linux-kernel - Re: [PATCH -next] modpost: move some defines to the file head

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <mhng-7876bf10-0477-4ca6-af22-e3479e8ad123@palmer-ri-x1c9a>
Date:   Wed, 12 Jul 2023 09:28:25 -0700 (PDT)
From:   Palmer Dabbelt <palmer@...osinc.com>
To:     masahiroy@...nel.org
CC:     wangkefeng.wang@...wei.com, mcgrof@...nel.org, nathan@...nel.org,
        ndesaulniers@...gle.com, nicolas@...sle.eu,
        linux-kbuild@...r.kernel.org, linux-kernel@...r.kernel.org
Subject:     Re: [PATCH -next] modpost: move some defines to the file head

On Wed, 12 Jul 2023 08:55:23 PDT (-0700), masahiroy@...nel.org wrote:
> +To: Luis Chamberlain, the commiter of the breakage
>
>
>
> On Wed, Jul 12, 2023 at 10:44 AM Kefeng Wang <wangkefeng.wang@...wei.com> wrote:
>>
>> with "module: Ignore RISC-V mapping symbols too", build error occurs,
>>
>> scripts/mod/modpost.c: In function ‘is_valid_name’:
>> scripts/mod/modpost.c:1055:57: error: ‘EM_RISCV’ undeclared (first use in this function)
>>   return !is_mapping_symbol(name, elf->hdr->e_machine == EM_RISCV);
>>
>> Fix it by moving the EM_RISCV to the file head, also some other
>> defines in case of similar problem in the future.
>
>
>
> BTW, why is the flag 'is_riscv' needed?
>
>
> All symbols starting with '$' look special to me.
>
>
>
> Why not like this?
>
>
>        if (str[0] == '$')
>                  return true;
>
>        return false;

There's a bit of commentary in the v1 
<https://lore.kernel.org/all/20230707054007.32591-1-palmer@rivosinc.com/>, 
but essentially it's not necessary.  I just wanted to play things safe 
and avoid changing the mapping symbol detection elsewhere in order to 
deal with RISC-V.

IIRC we decided $ was special in RISC-V because there were some other 
ports that behaved that way, but it wasn't universal.  If folks are OK 
treating $-prefixed symbols as special everywhere that's fine with me, I 
just wasn't sure what the right answer was.  

There's also some similar arch-specific-ness with the labels and such in 
here.

>> Signed-off-by: Kefeng Wang <wangkefeng.wang@...wei.com>
>> ---
>>  scripts/mod/modpost.c | 32 ++++++++++++++++----------------
>>  1 file changed, 16 insertions(+), 16 deletions(-)
>>
>> diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
>> index 7c71429d6502..885cca272eb8 100644
>> --- a/scripts/mod/modpost.c
>> +++ b/scripts/mod/modpost.c
>> @@ -60,6 +60,22 @@ static unsigned int nr_unresolved;
>>
>>  #define MODULE_NAME_LEN (64 - sizeof(Elf_Addr))
>>
>> +#ifndef EM_RISCV
>> +#define EM_RISCV               243
>> +#endif
>> +
>> +#ifndef R_RISCV_SUB32
>> +#define R_RISCV_SUB32          39
>> +#endif
>> +
>> +#ifndef EM_LOONGARCH
>> +#define EM_LOONGARCH           258
>> +#endif
>> +
>> +#ifndef R_LARCH_SUB32
>> +#define R_LARCH_SUB32          55
>> +#endif
>> +
>>  void __attribute__((format(printf, 2, 3)))
>>  modpost_log(enum loglevel loglevel, const char *fmt, ...)
>>  {
>> @@ -1428,22 +1444,6 @@ static int addend_mips_rel(uint32_t *location, Elf_Rela *r)
>>         return 0;
>>  }
>>
>> -#ifndef EM_RISCV
>> -#define EM_RISCV               243
>> -#endif
>> -
>> -#ifndef R_RISCV_SUB32
>> -#define R_RISCV_SUB32          39
>> -#endif
>> -
>> -#ifndef EM_LOONGARCH
>> -#define EM_LOONGARCH           258
>> -#endif
>> -
>> -#ifndef R_LARCH_SUB32
>> -#define R_LARCH_SUB32          55
>> -#endif
>> -
>>  static void section_rela(struct module *mod, struct elf_info *elf,
>>                          Elf_Shdr *sechdr)
>>  {
>> --
>> 2.41.0
>>