This is an automated email from the git hooks/post-receive script.
unknown user pushed a change to branch devel/omp/gcc-13 in repository gcc.
at 83983b4363a Implement LDPT_REGISTER_CLAIM_FILE_HOOK_V2 linker plugin ho [...]
This branch includes the following new commits:
new fabe84ca322 Identify OMP development branch in output of 'gcc --version' new c3298472780 Merge non-contiguous array support patches. new 18834f04651 Tweak error return value for acc_set_cuda_stream. new ef3e2cde820 Add OpenACC Fortran support for deviceptr and variable in c [...] new 5553bdf3828 Various OpenACC reduction enhancements - FE changes new c2583d509fa Various OpenACC reduction enhancements - ME and nvptx changes new 4f8126158a1 Various OpenACC reduction enhancements - test cases new 4ea892e7fc5 Adjustments and additions to testcases new 8e1d2dc2cdd Reinstate kernels-restrict behaviour new 97c19a9c9be Generate sequential loop for OpenACC loop directive inside kernels new 2d5938a10e4 Default compute dimensions (compile time) new 5acbbe5bcba Fix implicit mapping for array slices on lexically-enclosin [...] new 6c975073065 Enable GOMP_MAP_FIRSTPRIVATE_INT for OpenACC new 3f02bb71b3a Don't mark OpenACC auto loops as independent inside acc par [...] new 8c1ad6a2901 Enable firstprivate OpenACC reductions new b8097b604af Tweak target selector for libgomp.oacc-c-c++-common/lib-93.c. new a88dab97339 Disable libstdc++ dependency for libffi new 68504180ba5 Use "--param=openacc-kernels=parloops" to document "parloop [...] new c4ab26ed709 Make new OpenACC kernels conversion the default; adjust and [...] new ab14bdd1bc4 Update OpenACC version to 2.6 new 6245ea48f35 Add kernels for-index reuse testcase. new e9e96281977 Fix references declared in lexically-enclosing OpenACC data region new 565816face0 Fix lexically-nested data mappings for no_alloc or optional [...] new 959e28cdd1c Fix expected messages in goacc tests new ed3f79bb1cd Assumed-size arrays with non-lexical data mappings new a76681a9575 Add changes to profiling interface from OG8 branch new b9135c979f9 NVPTX GOMP_OFFLOAD_openacc_async_construct arg fix and gomp [...] new b4f6d722400 Add missing exec_params libgomp plugin entry points new 01988603d9e Fix libgomp.oacc-fortran/lib-13.f90 async bug new 663a024cafd Reference reduction localization new 2faea89bdab Fix tree check failure with reduction localization new 2f811f92aab Use more appropriate var in localize_reductions call new 5c7388475ba Update expected messages, errors and warnings for "kernels" tests new 68b6c64e135 Silence compiler warnings new 0e79aeba19e Handle references in OpenACC "private" clauses new 44d3956596d Fortran "declare create"/allocate support for OpenACC new b2d2470d789 Fix OpenACC "ephemeral" asynchronous host-to-device copies new 7d31e99e7cf Fix vector handling for firstprivate of <= pointer size new 15b1a57ec0e Add XFAIL for libgomp.oacc-c-c++-common/data-firstprivate-1.c new ed45f91a0b1 Kernels loops annotation: C and C++. new 63104803ff8 Add -fno-openacc-kernels-annotate-loops option to more testcases. new 6bc5f50de4f Kernels loops annotation: Fortran. new abc51748100 Additional Fortran testsuite fixes for kernels loops annota [...] new a14f2d110ff Fix bug in processing of