Fix use of COMPLETE_TYPE_P for -Wstrict-aliasing=1
Richard Sandiford
richard.sandiford@arm.com
Thu Mar 21 13:14:00 GMT 2019
The handling of -Wstrict-aliasing=1 applied COMPLETE_TYPE_P to the
pointer type rather than the pointer target, so missed the warnings
for "struct incomplete" in the testcase.
I couldn't find any existing C tests for -Wstrict-aliasing=1,
so I added a few extra tests besides the ones fixed by the patch.
I'm sure there's lots more we could test -- this is just supposed
to be better than the status quo (i.e. nothing).
Tested on aarch64-linux-gnu. Not sure this is a regression in any
meaningful sense, so I guess it should wait for GCC 10. OK for stage 1?
Richard
2019-03-21 Richard Sandiford <richard.sandiford@arm.com>
gcc/c-family/
* c-warn.c (strict_aliasing_warning): Apply COMPLETE_TYPE_P to
the pointer target rather than the pointer itself.
gcc/testsuite/
* gcc.dg/alias-16.c: New test.
Index: gcc/c-family/c-warn.c
===================================================================
--- gcc/c-family/c-warn.c 2019-03-21 13:01:18.000000000 +0000
+++ gcc/c-family/c-warn.c 2019-03-21 13:01:18.453582721 +0000
@@ -746,7 +746,7 @@ strict_aliasing_warning (location_t loc,
are not revealed at higher levels. */
alias_set_type set1 = get_alias_set (TREE_TYPE (otype));
alias_set_type set2 = get_alias_set (TREE_TYPE (type));
- if (!COMPLETE_TYPE_P (type)
+ if (!COMPLETE_TYPE_P (TREE_TYPE (type))
|| !alias_sets_must_conflict_p (set1, set2))
{
warning_at (loc, OPT_Wstrict_aliasing,
Index: gcc/testsuite/gcc.dg/alias-16.c
===================================================================
--- /dev/null 2019-03-08 11:40:14.606883727 +0000
+++ gcc/testsuite/gcc.dg/alias-16.c 2019-03-21 13:01:18.453582721 +0000
@@ -0,0 +1,46 @@
+/* { dg-do compile } */
+/* { dg-options "-Wstrict-aliasing=1 -fstrict-aliasing" } */
+
+struct incomplete;
+struct s1 { int i; };
+struct s2 { double d; };
+
+void
+f (int *i, double *d, struct s1 *s1, struct s2 *s2, char *c)
+{
+ (char *) i;
+ (char *) d;
+ (char *) s1;
+ (char *) s2;
+ (char *) c;
+
+ (int *) i;
+ (int *) d; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (int *) s1; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (int *) s2; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (int *) c;
+
+ (double *) i; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (double *) d;
+ (double *) s1; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (double *) s2; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (double *) c;
+
+ (struct incomplete *) i; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct incomplete *) d; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct incomplete *) s1; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct incomplete *) s2; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct incomplete *) c; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+
+ (struct s1 *) i; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct s1 *) d; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct s1 *) s1;
+ (struct s1 *) s2; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct s1 *) c;
+
+ (struct s2 *) i; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct s2 *) d; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct s2 *) s1; /* { dg-warning "dereferencing type-punned pointer might break strict-aliasing rules" } */
+ (struct s2 *) s2;
+ (struct s2 *) c;
+}
More information about the Gcc-patches
mailing list