Re: libatomic vs. __atomic_test_and_set()

Hello, I am not sure how this should be fixed. For example, clang generates for the test case a call to __atomic_exchange_1(): https://godbolt.org/z/EY49jPs78 >From my point of view it makes more sense, if the compiler generates calls to >_1, _2, ... variants. GCC does it for atomic_uint_leas

Re: Criteria for adding AIX as a secondary platform.

Hi All, Please suggest if changes are required in the patch. Thanks, Sangamesh On Fri, Jul 4, 2025 at 11:48 PM swamy sangamesh wrote: > Tried with the below change and able to bootstrap on AIX and X86. > > diff --git a/gcc/tree.cc b/gcc/tree.cc > index e9a83e4260b..b693a58ab9d 100644 > --- a/g