| Message ID | 20230601121001.1071533-8-masahiroy@kernel.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | modpost: fix section mismatch detection for ARM | expand |
On Thu, 1 Jun 2023 at 14:10, Masahiro Yamada <masahiroy@kernel.org> wrote: > > For ARM, modpost fails to detect some types of section mismatches. > > [test code] > > .section .init.data,"aw" > bar: > .long 0 > > .section .data,"aw" > .globl foo > foo: > .long bar - . > > It is apparently a bad reference, but modpost does not report anything. > > The test code above produces the following relocations. > > Relocation section '.rel.data' at offset 0xe8 contains 1 entry: > Offset Info Type Sym.Value Sym. Name > 00000000 00000403 R_ARM_REL32 00000000 .init.data > > Currently, R_ARM_REL32 is just skipped. > > Handle it like R_ARM_ABS32. OK, so the reason we can handle these in the same way is because we never calculate the resulting value, right? Because that value would be different for these cases. > > Signed-off-by: Masahiro Yamada <masahiroy@kernel.org> > --- > > scripts/mod/modpost.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c > index 55d142bb000b..9f0c87064ca5 100644 > --- a/scripts/mod/modpost.c > +++ b/scripts/mod/modpost.c > @@ -1281,6 +1281,7 @@ static int addend_arm_rel(struct elf_info *elf, Elf_Shdr *sechdr, Elf_Rela *r) > > switch (r_typ) { > case R_ARM_ABS32: > + case R_ARM_REL32: > inst = TO_NATIVE(*(uint32_t *)loc); > r->r_addend = inst + sym->st_value; > break; > -- > 2.39.2 >
On Thu, Jun 1, 2023 at 9:40 PM Ard Biesheuvel <ardb@kernel.org> wrote: > > On Thu, 1 Jun 2023 at 14:10, Masahiro Yamada <masahiroy@kernel.org> wrote: > > > > For ARM, modpost fails to detect some types of section mismatches. > > > > [test code] > > > > .section .init.data,"aw" > > bar: > > .long 0 > > > > .section .data,"aw" > > .globl foo > > foo: > > .long bar - . > > > > It is apparently a bad reference, but modpost does not report anything. > > > > The test code above produces the following relocations. > > > > Relocation section '.rel.data' at offset 0xe8 contains 1 entry: > > Offset Info Type Sym.Value Sym. Name > > 00000000 00000403 R_ARM_REL32 00000000 .init.data > > > > Currently, R_ARM_REL32 is just skipped. > > > > Handle it like R_ARM_ABS32. > > OK, so the reason we can handle these in the same way is because we > never calculate the resulting value, right? Because that value would > be different for these cases. Right. '- loc' is unnecessary here because modpost never calculates the resulting instruction. modpost wants to know the location of the referenced symbol. (the offset from the start of the section). For the same reason, I omitted '- loc' for PC-relative ones such as R_ARM_CALL, R_ARM_JUMP24, etc. -- Best Regards Masahiro Yamada
On Thu, 1 Jun 2023 at 16:36, Masahiro Yamada <masahiroy@kernel.org> wrote: > > On Thu, Jun 1, 2023 at 9:40 PM Ard Biesheuvel <ardb@kernel.org> wrote: > > > > On Thu, 1 Jun 2023 at 14:10, Masahiro Yamada <masahiroy@kernel.org> wrote: > > > > > > For ARM, modpost fails to detect some types of section mismatches. > > > > > > [test code] > > > > > > .section .init.data,"aw" > > > bar: > > > .long 0 > > > > > > .section .data,"aw" > > > .globl foo > > > foo: > > > .long bar - . > > > > > > It is apparently a bad reference, but modpost does not report anything. > > > > > > The test code above produces the following relocations. > > > > > > Relocation section '.rel.data' at offset 0xe8 contains 1 entry: > > > Offset Info Type Sym.Value Sym. Name > > > 00000000 00000403 R_ARM_REL32 00000000 .init.data > > > > > > Currently, R_ARM_REL32 is just skipped. > > > > > > Handle it like R_ARM_ABS32. > > > > OK, so the reason we can handle these in the same way is because we > > never calculate the resulting value, right? Because that value would > > be different for these cases. > > Right. > > '- loc' is unnecessary here because modpost never calculates the > resulting instruction. > > modpost wants to know the location of the referenced symbol. > (the offset from the start of the section). > > For the same reason, I omitted '- loc' for > PC-relative ones such as R_ARM_CALL, R_ARM_JUMP24, etc. > OK makes sense - I just wanted to double check
diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c index 55d142bb000b..9f0c87064ca5 100644 --- a/scripts/mod/modpost.c +++ b/scripts/mod/modpost.c @@ -1281,6 +1281,7 @@ static int addend_arm_rel(struct elf_info *elf, Elf_Shdr *sechdr, Elf_Rela *r) switch (r_typ) { case R_ARM_ABS32: + case R_ARM_REL32: inst = TO_NATIVE(*(uint32_t *)loc); r->r_addend = inst + sym->st_value; break;
For ARM, modpost fails to detect some types of section mismatches. [test code] .section .init.data,"aw" bar: .long 0 .section .data,"aw" .globl foo foo: .long bar - . It is apparently a bad reference, but modpost does not report anything. The test code above produces the following relocations. Relocation section '.rel.data' at offset 0xe8 contains 1 entry: Offset Info Type Sym.Value Sym. Name 00000000 00000403 R_ARM_REL32 00000000 .init.data Currently, R_ARM_REL32 is just skipped. Handle it like R_ARM_ABS32. Signed-off-by: Masahiro Yamada <masahiroy@kernel.org> --- scripts/mod/modpost.c | 1 + 1 file changed, 1 insertion(+)