[PATCH] always inline alloc_perturb.

[David Miller]

From: Roland McGrath <roland@hack.frob.com>
Date: Thu, 11 Dec 2014 12:52:09 -0800 (PST)

> What's the actual effect on code generation?  The warnings I've seen for
> this are because it had inlined, cloned, and outlined (so only the test was
> inlined), which is probably optimal.  In general, unless there is a strong
> demonstrated reason to force inlining, we should usually just drop the
> inline keyword and let the compiler decide.

+1