Hi Steve,
On Thu, Feb 07, 2019 at 05:52:39PM -0500, Steven Rostedt wrote:
> On Thu, 7 Feb 2019 16:11:01 -0500
> "Joel Fernandes (Google)" <joel@xxxxxxxxxxxxxxxxx> wrote:
>
> > +
> > +# Build a list of in-kernel headers for building kernel modules
> > +# Any other files will be stored in IKH_EXTRA variable.
> > +ikh_file_list := include/
> > +ikh_file_list += arch/$(ARCH)/Makefile
> > +ikh_file_list += arch/$(ARCH)/include/
> > +ikh_file_list += $(IKH_EXTRA)
> > +ikh_file_list += scripts/
> > +ikh_file_list += Makefile
> > +ikh_file_list += Module.symvers
> > +ifeq ($(CONFIG_STACK_VALIDATION), y)
> > +ikh_file_list += $(objtree)/tools/objtool/objtool
> > +endif
> > +
> > +$(obj)/kheaders.o: $(obj)/kheaders_data.h
> > +
> > +targets += kheaders_data.txz
> > +
> > +quiet_cmd_genikh = GEN $(obj)/kheaders_data.txz
> > +cmd_genikh = $(srctree)/scripts/gen_ikh_data.sh $@ $^ >/dev/null 2>&1
> > +$(obj)/kheaders_data.txz: $(ikh_file_list) FORCE
> > + $(call cmd,genikh)
> > +
> > +filechk_ikheadersxz = (echo "static const char kernel_headers_data[] __used = KH_MAGIC_START"; cat $< | scripts/bin2c; echo "KH_MAGIC_END;")
> > +
> > +targets += kheaders_data.h
> > +$(obj)/kheaders_data.h: $(obj)/kheaders_data.txz FORCE
> > + $(call filechk,ikheadersxz)
> > diff --git a/scripts/gen_ikh_data.sh b/scripts/gen_ikh_data.sh
> > new file mode 100755
> > index 000000000000..609196b5cea2
> > --- /dev/null
> > +++ b/scripts/gen_ikh_data.sh
> > @@ -0,0 +1,19 @@
> > +#!/bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +
> > +spath="$(dirname "$(readlink -f "$0")")"
> > +
> > +rm -rf $1.tmp
> > +mkdir $1.tmp
> > +
> > +for f in "${@:2}";
> > + do find "$f" ! -name "*.c" ! -name "*.o" ! -name "*.cmd" ! -name ".*";
>
> I wonder if it is a good idea to pick all files in the directories
> defined in ikh_file_list, and not just explicitly list what we want,
> with a '*.h' and such?
I also need few files in the archive that are not .h, these don't take up
much space but are needed to make an out-of-tree kernel module build succeed.
One of my goals with this was to make a self-contained module that could be
loaded to build other modules. Majority of the files are kernel headers, but
some are not, such as Module.symvers and other scripts. Then one can run
systemtap on Android which can be made to build modules using the embedded
headers.
> > +done | cpio -pd $1.tmp
> > +
> > +for f in $(find $1.tmp); do
> > + $spath/strip-comments.pl $f
> > +done
> > +
> > +tar -Jcf $1 -C $1.tmp/ . > /dev/null
> > +
> > +rm -rf $1.tmp
> > diff --git a/scripts/strip-comments.pl b/scripts/strip-comments.pl
> > new file mode 100755
> > index 000000000000..f8ada87c5802
> > --- /dev/null
> > +++ b/scripts/strip-comments.pl
> > @@ -0,0 +1,8 @@
> > +#!/usr/bin/perl -pi
> > +# SPDX-License-Identifier: GPL-2.0
> > +
> > +# This script removes /**/ comments from a file, unless such comments
> > +# contain "SPDX". It is used when building compressed in-kernel headers.
> > +
> > +BEGIN {undef $/;}
> > +s/\/\*((?!SPDX).)*?\*\///smg;
>
> Hmm, I'm also wondering if we could us the C pre-processor for the
> stripping of everything from the header file. We would then even get
> the header files only having what is necessary for the running kernel.
I thought about this too. An issue with that is it is going to be really slow
due to the large number of headers. The other is, I think it will actually
make the headers bigger and take up more space - because all the include
directives will also be expanded and have more duplication. Let me know if I
missed something though.
thanks,
- Joel