array dimensions in data clauses. new 40bc5d2eb40 OpenACC: fix privatization of by-reference arrays new db74e6ee042 Fix goacc/noncontig_array-1.c testcase new 3585ca3960d libgomp.oacc-fortran/firstprivate-int.f90 fix for nonexisti [...] new 8f3b4d2bad3 Fix failures in c-c++-common/goacc/note-parallelism-combine [...] new 3fd4e164aa4 XFAIL tests in gfortran.dg/goacc/loop-2-kernels.f95 new a951c5949e3 Fix gfortran.dg/goacc/pr70828.f90 testcase new 020f4009ee8 Add a "combined" flag for "acc kernels loop" etc directives. new 82abd6233bf Annotate inner loops in "acc kernels loop" directives (C/C++). new 826f16d586d Annotate inner loops in "acc kernels loop" directives (Fortran). new 4eb4a59c97b Update dg-* in gfortran.dg/gomp/pr67500.f90 new a24956e42e8 Permit calls to builtins and intrinsics in kernels loops. new 696eea7eb3f Fix patterns in Fortran tests for kernels loop annotation. new f693b12e0a7 Clean up loop variable extraction in OpenACC kernels loop a [...] new 04e861663ea Relax some restrictions on the loop bound in kernels loop a [...] new a54cf07165a [og10] openacc: Adjust loop lowering for AMD GCN new 819952663a6 [og10] vect: Add target hook to prefer gather/scatter instructions new 734cff0504f Fix offload dwarf info new 916e3424471 DWARF address space for variables new 0bb525cb867 Correct fix offload dwarf info new 325f085897e OpenMP 5.0: Allow multiple clauses mapping same variable new c24a3fbc238 DWARF: fix ICE caused by offload debug fix new b96d9239016 nvptx: remove erroneous stack deletion new 90d8139007c openmp: Scale type precision of collapsed iterator variable new a17b6dfc34d DWARF: late code range fixup new d6c760d4c6f Update expected messages in OpenACC tests new c0b28d4f5bc Fix is_oacc_parallel_or_serial for kernel regions new de040cb9812 Update expected messages in kernels-decompose-2 tests new 2159ee4a308 Update expected messages in data-clause-1 tests new 16e24c03f36 Add -Wopenacc-parallelism to tests only in OG11 new 9205073a97e libgomp amdgcn: Fix issues with dynamic OpenMP thread scaling new a7da4348366 openacc: fix ICE for non-decl expression in non-contiguous [...] new 830f04d4dd6 testsuite/libgomp.oacc-fortran/: Add -Wopenacc-parallelism new 96681deaaec Fix gimple_debug_cfg declaration new b547ec06215 Fortran: delinearize multi-dimensional array accesses new ed6a77b962e openacc: Move pass_oacc_device_lower after pass_graphite new 2da100e6e3f Fix branch prediction dump message new 3a9a5bcde04 Move compute_alias_check_pairs to tree-data-ref.c new 1a9d5c46c01 graphite: Add runtime alias checking new baf1aeb505d openacc: Use Graphite for dependence analysis in "kernels" regions new 81cade9632f openacc: Add "can_be_parallel" flag info to "graph" dumps new 3b08ffe1f96 openacc: Add further kernels tests new 8b6a7c4ba96 openacc: Remove unused partitioning in "kernels" regions new 6a8cf0e8ba9 Add function for printing a single OMP_CLAUSE new 9cc7461f318 openacc: Add data optimization pass new 6b7570182e3 openacc: Add runtime alias checking for OpenACC kernels new 5b5903b657b openacc: Warn about "independent" "kernels" loops with data [...] new 99889be3d31 openacc: Handle internal function calls in pass_lim new 5141b147f63 openacc: Disable pass_pre on outlined functions