-
 	value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
 
 	if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
-		value |= PADCFG0_RXEVCFG_EDGE_BOTH <<

PADCFG0_RXEVCFG_SHIFT;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 3 lines
drivers/pinctrl/intel/pinctrl-intel.c +1 -1
+		rxevcfg = PADCFG0_RXEVCFG_EDGE_BOTH;
 	} else if (type & IRQ_TYPE_EDGE_FALLING) {
-		value |= PADCFG0_RXEVCFG_EDGE <<

PADCFG0_RXEVCFG_SHIFT;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 4 lines
drivers/pinctrl/intel/pinctrl-intel.c +1 -2
-		value |= PADCFG0_RXINV;
+		rxevcfg = PADCFG0_RXEVCFG_EDGE;
 	} else if (type & IRQ_TYPE_EDGE_RISING) {
-		value |= PADCFG0_RXEVCFG_EDGE <<

PADCFG0_RXEVCFG_SHIFT;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 7 lines
drivers/pinctrl/intel/pinctrl-intel.c +2 -3
+		rxevcfg = PADCFG0_RXEVCFG_EDGE;
 	} else if (type & IRQ_TYPE_LEVEL_MASK) {
-		if (type & IRQ_TYPE_LEVEL_LOW)
-			value |= PADCFG0_RXINV;
+		rxevcfg = PADCFG0_RXEVCFG_LEVEL;
 	} else {
-		value |= PADCFG0_RXEVCFG_DISABLED <<

PADCFG0_RXEVCFG_SHIFT;

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type)

Quoted 4 lines
drivers/pinctrl/intel/pinctrl-intel.c +2 -0
+		rxevcfg = PADCFG0_RXEVCFG_DISABLED;
 	}
 
+	if (type == IRQ_TYPE_EDGE_FALLING || type ==

IRQ_TYPE_LEVEL_LOW)

Quoted 8 lines

drivers/pinctrl/intel/pinctrl-intel.c static int intel_gpio_irq_type(struct irq_data *d, unsigned int type) · line 1171

Quoted 4 lines
drivers/pinctrl/intel/pinctrl-intel.c +3 -0
+		value |= PADCFG0_RXINV;
+
+	value |= rxevcfg << PADCFG0_RXEVCFG_SHIFT;
 	writel(value, reg);

Looking at this I realized that entire temporary variable assignments can be done outside of spin lock. You probably would need another one for keeping rxinv value.

Something like this?

        u32 value, rxevcfg;
        u32 rxinv = 0;

        if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE_BOTH;
        } else if (type & IRQ_TYPE_EDGE_FALLING) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE;
        } else if (type & IRQ_TYPE_EDGE_RISING) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE;
        } else if (type & IRQ_TYPE_LEVEL_MASK) {
                rxevcfg = PADCFG0_RXEVCFG_LEVEL;
        } else {
                rxevcfg = PADCFG0_RXEVCFG_DISABLED;
        }

        if (type == IRQ_TYPE_EDGE_FALLING || type == IRQ_TYPE_LEVEL_LOW)
                rxinv = PADCFG0_RXINV;

        raw_spin_lock_irqsave(&pctrl->lock, flags);

        intel_gpio_set_gpio_mode(reg);
Quoted 7 lines
        value = readl(reg);

        value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
        value |= rxinv;
        value |= rxevcfg << PADCFG0_RXEVCFG_SHIFT;

        writel(value, reg);

This one looks better.

Quoted 9 lines
Quoted 1 line

Will it give us any memory reduction in comparison to the current code?

add/remove: 0/0 grow/shrink: 1/0 up/down: 4/0 (4)

Function                                     old     new   delta
intel_gpio_irq_type                          317     321      +4

Total: Before=10469, After=10473, chg +0.04%

Unfortunately gcc doesn't seem to consider this as best of the sequence, and I'm not entirely sure why.

It's fine as is, readability counts more than few bytes here.

