Re: [PATCH v2] hw/misc: use extract64 instead of 1 << i

[Тигран Согомонян]

27/12/24 18:16, Richard Henderson пишет:
> On 12/27/24 02:46, Tigran Sogomonian wrote:
> > 1 << i is casted to uint64_t while bitwise and with val.
> > So this value may become 0xffffffff80000000 but only
> > 31th "start" bit is required.
> > Use the bitfield extract() API instead.
>
> Again, I < 32.  There is no overflow.  The type of value is irrelevant.
s->num_leds maximum MPS2FPGAIO_MAX_LEDS (32). For bitwise AND, the 
result of shift 1 << i will be converted to uint64_t because value is of 
type uint64_t (integer promotion) and for i = 31 instead of the expected 
0b10000000000000000000000000000000  we get 
0b1111111111111111111111111111111100000000000000000000000000000000. Need 
to be corrected to value & (1U << i).
> > Found by Linux Verification Center (linuxtesting.org) with SVACE.
> >
> > Signed-off-by: Tigran Sogomonian <tsogomonian@astralinux.ru>
> > ---
> >   hw/misc/mps2-fpgaio.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/hw/misc/mps2-fpgaio.c b/hw/misc/mps2-fpgaio.c
> > index d07568248d..04a3da5db0 100644
> > --- a/hw/misc/mps2-fpgaio.c
> > +++ b/hw/misc/mps2-fpgaio.c
> > @@ -198,7 +198,7 @@ static void mps2_fpgaio_write(void *opaque, 
> > hwaddr offset, uint64_t value,
> >                 s->led0 = value & MAKE_64BIT_MASK(0, s->num_leds);
> >               for (i = 0; i < s->num_leds; i++) {
> > -                led_set_state(s->led[i], value & (1 << i));
> > +                led_set_state(s->led[i], extract64(value, i, 1));
> >               }
> >           }
> >           break;