[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ZO1rHGjch2s1Mitt@bergen.fjasle.eu>
Date: Tue, 29 Aug 2023 05:50:52 +0200
From: Nicolas Schier <nicolas@...sle.eu>
To: Masahiro Yamada <masahiroy@...nel.org>
Cc: linux-kbuild@...r.kernel.org, linux-kernel@...r.kernel.org,
Nathan Chancellor <nathan@...nel.org>,
Nick Desaulniers <ndesaulniers@...gle.com>
Subject: Re: [PATCH 6/8] kbuild: move more module installation code to
scripts/Makefile.modinst
On Tue 29 Aug 2023 11:35:41 GMT, Masahiro Yamada wrote:
> On Tue, Aug 29, 2023 at 11:15 AM Nicolas Schier <nicolas@...sle.eu> wrote:
> >
> > On Wed 23 Aug 2023 20:50:46 GMT, Masahiro Yamada wrote:
> > > Move more relevant code to scripts/Makefile.modinst.
> > >
> > > Signed-off-by: Masahiro Yamada <masahiroy@...nel.org>
> > > ---
> > >
> > > Makefile | 34 +++++++--------------------------
> > > scripts/Makefile.modinst | 41 +++++++++++++++++++++++++++++++++++++---
> > > 2 files changed, 45 insertions(+), 30 deletions(-)
> > >
> > > diff --git a/Makefile b/Makefile
> > > index 7d9cab3d2186..82d22debf6c9 100644
> > > --- a/Makefile
> > > +++ b/Makefile
> > > @@ -1477,24 +1477,6 @@ endif
> > >
> > > endif # CONFIG_MODULES
> > >
> > > -modinst_pre :=
> > > -ifneq ($(filter modules_install,$(MAKECMDGOALS)),)
> > > -modinst_pre := __modinst_pre
> > > -endif
> > > -
> > > -modules_install: $(modinst_pre)
> > > -PHONY += __modinst_pre
> > > -__modinst_pre:
> > > - @rm -rf $(MODLIB)/kernel
> > > - @rm -f $(MODLIB)/build
> > > - @mkdir -p $(MODLIB)
> > > -ifdef CONFIG_MODULES
> > > - @ln -s $(CURDIR) $(MODLIB)/build
> > > - @sed 's:^\(.*\)\.o$$:kernel/\1.ko:' modules.order > $(MODLIB)/modules.order
> > > -endif
> > > - @cp -f modules.builtin $(MODLIB)/
> > > - @cp -f $(objtree)/modules.builtin.modinfo $(MODLIB)/
> > > -
> > > ###
> > > # Cleaning is done on three levels.
> > > # make clean Delete most generated files
> > > @@ -1836,12 +1818,15 @@ help:
> > > @echo ' clean - remove generated files in module directory only'
> > > @echo ''
> > >
> > > +ifndef CONFIG_MODULES
> > > +modules modules_install: __external_modules_error
> > > __external_modules_error:
> > > @echo >&2 '***'
> > > @echo >&2 '*** The present kernel disabled CONFIG_MODULES.'
> > > @echo >&2 '*** You cannot build or install external modules.'
> > > @echo >&2 '***'
> > > @false
> > > +endif
> > >
> > > endif # KBUILD_EXTMOD
> > >
> > > @@ -1850,6 +1835,9 @@ endif # KBUILD_EXTMOD
> > >
> > > PHONY += modules modules_install modules_prepare
> > >
> > > +modules_install:
> > > + $(Q)$(MAKE) -f $(srctree)/scripts/Makefile.modinst
> >
> > I was a bit surprised to see 'modules_install' being allowed
> > unconditionally for in-tree usage (thus, even if CONFIG_MODULES=n), but
> > then realised that this is the same behaviour as we had before. Out of
> > curiosity: _why_ do we need to install
> > $(MODLIB)/modules.builtin{,.modinfo} also for configs w/
> > CONFIG_MODULES=n?
>
>
> I see your tags in commit
> 8ae071fc216a25f4f797f33c56857f4dd6b4408e :)
>
>
> Some drivers need to load firmware.
>
> To make such drivers working in initrd,
> mkinitramfs needs to copy necessary firmware files
> into the initrd.
> So, the tool needs to know which drivers are enabled.
>
> That is my understanding why modules.builtin(.modinfo)
> is needed even with CONFIG_MODULES=n.
Ups, yes. Thanks for the reminder!
Kind regards,
Nicolas
Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)
Powered by blists - more mailing lists