#7 lore raw

Re: [PATCH v3 2/3] pinctrl: intel: refine ->irq_set_type() hook

Quoted 2 lines · On Thu, Jun 15, 2023 at 12:55:17PM +0300, [email protected] wrote:
Quoted 1 line · Raag
Quoted 1 line · Raag Jadav

...

Quoted 5 lines
Quoted 5 lines
Quoted 3 lines

Looking at this I realized that entire temporary variable assignments can be done outside of spin lock. You probably would need another one for keeping rxinv value.

Something like this?

Almost, see below.

Quoted 2 lines
Quoted 2 lines
        u32 value, rxevcfg;
        u32 rxinv = 0;

No assignment here.

         u32 rxinv, rxevcfg;
         u32 value;
Quoted 11 lines
Quoted 11 lines
        if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE_BOTH;
        } else if (type & IRQ_TYPE_EDGE_FALLING) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE;
        } else if (type & IRQ_TYPE_EDGE_RISING) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE;
        } else if (type & IRQ_TYPE_LEVEL_MASK) {
                rxevcfg = PADCFG0_RXEVCFG_LEVEL;
        } else {
                rxevcfg = PADCFG0_RXEVCFG_DISABLED;
        }

Now, if it's fully included in the diff (even with --patience parameter), then you may drop {}.

Quoted 2 lines
Quoted 2 lines
        if (type == IRQ_TYPE_EDGE_FALLING || type == IRQ_TYPE_LEVEL_LOW)
                rxinv = PADCFG0_RXINV;
		else
			rxinv = 0;
Quoted 9 lines
Quoted 9 lines
        raw_spin_lock_irqsave(&pctrl->lock, flags);

        intel_gpio_set_gpio_mode(reg);

        value = readl(reg);

        value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
        value |= rxinv;
        value |= rxevcfg << PADCFG0_RXEVCFG_SHIFT;

And I would rewrite these to the standard patterns:

         value = (value & ~PADCFG0_RXINV) | rxinv;
         value = (value & ~PADCFG0_RXEVCFG_MASK) | (rxevcfg << PADCFG0_RXEVCFG_SHIFT);

And looking at this, perhaps do shift also outside the lock:

         } else {
                 rxevcfg = PADCFG0_RXEVCFG_DISABLED;
         }
         rxevcfg <<= PADCFG0_RXEVCFG_SHIFT;

But, taking into account scope of the _RXEVCFG_*, I would add shift directly to the definitions and kill that SHIFT entirely:

#define PADCFG0_RXEVCFG_LEVEL           (0 << 25)
#define PADCFG0_RXEVCFG_EDGE            (1 << 25)
#define PADCFG0_RXEVCFG_DISABLED        (2 << 25)
#define PADCFG0_RXEVCFG_EDGE_BOTH       (3 << 25)
	 ...

         value = (value & ~PADCFG0_RXINV) | rxinv;
         value = (value & ~PADCFG0_RXEVCFG_MASK) | rxevcfg;

Try that one and look if it looks better. It might even save bytes after all.

Quoted 15 lines
Quoted 1 line
        writel(value, reg);

This one looks better.

Quoted 9 lines
Quoted 1 line

Will it give us any memory reduction in comparison to the current code?

add/remove: 0/0 grow/shrink: 1/0 up/down: 4/0 (4)

Function                                     old     new   delta
intel_gpio_irq_type                          317     321      +4

Total: Before=10469, After=10473, chg +0.04%

Unfortunately gcc doesn't seem to consider this as best of the sequence, and I'm not entirely sure why.

It's fine as is, readability counts more than few bytes here.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#8 lore raw

RE: [PATCH v3 2/3] pinctrl: intel: refine ->irq_set_type() hook

Quoted 76 lines
Quoted 2 lines · On Thu, Jun 15, 2023 at 12:55:17PM +0300, [email protected] wrote:
Quoted 1 line · Raag
Quoted 1 line · Raag Jadav

...

Quoted 5 lines
Quoted 5 lines
Quoted 3 lines

Looking at this I realized that entire temporary variable assignments can be done outside of spin lock. You probably would need another one for keeping rxinv value.

