Re: [PATCH v1 net] net: dsa: microchip: fix wrong register write when masking interrupt

From: Vladimir Oltean
Date: Sun Jun 02 2024 - 10:15:39 EST


On Thu, May 30, 2024 at 06:39:13PM -0700, Tristram.Ha@xxxxxxxxxxxxx wrote:
> From: Tristram Ha <tristram.ha@xxxxxxxxxxxxx>
>
> Initially the macro REG_SW_PORT_INT_MASK__4 is defined as 0x001C in
> ksz9477_reg.h and REG_PORT_INT_MASK is defined as 0x#01F. Because the
> global and port interrupt handling is about the same the new
> REG_SW_PORT_INT_MASK__1 is defined as 0x1F in ksz_common.h. This works
> as only the least significant bits have effect. As a result the 32-bit
> write needs to be changed to 8-bit.
>
> Fixes: e1add7dd6183 ("net: dsa: microchip: use common irq routines for girq and pirq")
> Signed-off-by: Tristram Ha <tristram.ha@xxxxxxxxxxxxx>
> ---
> v1
> - clarify the reason to change the code

After v1 comes v2.

>
> drivers/net/dsa/microchip/ksz_common.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c
> index 1e0085cd9a9a..3ad0879b00cd 100644
> --- a/drivers/net/dsa/microchip/ksz_common.c
> +++ b/drivers/net/dsa/microchip/ksz_common.c
> @@ -2185,7 +2185,7 @@ static void ksz_irq_bus_sync_unlock(struct irq_data *d)
> struct ksz_device *dev = kirq->dev;
> int ret;
>
> - ret = ksz_write32(dev, kirq->reg_mask, kirq->masked);
> + ret = ksz_write8(dev, kirq->reg_mask, kirq->masked);
> if (ret)
> dev_err(dev->dev, "failed to change IRQ mask\n");
>
> --
> 2.34.1
>

What is the user-visible functional impact of the 32-bit access? Justify
why this is a bug worth sending to stable kernels please.

FWIW, struct ksz_irq operates on 16-bit registers.