On 3/24/19 6:21 PM, Martin Sebor wrote:
> The error issued when the aligned attribute argument is too big
> to be represented is incorrect: it says the maximum alignment
> is 1U << 31 when it should actually be 1 << 28. This was a typo
> introduced when the error message was enhanced earlier in GCC 9.
>
> The test I added to verify the fix for the typo exposed another
> bug introduced in the same commit as the incorrect value in
> the error message: assuming that the attribute aligned argument
> fits in SHWI.
>
> The attached patch corrects both problems. It has been tested
> on x86_64-linux. I will commit it as obvious sometime this week
> unless there are any objections or suggestions for changes.
>
> Martin
>
> PS I have a couple of questions related to the affected code:
> 1) Does GCC support building with compilers where int is not 32
> bits wide, or where BITS_PER_UNIT is not 3? (I.e., either is
> less or more?)
> 2) Is there a supported target that doesn't have __INT64_TYPE__?
> (And if so, how do I find it in a test? I couldn't find
> anythhing in target-supports.exp).
>
> gcc-89812.diff
>
> PR c/89812 - incorrect maximum in error: requested alignment ‘536870912’
> exceeds maximum 2147483648
>
> gcc/c-family/ChangeLog:
>
> PR c/89812
> * c-common.c (check_user_alignment): Rename local. Correct maximum
> alignment in diagnostic. Avoid assuming argument fits in SHWI,
> convert it to UHWI when it fits.
>
> gcc/testsuite/ChangeLog:
>
> PR c/89812
> * gcc.dg/attr-aligned-3.c: New test.
OK
jeff