Something like this?

Almost, see below.

Quoted 2 lines
Quoted 2 lines
        u32 value, rxevcfg;
        u32 rxinv = 0;

No assignment here.

         u32 rxinv, rxevcfg;
         u32 value;
Quoted 11 lines
Quoted 11 lines
        if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE_BOTH;
        } else if (type & IRQ_TYPE_EDGE_FALLING) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE;
        } else if (type & IRQ_TYPE_EDGE_RISING) {
                rxevcfg = PADCFG0_RXEVCFG_EDGE;
        } else if (type & IRQ_TYPE_LEVEL_MASK) {
                rxevcfg = PADCFG0_RXEVCFG_LEVEL;
        } else {
                rxevcfg = PADCFG0_RXEVCFG_DISABLED;
        }

Now, if it's fully included in the diff (even with --patience parameter), then you may drop {}.

Quoted 1 line
Quoted 1 line
        if (type == IRQ_TYPE_EDGE_FALLING || type ==

IRQ_TYPE_LEVEL_LOW)

Quoted 1 line
Quoted 1 line
                rxinv = PADCFG0_RXINV;
		else
			rxinv = 0;
Quoted 9 lines
Quoted 9 lines
        raw_spin_lock_irqsave(&pctrl->lock, flags);

        intel_gpio_set_gpio_mode(reg);

        value = readl(reg);

        value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
        value |= rxinv;
        value |= rxevcfg << PADCFG0_RXEVCFG_SHIFT;

And I would rewrite these to the standard patterns:

value = (value & ~PADCFG0_RXINV) | rxinv; value = (value & ~PADCFG0_RXEVCFG_MASK) | (rxevcfg <<
PADCFG0_RXEVCFG_SHIFT);

And looking at this, perhaps do shift also outside the lock:

         } else {
                 rxevcfg = PADCFG0_RXEVCFG_DISABLED;
         }
         rxevcfg <<= PADCFG0_RXEVCFG_SHIFT;

But, taking into account scope of the _RXEVCFG_*, I would add shift directly to the definitions and kill that SHIFT entirely:

#define PADCFG0_RXEVCFG_LEVEL           (0 << 25)
#define PADCFG0_RXEVCFG_EDGE            (1 << 25)
#define PADCFG0_RXEVCFG_DISABLED        (2 << 25)
#define PADCFG0_RXEVCFG_EDGE_BOTH       (3 << 25)
Quoted 6 lines
	 ...

         value = (value & ~PADCFG0_RXINV) | rxinv;
         value = (value & ~PADCFG0_RXEVCFG_MASK) | rxevcfg;

Try that one and look if it looks better. It might even save bytes after all.

Should I add all of this in original patch or send this as a separate patch on top this series?

#9 lore raw

Re: [PATCH v3 2/3] pinctrl: intel: refine ->irq_set_type() hook

On Thu, Jun 15, 2023 at 11:08:38AM +0000, Jadav, Raag wrote:

...

Quoted 2 lines

Should I add all of this in original patch or send this as a separate patch on top this series?

Always base the changes on the respective subsystem tree, don't forget to use --base when formatting patch with Git tools. Then send it separately. The 3rd patch in the series is questionable to me. I would like to look into it later on separately.

(The first implies that there is no changes as per this series in that
function).

Signature · 2 lines
With Best Regards,
Andy Shevchenko
author
#10 lore raw

PATCH v3 3/3 pinctrl: intel: simplify exit path of ->gpio_request_enable() hook

In reply to Raag Jadav

Simplify exit path of ->gpio_request_enable() hook and save a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-36 (-36)

Function                                     old     new   delta
intel_gpio_request_enable                    186     150     -36

