Re: [PATCH net-next 14/15] net: dsa: sja1105: replace mdiobus-pcs with xpcs-plat driver

[Andy Shevchenko <andriy.shevchenko@linux.intel.com>]

On Wed, Nov 19, 2025 at 10:35:45AM -0800, Jakub Kicinski wrote:
> On Wed, 19 Nov 2025 19:39:32 +0200 Andy Shevchenko wrote:
> > On Wed, Nov 19, 2025 at 05:23:48PM +0000, Russell King (Oracle) wrote:
> > > On Wed, Nov 19, 2025 at 06:17:48PM +0200, Andy Shevchenko wrote:  
> > > > On Wed, Nov 19, 2025 at 08:11:12AM -0800, Jakub Kicinski wrote:  
> > > I can say that stuff such as unused const variables gets found by
> > > nipa, via -Wunused-const-variable, which as I understand it is a
> > > W=1 thing.
> > > 
> > > I suspect CONFIG_WERROR=y isn't used (I don't know) as that would
> > > stop the build on warnings, whereas what is done instead is that
> > > the output of the build is analysed, and the number of warnings
> > > counted and reported in patchwork. Note that the "build" stuff
> > > reports number of errors/warnings before and after the patch.
> > > 
> > > Hope this answers your question.  
> > 
> > Pretty much, thanks!
> 
> We do:

A-ha, thanks!

> prep_config() {
>   make LLVM=1 O=$output_dir allmodconfig
>   ./scripts/config --file $output_dir/.config -d werror
>   ./scripts/config --file $output_dir/.config -d drm_werror
>   ./scripts/config --file $output_dir/.config -d kvm_werror
> }
> 
> I have strong feelings about people who think Werror is an acceptable
> practice in 2025, but let me not violate the CoC here..

WERROR=y by default was introduced you-know-by-whom :-)

-- 
With Best Regards,
Andy Shevchenko