RE: [PATCH] acpica: fix -Wnull-pointer-arithmetic warnings

[Moore, Robert]

-----Original Message-----
From: Qian Cai [mailto:cai@xxxxxx] 
Sent: Wednesday, July 17, 2019 5:50 PM
To: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
Cc: Wysocki, Rafael J <rafael.j.wysocki@xxxxxxxxx>; Moore, Robert <robert.moore@xxxxxxxxx>; Schmauss, Erik <erik.schmauss@xxxxxxxxx>; jkim@xxxxxxxxxxx; Len Brown <lenb@xxxxxxxxxx>; linux-acpi@xxxxxxxxxxxxxxx; devel@xxxxxxxxxx; clang-built-linux <clang-built-linux@xxxxxxxxxxxxxxxx>; LKML <linux-kernel@xxxxxxxxxxxxxxx>
Subject: Re: [PATCH] acpica: fix -Wnull-pointer-arithmetic warnings



> On Jul 17, 2019, at 6:01 PM, Nick Desaulniers <ndesaulniers@xxxxxxxxxx> wrote:
> 
> On Tue, Jul 16, 2019 at 8:38 PM Qian Cai <cai@xxxxxx> wrote:
>> 
>> Clang generate quite a few of those warnings.
>> 
>> drivers/acpi/scan.c:759:28: warning: arithmetic on a null pointer 
>> treated as a cast from integer to pointer is a GNU extension 
>> [-Wnull-pointer-arithmetic]
>>                status = acpi_get_handle(ACPI_ROOT_OBJECT,
>> obj->string.pointer,
>>                                         ^~~~~~~~~~~~~~~~
>> ./include/acpi/actypes.h:458:56: note: expanded from macro 
>> 'ACPI_ROOT_OBJECT'
>> #define ACPI_ROOT_OBJECT                ((acpi_handle) ACPI_TO_POINTER
>> (ACPI_MAX_PTR))
>>                                                        
>> ^~~~~~~~~~~~~~~
>> ./include/acpi/actypes.h:509:41: note: expanded from macro 
>> 'ACPI_TO_POINTER'
>> #define ACPI_TO_POINTER(i)              ACPI_ADD_PTR (void, (void *) 0,
>> (acpi_size) (i))
>>                                         
>> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> ./include/acpi/actypes.h:503:84: note: expanded from macro 
>> 'ACPI_ADD_PTR'
>> #define ACPI_ADD_PTR(t, a, b)           ACPI_CAST_PTR (t,
>> (ACPI_CAST_PTR (u8, (a)) + (acpi_size)(b)))
>>                                         ^~~~~~~~~~~~~~~~~
>> ./include/acpi/actypes.h:501:66: note: expanded from macro 
>> 'ACPI_CAST_PTR'
>> #define ACPI_CAST_PTR(t, p)             ((t *) (acpi_uintptr_t) (p))
>>                                                                  ^ 
>> This is because pointer arithmetic on a pointer not pointing to an 
>> array is an undefined behavior. Fix it by doing an integer arithmetic 
>> instead.
> 
> Hi Qian, thanks for the patch.  How do I reproduce this issue, 
> precisely?  I just tried:
> $ make CC=clang -j71 drivers/acpi/scan.o on linux-next today and don't 
> observe the warning.  My clang is ToT built sometime this week.  It 
> looks like drivers/acpi/scan.o when CONFIG_ACPI=y, which is set in the 
> defconfig.  Is there another set of configs to enable to observe the 
> warning?

# make W=1 -j 256

With the config,

https://raw.githubusercontent.com/cailca/linux-mm/master/arm64.config 

> 
> Also, the fix is curious.  Arithmetic on pointers to different 
> "objects" (with one element passed the end) may lead to provence 
> issues due to undefined behavior, but I would have expected some cases 
> to uintptr_t, then arithmetic on that type, as the solution (which is 
> what I suspect ACPI_CAST_PTR is doing).
> 
> Further, you seem to have modified ACPI_ADD_PTR but not ACPI_SUB_PTR; 
> I would have expected both to be afflicted together or not at all 
> based on their existing implementations.

Yes, I thought about that, but ACPI_SUB_PTR does not seem used anywhere, so I thought maybe just start a new discussion to remove it all together later.

ACPI_SUB_PTR is used in the iasl data table compiler.


> 
>> 
>> Signed-off-by: Qian Cai <cai@xxxxxx>
>> ---
>> include/acpi/actypes.h | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>> 
>> diff --git a/include/acpi/actypes.h b/include/acpi/actypes.h index 
>> ad6892a24015..25b4a32da177 100644
>> --- a/include/acpi/actypes.h
>> +++ b/include/acpi/actypes.h
>> @@ -500,13 +500,13 @@ typedef u64 acpi_integer;
>> 
>> #define ACPI_CAST_PTR(t, p)             ((t *) (acpi_uintptr_t) (p))
>> #define ACPI_CAST_INDIRECT_PTR(t, p)    ((t **) (acpi_uintptr_t) (p))
>> -#define ACPI_ADD_PTR(t, a, b)           ACPI_CAST_PTR (t, (ACPI_CAST_PTR (u8, (a)) + (acpi_size)(b)))
>> +#define ACPI_ADD_PTR(t, a, b)           ACPI_CAST_PTR (t, (a) + (acpi_size)(b))

We have some questions concerning this change. If (a) is not cast to a u8, the addition will be in whatever units are appropriate for (a) i.e., the type of (a). However, we want ACPI_ADD_PTR (And ACPI_SUB_PTR) to simply perform a byte addition or subtraction - thus the cast to u8. I believe that is the original thinking behind the macros.

>> #define ACPI_SUB_PTR(t, a, b)           ACPI_CAST_PTR (t, (ACPI_CAST_PTR (u8, (a)) - (acpi_size)(b)))
>> #define ACPI_PTR_DIFF(a, b)             ((acpi_size) (ACPI_CAST_PTR (u8, (a)) - ACPI_CAST_PTR (u8, (b))))
>> 
>> /* Pointer/Integer type conversions */
>> 
>> -#define ACPI_TO_POINTER(i)              ACPI_ADD_PTR (void, (void *) 0, (acpi_size) (i))
>> +#define ACPI_TO_POINTER(i)              ACPI_ADD_PTR (void, 0, (acpi_size) (i))
> 
> IIUC, these are adding `i` to NULL (or (void*)0)? X + 0 == X ?
> --
> Thanks,
> ~Nick Desaulniers