GCC Fortran: compiler crash on root and element pointer components in one map list

Legality: not independently confirmed. Independent readings are uncertain: Whether map(to: r) maps the allocated allocatable component nx and its elements, so that the later maps of r%nx(1)%a and r%nx(2)%a attach correctly, is not settled by the text; the allocatable-component rule only says the map may treat allocation status as unallocated. An ICE is undesirable but a required result cannot be established. How this is checked

GCC crashes in gimplify_adjust_omp_clauses on a root pointer component and element pointer components in one map list.

Affected variants

Compiler	MI300A gfx942 unified memory (XNACK 1)	MI300A gfx942 copy (XNACK 0)	MI250X gfx90a copy (XNACK 0)	MI300X gfx942 copy (XNACK 0)	MI350X gfx950 copy (XNACK 0)	A100 sm_80 separate memory	A100 sm_80 managed memory
AFAR 24.3.0	correct	correct	unknown not run	correct	correct	n/a	n/a
CCE 19.0.0	correct	correct	unknown not run	unknown not run	unknown not run	n/a	n/a
CCE 20.0.2	correct	correct	unknown not run	unknown not run	unknown not run	n/a	n/a
CCE 21.0.3	correct	correct	unknown not run	unknown not run	unknown not run	n/a	n/a
CCE 21.0.4	correct	correct	unknown not run	unknown not run	unknown not run	n/a	n/a
NVHPC 25.11	n/a	n/a	n/a	n/a	n/a	fails another way	correct
NVHPC 26.9	n/a	n/a	n/a	n/a	n/a	fails another way	correct
GCC 16.2.0	behaves incorrectly	behaves incorrectly	behaves incorrectly	behaves incorrectly	behaves incorrectly	behaves incorrectly	n/a

"behaves incorrectly" / "not supported": this issue, confirmed in up to four environments (another build directory, another node, a repeat). "fails another way": a different failure, not counted as this issue. "unknown": not run there, or the runs disagree.

Messages

Result	Compiler and GPU	Message
behaves incorrectly	GCC 16.2.0, MI300A unified GCC 16.2.0, MI300A copy	show site-nested--gcc-fortran-root-and-element-pointer-components-ice.f90:61:76: 61 \| !$omp target enter data map(to: r%a) map(to: r%nx(1)%a) map(to: r%nx(2)%a) \| ^ internal compiler error: in gimplify_adjust_omp_clauses, at gimplify.cc:16128 0x2072e2d internal_error(char const, ...) <dir>:787 0x7d943f fancy_abort(char const, int, char const*) <dir>:1813
behaves incorrectly	GCC 16.2.0, MI250X	show site-nested--gcc-fortran-root-and-element-pointer-components-ice.f90:56:76: 56 \| !$omp target enter data map(to: r%a) map(to: r%nx(1)%a) map(to: r%nx(2)%a) \| ^ internal compiler error: in gimplify_adjust_omp_clauses, at gimplify.cc:16128 0x2072e2d internal_error(char const, ...) <dir><host>-gcc-src/gcc-16.2.0/gcc/diagnostic-global-context.cc:787 0x7d943f fancy_abort(char const, int, char const*) <dir><host>-gcc-src/gcc-16.2.0/gcc/diagnostics/context.cc:1813
behaves incorrectly	GCC 16.2.0, MI300X GCC 16.2.0, MI350X	show site-nested--gcc-fortran-root-and-element-pointer-components-ice.f90:61:76: 61 \| !$omp target enter data map(to: r%a) map(to: r%nx(1)%a) map(to: r%nx(2)%a) \| ^ internal compiler error: in gimplify_adjust_omp_clauses, at gimplify.cc:16128 0x2083fcd internal_error(char const, ...) <home>/<user>/gcc-work/gcc-16.2.0/gcc/diagnostic-global-context.cc:787 0x7cebc6 fancy_abort(char const, int, char const*) <home>/<user>/gcc-work/gcc-16.2.0/gcc/diagnostics/context.cc:1813
behaves incorrectly	GCC 16.2.0, A100 separate	show site-nested--gcc-fortran-root-and-element-pointer-components-ice.f90:56:76: 56 \| !$omp target enter data map(to: r%a) map(to: r%nx(1)%a) map(to: r%nx(2)%a) \| ^ internal compiler error: in gimplify_adjust_omp_clauses, at gimplify.cc:16128 0x214aceb internal_error(char const, ...) <home>/gcc-build/gcc-16.2.0/gcc/diagnostic-global-context.cc:787 0x7db5a3 fancy_abort(char const, int, char const*) <home>/gcc-build/gcc-16.2.0/gcc/diagnostics/context.cc:1813
fails another way	NVHPC 25.11, A100 separate NVHPC 26.9, A100 separate	show Failing in Thread:1 Accelerator Fatal Error: call to cuMemcpyDtoHAsync returned error 700 (CUDA_ERROR_ILLEGAL_ADDRESS): Illegal address during kernel execution File: <dir> Function: t:31 Line: 59

