Re: [PATCH] kbuild: install-extmod-build: Add missing python libraries

[Nicolas Schier <nsc@kernel.org>]

On Fri, Jan 30, 2026 at 09:32:03AM +0100, Mauro Carvalho Chehab wrote:
> On Thu, 29 Jan 2026 23:40:11 -0700
> Nathan Chancellor <nathan@kernel.org> wrote:
> 
> > On Fri, Jan 30, 2026 at 06:30:56AM +0100, Mauro Carvalho Chehab wrote:
> > > On Thu, 29 Jan 2026 18:11:06 -0700
> > > Nathan Chancellor <nathan@kernel.org> wrote:  
> > > > On Fri, Jan 30, 2026 at 01:49:55AM +0800, Rong Zhang wrote:  
> > ...
> > > > >   $ make -C /lib/modules/6.19.0-rc6/build/ M="$(pwd)" modules V=1 W=1
> > > > >   [...]
> > > > >   make -f /usr/src/linux-headers-6.19.0-rc6/scripts/Makefile.build obj=. need-builtin=1 need-modorder=1
> > > > >   # CC [M]  mod.o
> > > > >   [...]
> > > > >   # cmd_checkdoc mod.o
> > > > >   PYTHONDONTWRITEBYTECODE=1 python3 /usr/src/linux-headers-6.19.0-rc6/scripts/kernel-doc.py -none mod.c  
> > > 
> > > This sounds really weird, as it is trying to run scripts/kernel-doc.py
> > > instead of tools/docs/kernel-doc. Does the out-of-tree Makefile
> > > override KERNELDOC variable? The current version contains:
> > > 	
> > > 	KERNELDOC       = $(srctree)/tools/docs/kernel-doc
> > > 
> > > But somehow it is using the old version before the renames:
> > > 
> > > 	KERNELDOC       = $(srctree)/scripts/kernel-doc.py  
> > 
> > Well I think based on the "6.19.0-rc6" in the path above, this is
> > mainline, not -next, so the rename has has not happend there yet.
> 
> Ah, ok. On your e-mail you mentioned the renaming patch, so I
> ended assuming that was the case.
> 
> > 
> > > Btw, I did a very quick test here, using an old OOT project I have
> > > at github:
> > > 
> > > 	https://github.com/mchehab/xr_serial  
> > ...
> > > It sounds to me that Rong may be using a Makefile on his OOT project
> > > that was not updated to pick the right kernel-doc tool.  
> > 
> > If I use that project with Rong's original command, the
> > linux-upstream-headers package from the pacman-pkg target (which uses
> > install-extmod-build), and the following fix up for a more modern kernel
> > version, I see the following error:
> > 
> >   $ make -C /usr/lib/modules/6.19.0-rc7-next-20260129/build M=/tmp/xr_serial modules V=1 W=1
> >   ...
> >   # CC [M]  xr_serial.o
> >     gcc ...
> >   # cmd_checkdoc xr_serial.o
> >     PYTHONDONTWRITEBYTECODE=1 python3 /usr/lib/modules/6.19.0-rc7-next-20260129/build/tools/docs/kernel-doc -none   xr_serial.c
> >   python3: can't open file '/usr/lib/modules/6.19.0-rc7-next-20260129/build/tools/docs/kernel-doc': [Errno 2] No such file or directory
> >   make[3]: *** [/usr/lib/modules/6.19.0-rc7-next-20260129/build/scripts/Makefile.build:287: xr_serial.o] Error 2
> > 
> > If it is not expected that kernel-doc runs for external modules, then
> > maybe cmd_checkdoc should also be wrapped in a check for KBUILD_EXTMOD?
> 
> I don't particularly see any reason why running kernel-doc for external
> modules, as the goal of running it there (outside make htmldocs) is to
> detect early problems at linux-next and other CIs.
> 
> On the other hand, I also don't see any problem on encapsulating
> kernel-doc at the tarball created by install-extmod-build with something
> like:
> 
> 	ifeq ($(KBUILD_EXTMOD),)
> 		ifneq ($(KBUILD_EXTRA_WARN),)
> 		  cmd_checkdoc = PYTHONDONTWRITEBYTECODE=1 $(PYTHON3) $(KERNELDOC) -none $(KDOCFLAGS) \
> 	        	$(if $(findstring 2, $(KBUILD_EXTRA_WARN)), -Wall) \
> 		        $<
> 		endif
> 	endif
> 
> at scripts/Makefile.build.
> 
> 
> So, feel free to add to either one of the approaches:
> 
> Reviewed-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Thanks everybody for report and analysis!  I put Mauro's suggestion into
a new patch for kbuild-next:

https://lore.kernel.org/linux-kbuild/20260131-run-kernel-doc-only-in-tree-v1-1-93a9c695dfb1@kernel.org/

I plan to send the pull request tonight.

> That's said, it should be noticed that the DRM subsystem also has
> some similar code:
> 
> 	drivers/gpu/drm/Makefile:                PYTHONDONTWRITEBYTECODE=1 $(PYTHON3) $(KERNELDOC) -none $(if $(CONFIG_WERROR)$(CONFIG_DRM_WERROR),-Werror) $<; \
> 	drivers/gpu/drm/i915/Makefile:    cmd_checkdoc = PYTHONDONTWRITEBYTECODE=1 $(PYTHON3) $(KERNELDOC) -none -Werror $<
> 	drivers/gpu/drm/i915/Makefile:          $(KERNELDOC) -none -Werror $<; touch $@
> 	include/drm/Makefile:           PYTHONDONTWRITEBYTECODE=1 $(PYTHON3) $(KERNELDOC) -none $(if $(CONFIG_WERROR)$(CONFIG_DRM_WERROR),-Werror) $<; \
> 	scripts/Makefile.build:  cmd_checkdoc = PYTHONDONTWRITEBYTECODE=1 $(PYTHON3) $(KERNELDOC) -none $(KDOCFLAGS) \
> 
> but I suspect that only scripts/Makefile.build probably need
> such check.

yes, I think so too, as the definitions in drm should only take effect
on hdrtest target.

Kind regards,
Nicolas