Revert "GCN, nvptx: Basic '__cxa_guard_{acquire,abort,release}' for C++ static local variables support" (was: OpenMP offloading vs. C++ static local variables)

Hi! On 2023-12-23T10:17:57+0100, I wrote: > On 2023-12-21T13:58:23+0100, Jakub Jelinek wrote: >> On Thu, Dec 21, 2023 at 01:31:19PM +0100, Thomas Schwinge wrote: >>> OK to push, for a start, the attached >>> "GCN, nvptx: Basic '__cxa_guard_{acquire,abort,release}' for C++ static >>> local variab

Re: OpenMP offloading vs. C++ static local variables

Hi! On 2023-12-21T13:58:23+0100, Jakub Jelinek wrote: > On Thu, Dec 21, 2023 at 01:31:19PM +0100, Thomas Schwinge wrote: >> [...] the gimplification-level code re >> 'Static locals [...] need to be "omp declare target"' runs *after* >> 'omp_discover_implicit_declare_target'. Thus my "move" idea

Re: OpenMP offloading vs. C++ static local variables

On Thu, Dec 21, 2023 at 01:31:19PM +0100, Thomas Schwinge wrote: > These three: implicitly, or explicit '#pragma omp declare target' etc., > or inside '#pragma omp begin declare target' region are the only OpenMP > facilities to get things 'omp declare target'ed, right? I think so. > That doesn't

Re: OpenMP offloading vs. C++ static local variables

Hi Jakub! On 2023-12-07T16:33:08+0100, Jakub Jelinek wrote: > On Thu, Dec 07, 2023 at 04:09:04PM +0100, Thomas Schwinge wrote: >> > Yeah, I believe we should in the omp_discover_* sub-pass handle with >> > a help of a langhook automatically mark the guard variables (possibly >> > iff the guarded

Re: OpenMP offloading vs. C++ static local variables

On Thu, Dec 07, 2023 at 04:09:04PM +0100, Thomas Schwinge wrote: > > Yeah, I believe we should in the omp_discover_* sub-pass handle with > > a help of a langhook automatically mark the guard variables (possibly > > iff the guarded variable is marked?), > > Looking at 'gcc/omp-offload.cc:omp_disco

Re: OpenMP offloading vs. C++ static local variables

Hi! Jakub, would you please provide guidance? Elsewhere, I wrote: || I'm working on implementing (some) C++ standard library support for code || offloading in GCC, and ran into the following issue: per || , ||