smithp35 wrote: I can't spot anything obviously wrong. I'm not familiar enough with the clang code-base to give you much useful feedback though. I can check specific things against the ABI. Best to call on me if you've got a specific question.
As an aside from this review, internally we used a random calling convention generator program that mixed GCC and Clang generated functions, with the callee checking the expected arguments from the caller. Although not something that can be done now, when clang is able to generate code from this ABI library it will be good to try something like that with clang against itself. Issues often occurred when registers ran out and the stack had to be used. Issues that it found relating to HFAs, the issue I think still hasn't been resolved: * https://github.com/llvm/llvm-project/issues/99644 * https://reviews.llvm.org/D98794 * https://reviews.llvm.org/D127197 https://github.com/llvm/llvm-project/pull/218799 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