analyzed by [...] new 67badb50b1f graphite: Tune parameters for OpenACC use new 0859356c320 graphite: Adjust scop loop-nest choice new 320fbf11ce2 graphite: Accept loops without data references new de46b33a209 openacc: Adjust test expectations to new "kernels" handling new a56cb34f65c libgomp, nvptx: Update bundled CUDA header file new 263573e9423 libgomp, nvptx: low-latency memory allocator new 4ddca7088cf openmp: Add C support for parsing metadirectives new 1a3fe117506 openmp: Add middle-end support for metadirectives new a629207a1a0 openmp: Add support for resolving metadirectives during par [...] new 6d80b3e0c23 openmp: Add support for streaming metadirectives and resolv [...] new a4c8880710e openmp: Add C++ support for parsing metadirectives new 8ce26c7a14a openmp, fortran: Add Fortran support for parsing metadirectives new 12ca40a57b0 openmp: Add testcases for metadirectives new 512a1e46e77 openmp: Add support for 'target_device' context selector set new 1372ba8a9d9 openmp: Add warning when functions containing metadirective [...] new 38f874900fd openmp: Eliminate non-matching metadirective variants early [...] new f0227c021f3 openmp: Improve handling of nested OpenMP metadirectives in [...] new 803869086d5 openmp: Handle C/C++ array reference base-pointers in array [...] new 70008a509c5 Fortran/OpenMP: Support mapping of DT with allocatable components new 6ce5ee77f73 Add parsing support for allocate directive (OpenMP 5.0) new a4a8582a9f8 Translate allocate directive (OpenMP 5.0). new ecb8c1e1a12 Handle cleanup of omp allocated variables (OpenMP 5.0). new db946de50f4 Gimplify allocate directive (OpenMP 5.0). new afdf93a169d Lower allocate directive (OpenMP 5.0). new e120797b7c2 libgomp: pinned memory new 9c457cff563 libgomp, openmp: Add ompx_pinned_mem_alloc new fa4e696600f openmp, nvptx: low-lat memory access traits new 0bc47d428fe openmp: Add -foffload-memory new d0de00ae1f5 openmp: allow requires unified_shared_memory new a2d948dda3c openmp, nvptx: ompx_unified_shared_mem_alloc new 9e0add2f9e5 openmp: Use libgomp memory allocation functions with unifie [...] new 48f202dd79e openmp: -foffload-memory=pinned new 44c9ad04d66 libgomp, nvptx: report USM supported new 043b062e083 Fortran: Fix finalization resolution with deep copy new d087b95598f Fortran: Fix finalization resolution with deep copy (cont) new 452de20dc2f Fortran: Fix proc pointer as elemental arg handling new f69cc248a5d openmp: Implement uses_allocators clause new eaf97e83f52 Fix ICE when cache-3-1.c testcase is run new 23fb8f963f0 Fix ICE in kernels-decompose-pr100280-1.c testcase new 76415a305dc Fix ICE in nested-function-1.c testcase new ae1ada62bc4 Fix gfortran.dg/gomp/num-teams-2.f90 new 0229066ecb2 amdgcn: Support XNACK mode new c0b328211a6 amdgcn, openmp: Auto-detect USM mode and set HSA_XNACK new 3ddf3565fae amdgcn: libgomp plugin USM implementation new 53affbc28a8 openmp: fix max_vf setting for amdgcn offloading new 1954e9d3f12 gfortran.dg/gomp/depend-4.f90: minor fix + dump update new dea11975d2a gfortran.dg/gomp/depend-6.f90: minor fix + dump update new 6b683f644b7 OpenMP: Fix ICE with OMP metadirectives new 64b24e6b8f0 Fix dg- pattern for gomp/{affinity-clause-1.f90,uses_alloca [...] new 1a759b78eb5 