Re: [PATCH] Performance regression in 2.6.7-rc3

[Linus Torvalds]

On Wed, 16 Jun 2004, Nick Piggin wrote:
> 
> I think balance on clone probably needs to be turned off by default
> presently.
> 
> It slows down a simple thread creation test by a factor of 7(!) here,
> and has quite a few not too difficult to imagine performance problems.

I agree. However, I still think we should do my suggested
"wake_up_new(p,clone_flags)" thing, and then have the logic on whether to 
try to care about threading or not be in schedule.c, not in kernel/fork.c.

The fact is, fork.c shouldn't try to make scheduling decisions. But it 
could inform the scheduler about the new process, and THAT can then make 
the decisions.

		Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/