On Mon, Apr 17, 2006 at 09:00:39PM +0900, Atsushi Nemoto wrote: > 64bit mips has different r_info layout. This patch fixes modpost > segfault for 64bit little endian mips kernel. > > Signed-off-by: Atsushi Nemoto <anemo@xxxxxxxxxxxxx> > > diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c > index cd00e9f..7846600 100644 > --- a/scripts/mod/modpost.c > +++ b/scripts/mod/modpost.c > @@ -712,7 +712,13 @@ static void check_sec_ref(struct module > r.r_offset = TO_NATIVE(rela->r_offset); > r.r_info = TO_NATIVE(rela->r_info); > r.r_addend = TO_NATIVE(rela->r_addend); > +#if KERNEL_ELFCLASS == ELFCLASS64 && KERNEL_ELFDATA == ELFDATA2LSB > + sym = elf->symtab_start + > + (hdr->e_machine == EM_MIPS ? > + (Elf32_Word)r.r_info : ELF_R_SYM(r.r_info)); > +#else > sym = elf->symtab_start + ELF_R_SYM(r.r_info); > +#endif > /* Skip special sections */ > if (sym->st_shndx >= SHN_LORESERVE) > continue; Hi Atsushi So with 64bit mips with the above we read r.r_info with no modifications. But the elf.h I have on my system looks like this: /* How to extract and insert information held in the r_info field. */ #define ELF32_R_SYM(val) ((val) >> 8) #define ELF32_R_TYPE(val) ((val) & 0xff) #define ELF32_R_INFO(sym, type) (((sym) << 8) + ((type) & 0xff)) #define ELF64_R_SYM(i) ((i) >> 32) #define ELF64_R_TYPE(i) ((i) & 0xffffffff) #define ELF64_R_INFO(sym,type) ((((Elf64_Xword) (sym)) << 32) + (type)) So the difference between 64bit and 32bit is only the number of right shifts. Maybe we need to do some simple range check also so it becomes: if ((sym->st_shndx >= SHN_LORESERVE) || (sym > elf->symtab_stop)) continue; Could you try this - and if you still see the SEGVAL then try to print out the value of all members in 'r'. You can also try to add '-g' to HOSTCFLAGS in toplevel makefile. Then use make V=1 to see what arguments modpost is caleld with and call it from the commandline or from a debugger. Sam
