On Mon, Aug 22, 2016 at 8:35 AM Barret Rhoden <[email protected]> wrote:
>
>
> yikes. was the driver doing stuff like a u32 write to a non-32 byte
> alignment?
>
32 bit writes to 8 byte fields, as I struggled to make it work in a
non-ken-C world.
>
> i'd be happy to merge this, though my slight concern is that these
> helpers were linux code, and there might have been a reason it was the
> way it was. (like painful casting, or something, in a lot of driver
> code).
>
They usually do that when they're going to put them in some struct
somewhere, which we don't have. I'd rather leave it type safe here and cast
it in the struct initializer, after having been hurt by this.
>
> also:
>
> > -static inline uint8_t __raw_read8(const volatile void __iomem *addr)
> > +static inline uint8_t __raw_read8(const volatile uint8_t __iomem *addr)
> {
> > return *(const volatile uint8_t __force *)addr;
>
> Do these still need the cast, since they now take a const vol u8 *?
>
>
>
yeah, because we might someday want __iomem and __force.
ron
--
You received this message because you are subscribed to the Google Groups
"Akaros" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
For more options, visit https://groups.google.com/d/optout.
