From: Thiemo Seufer <ths@networkno.de>
To: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: linux-mips@linux-mips.org, ralf@linux-mips.org, sam@ravnborg.org
Subject: Re: [PATCH] fix modpost segfault for 64bit mipsel kernel
Date: Mon, 17 Apr 2006 17:27:42 +0100 [thread overview]
Message-ID: <20060417162742.GD28935@networkno.de> (raw)
In-Reply-To: <20060418.004718.108120830.anemo@mba.ocn.ne.jp>
Atsushi Nemoto wrote:
> On Mon, 17 Apr 2006 15:07:35 +0100, Thiemo Seufer <ths@networkno.de> wrote:
> > I should have read more carefully. The ELF_R_SYM seems to be correct, if
> > this patch makes it work fo you then the toolchain you use creates broken
> > (word-swapped ?) relocation entries for mips64el.
>
> Looking at following codes in glibc source
> (sysdeps/mips/elf/ldsodefs.h), I thought r_info on 64bit mips needs
> special handling. Is not this structure used for 64bit kernel
> modules?
>
> typedef struct
> {
> Elf32_Word r_sym; /* Symbol index */
> unsigned char r_ssym; /* Special symbol for 2nd relocation */
> unsigned char r_type3; /* 3rd relocation type */
> unsigned char r_type2; /* 2nd relocation type */
> unsigned char r_type1; /* 1st relocation type */
> } _Elf64_Mips_R_Info;
Hm, binutils uses generically 64bit quantities:
#define ELF32_R_SYM(i) ((i) >> 8)
#define ELF32_R_TYPE(i) ((i) & 0xff)
#define ELF32_R_INFO(s,t) (((s) << 8) + ((t) & 0xff))
#define ELF64_R_SYM(i) ((i) >> 32)
#define ELF64_R_TYPE(i) ((i) & 0xffffffff)
#define ELF64_R_INFO(s,t) (((bfd_vma) (s) << 31 << 1) + (bfd_vma) (t))
But for MIPS64 the same as glibc:
typedef struct
{
/* Address of relocation. */
unsigned char r_offset[8];
/* Symbol index. */
unsigned char r_sym[4];
/* Special symbol. */
unsigned char r_ssym[1];
/* Third relocation. */
unsigned char r_type3[1];
/* Second relocation. */
unsigned char r_type2[1];
/* First relocation. */
unsigned char r_type[1];
/* Addend. */
unsigned char r_addend[8];
} Elf64_Mips_External_Rela;
/* MIPS ELF 64 relocation info access macros. */
#define ELF64_MIPS_R_SSYM(i) (((i) >> 24) & 0xff)
#define ELF64_MIPS_R_TYPE3(i) (((i) >> 16) & 0xff)
#define ELF64_MIPS_R_TYPE2(i) (((i) >> 8) & 0xff)
#define ELF64_MIPS_R_TYPE(i) ((i) & 0xff)
So it is the
r.r_info = TO_NATIVE(rela->r_info);
in modpost.c which breaks both SYM and TYPE because it assumes a
64bit integer. The proper solution would be to add a Elf64_Mips_Rela
structure (with lots of nearly identical duplicated code), the hack
would be to cast r_info to a 32bit integer for mips, before feeding
it to TO_NATIVE (which works until somebody asks for the TYPE, then
a separate mips64 version becomes inevitable.)
Thiemo
next prev parent reply other threads:[~2006-04-17 16:15 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-04-17 12:00 [PATCH] fix modpost segfault for 64bit mipsel kernel Atsushi Nemoto
2006-04-17 12:53 ` Thiemo Seufer
2006-04-17 14:07 ` Thiemo Seufer
2006-04-17 15:47 ` Atsushi Nemoto
2006-04-17 16:27 ` Thiemo Seufer [this message]
2006-04-19 2:22 ` Atsushi Nemoto
2006-04-20 0:19 ` Thiemo Seufer
2006-04-20 16:02 ` Atsushi Nemoto
2006-04-20 16:23 ` Thiemo Seufer
2006-04-20 17:05 ` Atsushi Nemoto
2006-04-21 13:57 ` Thiemo Seufer
2006-04-17 13:40 ` Sam Ravnborg
2006-04-17 16:03 ` Atsushi Nemoto
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20060417162742.GD28935@networkno.de \
--to=ths@networkno.de \
--cc=anemo@mba.ocn.ne.jp \
--cc=linux-mips@linux-mips.org \
--cc=ralf@linux-mips.org \
--cc=sam@ravnborg.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox