On Fri, 28 Dec 2018, Daniel Vetter daniel@ffwll.ch wrote:
On Fri, Dec 28, 2018 at 10:23:00AM +0200, Jani Nikula wrote:
On Thu, 27 Dec 2018, Daniel Vetter daniel@ffwll.ch wrote:
On Thu, Dec 27, 2018 at 05:04:20PM +0100, Daniel Vetter wrote:
On Thu, Dec 27, 2018 at 02:56:36PM +0200, Jani Nikula wrote:
Make it easier to drop drmP.h includes.
Cc: Sam Ravnborg sam@ravnborg.org Cc: Daniel Vetter daniel@ffwll.ch Cc: Laurent Pinchart laurent.pinchart@ideasonboard.com Signed-off-by: Jani Nikula jani.nikula@intel.com
include/drm/drm_legacy.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/include/drm/drm_legacy.h b/include/drm/drm_legacy.h index 8fad66f88e4f..743d7e70c896 100644 --- a/include/drm/drm_legacy.h +++ b/include/drm/drm_legacy.h @@ -2,6 +2,7 @@ #define __DRM_DRM_LEGACY_H__
#include <drm/drm_auth.h> +#include <drm/drm_device.h>
From a quick look, shouldn't a
struct drm_device;
forward decl be enough? You might need a pile more forward decl, but that's all drm_device.h seems to pull in that drm_legacy.h needs.
Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch with the forward decl (assuming it all works out).
The forward declaration is not enough, there's a &dev->maplist reference in the static inline drm_legacy_findmap().
Missed that, despite that I tried checking for static inlines. And a static inline for a O(n) lookup ... tempted to just deinline that.
Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch on either your v1 or a patch that deinlines the offending function and just uses a forward decl.
It'll still require drm_hashtab.h include, but less intrusive than drm_device.h I guess. Patch sent.
BR, Jani.
Cheers, Daniel
BR, Jani.
-Daniel
-Daniel
/*
- Legacy driver interfaces for the Direct Rendering Manager
-- 2.11.0
-- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch
-- Jani Nikula, Intel Open Source Graphics Center