Re: [PATCHv2] cpumask: work around false-postive stringop-overread errors

[Nilay Shroff <nilay@linux.ibm.com>]

On 12/8/24 18:58, Greg KH wrote:
> On Sun, Dec 08, 2024 at 03:51:10PM +0530, Nilay Shroff wrote:
>>
>>
>> On 12/7/24 17:14, Greg KH wrote:
>>> On Sat, Dec 07, 2024 at 12:43:19PM +0100, Greg KH wrote:
>>>> On Thu, Dec 05, 2024 at 06:04:09PM +0530, Nilay Shroff wrote:
>>>>> While building the powerpc code using gcc 13, I came across following
>>>>> errors generated for kernel/padata.c file:
>>>>>
>>>>>   CC      kernel/padata.o
>>>>> In file included from ./include/linux/string.h:390,
>>>>>                  from ./arch/powerpc/include/asm/paca.h:16,
>>>>>                  from ./arch/powerpc/include/asm/current.h:13,
>>>>>                  from ./include/linux/thread_info.h:23,
>>>>>                  from ./include/asm-generic/preempt.h:5,
>>>>>                  from ./arch/powerpc/include/generated/asm/preempt.h:1,
>>>>>                  from ./include/linux/preempt.h:79,
>>>>>                  from ./include/linux/spinlock.h:56,
>>>>>                  from ./include/linux/swait.h:7,
>>>>>                  from ./include/linux/completion.h:12,
>>>>>                  from kernel/padata.c:14:
>>>>> In function ‘bitmap_copy’,
>>>>>     inlined from ‘cpumask_copy’ at ./include/linux/cpumask.h:839:2,
>>>>>     inlined from ‘__padata_set_cpumasks’ at kernel/padata.c:730:2:
>>>>> ./include/linux/fortify-string.h:114:33: error: ‘__builtin_memcpy’ reading between 257 and 536870904 bytes from a region of size 256 [-Werror=stringop-overread]
>>>>>   114 | #define __underlying_memcpy     __builtin_memcpy
>>>>>       |                                 ^
>>>>> ./include/linux/fortify-string.h:633:9: note: in expansion of macro ‘__underlying_memcpy’
>>>>>   633 |         __underlying_##op(p, q, __fortify_size);                        \
>>>>>       |         ^~~~~~~~~~~~~
>>>>> ./include/linux/fortify-string.h:678:26: note: in expansion of macro ‘__fortify_memcpy_chk’
>>>>>   678 | #define memcpy(p, q, s)  __fortify_memcpy_chk(p, q, s,                  \
>>>>>       |                          ^~~~~~~~~~~~~~~~~~~~
>>>>> ./include/linux/bitmap.h:259:17: note: in expansion of macro ‘memcpy’
>>>>>   259 |                 memcpy(dst, src, len);
>>>>>       |                 ^~~~~~
>>>>> kernel/padata.c: In function ‘__padata_set_cpumasks’:
>>>>> kernel/padata.c:713:48: note: source object ‘pcpumask’ of size [0, 256]
>>>>>   713 |                                  cpumask_var_t pcpumask,
>>>>>       |                                  ~~~~~~~~~~~~~~^~~~~~~~
>>>>> In function ‘bitmap_copy’,
>>>>>     inlined from ‘cpumask_copy’ at ./include/linux/cpumask.h:839:2,
>>>>>     inlined from ‘__padata_set_cpumasks’ at kernel/padata.c:730:2:
>>>>> ./include/linux/fortify-string.h:114:33: error: ‘__builtin_memcpy’ reading between 257 and 536870904 bytes from a region of size 256 [-Werror=stringop-overread]
>>>>>   114 | #define __underlying_memcpy     __builtin_memcpy
>>>>>       |                                 ^
>>>>> ./include/linux/fortify-string.h:633:9: note: in expansion of macro ‘__underlying_memcpy’
>>>>>   633 |         __underlying_##op(p, q, __fortify_size);                        \
>>>>>       |         ^~~~~~~~~~~~~
>>>>> ./include/linux/fortify-string.h:678:26: note: in expansion of macro ‘__fortify_memcpy_chk’
>>>>>   678 | #define memcpy(p, q, s)  __fortify_memcpy_chk(p, q, s,                  \
>>>>>       |                          ^~~~~~~~~~~~~~~~~~~~
>>>>> ./include/linux/bitmap.h:259:17: note: in expansion of macro ‘memcpy’
>>>>>   259 |                 memcpy(dst, src, len);
>>>>>       |                 ^~~~~~
>>>>> kernel/padata.c: In function ‘__padata_set_cpumasks’:
>>>>> kernel/padata.c:713:48: note: source object ‘pcpumask’ of size [0, 256]
>>>>>   713 |                                  cpumask_var_t pcpumask,
>>>>>       |                                  ~~~~~~~~~~~~~~^~~~~~~~
>>>>>
>>>>> Apparently, above errors only manifests with GCC 13.x and config option
>>>>> CONFIG_FORTIFY_SOURCE. Furthermore, if I use gcc 11.x or gcc 12.x then I
>>>>> don't encounter above errors. Prima facie, these errors appear to be false-
>>>>> positive. Brian informed me that currently some efforts are underway by
>>>>> GCC developers to emit more verbose information when GCC detects string
>>>>> overflow errors and that might help to further narrow down the root cause
>>>>> of this error. So for now, silence these errors using -Wno-stringop-
>>>>> overread gcc option while building kernel/padata.c file until we find the
>>>>> root cause.
>>>>
>>>> I'm hitting this now on Linus's tree using gcc14 on x86-64 so this isn't
>>>> just a problem with your arch.
>> Thanks Greg for confirming that this is not isolated to PowerPC!!
>>>>
>>>> Let me try this patch locally and see if it helps...
>>>
>>> Yes, fixes the build for me, so either this is a real fix, or something
>>> else needs to be done for it, so I'll give a:
>>>
>>> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>>>
>>> for now.
>> Okay so now we've an evidence confirming that this is no longer PowerPC specific 
>> issue. Hence as Yury suggested, in another mail[1], fixing this error by disabling
>> stringop-overread globally for GCC13+ and CONFIG_FORTIFY_SOURCE=n, I will spin a 
>> new patch and submit it.
>>
>> [1] https://lore.kernel.org/all/Z1HTdtvNjm-nZSvJ@yury-ThinkPad/
> 
> That feels wrong, unless this is a compiler bug.  And if it's a compiler
> bug, can we fix the compiler please or at least submit a bug to the gcc
> developers?
> 
Yes this seems to be a compiler bug. Please see here : 
[1] https://lore.kernel.org/all/202411021337.85E9BB06@keescook/
[2] https://gcc.gnu.org/pipermail/gcc-patches/2024-October/666872.html

> I'm slowly moving all my boxes/builds over to using clang to build the
> kernel due to rust kernel work, so I guess I can do that here as well as
> I don't think this issue shows up for that compiler, right?
> 
Yes this error doesn't show up for LLVM/clang. We've two options here:
1) To disable this error globally for GCC-13+ until we find the root cause. Maybe when 
   GCC folks add more diagnostics and contexts around -Wstringop-* compiler warning as
   discussed in [2] above.
or 
2) Silence this error only for file kernel/padata.c compiling which this error manifests
   as of today.

Yury suggested option #1 above so that we may avoid random victims hitting this error. 
What do you suggest? 

Thanks,
--Nilay