linux-kernel - Re: [PATCH] kbuild: do not automatically add -w option to modpost

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <Y9AXytjc61Le0PQZ@google.com>
Date:   Tue, 24 Jan 2023 09:39:22 -0800
From:   William McVicker <willmcvicker@...gle.com>
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>,
        Nicolas Schier <nicolas@...sle.eu>
Subject: Re: [PATCH] kbuild: do not automatically add -w option to modpost

On 01/24/2023, Masahiro Yamada wrote:
> On Tue, Jan 24, 2023 at 7:42 AM William McVicker
> <willmcvicker@...gle.com> wrote:
> >
> > On 01/23/2023, Masahiro Yamada wrote:
> > > When there is a missing input file (vmlinux.o or Module.symvers), you
> > > are likely to get a ton of unresolved symbols.
> > >
> > > Currently, Kbuild automatically adds the -w option to allow module builds
> > > to continue with warnings instead of errors.
> > >
> > > This may not be what the user expects because it is generally more useful
> > > to catch all possible issues at build time instead of at run time.
> > >
> > > Let's not do what the user did not ask.
> > >
> > > If you still want to build modules anyway, you can proceed by explicitly
> > > setting KBUILD_MODPOST_WARN=1. Since you may miss a real issue, you need
> > > to be aware of what you are doing.
> > >
> > > Suggested-by: William McVicker <willmcvicker@...gle.com>
> > > Signed-off-by: Masahiro Yamada <masahiroy@...nel.org>
> > > ---
> > >
> > >  scripts/Makefile.modpost | 8 +++-----
> > >  1 file changed, 3 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/scripts/Makefile.modpost b/scripts/Makefile.modpost
> > > index 43343e13c542..9254ed811ddd 100644
> > > --- a/scripts/Makefile.modpost
> > > +++ b/scripts/Makefile.modpost
> > > @@ -121,16 +121,14 @@ modpost-args += -e $(addprefix -i , $(KBUILD_EXTRA_SYMBOLS))
> > >
> > >  endif # ($(KBUILD_EXTMOD),)
> > >
> > > -ifneq ($(missing-input),)
> > > -modpost-args += -w
> > > -endif
> > > -
> > >  quiet_cmd_modpost = MODPOST $@
> > >        cmd_modpost = \
> > >       $(if $(missing-input), \
> > >               echo >&2 "WARNING: $(missing-input) is missing."; \
> > >               echo >&2 "         Modules may not have dependencies or modversions."; \
> > > -             echo >&2 "         You may get many unresolved symbol warnings.";) \
> > > +             echo >&2 "         You may get many unresolved symbol errors.";) \
> >
> > You need to move the closing parenthesis to come at the end of these
> > echo messages. Otherwise you get this new message unconditionally.
> 
> 
> Ah, thanks for catching it.
> 
> 
> > I also found during testing that the refactoring in commit f73edc8951b2
> > ("kbuild: unify two modpost invocations") dropped the check for missing
> > KBUILD_EXTRA_SYMBOLS. That means if an external module depends on
> > another external module and sets:
> >
> >   KBUILD_EXTRA_SYMBOLS=/path/to/ext_module/Module.symvers
> >
> > ... then make will fail even with KBUILD_MODPOST_WARN=1 since we
> > unconditionally add KBUILD_EXTRA_SYMBOLS to the modpost-args like this:
> >
> >   modpost-args += -e $(addprefix -i , $(KBUILD_EXTRA_SYMBOLS))
> >
> > To fix this, I suggest you also take the following patch so that
> > KBUILD_MODPOST_WARN=1 will allow you to skip those unresolved symbols as
> > well:
> 
> 
> How is this useful?
> 
> KBUILD_EXTRA_SYMBOLS is explicitly specified by the user
> via the command line or the environment variable.
> 
> If $(KBUILD_EXTRA_SYMBOLS) does not exist,
> it is a user's fault, isn't it?

Sort of, yes. One could argue that it's the same situation as missing
the in-tree Module.symvers or the vmlinuux.o/vmlinux.symvers. Basically,
if you keep it as is, then KBUILD_MODPOST_WARN=1 would only work if the
user edits the Makefile to remove the KBUILD_EXTRA_SYMBOLS line.  With
my suggested patch, the user could build the module as is without any
dependencies by setting KBUILD_MODPOST_WARN=1.

I'm fine with whichever way you choose to support since I know this is
an external modules edge-case. I personally like to build my modules
with all the dependencies present to try an catch any issues at build
time.

Thanks,
Will

> 
> 
> 
> 
> 
> 
> -- 
> Best Regards
> Masahiro Yamada