On Mon, Mar 11, 2013 at 11:54 AM, Michel Dänzer <mic...@daenzer.net> wrote: > On Son, 2013-03-10 at 23:05 +0100, Martin Andersson wrote: >> --- >> src/mesa/main/readpix.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/src/mesa/main/readpix.c b/src/mesa/main/readpix.c >> index 2f130ae..349b0bc 100644 >> --- a/src/mesa/main/readpix.c >> +++ b/src/mesa/main/readpix.c >> @@ -238,7 +238,7 @@ fast_read_rgba_pixels_memcpy( struct gl_context *ctx, >> } >> else if (rb->Format == MESA_FORMAT_XRGB8888 && >> format == GL_BGRA && >> - type == GL_UNSIGNED_INT_8_8_8_8_REV && >> + (type == GL_UNSIGNED_INT_8_8_8_8_REV || type == GL_UNSIGNED_BYTE) && > > This cannot be equivalent on little endian and big endian hosts at the > same time. As it works for you, it's apparently equivalent on little > endian.
ok, I guess it is also undesirable to have lots of special cases there, with checks for lots of different combos of types and endianness. > > I suspect ReadPixels could be made even faster with similar treatment as > Marek has applied to TexSubImage etc. I have looked at TexSubImage and how the radeon dri and intel dri drivers implement glReadPixels (they use a blit call). But I did not understand how I could use it for glReadPixels. I will look at it some more, thanks. > > -- > Earthling Michel Dänzer | http://www.amd.com > Libre software enthusiast | Debian, X and DRI developer _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