Reproduce

GCC 16.2.0, MI300A unified

export LD_LIBRARY_PATH=$GCC_16_2/lib64:/opt/rocm-7.2.1/lib
$GCC_16_2/bin/gfortran -fopenmp -foffload=amdgcn-amdhsa -foffload-options=amdgcn-amdhsa=-march=gfx942 -O2 repro.f90 -o repro
HSA_XNACK=1 OMP_TARGET_OFFLOAD=mandatory ./repro

GCC 16.2.0, A100 separate

export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=$GCC_16_2/lib64
$GCC_16_2/bin/gfortran -fopenmp -foffload=nvptx-none -foffload-options=nvptx-none=-march=sm_80 -foffload-options=nvptx-none=-latomic -O2 repro.f90 -o repro
OMP_TARGET_OFFLOAD=mandatory ./repro

AFAR 24.3.0, MI300A unified (runs correctly)

$AFAR/bin/amdflang -fopenmp --offload-arch=gfx942 -O2 repro.f90 -o repro
HSA_XNACK=1 OMP_TARGET_OFFLOAD=mandatory ./repro

Program (70 lines)

! # Finding (nested-data): GCC ICE in gimplify_adjust_omp_clauses: a pointer component of the root and the pointer components
! # of the elements of an allocatable component array in one map list
!
! - toolchains and modes that fail: gcc_g90 (GCC 16.2.0), gcc_sm80 (GCC 16.2.0); modes: copy; -O2
! - build: gfortran -fopenmp -foffload=amdgcn-amdhsa -foffload-options=amdgcn-amdhsa=-march=gfx90a -O2 (gcc_g90); gfortran -fopenmp
!   -foffload=nvptx-none -foffload-options=nvptx-none=-march=sm_80 -foffload-options=nvptx-none=-latomic -O2 (gcc_sm80)
! - failure: ice
! - observed: `Please submit a full bug report, with preprocessed source (by using -freport-bug).`; identical signature on 2 of 2
!   failing cells
! - expected: compiles and prints `tot= 24` with `check=pass`
! - rule: pointer components are list items when they appear in a map clause ("The list items that appear in a map
!   clause may include array sections and structure elements", 5.2 §5.8.3 p.151 l.13); the construct names a pointer
!   component of the root and the pointer components of the elements of an allocatable array component, none of them the
!   base pointer of another item (5.2 §5.8.3 p.152 l.1-3)
! - same-compiler control: controls/g5-ctl-element-pointers-only.f90 (one change from this file; correct on gcc_g90 (GCC 16.2.0),
!   gcc_sm80 (GCC 16.2.0); rows results/nested/red/gate_abd.jsonl, results/nested/red/gate_c.jsonl, results/nested/red/red3.jsonl,
!   results/nested/red/tuo7.jsonl, results/nested/red/wgs.jsonl, results/nested/red/wnv.jsonl);
!   controls/g5-ctl-root-pointer-only.f90 (one change from this file; correct on gcc_g90 (GCC 16.2.0), gcc_sm80 (GCC 16.2.0); rows
!   results/nested/red/gate_abd.jsonl, results/nested/red/gate_c.jsonl, results/nested/red/red3.jsonl,
!   results/nested/red/tuo7.jsonl, results/nested/red/wgs.jsonl, results/nested/red/wnv.jsonl)
! - other families: afar (AFAR 24.3.0), cce19 (CCE 19.0.0), cce20 (CCE 20.0.2), cce21 (CCE 21.0.3), cce21_4 (CCE 21.0.4), nvhpc
!   (NVHPC 25.11), nvhpc269 (NVHPC 26.9) run the file correctly in managed mode (rows results/nested/red/tuo7.jsonl,
!   results/nested/red/wnv.jsonl, site rows, check case site-nested--gcc-fortran-root-and-element-pointer-components-ice); nvhpc
!   (NVHPC 25.11), nvhpc269 (NVHPC 26.9) fail it another way in copy mode, e.g. `Failing in Thread:1`
! - gate and confirmer: campaign gate of the programs below: 4 cells, 4 confirmed in the four environments (a-d; (c) on another node
!   or GPU; rows results/nested/main/gate_reruns.part.<host>.2711328.jsonl,
!   results/nested/main/gate_reruns.part.<host>.351539.jsonl,
!   results/nested/main/gate_reruns.part.wingtip.3374144.jsonl,
!   results/nested/main/gate_reruns.part.wingtip.3386107.jsonl,
!   results/nested/main/gate_reruns.part.wingtip.3582436.jsonl,
!   results/nested/main/gate_reruns.part.wingtip.3637854.jsonl,
!   results/nested/main/rows.part.<host>.2691037.jsonl, results/nested/main/rows.part.wingtip.3363229.jsonl,
!   results/nested/main/rows.part.wingtip.3363230.jsonl)
! - programs (2): nd-f-omp-tdata_nested-chain-d3-0fae50d2, nd-f-omp-tdata_nested-chain-d3-b0edb43f
! - legality: the rules (offload.engine.rules) accept it and AFAR's flang front end accepts it
program t
  use omp_lib
  implicit none
  type :: n1
    integer, pointer :: a(:) => null()
  end type n1
  type :: n0
    integer, pointer :: a(:) => null()
    type(n1), allocatable :: nx(:)
  end type n0
  type(n0) :: r
  integer :: q, hf, tot
  logical :: ok
  hf = 1
  !$omp target map(from: hf)
  hf = merge(1, 0, omp_is_initial_device())
  !$omp end target
  allocate(r%nx(2))
  allocate(r%a(4))
  allocate(r%nx(1)%a(4))
  allocate(r%nx(2)%a(4))
  r%a = 1
  r%nx(1)%a = 2
  r%nx(2)%a = 3
  !$omp target enter data map(to: r)
  !$omp target enter data map(to: r%a) map(to: r%nx(1)%a) map(to: r%nx(2)%a)
  !$omp target map(tofrom: tot)
    tot = sum(r%nx(1)%a) + sum(r%nx(2)%a) + sum(r%a)
  !$omp end target
  !$omp target exit data map(release: r%a) map(release: r%nx(1)%a) map(release: r%nx(2)%a)
  !$omp target exit data map(release: r)
  ok = tot == 24
  print '(a,i0)', 'tot= ', tot
  print '(a,i0,a,a)', 'RESULT ran_on_host=', hf, ' check=', merge('pass', 'fail', ok)
end program t

