This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[hsa] Fix segfault in fixup_child_record_type
- From: Martin Jambor <mjambor at suse dot cz>
- To: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Thu, 8 Oct 2015 19:11:17 +0200
- Subject: [hsa] Fix segfault in fixup_child_record_type
- Authentication-results: sourceware.org; auth=none
Hi,
I have committed the following patch to the hsa branch, toput the
bail-out test in fixup_child_record_type all the way to the top where
it should be, otherwise we get segfaults when gridifying kernels with
variable-sized variables.
Martin
2015-10-08 Martin Jambor <mjambor@suse.cz>
* omp-low.c (fixup_child_record_type): Bail out early if
ctx->receiver_decl is NULL.
diff --git a/gcc/omp-low.c b/gcc/omp-low.c
index 65cc5c3..a21d301 100644
--- a/gcc/omp-low.c
+++ b/gcc/omp-low.c
@@ -1613,6 +1613,8 @@ fixup_child_record_type (omp_context *ctx)
{
tree f, type = ctx->record_type;
+ if (!ctx->receiver_decl)
+ return;
/* ??? It isn't sufficient to just call remap_type here, because
variably_modified_type_p doesn't work the way we expect for
record types. Testing each field for whether it needs remapping
@@ -1652,9 +1654,8 @@ fixup_child_record_type (omp_context *ctx)
layout_type (type);
}
- if (ctx->receiver_decl)
- TREE_TYPE (ctx->receiver_decl)
- = build_qualified_type (build_reference_type (type), TYPE_QUAL_RESTRICT);
+ TREE_TYPE (ctx->receiver_decl)
+ = build_qualified_type (build_reference_type (type), TYPE_QUAL_RESTRICT);
}
/* Instantiate decls as necessary in CTX to satisfy the data sharing