gfortran.dg/gomp/{depend-5,scope-6}.f90: Update scan-tree-dump new f4c5565ccbc Fortran: Fix delinearization regression new 907ec8f7619 Fix omp-expand.cc's expand_omp_target for OpenACC new 0c85b75243c libgomp.c-c++-common/requires-4.c: dg-xfail-run-if for USM [...] new 44579e0298e Added "noclone" to scan-tree-dump for several OpenAcc tests. new 2dcca15948c omp-oacc-kernels-decompose.cc: fix -fcompare-debug with GIM [...] new d15cd9fc03f gcc/testsuite: Change 'cunrolli' to 'cunrolli1' in dump sca [...] new b26bad9539a amdgcn, libgomp: USM allocation update new 96c9528a2fa amdgcn: disallow USM on gfx908 new d5041478b24 vect: WORKAROUND vectorizer bug new f52701e757f Handle operator new with alignment in usm transform. new 3ae3f87962b openmp: Bugfix in omp_expand_metadirective for same blocks/ [...] new 2f3bdc6c68d OpenMP/Fortran: 'target update' with strides + DT components new a8bb527eb34 Fortran/OpenMP: Fix DT struct-component with 'alloc' and ar [...] new 300c350fbee XFAIL some OpenACC 'kernels' confusion in 'libgomp.oacc-for [...] new 0f11094c1a7 Fortran "declare create"/allocate support for OpenACC: adju [...] new f0dcab86ba5 Fortran "declare create"/allocate support for OpenACC: adju [...] new 2a9d8670157 amdgcn: Support AMD-specific 'isa' and 'arch' traits in Ope [...] new 33e374dbf01 OpenMP, libgomp: Handle unified shared memory in omp_target [...] new 3e9112c2bea libgomp: Fix USM bugs new 3058aac0d37 amdgcn, libgomp: custom USM allocator new cb41a9ae655 libgomp, amdgcn: Switch USM to 128-byte alignment new b89599c2e47 Force '--param openacc-kernels=parloops' in 'libgomp.oacc-c [...] new b9c80c5b2c0 Clean up after newlib "nvptx: In offloading execution, map [...] new 473a3169dee nvptx: Make 'nvptx_uniform_warp_check' fit for non-full-war [...] new 308d4efc00d Add 'gcc.target/nvptx/softstack-decl-1.c', 'gcc.target/nvpt [...] new 15507794a37 nvptx: Prevent emitting duplicate declarations for '__nvptx [...] new 51c04bc257e nvptx: Support global constructors/destructors via 'collect2' new 0bef4fb9646 nvptx: Support global constructors/destructors via 'collect [...] new 45396d92c63 nvptx, libgcc: Stub unwinding implementation new 99e450cd251 nvptx, libgfortran: Switch out of "minimal" mode new 489adab23e7 Make 'libgcc/config/nvptx/crt0.c' build '--without-headers' new 895822e3388 Update 'libgomp/libgomp.texi' for 'nvptx, libgfortran: Swit [...] new ad9e36ae50e Update 'libgomp/libgomp.texi' for 'nvptx, libgfortran: Swit [...] new bfc54a2146d Fix 'omp_allocator_handle_kind' example in 'gfortran.dg/gom [...] new 5dc45dc262d 'c-c++-common/gomp/alloc-pinned-1.c' -> 'libgomp.c-c++-comm [...] new 3f3fd7e97a0 '{c-c++-common,gfortran.dg}/gomp/uses_allocators-*' -> 'lib [...] new 0f0634e7d7d 'gfortran.dg/gomp/allocate-4.f90' -> 'libgomp.fortran/alloc [...] new 6e3816fa47c openmp: Add support for the 'present' modifier new d7488e05624 gomp/openmp-simd-8.f90: Remove .ASSUME tree-dump check new f719ab9a3ac Fix 'libgomp.