Re: [PATCH 5/5] gcc-plugins/stackleak: Don't instrument vgettimeofday.c in arm64 VDSO

[Alexander Popov]

On 04.06.2020 17:14, Jann Horn wrote:
> On Thu, Jun 4, 2020 at 3:58 PM Will Deacon <will@xxxxxxxxxx> wrote:
>> On Thu, Jun 04, 2020 at 04:49:57PM +0300, Alexander Popov wrote:
>>> Don't try instrumenting functions in arch/arm64/kernel/vdso/vgettimeofday.c.
>>> Otherwise that can cause issues if the cleanup pass of stackleak gcc plugin
>>> is disabled.
>>>
>>> Signed-off-by: Alexander Popov <alex.popov@xxxxxxxxx>
>>> ---
>>>  arch/arm64/kernel/vdso/Makefile | 3 ++-
>>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/arch/arm64/kernel/vdso/Makefile b/arch/arm64/kernel/vdso/Makefile
>>> index 3862cad2410c..9b84cafbd2da 100644
>>> --- a/arch/arm64/kernel/vdso/Makefile
>>> +++ b/arch/arm64/kernel/vdso/Makefile
>>> @@ -32,7 +32,8 @@ UBSAN_SANITIZE                      := n
>>>  OBJECT_FILES_NON_STANDARD    := y
>>>  KCOV_INSTRUMENT                      := n
>>>
>>> -CFLAGS_vgettimeofday.o = -O2 -mcmodel=tiny -fasynchronous-unwind-tables
>>> +CFLAGS_vgettimeofday.o = -O2 -mcmodel=tiny -fasynchronous-unwind-tables \
>>> +             $(DISABLE_STACKLEAK_PLUGIN)
>>
>> I can pick this one up via arm64, thanks. Are there any other plugins we
>> should be wary of? 

I can't tell exactly. I'm sure Kees has the whole picture.

>> It looks like x86 filters out $(GCC_PLUGINS_CFLAGS)
>> when building the vDSO.

Yes, that's why building x86 vDSO doesn't need $(DISABLE_STACKLEAK_PLUGIN).

> Maybe at some point we should replace exclusions based on
> GCC_PLUGINS_CFLAGS and KASAN_SANITIZE and UBSAN_SANITIZE and
> OBJECT_FILES_NON_STANDARD and so on with something more generic...
> something that says "this file will not be built into the normal
> kernel, it contains code that runs in realmode / userspace / some
> similarly weird context, and none of our instrumentation
> infrastructure is available there"...

Good idea. I would also add 'notrace' to that list.

Best regards,
Alexander