public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "burnus at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug fortran/99400] New: OpenMP: ICE in install_var_field, at omp-low.c:789 with "map(alloc: A) map(to: A)"
Date: Fri, 05 Mar 2021 07:39:26 +0000 [thread overview]
Message-ID: <bug-99400-4@http.gcc.gnu.org/bugzilla/> (raw)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99400
Bug ID: 99400
Summary: OpenMP: ICE in install_var_field, at omp-low.c:789
with "map(alloc: A) map(to: A)"
Product: gcc
Version: 11.0
Status: UNCONFIRMED
Keywords: openmp, patch
Severity: normal
Priority: P3
Component: fortran
Assignee: unassigned at gcc dot gnu.org
Reporter: burnus at gcc dot gnu.org
Target Milestone: ---
SOLLVE_VV Fortran test – this one is
https://github.com/SOLLVE/sollve_vv/pull/306
TRACKING: Patch exists on a branch (and was possibly submitted for mainline at
some point) – but it is not yet on mainline.
The testcase does "map(alloc: A) map(to: A)" which should be valid OpenMP 5 and
invalid OpenMP 4.5
The C testcase fails on *mainline* with:
tests/5.0/target/test_target_mapping_before_alloc.c:40:41: error: ‘a’ appears
more than once in map clauses
On the *OG10* (devel/omp/gcc-10) branch, the C and Fortran testcases compile.
(I have no done a run-time check.) Hence, one of Chung-Lin's mapping patches
did fix it...
TODO: Wait until it lands and then commit a Fortran testcase as well.
On mainline, gfortran fails as follows:
gfortran -fopenmp -I ompvv test_target_mapping_before_alloc.F90
during GIMPLE pass: omplower
test_target_mapping_before_alloc.F90:51:43:
51 | !$omp& test_struct) map(tofrom: errors)
| ^
internal compiler error: in install_var_field, at omp-low.c:789
0x6ed020 install_var_field
../../repos/gcc/gcc/omp-low.c:789
0xe06466 scan_sharing_clauses
../../repos/gcc/gcc/omp-low.c:1543
0xe07cdb scan_omp_target
../../repos/gcc/gcc/omp-low.c:2946
0xe08c14 scan_omp_1_stmt
../../repos/gcc/gcc/omp-low.c:4029
* * *
INTEGER :: errors, x, scalar
INTEGER,DIMENSION(N) :: a
TYPE structure
INTEGER :: var
INTEGER,DIMENSION(N) :: b
END TYPE structure
TYPE(structure) :: test_struct
test_struct%var = 1
DO x = 1, N
a(x) = x
test_struct%b(x) = x
END DO
scalar = 80
errors = 0
!$omp target map(alloc: scalar, a, test_struct) map(to: scalar, a, &
!$omp& test_struct) map(tofrom: errors)
IF (scalar .ne. 80 .OR. a(2) .ne. 2 .OR. test_struct%var .ne. 1 &
& .OR. test_struct%b(2) .ne. 2) THEN
errors = errors + 1
END IF
!$omp end target
next reply other threads:[~2021-03-05 7:39 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-05 7:39 burnus at gcc dot gnu.org [this message]
2021-07-20 16:56 ` [Bug fortran/99400] " gscfq@t-online.de
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=bug-99400-4@http.gcc.gnu.org/bugzilla/ \
--to=gcc-bugzilla@gcc.gnu.org \
--cc=gcc-bugs@gcc.gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).