Hi Shawn,
On 08/15/2016 03:28 AM, Uwe Kleine-König wrote:
Hello,
Cc += linux-gpio
On Mon, Aug 15, 2016 at 06:18:07AM +0000, Ken.Lin wrote:
1. As you can see in the datasheet(that I attached to this email), at
power on, P00 - P17 are configured as inputs (by chip default setting,
datasheet P.3) and an interrupt is generated by any rising or falling
edge of the port inputs in the input mode (datasheet P.10)
ah, I see. All input pins generate an irq. The driver should handle this
just fine however. So I hope you don't see an oops or messages about
unhandled irqs, right?
2. The unhandled interrupts ( return IRQ_NONE in pca953x_irq_handler)
in GPIO expander driver (drivers/gpio/gpio-pca953x.c) could happen
when unused GPIO pins in GPIO expander are not well configured.
Hmm, maybe the driver should return IRQ_HANDLED in this case, too. After
all the irq reason is known and it was expected that an irq happend,
right?
3. As a result, we tried to configure the unused (default floating
inputs) pca953x GPIO pins as GPO low to avoid the unexpected
interrupts (could be observed during the interference situation e.g.
putting fingers on the unused pin to make the signal level change)
So changing these to outputs might still make sense to prevent these
from floating or triggering unused irqs, but this should not make a
semantical difference for your machine. (It might have more load and a
bigger irq count, but apart from that everything should be the same.)
As Uwe mentioned above, independent of what the pca953x interrupt
handler returns (IRQ_HANDLED or IRQ_NONE), I feel we should still change
the pins to outputs to avoid trigger unused irqs. Is there any concern
in accepting the patch?
Thanks,
Akshay
Best regards
Uwe
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
