[PATCH] modpost: Get proper section index by get_secindex() instead of st_shndx

[PATCH] modpost: Get proper section index by get_secindex() instead of st_shndx

[Xiao Yang]

(uint16_t) st_shndx is limited to 65535(i.e. SHN_XINDEX) so sym_get_data() gets
wrong section index by st_shndx if object file(e.g. vmlinux.o) has more than
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  ▒▒▒▒▒▒▒▒        vmlinux EXPORT_SYMBOL
...
[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 0345a2 00   A  0   0  1
-------------------------------------------------------------------

Fixes: afa0459daa7b ("modpost: add a helper to get data pointed by a symbol")
Fixes: 5545322c86d9 ("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, int secindex)
 
 static void *sym_get_data(const struct elf_info *info, const Elf_Sym *sym)
 {
-	Elf_Shdr *sechdr = &info->sechdrs[sym->st_shndx];
+	unsigned int secindex = get_secindex(info, sym);
+	Elf_Shdr *sechdr = &info->sechdrs[secindex];
 	unsigned long offset;
 
 	offset = sym->st_value;
-- 
2.21.0

Re: [PATCH] modpost: Get proper section index by get_secindex() instead of st_shndx

[Xiao Yang]

On 2020/3/16 20:28, Xiao Yang wrote:
> (uint16_t) st_shndx is limited to 65535(i.e. SHN_XINDEX) so sym_get_data() 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.

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  ▒▒▒▒▒▒▒▒        vmlinux EXPORT_SYMBOL
> ...
> [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 0345a2 00   A  0   0  1
> -------------------------------------------------------------------
>
> Fixes: afa0459daa7b ("modpost: add a helper to get data pointed by a symbol")
> Fixes: 5545322c86d9 ("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, int secindex)
>
>   static void *sym_get_data(const struct elf_info *info, const Elf_Sym *sym)
>   {
> -	Elf_Shdr *sechdr =&info->sechdrs[sym->st_shndx];
> +	unsigned int secindex = get_secindex(info, sym);
> +	Elf_Shdr *sechdr =&info->sechdrs[secindex];
>   	unsigned long offset;
>
>   	offset = sym->st_value;

Re: [PATCH] modpost: Get proper section index by get_secindex() instead of st_shndx

[Masahiro Yamada]

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_data() 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  ▒▒▒▒▒▒▒▒        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 0345a2 00   A  0   0  1
> > -------------------------------------------------------------------
> >
> > Fixes: afa0459daa7b ("modpost: add a helper to get data pointed by a symbol")

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 not 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, int secindex)
> >
> >   static void *sym_get_data(const struct elf_info *info, const Elf_Sym *sym)
> >   {
> > -     Elf_Shdr *sechdr =&info->sechdrs[sym->st_shndx];
> > +     unsigned int secindex = get_secindex(info, sym);
> > +     Elf_Shdr *sechdr =&info->sechdrs[secindex];
> >       unsigned long offset;
> >
> >       offset = sym->st_value;
>
>
>


--
Best Regards
Masahiro Yamada

Re: [PATCH] modpost: Get proper section index by get_secindex() instead of st_shndx

[Xiao Yang]

On 2020/3/18 17:49, Masahiro Yamada wrote:
> 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_data() 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  ▒▒▒▒▒▒▒▒        vmlinux EXPORT_SYMBOL
>
>
> Could you delete this dump?
> I'd like to avoid garbling where possible.
>
Hi Masahiro,

Sure, it is OK to delete it.

>
>>> ...
>>> [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 0345a2 00   A  0   0  1
>>> -------------------------------------------------------------------
>>>
>>> Fixes: afa0459daa7b ("modpost: add a helper to get data pointed by a symbol")
>
> 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")
>

Will replace it with your suggested commit.

>
>>> Fixes: 5545322c86d9 ("modpost: refactor namespace_from_kstrtabns() to not 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")
>
>

Sorry for the wrong commit hash.

Thanks a lot for your quick review.
I will send v2 patch soon.

Best Regards,
Xiao Yang

>
>
>
>>> 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, int secindex)
>>>
>>>    static void *sym_get_data(const struct elf_info *info, const Elf_Sym *sym)
>>>    {
>>> -     Elf_Shdr *sechdr =&info->sechdrs[sym->st_shndx];
>>> +     unsigned int secindex = get_secindex(info, sym);
>>> +     Elf_Shdr *sechdr =&info->sechdrs[secindex];
>>>        unsigned long offset;
>>>
>>>        offset = sym->st_value;
>>
>>
>>
>
>
> --
> Best Regards
> Masahiro Yamada
>
>
> .
>