On Fri, Apr 14, 2017 at 8:47 PM, Eric Anholt <e...@anholt.net> wrote:
> NEON is sufficiently different on arm64 that we can't just reuse this
> code. Disable it on arm64 for now.
>
> Signed-off-by: Eric Anholt <e...@anholt.net>
> ---
> src/gallium/drivers/vc4/vc4_tiling_lt.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/src/gallium/drivers/vc4/vc4_tiling_lt.c
> b/src/gallium/drivers/vc4/vc4_tiling_lt.c
> index c9cbc65e2dbc..7de67b652daa 100644
> --- a/src/gallium/drivers/vc4/vc4_tiling_lt.c
> +++ b/src/gallium/drivers/vc4/vc4_tiling_lt.c
> @@ -61,7 +61,7 @@ static void
> vc4_load_utile(void *cpu, void *gpu, uint32_t cpu_stride, uint32_t cpp)
> {
> uint32_t gpu_stride = vc4_utile_stride(cpp);
> -#if defined(VC4_BUILD_NEON) && defined(__ARM_ARCH)
> +#if defined(VC4_BUILD_NEON) && defined(__ARM_ARCH) && __ARM_ARCH <= 7
>
This doesn't look right, __ARM_ARCH can be 8 on 32bit, try -mcpu=cortex-a57
on a 32bit toolchain for example. Could use PIPE_ARCH_ARM I guess, or
!PIPE_ARCH_AARCH64, or maybe use the intrinsics...
GraÅžvydas
if (gpu_stride == 8) {
> __asm__ volatile (
> /* Load from the GPU in one shot, no interleave,
> to
> @@ -118,7 +118,7 @@ vc4_store_utile(void *gpu, void *cpu, uint32_t
> cpu_stride, uint32_t cpp)
> {
> uint32_t gpu_stride = vc4_utile_stride(cpp);
>
> -#if defined(VC4_BUILD_NEON) && defined(__ARM_ARCH)
> +#if defined(VC4_BUILD_NEON) && defined(__ARM_ARCH) && __ARM_ARCH <= 7
> if (gpu_stride == 8) {
> __asm__ volatile (
> /* Load each 8-byte line from cpu-side source,
> --
> 2.11.0
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
