Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20200316122820.11032-1-yangx.jy@cn.fujitsu.com> <5E702236.8070109@cn.fujitsu.com>
In-Reply-To: <5E702236.8070109@cn.fujitsu.com>
From:   Masahiro Yamada <masahiroy@kernel.org>
Date:   Wed, 18 Mar 2020 18:49:43 +0900
Message-ID: <CAK7LNARcs_ouquXPWJfKkfpikXV-4t=2YVCjEWeDzaNhFjmaeA@mail.gmail.com>
Subject: Re: [PATCH] modpost: Get proper section index by get_secindex()
 instead of st_shndx
To:     Xiao Yang <yangx.jy@cn.fujitsu.com>
Cc:     Michal Marek <michal.lkml@markovi.net>,
        Linux Kbuild mailing list <linux-kbuild@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Mar 17, 2020 at 10:05 AM Xiao Yang <yangx.jy@cn.fujitsu.com> wrote:
>
> On 2020/3/16 20:28, Xiao Yang wrote:

Thanks for catching this bug.


> > (uint16_t) st_shndx is limited to 65535(i.e. SHN_XINDEX) so sym_get_dat=
a() gets
> > wrong section index by st_shndx if object file(e.g. vmlinux.o) has more=
 than
> Hi,
>
> It seems better to say that sym_get_data() gets wrong section index by
> st_shndx if requested symbol contains extended section index that is
> more than 65535.


Sounds good to me.


> Thanks,
> Xiao Yang
> > 65535 sessions.  In this case, we need to get proper section index by .=
symtab_shndx
> > section.
> >
> > Module.symvers generated by building kernel with "-ffunction-sections -=
fdata-sections"
> > shows the issue(i.e. cannot get 89902 by st_shndx):
> > -------------------------------------------------------------------
> > [root@Fedora-30 linux]# file Module.symvers
> > Module.symvers: data
> > [root@Fedora-30 linux]# head -n1 Module.symvers
> > 0x5caf3011      ipv6_chk_custom_prefix  =E2=96=92=E2=96=92=E2=96=92=E2=
=96=92=E2=96=92=E2=96=92=E2=96=92=E2=96=92        vmlinux EXPORT_SYMBOL


Could you delete this dump?
I'd like to avoid garbling where possible.


> > ...
> > [root@Fedora-30 linux]# readelf -s -W vmlinux.o | grep __kstrtabns_ipv6=
_chk_custom_prefix
> > 199174: 0000000000032578     1 OBJECT  LOCAL  DEFAULT 89902 __kstrtabns=
_ipv6_chk_custom_prefix
> > [root@Fedora-30 linux]# readelf -S -W vmlinux.o | grep 89902
> >    [89902] __ksymtab_strings PROGBITS        0000000000000000 a94e00 03=
45a2 00   A  0   0  1
> > -------------------------------------------------------------------
> >
> > Fixes: afa0459daa7b ("modpost: add a helper to get data pointed by a sy=
mbol")

Strictly speaking, this commit does not introduce any bug.

The CRC bug for MODVERSIONS exists since
56067812d5b0 ("kbuild: modversions: add infrastructure for emitting
relative CRCs")


> > Fixes: 5545322c86d9 ("modpost: refactor namespace_from_kstrtabns() to n=
ot hard-code section name")

This commit hash is wrong.
The correct one is

Fixes: e84f9fbbece1 ("modpost: refactor namespace_from_kstrtabns() to
not hard-code section name")


> > Signed-off-by: Xiao Yang<yangx.jy@cn.fujitsu.com>
> > ---
> >   scripts/mod/modpost.c | 3 ++-
> >   1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
> > index d9418c58a8c0..c1fec8cac257 100644
> > --- a/scripts/mod/modpost.c
> > +++ b/scripts/mod/modpost.c
> > @@ -310,7 +310,8 @@ static const char *sec_name(struct elf_info *elf, i=
nt secindex)
> >
> >   static void *sym_get_data(const struct elf_info *info, const Elf_Sym =
*sym)
> >   {
> > -     Elf_Shdr *sechdr =3D&info->sechdrs[sym->st_shndx];
> > +     unsigned int secindex =3D get_secindex(info, sym);
> > +     Elf_Shdr *sechdr =3D&info->sechdrs[secindex];
> >       unsigned long offset;
> >
> >       offset =3D sym->st_value;
>
>
>


--
Best Regards
Masahiro Yamada