Re: [PATCH] drivers/firmware: const-ify DMI API and internals
From: Satyam Sharma <hidden>
Date: 2007-09-01 21:09:33
Also in:
lkml
From: Satyam Sharma <hidden>
Date: 2007-09-01 21:09:33
Also in:
lkml
On Sat, 1 Sep 2007, Andi Kleen wrote:
quoted
And if we're really lucky, this might enable some additional optimizations on the part of the compiler.Only if the kernel was compiled C++. C compilers generally ignore constness for optimization purposes because it can be so easily casted away.
??? I can't speak for the exact const-ification that Jeff did in this particular patch, but if you intended your comment up there to be a blanket statement, then you're totally and horribly wrong. There's a whole plethora of optimizations that gcc does when dealing with const- marked data. If you've seen / experienced otherwise, please post some testcases.