Re: [PATCH] pinctrl: bcm2835: Use raw spinlock for RT compatibility

[Linus Walleij <linus.walleij@xxxxxxxxxx>]

On Mon, Oct 29, 2018 at 3:45 PM Julia Cartwright <julia@xxxxxx> wrote:

> The change seems reasonable to me.

Thanks. (I record that as an Acked-by.)

> As a side note: the bitops used to manage the enabled_irq_map are the
> atomic variants (set_bit, etc), is this actually necessary?  In other
> words, are they ever accessed outside of the spinlock?

Lukas can you look into this?

On a further side note I have developed an aversion toward the
implicit semantics in __set_bit() for non-atomic set_bit(), I was
thinking to send Torvalds something like a sed script to simply
rename all the __set_bit() and friends to set_bit_nonatomic() etc
or set_bit_na() so it becomes human readable that he could
run before finalizing the next kernel. If people think it is a good
idea.

Yours,
Linus Walleij