Messages in this thread | | | Date | Thu, 20 Mar 2014 05:41:00 -0700 | Subject | Re: [PATCH 1/5] mtd: st_spi_fsm: Remove useless consts from function arguments | From | Brian Norris <> |
| |
On Thu, Mar 20, 2014 at 5:13 AM, Joe Perches <joe@perches.com> wrote: > On Thu, 2014-03-20 at 12:03 +0000, Lee Jones wrote: >> > Good catch. That wasn't my intention. >> > > Why? >> An oversight. > > That's still not an explanation. > > Why, unless cast away by the code itself, is > const removal a good thing?
It's not so much removal as it is review of the initial driver merge. I'd contend that const was applied somewhat thoughtlessly originally, and it didn't really serve a good purpose.
> It does serve as an indication to a reader what > the code does with the argument. > > About the only reason I can think of arguing in > favor of removal is inconsistent application of > const within the module.
That's one good reason. And not only consistency within the modules, but consistency within the subsystem (and the kernel at large, really). There's rarely a case of a const function parameter. And I'm sure there are numerous function parameters which could potentially be marked 'const'.
I also don't think that a function parameter is the right place to mark const like this. Function arguments are always pass-by-value, so this 'const' tells users (callers) nothing useful. It only provides useless constraints on what the function can do with its copy of the parameter.
Brian
| |