[PATCH] glamor: Use GL_STREAM_READ also for read/write access to a PBO

[Markus Wick]

Am 2014-09-25 08:27, schrieb Michel Dänzer:
> From: Michel Dänzer <michel.daenzer at amd.com>
> 
> Otherwise the CPU may end up reading from non-cacheable memory, which 
> is
> very slow.
> 
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84178
> Signed-off-by: Michel Dänzer <michel.daenzer at amd.com>
> ---
> 
> Keeping gl_usage in case we need to add back GLAMOR_ACCESS_WO.
> 
>  glamor/glamor_prepare.c | 5 +----
>  1 file changed, 1 insertion(+), 4 deletions(-)
> 
> diff --git a/glamor/glamor_prepare.c b/glamor/glamor_prepare.c
> index 561c55d..fb85d90 100644
> --- a/glamor/glamor_prepare.c
> +++ b/glamor/glamor_prepare.c
> @@ -84,10 +84,7 @@ glamor_prep_pixmap_box(PixmapPtr pixmap,
> glamor_access_t access, BoxPtr box)
>              if (priv->base.pbo == 0)
>                  glGenBuffers(1, &priv->base.pbo);
> 
> -            if (access == GLAMOR_ACCESS_RW)
> -                gl_usage = GL_DYNAMIC_DRAW;
> -            else
> -                gl_usage = GL_STREAM_READ;
> +            gl_usage = GL_STREAM_READ;
> 
>              glBindBuffer(GL_PIXEL_PACK_BUFFER, priv->base.pbo);
>              glBufferData(GL_PIXEL_PACK_BUFFER,

Was the write only patch merged? If so, this should be changed to use 
GL_STREAM_DRAW for WO.

But you're right, for RW, it should be STREAM (used once) READ (accessed 
by the CPU), so

Reviewed-by: Markus Wick <markus at selfnet.de>