Re: [patch] IBM HDAPS accelerometer driver, with probing.

[Andi Kleen <ak@suse.de>]

"David S. Miller" <davem@davemloft.net> writes:

> From: Alexey Dobriyan <adobriyan@gmail.com>
> Date: Sat, 27 Aug 2005 02:58:48 +0400
> 
> > What's the point of having unlikely() attached to every possible if ()?
> 
> If can result in smaller code, for one thing, even if it
> isn't a performance critical path.

Really? At least on x86 it tends to generate bigger code when 
block reordering is enabled because a jump forward and a jump
backward and a possible label alignment are bigger than just
a single jump forward. But then it doesn't make that much difference
because the compiler does it on its own for every block.

On x86-64 I keep it disabled because:
- it generates bigger code
- it makes the assembly code unreadable
- it doesn't seem to help that much on modern CPUs with good
branch prediction and big icaches anyways.

-Andi (who originally introduced likely/unlikely, but regrets it these
days because it's far overused and makes code uglier everywhere)