Re: [PATCH] vmlinux.lds.h: Handle clang's module.{c,d}tor sections

[Fangrui Song <maskray@google.com>]

On 2021-07-30, Nick Desaulniers wrote:
>On Fri, Jul 30, 2021 at 3:38 PM Nathan Chancellor <nathan@kernel.org> wrote:
>>
>> A recent change in LLVM causes module_{c,d}tor sections to appear when
>> CONFIG_K{A,C}SAN are enabled, which results in orphan section warnings
>> because these are not handled anywhere:
>>
>> ld.lld: warning: arch/x86/pci/built-in.a(legacy.o):(.text.asan.module_ctor) is being placed in '.text.asan.module_ctor'
>> ld.lld: warning: arch/x86/pci/built-in.a(legacy.o):(.text.asan.module_dtor) is being placed in '.text.asan.module_dtor'
>> ld.lld: warning: arch/x86/pci/built-in.a(legacy.o):(.text.tsan.module_ctor) is being placed in '.text.tsan.module_ctor'
>
>^ .text.tsan.*

I was wondering why the orphan section warning only arose recently.
Now I see: the function asan.module_ctor has the SHF_GNU_RETAIN flag, so
it is in a separate section even with -fno-function-sections (default).

It seems that with -ffunction-sections the issue should have been caught
much earlier.

>>
>> Place them in the TEXT_TEXT section so that these technologies continue
>> to work with the newer compiler versions. All of the KASAN and KCSAN
>> KUnit tests continue to pass after this change.
>>
>> Cc: stable@vger.kernel.org
>> Link: https://github.com/ClangBuiltLinux/linux/issues/1432
>> Link: https://github.com/llvm/llvm-project/commit/7b789562244ee941b7bf2cefeb3fc08a59a01865
>> Signed-off-by: Nathan Chancellor <nathan@kernel.org>
>> ---
>>  include/asm-generic/vmlinux.lds.h | 1 +
>>  1 file changed, 1 insertion(+)
>>
>> diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
>> index 17325416e2de..3b79b1e76556 100644
>> --- a/include/asm-generic/vmlinux.lds.h
>> +++ b/include/asm-generic/vmlinux.lds.h
>> @@ -586,6 +586,7 @@
>>                 NOINSTR_TEXT                                            \
>>                 *(.text..refcount)                                      \
>>                 *(.ref.text)                                            \
>> +               *(.text.asan .text.asan.*)                              \
>
>Will this match .text.tsan.module_ctor?

asan.module_ctor is the only function AddressSanitizer synthesizes in the instrumented translation unit.
There is no function called "asan".

(Even if a function "asan" exists due to -ffunction-sections
-funique-section-names, TEXT_MAIN will match .text.asan, so the
.text.asan pattern will match nothing.)

>Do we want to add these conditionally on
>CONFIG_KASAN_GENERIC/CONFIG_KCSAN like we do for SANITIZER_DISCARDS?
>
>>                 TEXT_CFI_JT                                             \
>>         MEM_KEEP(init.text*)                                            \
>>         MEM_KEEP(exit.text*)                                            \
>>
>> base-commit: 4669e13cd67f8532be12815ed3d37e775a9bdc16
>> --
>
>
>-- 
>Thanks,
>~Nick Desaulniers