On 09/25/2013 01:00 PM, Chris Forbes wrote:
> This is useful when working on the drivers, since falling back to
> indirect rendering means we broke our driver's ability to load somehow.
>
> Signed-off-by: Chris Forbes <chr...@ijw.co.nz>
> ---
> docs/envvars.html | 1 +
> src/glx/glxext.c | 10 ++++++++--
> 2 files changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/docs/envvars.html b/docs/envvars.html
> index 67d8e29..07e82b5 100644
> --- a/docs/envvars.html
> +++ b/docs/envvars.html
> @@ -30,6 +30,7 @@ sometimes be useful for debugging end-user issues.
> If set to 'verbose' additional information will be printed.
> <li>LIBGL_DRIVERS_PATH - colon-separated list of paths to search for DRI
> drivers
> <li>LIBGL_ALWAYS_INDIRECT - forces an indirect rendering context/connection.
> +<li>LIBGL_NEVER_INDIRECT - disables fallback to indirect rendering (will
> abort instead)
> <li>LIBGL_ALWAYS_SOFTWARE - if set, always use software rendering
> <li>LIBGL_NO_DRAWARRAYS - if set do not use DrawArrays GLX protocol (for
> debugging)
> <li>LIBGL_SHOW_FPS - print framerate to stdout based on the number of
> glXSwapBuffers
> diff --git a/src/glx/glxext.c b/src/glx/glxext.c
> index bea1ccb..491f3e2 100644
> --- a/src/glx/glxext.c
> +++ b/src/glx/glxext.c
> @@ -781,8 +781,14 @@ AllocAndFetchScreenConfigs(Display * dpy, struct
> glx_display * priv)
> if (psc == NULL)
> psc = applegl_create_screen(i, priv);
> #else
> - if (psc == NULL)
> - psc = indirect_create_screen(i, priv);
> + if (psc == NULL) {
> + if (getenv("LIBGL_NEVER_INDIRECT") != NULL) {
> + printf("Would fall back to indirect rendering, but giving up"
> + " due to LIBGL_NEVER_INDIRECT.\n");
> + abort();
> + }
> + psc = indirect_create_screen(i, priv);
> + }
> #endif
> priv->screens[i] = psc;
> }
>
I would add a "glx: " prefix to the commit title.
Otherwise, this looks great. I tested it by creating bogus i965_dri.so
and swrast_dri.so files in my $LIBGL_DRIVERS_PATH directory and running
glxinfo. It fell back to indirect. Then I used LIBGL_NEVER_INDIRECT=1
and it aborted, as I expected.
Tested-by: Kenneth Graunke <kenn...@whitecape.org>
Thanks!
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
