On Sun, 2010-08-15 at 21:23 +0200, Sam Ravnborg wrote:
When I ran "make headers_check" on upstream I got following set of warnings for the drm headers:
usr/include/drm/drm_mode.h:85: found __[us]{8,16,32,64} type without #include <linux/types.h> usr/include/drm/i915_drm.h:120: found __[us]{8,16,32,64} type without #include <linux/types.h> usr/include/drm/mga_drm.h:260: found __[us]{8,16,32,64} type without #include <linux/types.h> usr/include/drm/radeon_drm.h:758: found __[us]{8,16,32,64} type without #include <linux/types.h> usr/include/drm/via_drm.h:117: found __[us]{8,16,32,64} type without #include <linux/types.h>
When I looked through the files I noticed a few things:
- Several files uses a mixture of native types and linux types. For example "unsigned int" and "__u32" in the same file.
- drm.h contains:
#if defined(__linux__)
#include <linux/types.h> #include <asm/ioctl.h> typedef unsigned int drm_handle_t;
#else /* One of the BSDs */
#include <sys/ioccom.h> #include <sys/types.h> typedef int8_t __s8; typedef uint8_t __u8; typedef int16_t __s16; typedef uint16_t __u16; typedef int32_t __s32; typedef uint32_t __u32; typedef int64_t __s64; typedef uint64_t __u64; typedef unsigned long drm_handle_t;
#endif
Why this when no-one else require this treatment?
- Lots of typedefs...
All in all - too much to dig into without knowing the background for the current state.
At a minimum please get the warnings fixed.
At a minimum please fix the checker to work.
It doesn't recurse, all of the .h files include drm.h, which includes linux/types.h, this is how its meant to work. If the checker does recurse then doesn't define __linux__ then fix that. This isn't a bug in the drm header files.
We copy these files as is into a userspace that is shared with FreeBSD, the ifdef is the minimum down from about 10 ifdefs 2-3 years ago concession we've made.
Dave.