Re: [PATCH rtw-next] wifi: rtw89: Replace hardcoded strings with helper functions

[Yu-Chun Lin <eleanor15x@gmail.com>]

On Thu, Sep 25, 2025 at 04:28:40AM +0000, Ping-Ke Shih wrote:
> Yu-Chun Lin <eleanor15x@gmail.com> wrote:
> > On Wed, Sep 24, 2025 at 12:43:56AM +0000, Ping-Ke Shih wrote:
> > > Yu-Chun Lin <eleanor15x@gmail.com> wrote:
> > > > On Mon, Sep 22, 2025 at 05:50:50AM +0000, Ping-Ke Shih wrote:
> > > > > Yu-Chun Lin <eleanor15x@gmail.com> wrote:
> > > > > > Replace hardcoded strings with 'str_on_off()', 'str_enable_disable()',
> > > > > > and 'str_read_write()'.
> > > > > >
> > > > > > The change improves readability and enables potential string deduplication
> > > > > > by the linker, which may slightly reduce binary size.
> > > > >
> > > > > Have you measured the change of binary size? Maybe you can share the result
> > > > > here.
> > > > >
> > > >
> > > > I tested the patch and did not observe any measurable change in the binary size.
> > > > The commit message was refers to the description in string-choice.h:
> > > >
> > > > "3) Deduping by the linker, which results in a smaller binary file."
> > >
> > > It might rely on LTO (link-time optimization).
> > >
> > 
> > I re-tested with 'LTO_CLANG=y', and the code size remains unchanged before
> > and after the patch. I will remove the statement
> > ("may slightly reduce binary size") from the v2 commit message.
> > 
> 
> What is the .o you checked? Have you also checked size of .ko?
> I guess (sorry I'm not familiar with LTO) that LTO only takes effect
> at least doing partial link. 
> 
> As the changes of your patch, maybe you can check rtw89_8851b.o or
> rtw89_8851b.ko.
>

I built all related files directly into the kernel, so my measurement was
on the size of the entire vmlinux, not individual .o files or modules.