[Bug c/24455] [gomp] Trouble with threadprivate and extern
--- Comment #6 from cvs-commit at gcc dot gnu dot org 2005-10-21 18:57 --- Subject: Bug 24455 CVSROOT:/cvs/gcc Module name:gcc Branch: gomp-20050608-branch Changes by: [EMAIL PROTECTED]2005-10-21 18:57:39 Modified files: gcc: ChangeLog.gomp c-decl.c c-parser.c c-tree.h gcc/cp : ChangeLog.gomp cp-tree.h decl.c semantics.c libgomp: ChangeLog Added files: libgomp/testsuite/libgomp.c++: pr24455-1.C pr24455.C libgomp/testsuite/libgomp.dg: pr24455-1.c pr24455.c Log message: gcc/ PR 24455 * c-tree.h (C_DECL_THREADPRIVATE_P): Define. * c-parser.c (c_parser_omp_threadprivate): Set. Do not error out if C_DECL_THREADPRIVATE_P is set already. * c-decl.c (diagnose_mismatched_decls): Do not check for mismatched thread-local attributes when OLDDECL is marked threadprivate and NEWDECL has no thread-local attributes. (merge_decls): Merge C_DECL_THREADPRIVATE_P. gcc/cp/ PR 24455 * cp/cp-tree.h (struct lang_decl_flags): Add field threadprivate_p. (CP_DECL_IS_THREADPRIVATE): Define. * cp/semantics.c (finish_omp_threadprivate): Set. Do not error out if CP_DECL_IS_THREADPRIVATE is set already. * cp/decl.c (duplicate_decls): Merge CP_DECL_THREADPRIVATE_P. libgomp/ PR 24455 * testsuite/libgomp.c++/pr24455-1.C: New test. * testsuite/libgomp.c++/pr24455.C: New test. * testsuite/libgomp.dg/pr24455-1.c: New test. * testsuite/libgomp.dg/pr24455.c: New test. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.gomp.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=1.1.6.97r2=1.1.6.98 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/c-decl.c.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=1.662.4.13r2=1.662.4.14 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/c-parser.c.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=2.17.4.37r2=2.17.4.38 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/c-tree.h.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=1.203.4.10r2=1.203.4.11 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.gomp.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=1.1.8.14r2=1.1.8.15 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/cp-tree.h.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=1.1144.4.13r2=1.1144.4.14 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/decl.c.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=1.1403.4.9r2=1.1403.4.10 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/semantics.c.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=1.475.4.12r2=1.475.4.13 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libgomp/ChangeLog.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=1.1.6.49r2=1.1.6.50 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libgomp/testsuite/libgomp.c++/pr24455-1.C.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=NONEr2=1.1.2.1 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libgomp/testsuite/libgomp.c++/pr24455.C.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=NONEr2=1.1.2.1 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libgomp/testsuite/libgomp.dg/pr24455-1.c.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=NONEr2=1.1.2.1 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libgomp/testsuite/libgomp.dg/pr24455.c.diff?cvsroot=gcconly_with_tag=gomp-20050608-branchr1=NONEr2=1.1.2.1 -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24455
[Bug c/24455] [gomp] Trouble with threadprivate and extern
--- Comment #7 from dnovillo at gcc dot gnu dot org 2005-10-21 18:59 --- Fixed. http://gcc.gnu.org/ml/gcc-patches/2005-10/msg01311.html -- dnovillo at gcc dot gnu dot org changed: What|Removed |Added Status|UNCONFIRMED |RESOLVED Resolution||FIXED http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24455
[Bug c/24455] [gomp] Trouble with threadprivate and extern
--- Comment #1 from dnovillo at redhat dot com 2005-10-20 16:45 --- Subject: Re: New: [gomp] Trouble with threadprivate and extern On Thursday 20 October 2005 12:34, reichelt at gcc dot gnu dot org wrote: I'd expect that i is threadprivate in file1.c and file2.c. But you have to mark it so in both places. Am I misreading the OpenMP spec or is this a bug in the frontends? I think you're misreading the spec, in 2.8.2. page 69, lines 4-6: If a variable is specified in a threadprivate directive in one translation unit, it must be specified in a threadprivate directive in every translation unit in which it is declared. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24455
[Bug c/24455] [gomp] Trouble with threadprivate and extern
--- Comment #2 from pinskia at gcc dot gnu dot org 2005-10-20 16:49 ---
(In reply to comment #1)
Subject: Re: New: [gomp] Trouble with threadprivate and extern
On Thursday 20 October 2005 12:34, reichelt at gcc dot gnu dot org wrote:
I'd expect that i is threadprivate in file1.c and file2.c.
But you have to mark it so in both places.
Huh, to me it is marked in both TUs.
---
file1.c is equivalant to:
extern int i;
#pragma omp threadprivate (i)
int main()
{
return i;
}
---
While file2.c is equivalant to:
extern int i;
#pragma omp threadprivate (i)
int i;
---
I don't see anything wrong here at all.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24455
[Bug c/24455] [gomp] Trouble with threadprivate and extern
--- Comment #3 from reichelt at igpm dot rwth-aachen dot de 2005-10-20 16:50 --- Subject: Re: [gomp] Trouble with threadprivate and extern On 20 Oct, dnovillo at redhat dot com wrote: --- Comment #1 from dnovillo at redhat dot com 2005-10-20 16:45 --- Subject: Re: New: [gomp] Trouble with threadprivate and extern On Thursday 20 October 2005 12:34, reichelt at gcc dot gnu dot org wrote: I'd expect that i is threadprivate in file1.c and file2.c. But you have to mark it so in both places. Am I misreading the OpenMP spec or is this a bug in the frontends? I think you're misreading the spec, in 2.8.2. page 69, lines 4-6: If a variable is specified in a threadprivate directive in one translation unit, it must be specified in a threadprivate directive in every translation unit in which it is declared. Doesn't translation unit cover the include file? But anyway. How should I mark it threadprivate in file2.c? Adding #pragma omp threadprivate (i) before or after int i; doesn't work. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24455
Re: [Bug c/24455] [gomp] Trouble with threadprivate and extern
On Thursday 20 October 2005 12:50, reichelt at igpm dot rwth-aachen dot de Doesn't translation unit cover the include file? But anyway. How should I mark it threadprivate in file2.c? Adding #pragma omp threadprivate (i) before or after int i; doesn't work. Indeed, sorry about that. I misread your code. I'll take a look.
[Bug c/24455] [gomp] Trouble with threadprivate and extern
--- Comment #4 from dnovillo at redhat dot com 2005-10-20 17:07 --- Subject: Re: [gomp] Trouble with threadprivate and extern On Thursday 20 October 2005 12:50, reichelt at igpm dot rwth-aachen dot de Doesn't translation unit cover the include file? But anyway. How should I mark it threadprivate in file2.c? Adding #pragma omp threadprivate (i) before or after int i; doesn't work. Indeed, sorry about that. I misread your code. I'll take a look. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24455
[Bug c/24455] [gomp] Trouble with threadprivate and extern
--- Comment #5 from dnovillo at gcc dot gnu dot org 2005-10-20 17:52 --- The problem here is that when we are parsing the second declaration, we call start_decl before having parsed the threadprivate clause. This causes us to miss the attribute. One solution I'm considering is to look-ahead in c_parser_declaration_or_fndef and build a regular tls_model attribute instead of lazily setting DECL_TLS_MODEL when we get to the threadprivate #pragma. That seems a more natural way of setting threadprivate attributes. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24455
