jdoerfert added a comment. (answered now, I forgot to submit this yesterday -.- )
In D102107#2819869 <https://reviews.llvm.org/D102107#2819869>, @lebedev.ri wrote: > (This is not offload-specific, right?) It is not. It applies to all parallel regions. We still use the variadic call on the CPU but we can now replace it with a non-variadic interface. > This does not bring any compatibility issues, right? We'll keep the old interface (kmpc_fork_call) for the host (libomp) and code that uses it will continue to work. New code will use it also in a compatible way because we change caller and callee of kmpc_fork_call to adjust for the aggregate. This will also hold as the new non-variadic interface is introduced. The device side is simpler, we are only compatible with a device runtime that machtes (as we link it into the app), so there is no compatibility issue to be expected. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D102107/new/ https://reviews.llvm.org/D102107 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
