On 2017-08-07 18:22, Danilo Krummrich wrote:
I couldn't find any guarantee that the mode and tx_byte change is
covered by a barrier in this case. E.g. the bcm2835 driver does not make
stores are completed before the particular interrupt is enabled, except
fact that writel on ARM contains a wmb(). But this is nothing to rely
Yes, seems it can be removed. I didn't saw any explicit barriers in the
driver (I'm testing on bcm2835), but it seems MMIO operations on SMP
does contain barriers. Not sure if all do. If some do not this barrier
be needed to ensure ISR on other CPU notice the correct mode and byte
> +static int ps2_gpio_write(struct serio *serio, unsigned char val)
> + struct ps2_gpio_data *drvdata = serio->port_data;
> + drvdata->mode = PS2_MODE_TX;
> + drvdata->tx_byte = val;
> + /* Make sure ISR running on other CPU notice changes. */
> + barrier();
This seems overengineered, is this really needed?
If we have races like this, the error is likely elsewhere, and should
fixed in the GPIO driver MMIO access or so.
tell me if I miss something.)
Therefore I would like to keep this barrier and replace it with
smp_wmb() if you
are fine with that.