Re: [PATCH] correct -Wrestrict handling of arrays of arrays (PR 84095)

[Jeff Law <law@redhat.com>]

On 02/15/2018 10:47 AM, Martin Sebor wrote:
> On 02/13/2018 11:14 PM, Jeff Law wrote:
>> On 02/01/2018 04:45 PM, Martin Sebor wrote:
>>> The previous patch didn't resolve all the false positives
>>> in the Linux kernel.  The attached is an update that fixes
>>> the remaining one having to do with multidimensional array
>>> members:
>>>
>>>   struct S { char a[2][4]; };
>>>
>>>   void f (struct S *p, int i)
>>>   {
>>>     strcpy (p->a[0], "012");
>>>     strcpy (p->a[i] + 1, p->a[0]);   // false positive here
>>>   }
>>>
>>> In the process of fixing this I also made a couple of minor
>>> restructuring changes to the builtin_memref constructor to
>>> in order to make the code easier to follow: I broke it out
>>> into a couple of helper functions and called those.
>>>
>>> As with the first revision of the patch, this one is also
>>> meant to be applied on top of
>>>
>>>   https://gcc.gnu.org/ml/gcc-patches/2018-01/msg01488.html
>>>
>>> Sorry about the late churn.  Even though I tested the original
>>> implementation with the Linux kernel the bugs were only exposed
>>> non-default configurations that I didn't build.
>>>
>>> Jakub, you had concerns about the code in the constructor
>>> and about interpreting the offsets in the diagnostics.
>>> I tried to address those in the patch.  Please review
>>> the changes and let me know if you have any further comments.
>>>
>>> Thanks
>>> Martin
>>>
>>> On 01/30/2018 04:19 PM, Martin Sebor wrote:
>>>> Testing GCC 8 with recent Linux kernel sources has uncovered
>>>> a bug in the handling of arrays of arrays by the -Wrestrict
>>>> checker where it fails to take references to different array
>>>> elements into consideration, issuing false positives.
>>>>
>>>> The attached patch corrects this mistake.
>>>>
>>>> In addition, to make warnings involving excessive offset bounds
>>>> more meaningful (less confusing), I've made a cosmetic change
>>>> to constrain them to the bounds of the accessed object.  I've
>>>> done this in response to multiple comments indicating that
>>>> the warnings are hard to interpret.  This change is meant to
>>>> be applied on top of the patch for bug 83698 (submitted mainly
>>>> to improve the readability of the offsets):
>>>>
>>>>   https://gcc.gnu.org/ml/gcc-patches/2018-01/msg01488.html
>>>>
>>>> Martin
>>>
>>>
>>> gcc-84095.diff
>>>
>>>
>>> PR middle-end/84095 - false-positive -Wrestrict warnings for memcpy
>>> within array
>>>
>>> gcc/ChangeLog:
>>>
>>>     PR middle-end/84095
>>>     * gimple-ssa-warn-restrict.c
>>> (builtin_memref::extend_offset_range): New.
>>>     (builtin_memref::set_base_and_offset): Same.  Handle inner
>>> references.
>>>     (builtin_memref::builtin_memref): Factor out parts into
>>>     set_base_and_offset and call it.
>>>
>>> gcc/testsuite/ChangeLog:
>>>
>>>     PR middle-end/84095
>>>     * c-c++-common/Warray-bounds-3.c: Adjust text of expected warnings.
>>>     * c-c++-common/Wrestrict.c: Same.
>>>     * gcc.dg/Wrestrict-6.c: Same.
>>>     * gcc.dg/Warray-bounds-27.c: New test.
>>>     * gcc.dg/Wrestrict-8.c: New test.
>>>     * gcc.dg/Wrestrict-9.c: New test.
>>>     * gcc.dg/pr84095.c: New test.
>>>
>>> diff --git a/gcc/gimple-ssa-warn-restrict.c
>>> b/gcc/gimple-ssa-warn-restrict.c
>>> index 528eb5b..367e05f 100644
>>> --- a/gcc/gimple-ssa-warn-restrict.c
>>> +++ b/gcc/gimple-ssa-warn-restrict.c
>>
>>> +      else if (gimple_nop_p (stmt))
>>> +    expr = SSA_NAME_VAR (expr);
>>> +      else
>>> +    {
>>> +      base = expr;
>>> +      return;
>>>      }
>> This looks odd.  Can you explain what you're trying to do here?
>>
>> I'm not offhand why you'd ever want to extract SSA_NAME_VAR.  In general
>> it's primary use is for dumps and debugging info.  I won't quite go so
>> far as to say using it for anything else is wrong, but it's certainly
>> something you ought to explain.
> 
> It appears to be dead code.  Nothing in the GCC test suite hits
> this code.  It's most likely a vestige of an approach I tried
> that didn't work and that I ended up doing differently and forgot
> to remove.  I'll remove it before committing.
> 
>> The rest looks fairly reasonable.  It's a bit hard to follow, but I
>> don't think we should do another round of refactoring at this stage.
> 
> Is the patch good to commit then with the unused code above
> removed?
Yes.  Again, sorry for the delays.

jeff