On Tue, Nov 19, 2019 at 06:22:28PM +0900, Namjae Jeon wrote:
> > No idea what the code does. But I was just skimming over and find the
> > above pattern somehow strange. Shouldn't this be something like
> Right.
>
> >
> > if (!READ_ONCE(sbi->s_dirt)) {
> > WRITE_ONCE(sbi->s_dirt, true);
>
> It should be :
> if (READ_ONCE(sbi->s_dirt)) {
> WRITE_ONCE(sbi->s_dirt, false);
> I will fix it on v3.
The other option would be to an unsigned long flags field and define
bits flags on it, then use test_and_set_bit, test_and_clear_bit etc.
Which might be closer to the pattern we use elsewhere in the kernel.
