On Thu, 2017-03-09 at 19:48 +0100, Hans de Goede wrote:
> Hi,
>
> On 09-03-17 15:45, Hans de Goede wrote:
> > Hi,
> >
> > On 09-03-17 15:03, Andy Shevchenko wrote:
> > > On Thu, 2017-03-09 at 14:57 +0100, Hans de Goede wrote:
> Ok, so there is one problem with the extcon-intel-int3496.c
> together with your gpio patches:
>
> [ 2382.484415] gpio gpiochip1: (INT33FF:01): gpiochip_lock_as_irq:
> tried to flag a GPIO set as output for IRQ
> [ 2382.484425] gpio gpiochip1: (INT33FF:01): unable to lock HW IRQ 3
> for IRQ
> [ 2382.484429] genirq: Failed to request resources for INT3496:00 (irq
> 174) on irqchip chv-gpio
> [ 2382.484518] intel-int3496 INT3496:00: can't request IRQ for USB ID
> GPIO: -22
> [ 2382.500359] intel-int3496: probe of INT3496:00 failed with error
> -22
By the way did not you see a warning from gpiolib-acpi.c that flags are
overridden by firmware?
>
> This can be fixed / worked around by doing this:
>
>
> --- a/drivers/extcon/extcon-intel-int3496.c
> +++ b/drivers/extcon/extcon-intel-int3496.c
> @@ -113,6 +113,7 @@ static int int3496_probe(struct platform_device
> *pdev)
> dev_err(dev, "can't request USB ID GPIO: %d\n", ret);
> return ret;
> }
> + gpiod_direction_input(data->gpio_usb_id);
(1)
>
> data->usb_id_irq = gpiod_to_irq(data->gpio_usb_id);
> if (data->usb_id_irq < 0) {
>
> But the gpio is requested as:
>
> data->gpio_usb_id = devm_gpiod_get(dev, "id", GPIOD_IN);
>
> Note the GPIOD_IN I think the new behavior is caused by:
>
> https://bitbucket.org/andy-shev/linux/commits/ba458eb945879a66eac75de6
> 9a4e7312c4f44cc7
>
> Of which I'm not sure it is a good idea, as the dsdt of my
> cube iwork8 air shows:
>
> Device (OTG2)
> {
> Name (_HID, "INT3496") // _HID: Hardware ID
> Name (_CID, "INT3496") // _CID: Compatible ID
>
> Method (_CRS, 0, NotSerialized) // _CRS: Current Resource Settings
> {
> Name (ABUF, ResourceTemplate ()
> {
> GpioIo (Exclusive, PullDefault, 0x0000, 0x0000,
> IoRestrictionOutputOnly,
> "\\_SB.GPO1", 0x00, ResourceConsumer, ,
> )
> { // Pin list
> 0x0003
> }
This is definitely bug in firmware.
> GpioIo (Exclusive, PullDefault, 0x0000, 0x0000,
> IoRestrictionOutputOnly,
> "\\_SB.GPO3", 0x00, ResourceConsumer, ,
> )
> { // Pin list
> 0x002E
> }
> })
> Return (ABUF) /* \_SB_.PCI0.OTG2._CRS.ABUF */
> }
>
> }
>
> The dstd-s out there in the wild do not always get this right.
Yeah and here is the dilemma of making life easier to the user and
provoke vendors to abuse standards, or say "no".
In this particular case it is even worse since driver is in upstream and
it's working. (2)
>
> With my workaround from above the extcon-intel-int3496 driver
> does work properly again, so I see 2 options here:
>
> 1) Let drivers workaround known broken IoRestriction in dstd-s
> by using gpiod_direction_...
I see two options here (at least):
a) override DSDT completely (or in the future perhaps partially);
b) use workaround in the driver.
Taking into consideration (2) I would go with option b) and warn LOUDLY
in (1) that firmware has a bug which we aware of.
> 2) Drop the patch which makes gpiod_get honor the IoRestrictions
I would go 1).
> I've also tested the silead driver and with the new more strict
> gpio code it works as it should as-is (as expected).
Thanks!
--
Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
Intel Finland Oy
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
