Re: [PATCH] drm: Allow drm_fourcc.h without including drm.h

[James Park <james.park@xxxxxxxxxxxxxxxx>]

I suppose I should do this to avoid fighting with <linux/types.h>

#ifdef DRM_FOURCC_STANDALONE
#if defined(__linux__)

#include <linux/types.h>

#else
#include <stdint.h>
typedef uint32_t __u32;
typedef uint64_t __u64;

#endif
#else
#include "drm.h"
#endif

I'll wait for more feedback before updating the patch though.

On Fri, Dec 4, 2020 at 1:47 AM James Park <james.park@xxxxxxxxxxxxxxxx> wrote:

The typedefs might also conflict on Linux if DRM_FOURCC_STANDALONE is enabled with whatever LInux declared __u32/__u64 as, but I think the implication is that once DRM_FOURCC_STANDALONE has been declared, that's kind of a promise not to include drm.h.

I'm fine with this, but I'm not married to it if someone has a problem where they want to define DRM_FOURCC_STANDALONE, but also can't avoid including drm.h for some reason.

On Fri, Dec 4, 2020 at 12:53 AM Simon Ser <contact@xxxxxxxxxxx> wrote:

On Friday, December 4, 2020 5:53 AM, James Park <jpark37@xxxxxxxxxxxxxxxx> wrote:

> +#ifdef DRM_FOURCC_STANDALONE
> +#include <stdint.h>
>
> +typedef uint32_t __u32;
> +typedef uint64_t __u64;
> +#else
> #include "drm.h"
> +#endif

C11 allows duplicate typedefs, but older versions of the standard
don't AFAIK. If this is a concern, a solution would be to guard the
typedefs.

_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/dri-devel