{c-c++-common,fortran}/target-present-*' test cases new 4fb2d4548d6 nvptx, libgomp: Move the low-latency allocator code new e63108c8ca8 amdgcn, libgomp: low-latency allocator new 52dbc3eee0e 'libgomp.c/usm-{1,2,3,4}.c': Re-enable non-GCN offloading c [...] new fc6bfa8f6d5 Miscellaneous clean-up re OpenMP 'ompx_unified_shared_mem_s [...] new 7e943c88f1e Clarify/verify OpenMP 'omp_calloc' zero-initialization for [...] new 86ba358a39b In 'libgomp/allocator.c:omp_realloc', route 'free' through [...] new 23fa6a8ab0f Attempt to register OpenMP pinned memory using a device ins [...] new c9e88c6906e Attempt to not just register but allocate OpenMP pinned mem [...] new 16fc5f61bc3 Prototype 'GOMP_enable_pinned_mode' new 3f0a8cdf9fc Fortran/OpenMP: Fix mapping of array descriptors and deferr [...] new a76e5f70152 libgomp: no need to attach USM pointers new a26760e459d Update dg-dump-scan for "Fortran/OpenMP: Fix mapping of arr [...] new 1127e54505a libgomp: Merge 'gomp_map_vars_openacc' into 'goacc_map_vars [...] new 18d5e4c6132 amdgcn: vec_extract no-op insns new 8c324588cb6 amdgcn: gather/scatter with DImode offsets new 5b13e48eb7d amdgcn: Fix register size bug new c3d63305f75 OpenMP/Fortran: Fix unmapping of GOMP_MAP_POINTER for scala [...] new 15cc37fdb02 Fortran: Add attr.class_ok check for generate_callback_wrapper new 3a5f9cda7d2 Fortran/OpenMP: Fix 'alloc' and 'from' mapping for allocata [...] new a9b887f8704 Given OpenACC 'async', defer 'free' of non-contiguous array [...] new d8235b10642 In 'libgomp/target.c:gomp_unmap_vars_internal', defer 'gomp [...] new 56e719f76c8 libgomp: Simplify OpenMP reverse offload host <-> device me [...] new 6ce611a6952 Miscellaneous clean-up re OpenMP 'ompx_host_mem_space' new 14daae29fa9 Add 'libgomp.c/alloc-ompx_host_mem_alloc-1.c' new ea770a4a9f6 openmp: Add Fortran support for "omp unroll" directive new c17983cd052 openmp: Add C/C++ support for "omp unroll" directive new 5e11b7ffc84 openacc: Rename OMP_CLAUSE_TILE to OMP_CLAUSE_OACC_TILE new c6f772b089d openmp: Add Fortran support for "omp tile" new 5b9c9f10892 openmp: Add C/C++ support for "omp tile" new 5c71451f714 openmp: Add Fortran support for loop transformations on inn [...] new 3baddf440b7 openmp: Add C/C++ support for loop transformations on inner loops new 44bb7b07de8 openmp: Handle GIMPLE_OMP_METADIRECTIVE in walk_omp_for_loops new 7f72284eef2 libgomp: Document OpenMP 'pinned' memory new f21ec30660e OpenACC: Pass pre-allocated 'ptrblock' to 'goacc_noncontig_ [...] new d4c585eeaa1 '-foffload-memory=pinned' using offloading device interfaces new 30e772510d7 '-foffload-memory=pinned' using offloading device interface [...] new 5410e14552b amdgcn, openmp: Fix concurrency in low-latency allocator new c89addc938d openmp: Fix loop transformation tests new ae8809a62fb Fix ICE in libgomp.oacc-fortran/declare-allocatable*.f90 tests new fbb0062ece7 Fix ICE in libgomp.oacc-c-c++-common/noncontig_array-* tests new 6ea27d1228a Fix ICE in gfortran.dg/goacc/omp_data_optimize-1.f90 new 03d7f264d7b Fix expected output in map-10a.f90 new f0263b7785f Fix test failures in gfortran.dg/gomp/target-exit-data.f90 new 94a75fb12e6 openmp: Fix initialization for 'unroll full' new bf0ac7c1ff1 OpenMP: Constructors and destructors for "declare target" s [...] new 83983b4363a Implement LDPT_REGISTER_CLAIM_FILE_HOOK_V2 linker plugin ho [...]
The 228 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "adds" were already present in the repository and have only been added to this reference.