[RFC] eal: use same atomic intrinsics for gcc and clang
Tyler Retzlaff
roretzla at linux.microsoft.com
Mon Feb 13 21:01:05 CET 2023
On Fri, Feb 10, 2023 at 07:56:22PM -0600, Honnappa Nagarahalli wrote:
> The size generic atomic intrinsics generate the same
> code as the size specific intrinsics for gcc. Use size
> generic intrinsics for both gcc and clang.
>
> Fixes: 7bdccb93078e ("eal: fix ARM build with clang")
> Cc: stable at dpdk.org
> Cc: pbhagavatula at marvell.com
>
> Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli at arm.com>
Acked-by: Tyler Retzlaff <roretzla at linux.microsoft.com>
so first of all thanks for fixing this up. it was confusing me why it
was being conditionally compiled (there was no documented rationale).
i do have a slight concern that perhaps the reason this was done was
because __atomic_exchange_n on some gcc versions or for some specific
target processor was generating less than desirable code but if nobody
else has this concern i won't get in the way since it simplifies things.
thanks!
> ---
> lib/eal/include/generic/rte_atomic.h | 12 ------------
> 1 file changed, 12 deletions(-)
>
> diff --git a/lib/eal/include/generic/rte_atomic.h b/lib/eal/include/generic/rte_atomic.h
> index f5c49a9870..234b268b91 100644
> --- a/lib/eal/include/generic/rte_atomic.h
> +++ b/lib/eal/include/generic/rte_atomic.h
> @@ -176,11 +176,7 @@ rte_atomic16_exchange(volatile uint16_t *dst, uint16_t val);
> static inline uint16_t
> rte_atomic16_exchange(volatile uint16_t *dst, uint16_t val)
> {
> -#if defined(__clang__)
> return __atomic_exchange_n(dst, val, __ATOMIC_SEQ_CST);
> -#else
> - return __atomic_exchange_2(dst, val, __ATOMIC_SEQ_CST);
> -#endif
> }
> #endif
>
> @@ -459,11 +455,7 @@ rte_atomic32_exchange(volatile uint32_t *dst, uint32_t val);
> static inline uint32_t
> rte_atomic32_exchange(volatile uint32_t *dst, uint32_t val)
> {
> -#if defined(__clang__)
> return __atomic_exchange_n(dst, val, __ATOMIC_SEQ_CST);
> -#else
> - return __atomic_exchange_4(dst, val, __ATOMIC_SEQ_CST);
> -#endif
> }
> #endif
>
> @@ -741,11 +733,7 @@ rte_atomic64_exchange(volatile uint64_t *dst, uint64_t val);
> static inline uint64_t
> rte_atomic64_exchange(volatile uint64_t *dst, uint64_t val)
> {
> -#if defined(__clang__)
> return __atomic_exchange_n(dst, val, __ATOMIC_SEQ_CST);
> -#else
> - return __atomic_exchange_8(dst, val, __ATOMIC_SEQ_CST);
> -#endif
> }
> #endif
>
> --
> 2.25.1
More information about the stable
mailing list