Total: Before=10453, After=10417, chg -0.34%

---
 drivers/pinctrl/intel/pinctrl-intel.c | 20 +++++++++-----------
 1 file changed, 9 insertions(+), 11 deletions(-)
drivers/pinctrl/intel/pinctrl-intel.c +9 -11
index 3f78066b1837..43e17bc177d0 100644
@@ -485,20 +485,19 @@ static int intel_gpio_request_enable(struct pinctrl_dev *pctldev,
 	struct intel_pinctrl *pctrl = pinctrl_dev_get_drvdata(pctldev);
 	void __iomem *padcfg0;
 	unsigned long flags;
+	int ret = 0;
 
 	padcfg0 = intel_get_padcfg(pctrl, pin, PADCFG0);
 
 	raw_spin_lock_irqsave(&pctrl->lock, flags);
 
 	if (!intel_pad_owned_by_host(pctrl, pin)) {
-		raw_spin_unlock_irqrestore(&pctrl->lock, flags);
-		return -EBUSY;
+		ret = -EBUSY;
+		goto out_unlock;
 	}
 
-	if (!intel_pad_is_unlocked(pctrl, pin)) {
-		raw_spin_unlock_irqrestore(&pctrl->lock, flags);
-		return 0;
-	}
+	if (!intel_pad_is_unlocked(pctrl, pin))
+		goto out_unlock;
 
 	/*
 	 * If pin is already configured in GPIO mode, we assume that
@@ -506,16 +505,15 @@ static int intel_gpio_request_enable(struct pinctrl_dev *pctldev,
 	 * potential glitches on the pin. Otherwise, for the pin in
 	 * alternative mode, consumer has to supply respective flags.
 	 */
-	if (intel_gpio_get_gpio_mode(padcfg0) == PADCFG0_PMODE_GPIO) {
-		raw_spin_unlock_irqrestore(&pctrl->lock, flags);
-		return 0;
-	}
+	if (intel_gpio_get_gpio_mode(padcfg0) == PADCFG0_PMODE_GPIO)
+		goto out_unlock;
 
 	intel_gpio_set_gpio_mode(padcfg0);
 
+out_unlock:
 	raw_spin_unlock_irqrestore(&pctrl->lock, flags);
 
-	return 0;
+	return ret;
 }
 
 static int intel_gpio_set_direction(struct pinctrl_dev *pctldev,
Signature · 1 line
2.17.1
#11 lore raw

Re: [PATCH v3 3/3] pinctrl: intel: simplify exit path of ->gpio_request_enable() hook

Quoted 6 lines · Raag Jadav

Simplify exit path of ->gpio_request_enable() hook and save a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-36 (-36)

Function                                     old     new   delta
intel_gpio_request_enable                    186     150     -36

Total: Before=10453, After=10417, chg -0.34%

I believe Mika's comment against patch 3 of v2 of this series applies here as well, meaning this patch won't be accepted.

But let Mika express himself.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#12 lore raw

Re: [PATCH v3 3/3] pinctrl: intel: simplify exit path of ->gpio_request_enable() hook

In reply to Raag Jadav
Quoted 6 lines · Raag Jadav

Simplify exit path of ->gpio_request_enable() hook and save a few bytes.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-36 (-36)

Function                                     old     new   delta
intel_gpio_request_enable                    186     150     -36

Total: Before=10453, After=10417, chg -0.34%

While it seems a good footprint improvement, it looks like we do not use goto in entire drivers/pinctrl/intel. So, I would keep it that way.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
maintainer Acked-by on the v3 series
#13 lore raw
In reply to Raag Jadav
Quoted 13 lines · Raag Jadav

This series implements minor improvements for Intel pinctrl driver.

The optimizations are as tested with gcc 7.5.0 with default -O2.

Changes since v2:
- Drop redundant patches.
- Update commit message.

Changes since v1:
- Update commit message.
- Specify compiler options used to measure optimizations.
- Drop redundant comments.

Quoted 4 lines

Raag Jadav (3):
pinctrl: intel: refine ->set_mux() hook
pinctrl: intel: refine ->irq_set_type() hook
pinctrl: intel: simplify exit path of ->gpio_request_enable() hook

For the series,