Re: [PATCH v11 RESEND 6/9] arm64: futex: support futex with FEAT_LSUI

[Will Deacon <will@kernel.org>]

On Mon, Jan 19, 2026 at 10:17:47PM +0000, Yeoreum Yun wrote:
> > > +"2:\n"
> > > +	_ASM_EXTABLE_UACCESS_ERR(1b, 2b, %w0)
> > > +	: "+r" (ret), "+Q" (*uaddr), "+r" (*oldval)
> > > +	: "r" (newval)
> > > +	: "memory");
> >
> > Don't you need to update *oldval here if the CAS didn't fault?
> 
> No. if CAS doesn't make fault the oldval update already.

Sorry, it was the "+r" constraint with a pointer dereference that threw
me but you have the "memory" clobber so it looks like this will work.

> > > +
> > > +	for (i = 0; i < FUTEX_MAX_LOOPS; i++) {
> > > +		if (get_user(oval64.raw, uaddr64))
> > > +			return -EFAULT;
> >
> > Since oldval is passed to us as an argument, can we get away with a
> > 32-bit get_user() here?
> 
> It's not a probelm. but is there any sigificant difference?

I think the code would be clearer if you only read what you actually
use.

> > > +		nval64.raw = oval64.raw;
> > > +
> > > +		if (futex_on_lo) {
> > > +			oval64.lo_futex.val = oldval;
> > > +			nval64.lo_futex.val = newval;
> > > +		} else {
> > > +			oval64.hi_futex.val = oldval;
> > > +			nval64.hi_futex.val = newval;
> > > +		}
> > > +
> > > +		orig64.raw = oval64.raw;
> > > +
> > > +		if (__lsui_cmpxchg64(uaddr64, &oval64.raw, nval64.raw))
> > > +			return -EFAULT;
> > > +
> > > +		if (futex_on_lo) {
> > > +			oldval = oval64.lo_futex.val;
> > > +			other = oval64.lo_futex.other;
> > > +			orig_other = orig64.lo_futex.other;
> > > +		} else {
> > > +			oldval = oval64.hi_futex.val;
> > > +			other = oval64.hi_futex.other;
> > > +			orig_other = orig64.hi_futex.other;
> > > +		}
> > > +
> > > +		if (other == orig_other) {
> > > +			ret = 0;
> > > +			break;
> > > +		}
> > > +	}
> > > +
> > > +	if (!ret)
> > > +		*oval = oldval;
> >
> > Shouldn't we set *oval to the value we got back from the CAS?
> 
> Since it's a "success" case, the CAS return and oldval must be the same.
> That's why it doesn't matter to use got back from the CAS.
> Otherwise, it returns error and *oval doesn't matter for
> futex_atomic_cmpxchg_inatomic().

Got it, but then the caller you have is very weird because e.g.
__lsui_futex_atomic_eor() goes and does another get_user() on the next
iteration instead of using the value returned by the CAS.

It would probably be clearer if you restructured your CAS helper to look
more like try_cmpxchg() and then the loop around it would be minimal.
You might need to distinguish the faulting case from the comparison
failure case with e.g. -EFAULT vs -EAGAIN.

Will