Source in the repository

Something to add? Dispute its legality · Report it fixed in a newer release · Report a mistake on this page

Expected and observed

Observed where it behaves incorrectly

site-nested--gcc-fortran-root-and-element-pointer-components-ice.f90:61:76:
   61 |   !$omp target enter data map(to: r%a) map(to: r%nx(1)%a) map(to: r%nx(2)%a)
      |                                                                            ^
internal compiler error: in gimplify_adjust_omp_clauses, at gimplify.cc:16128
0x2072e2d internal_error(char const*, ...)
	<dir>:787
0x7d943f fancy_abort(char const*, int, char const*)
	<dir>:1813

What the specification requires

The specification text this program relies on and the compiler is expected to support.

OpenMP 5.2 §5.8.3 p.151 l.13

The list items that appear in a map clause may include array sections and structure elements

OpenMP 5.2 §5.8.3 p.152 l.1-3

Evidence

Same compiler, one change

one change from the finding; correct on gcc_g90, gcc_sm80: not run yet

Control program

! control: the element pointers only
program t
  use omp_lib
  implicit none
  type :: n1
    integer, pointer :: a(:) => null()
  end type n1
  type :: n0
    integer, pointer :: a(:) => null()
    type(n1), allocatable :: nx(:)
  end type n0
  type(n0) :: r
  integer :: q, hf, tot
  logical :: ok
  hf = 1
  !$omp target map(from: hf)
  hf = merge(1, 0, omp_is_initial_device())
  !$omp end target
  allocate(r%nx(2))
  allocate(r%a(4))
  allocate(r%nx(1)%a(4))
  allocate(r%nx(2)%a(4))
  r%a = 1
  r%nx(1)%a = 2
  r%nx(2)%a = 3
  !$omp target enter data map(to: r)
  !$omp target enter data map(to: r%nx(1)%a) map(to: r%nx(2)%a)
  !$omp target map(tofrom: tot)
    tot = sum(r%nx(1)%a) + sum(r%nx(2)%a)
  !$omp end target
  !$omp target exit data map(release: r%nx(1)%a) map(release: r%nx(2)%a)
  !$omp target exit data map(release: r)
  ok = tot == 20
  print '(a,i0)', 'tot= ', tot
  print '(a,i0,a,a)', 'RESULT ran_on_host=', hf, ' check=', merge('pass', 'fail', ok)
end program t

one change from the finding; correct on gcc_g90, gcc_sm80: not run yet

Control program

! control: the root pointer only (elements not read)
program t
  use omp_lib
  implicit none
  type :: n1
    integer, pointer :: a(:) => null()
  end type n1
  type :: n0
    integer, pointer :: a(:) => null()
    type(n1), allocatable :: nx(:)
  end type n0
  type(n0) :: r
  integer :: q, hf, tot
  logical :: ok
  hf = 1
  !$omp target map(from: hf)
  hf = merge(1, 0, omp_is_initial_device())
  !$omp end target
  allocate(r%nx(2))
  allocate(r%a(4))
  allocate(r%nx(1)%a(4))
  allocate(r%nx(2)%a(4))
  r%a = 1
  r%nx(1)%a = 2
  r%nx(2)%a = 3
  !$omp target enter data map(to: r)
  !$omp target enter data map(to: r%a)
  !$omp target map(tofrom: tot)
    tot = sum(r%a)
  !$omp end target
  !$omp target exit data map(release: r%a)
  !$omp target exit data map(release: r)
  ok = tot == 4
  print '(a,i0)', 'tot= ', tot
  print '(a,i0,a,a)', 'RESULT ran_on_host=', hf, ' check=', merge('pass', 'fail', ok)
end program t

Other compilers run it correctly

AFAR 24.3.0, CCE 19.0.0, CCE 20.0.2, CCE 21.0.3, CCE 21.0.4, NVHPC 25.11, NVHPC 26.9