[PATCH] pinctrl: samsung: use 'int' for register masks in Exynos

[PATCH] pinctrl: samsung: use 'int' for register masks in Exynos

[Krzysztof Kozlowski]

The Special Function Registers on all Exynos SoC, including ARM64, are
32-bit wide, so entire driver uses matching functions like readl() or
writel().  On 64-bit ARM using unsigned long for register masks:
1. makes little sense as immediately after bitwise operation it will be
   cast to 32-bit value when calling writel(),
2. is actually error-prone because it might promote other operands to
   64-bit.

Addresses-Coverity: Unintentional integer overflow
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>

---

Not tested on ARM64.

Dear Linus,

Please apply it directly, I don't have any patches for Samsung pinctrl
in my tree.
---
 drivers/pinctrl/samsung/pinctrl-exynos.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/pinctrl/samsung/pinctrl-exynos.c b/drivers/pinctrl/samsung/pinctrl-exynos.c
index 0cd7f33cdf25..2b99f4130e1e 100644
--- a/drivers/pinctrl/samsung/pinctrl-exynos.c
+++ b/drivers/pinctrl/samsung/pinctrl-exynos.c
@@ -55,7 +55,7 @@ static void exynos_irq_mask(struct irq_data *irqd)
 	struct exynos_irq_chip *our_chip = to_exynos_irq_chip(chip);
 	struct samsung_pin_bank *bank = irq_data_get_irq_chip_data(irqd);
 	unsigned long reg_mask = our_chip->eint_mask + bank->eint_offset;
-	unsigned long mask;
+	unsigned int mask;
 	unsigned long flags;
 
 	raw_spin_lock_irqsave(&bank->slock, flags);
@@ -83,7 +83,7 @@ static void exynos_irq_unmask(struct irq_data *irqd)
 	struct exynos_irq_chip *our_chip = to_exynos_irq_chip(chip);
 	struct samsung_pin_bank *bank = irq_data_get_irq_chip_data(irqd);
 	unsigned long reg_mask = our_chip->eint_mask + bank->eint_offset;
-	unsigned long mask;
+	unsigned int mask;
 	unsigned long flags;
 
 	/*
@@ -483,7 +483,7 @@ static void exynos_irq_eint0_15(struct irq_desc *desc)
 	chained_irq_exit(chip, desc);
 }
 
-static inline void exynos_irq_demux_eint(unsigned long pend,
+static inline void exynos_irq_demux_eint(unsigned int pend,
 						struct irq_domain *domain)
 {
 	unsigned int irq;
@@ -500,8 +500,8 @@ static void exynos_irq_demux_eint16_31(struct irq_desc *desc)
 {
 	struct irq_chip *chip = irq_desc_get_chip(desc);
 	struct exynos_muxed_weint_data *eintd = irq_desc_get_handler_data(desc);
-	unsigned long pend;
-	unsigned long mask;
+	unsigned int pend;
+	unsigned int mask;
 	int i;
 
 	chained_irq_enter(chip, desc);
-- 
2.25.1

Re: [PATCH] pinctrl: samsung: use 'int' for register masks in Exynos

[Sylwester Nawrocki]

On 08.04.2021 21:50, Krzysztof Kozlowski wrote:
> The Special Function Registers on all Exynos SoC, including ARM64, are
> 32-bit wide, so entire driver uses matching functions like readl() or
> writel().  On 64-bit ARM using unsigned long for register masks:
> 1. makes little sense as immediately after bitwise operation it will be
>    cast to 32-bit value when calling writel(),
> 2. is actually error-prone because it might promote other operands to
>    64-bit.
> 
> Addresses-Coverity: Unintentional integer overflow
> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>

Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>

> ---
> 
> Not tested on ARM64.

I have tested it on exynos5433/tm2e and didn't notice any issues
as we could expect. 
The patch looks good to me, however I would personally use u32
rather than "unsigned int", like in other places for the register
value variables.

--
Regards, 
Sylwester

Re: [PATCH] pinctrl: samsung: use 'int' for register masks in Exynos

[Linus Walleij]

On Thu, Apr 8, 2021 at 9:50 PM Krzysztof Kozlowski
<krzysztof.kozlowski@canonical.com> wrote:

> The Special Function Registers on all Exynos SoC, including ARM64, are
> 32-bit wide, so entire driver uses matching functions like readl() or
> writel().  On 64-bit ARM using unsigned long for register masks:
> 1. makes little sense as immediately after bitwise operation it will be
>    cast to 32-bit value when calling writel(),
> 2. is actually error-prone because it might promote other operands to
>    64-bit.
>
> Addresses-Coverity: Unintentional integer overflow
> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
(...)
> Please apply it directly, I don't have any patches for Samsung pinctrl
> in my tree.

OK! Patch applied!

Yours,
Linus Walleij