On Mon, 2011-02-21 at 18:54 +0200, Felipe Balbi wrote: > What you seem to have missed is that sched.h doesn't include wait.h, it > includes completion.h and completion.h needs wait.h due the > wait_queue_head_t it uses. Yeah, so? sched.h doesn't need completion.h, but like with wait.h I'd argue the other way around, completion.h would want to include sched.h > If someone finds a cleaner way to drop that need, then I'm all for it as > my original suggestion to the original patch was to include sched.h in > wait.h, but it turned out that it's not possible due to the reasons > already explained. Feh,. I'm saying the proposed solution stinks and if you want to make things better you need to work on fixing whatever is in the way of including sched.h from wait.h. 1) remove the inclusion of completion.h -- easy we can live with an incomplete type. 2) move the other wait_queue_head_t users (signal_struct sighand_struct) out of sched.h 3) ... 4) profit! Just isolating the TASK_state bits isn't going to be enough, wait.h also wants wake_up goo and schedule*(), therefore either include sched.h from whatever .c file you're using wait.h bits or do the above cleanup. -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
