Re: [PATCH] script: modpost: require a MODULE_DESCRIPTION()

[Guenter Roeck]

On 4/6/25 09:06, Arnd Bergmann wrote:
> On Sun, Apr 6, 2025, at 17:59, Guenter Roeck wrote:
> > On Tue, Mar 11, 2025 at 12:49:02PM -0700, Jeff Johnson wrote:
> > > Since commit 1fffe7a34c89 ("script: modpost: emit a warning when the
> > > description is missing"), a module without a MODULE_DESCRIPTION() has
> > > resulted in a warning with make W=1. Since that time, all known
> > > instances of this issue have been fixed. Therefore, now make it an
> > > error if a MODULE_DESCRIPTION() is not present.
> > >
> > > Signed-off-by: Jeff Johnson <jeff.johnson@xxxxxxxxxxxxxxxx>
> > > ---
> > > did my treewide cleanup for v6.11, Arnd had a few more stragglers that
> > > he was going to fix. I hope that by posting, some of the 0-day bots
> > > will pick it up and hopefully provide some feedback.
> > >
> > > Note: I'm not really sure if *all* of these have been fixed. After I
> >
> > FWIW, I ran
> >
> > for f in $(find . -name '*.c'); do grep -q MODULE_LICENSE $f && (grep
> > -q MODULE_DESCRIPTION $f || echo $f); done
> >
> > That reports a large number of files (111, to be exact) with MODULE_LICENSE
> > but not MODULE_DESCRIPTION. I cross-checked a few, and found that many can
> > be built as module. The fall-out from this patch might be interesting.
>
> The ones I sent were only the result of what I found from randconfig
> builds on x86, arm32 and arm64, so I likely missed drivers that are
> specific to other architectures, or that are very unlikely to be
> in loadable modules because of random 'select FOO' Kconfig statements
> leading to them being always built-in in practice.
>
>          Arnd

ERROR: modpost: missing MODULE_DESCRIPTION() in drivers/block/swim3.o

Guess I'll just keep disabling affected drivers.

Guenter