Re: [PATCH] break_lock forever broken

[Nick Piggin <nickpiggin@yahoo.com.au>]

Arjan van de Ven wrote:
>>Yes that's the tradeoff. I just feel that the former may be better,
>>especially because the latter can be timing dependant (so you may get
>>things randomly "happening"), and the former is apparently very low
>>overhead compared with the cost of taking the lock. Any numbers,
>>anyone?
> 
> 
> as I said, since the cacheline just got dirtied, the write is just half
> a cycle which is so much in the noise that it really doesn't matter.
> 

Yes, you were the "apparently" that I cited :)

I just wondered if Ingo has or has seen numbers that make
him dislike this way of doing it.

I would have thought that the spinlock structure and code bloat,
and the lock break checks in fast paths would be the far greater
cost of lockbreak than what Hugh's patch adds. But real numbers
are pretty important when it comes to this kind of thing.