[Mesa-dev] [PATCH 1/7] radv/winsys: request high address

[Bas Nieuwenhuizen]

On Fri, May 18, 2018 at 6:00 PM, Samuel Pitoiset
<samuel.pitoiset at gmail.com> wrote:
> This is needed for 32-bit GPU pointers. Ported from RadeonSI.
>
> Signed-off-by: Samuel Pitoiset <samuel.pitoiset at gmail.com>
> ---
>  src/amd/vulkan/winsys/amdgpu/radv_amdgpu_bo.c | 12 +++++++++---
>  1 file changed, 9 insertions(+), 3 deletions(-)
>
> diff --git a/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_bo.c b/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_bo.c
> index 8ed3e53e2d..442940eb43 100644
> --- a/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_bo.c
> +++ b/src/amd/vulkan/winsys/amdgpu/radv_amdgpu_bo.c
> @@ -38,6 +38,9 @@
>
>  #include "util/u_atomic.h"
>
> +#ifndef AMDGPU_VA_RANGE_HIGH
> +#define AMDGPU_VA_RANGE_HIGH   0x2
> +#endif

I don't think this is required since libdrm_amdgpu version 2.4.91,
which we require.

Besides the comment on this patch and patch 7 I think this series looks okay.

>
>  static void radv_amdgpu_winsys_bo_destroy(struct radeon_winsys_bo *_bo);
>
> @@ -306,7 +309,8 @@ radv_amdgpu_winsys_bo_create(struct radeon_winsys *_ws,
>         }
>
>         r = amdgpu_va_range_alloc(ws->dev, amdgpu_gpu_va_range_general,
> -                                 size, alignment, 0, &va, &va_handle, 0);
> +                                 size, alignment, 0, &va, &va_handle,
> +                                 AMDGPU_VA_RANGE_HIGH);
>         if (r)
>                 goto error_va_alloc;
>
> @@ -424,7 +428,8 @@ radv_amdgpu_winsys_bo_from_ptr(struct radeon_winsys *_ws,
>                 goto error;
>
>         if (amdgpu_va_range_alloc(ws->dev, amdgpu_gpu_va_range_general,
> -                                 size, 1 << 12, 0, &va, &va_handle, 0))
> +                                 size, 1 << 12, 0, &va, &va_handle,
> +                                 AMDGPU_VA_RANGE_HIGH))
>                 goto error_va_alloc;
>
>         if (amdgpu_bo_va_op(buf_handle, 0, size, va, 0, AMDGPU_VA_OP_MAP))
> @@ -480,7 +485,8 @@ radv_amdgpu_winsys_bo_from_fd(struct radeon_winsys *_ws,
>                 goto error_query;
>
>         r = amdgpu_va_range_alloc(ws->dev, amdgpu_gpu_va_range_general,
> -                                 result.alloc_size, 1 << 20, 0, &va, &va_handle, 0);
> +                                 result.alloc_size, 1 << 20, 0, &va, &va_handle,
> +                                 AMDGPU_VA_RANGE_HIGH);
>         if (r)
>                 goto error_query;
>
> --
> 2.17.0
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev