Re: [brgl:gpio/for-next 42/47] gpio-mmio.c:undefined reference to `iowrite64'

["Arnd Bergmann" <arnd@xxxxxxxx>]

On Mon, Feb 26, 2024, at 03:17, Andy Shevchenko wrote:
> On Sun, Feb 25, 2024 at 11:09 PM Arnd Bergmann <arnd@xxxxxxxx> wrote:
>> On Sun, Feb 25, 2024, at 00:23, Andy Shevchenko wrote:
>>
>> readq()/writeq() are not generally a replace for
>> ioread64()/iowrite64() because they can't deal with ioport_map()
>> type mappings, though the reverse is true and you can always
>> replace readl()/writel() with ioread32()/iowrite32() if you
>> can live with the performance overhead on x86.
>
> The driver in question by name assumes that it won't perform IO port
> access.

Do you have a link to the driver? Maybe it can just be
made 'depends on !GENERIC_IOMAP' if it doesn't run on x86
or um.

bgpio_setup_accessors() already has a special case for 64-bit
configs to allow bgpio_read64(), so it wouldn't be any
weirder to also require !GENERIC_IOMAP here.

> Perhaps use of ioread*()/iowrite*() is not what we should even
> consider there, Linus, Bart, do you know if gpio-mmio was ever used
> for devices that want IO port rather than MMIO accesses?

It looks like there is only one PCI driver using pci_iomap
with bgpio_init(), and this one is x86 specific:
drivers/gpio/gpio-sodaville.c. This one is little-endian
and 32-bit only.

    Arnd