Re: Use of yield() in the kernel

[Duncan Sands <baldrick@wanadoo.fr>]

On Tuesday 22 October 2002 19:24, Mark Mielke wrote:
> Would it be sensible to add a "yield_short()" function to the kernel?

You can do:

set_current_state(TASK_RUNNING);
schedule();

I'm not clear on what you are guaranteed to get - for example,
it looks as if it can sometimes return without sleeping at all.
There is also cond_resched(), which is

static inline void cond_resched(void)
{
        if (need_resched())
                __cond_resched();
}

void __cond_resched(void)
{
        set_current_state(TASK_RUNNING);
        schedule();
}

I don't know when need_resched evaluates to true, I haven't
had time to look into this yet.

Ciao, Duncan.