Re: [RFC PATCH] sparse_irq aka dyn_irq

[Ingo Molnar <mingo@elte.hu>]

* Yinghai Lu <yinghai@kernel.org> wrote:

> Ingo Molnar wrote:
> > * Yinghai Lu <yinghai@kernel.org> wrote:
> > 
> >> Andrew Morton wrote:
> >>> On Mon, 10 Nov 2008 10:40:33 +0100 Ingo Molnar <mingo@elte.hu> wrote:
> >>>
> >>>>>>> @@ -987,6 +988,8 @@ void __init mem_init(void)
> >>>>>>>
> >>>>>>>       set_highmem_pages_init();
> >>>>>>>
> >>>>>>> +     after_bootmem = 1;
> >>>>>> this hack can go away once we have a proper percpu_alloc() that can be
> >>>>>> used early enough.
> >>>>> where is that fancy patch? current percpu_alloc(), will keep big 
> >>>>> pointer in array..., instead of put that pointer in percpu_area
> >>>>>
> >>>>> 64bit has that after_bootmem already.
> >>>> or at least introduce a "bootmem agnostic" allocator instead of 
> >>>> open-coding the after_bootmem flag.
> >>>>
> >>>> Something like:
> >>>>
> >>>>   early_kzalloc()
> >>>>
> >>>> ?
> >>>>
> >>>> Andrew, any preferences?
> >>> My mind reading ain't what it was, and this after_bootmem flag is
> >>> write-only in this patch.
> >>>
> >>> So what's all this about?
> >> if i use alloc_bootmem to get some memory, and later after_bootmem, 
> >> can I use kfree to free it?
> > 
> > hm, no. If we used alloc_bootmem(), then we must not free it after 
> > after_bootmem has been set.
> 
> ok, let keep irq_desc for legacy irqs not movable...

most of them are movable right now, correct? If we restrict their 
movability now that might surprise existing usecases negatively.

	Ingo