On Sun, 2003-12-07 at 13:01, Shinpei Kato wrote:
> I see! Can I ask one more thing. Look at following codes.
>
> ++++
> #define BITMAP_SIZE ((((MAX_PRIO+1+7)/8)+sizeof(long)-1)/sizeof(long))
>
> struct prio_array {
> int nr_active;
> unsigned long bitmap[BITMAP_SIZE];
> struct list_head queue[MAX_PRIO];
> };
> ++++
>
> The default MAX_PRIO is defined 140 and if we assume the size of long is
> 32 bits, BITMAP_SIZE comes to be ((((140+1+7)/8)+4-1)/4)) = 5. So a
> range of b[0] to b[4] is related to MAX_PRIO, isn't it?
Yup. What BITMAP_SIZE does is find the smallest array of longs that
would fit MAX_PRIO bits. That size is 5 on 32-bit machines, since
(5*32=160) > (MAX_PRIO=140) > (4*32=128).
And that b[0]..b[4] is exactly what sched_find_first_bit() checks. So
b[0] is the first 32-bits, corresponding to priority 0..31, etc.
> > Looking at the source helps a lot.
>
> Certainly.
>
> Thank you for your sharing your time to explain details.
No problem.
Good luck.
Rob Love
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive: http://mail.nl.linux.org/kernelnewbies/
FAQ: http://kernelnewbies.org/faq/
