one more reason to have common code for boolean flags. there is a bigish series to do that
On Thu, Oct 29, 2026, 22:17 Nawal Kishor <[email protected]> wrote: > parse_halo_ena_handler() kept the parsed value in a uint8_t and stored > it through a uint8_t pointer, while the caller passes a uint32_t. The > narrowing happened before the 0/1 check, so 256 was accepted as 0 and > 257 as 1. The one-byte store also updated only the first byte of that > uint32_t, which is correct only on a little-endian host when the local > is already zero. > > Parse and store a uint32_t so the check sees the full value and the > caller's object is written completely. > > Fixes: 45abbfc836b1 ("mempool/cnxk: support HALO in mempool") > > Signed-off-by: Nawal Kishor <[email protected]> > --- > drivers/mempool/cnxk/cnxk_mempool.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/mempool/cnxk/cnxk_mempool.c > b/drivers/mempool/cnxk/cnxk_mempool.c > index 6939fccff4..f1cdf9cbb5 100644 > --- a/drivers/mempool/cnxk/cnxk_mempool.c > +++ b/drivers/mempool/cnxk/cnxk_mempool.c > @@ -51,13 +51,13 @@ static int > parse_halo_ena_handler(const char *key, const char *value, void > *extra_args) > { > RTE_SET_USED(key); > - uint8_t val; > + uint32_t val; > > val = atoi(value); > if (val != 0 && val != 1) > return -EINVAL; > > - *(uint8_t *)extra_args = val; > + *(uint32_t *)extra_args = val; > return 0; > } > > -- > 2.48.1 > >
