[gcc(refs/users/marxin/heads/marxin-gcc-benchmark-branch)] analyzer: round-trip pointer-equality through intptr_t

Martin Liska marxin@gcc.gnu.org
Mon Mar 30 10:51:13 GMT 2020


https://gcc.gnu.org/g:cb273d81a45092ceee793f0357526e291f03c7b7

commit cb273d81a45092ceee793f0357526e291f03c7b7
Author: David Malcolm <dmalcolm@redhat.com>
Date:   Wed Feb 5 12:09:23 2020 -0500

    analyzer: round-trip pointer-equality through intptr_t
    
    When investigating how the analyzer handles malloc/free of Cray pointers
    in gfortran I noticed that that analyzer was losing information on
    pointers that were cast to an integer type, and then back to a pointer
    type again.
    
    The root cause is that region_model::maybe_cast_1 was only preserving
    the region_svalue-ness of the result if both types were pointers,
    instead returning an unknown_svalue for a pointer-to-int cast.
    
    This patch updates the above code so that it attempts to use a
    region_svalue if *either* type is a pointer
    
    Doing so allows the analyzer to recognize that the same underlying
    region is in use through various casts through integer types.
    
    gcc/analyzer/ChangeLog:
            * region-model.cc (region_model::maybe_cast_1): Attempt to provide
            a region_svalue if either type is a pointer, rather than if both
            types are pointers.
    
    gcc/testsuite/ChangeLog:
            * gcc.dg/analyzer/torture/intptr_t.c: New test.

Diff:
---
 gcc/analyzer/ChangeLog                           |  6 +++++
 gcc/analyzer/region-model.cc                     |  2 +-
 gcc/testsuite/ChangeLog                          |  4 ++++
 gcc/testsuite/gcc.dg/analyzer/torture/intptr_t.c | 28 ++++++++++++++++++++++++
 4 files changed, 39 insertions(+), 1 deletion(-)

diff --git a/gcc/analyzer/ChangeLog b/gcc/analyzer/ChangeLog
index 0666d0028c6..ba59131ce4b 100644
--- a/gcc/analyzer/ChangeLog
+++ b/gcc/analyzer/ChangeLog
@@ -1,3 +1,9 @@
+2020-02-06  David Malcolm  <dmalcolm@redhat.com>
+
+	* region-model.cc (region_model::maybe_cast_1): Attempt to provide
+	a region_svalue if either type is a pointer, rather than if both
+	types are pointers.
+
 2020-02-05  David Malcolm  <dmalcolm@redhat.com>
 
 	* engine.cc (exploded_node::dump_dot): Show merger enodes.
diff --git a/gcc/analyzer/region-model.cc b/gcc/analyzer/region-model.cc
index c837ec6ed3b..60363c726d7 100644
--- a/gcc/analyzer/region-model.cc
+++ b/gcc/analyzer/region-model.cc
@@ -4977,7 +4977,7 @@ region_model::maybe_cast_1 (tree dst_type, svalue_id sid)
     return sid;
 
   if (POINTER_TYPE_P (dst_type)
-      && POINTER_TYPE_P (src_type))
+      || POINTER_TYPE_P (src_type))
     {
       /* Pointer to region.  */
       if (region_svalue *ptr_sval = sval->dyn_cast_region_svalue ())
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index fd80c8d0f32..f6f12dd764c 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,7 @@
+2020-02-06  David Malcolm  <dmalcolm@redhat.com>
+
+	* gcc.dg/analyzer/torture/intptr_t.c: New test.
+
 2020-02-06  Segher Boessenkool  <segher@kernel.crashing.org>
 
 	* gcc.target/powerpc/pr93012.c: New.
diff --git a/gcc/testsuite/gcc.dg/analyzer/torture/intptr_t.c b/gcc/testsuite/gcc.dg/analyzer/torture/intptr_t.c
new file mode 100644
index 00000000000..847ba626350
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/analyzer/torture/intptr_t.c
@@ -0,0 +1,28 @@
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } { "" } } */
+
+#include <stdlib.h>
+
+typedef __INTPTR_TYPE__ intptr_t;
+typedef __UINTPTR_TYPE__ uintptr_t;
+
+void test_1 (void)
+{
+  intptr_t ip;
+  void *p = malloc (1024);
+  ip = (intptr_t)p;
+  free ((void *)ip);
+} /* { dg-bogus "leak" } */
+
+void test_2 (void)
+{
+  uintptr_t uip;
+  void *p = malloc (1024);
+  uip = (uintptr_t)p;
+  free ((void *)uip);
+} /* { dg-bogus "leak" } */
+
+void test_3 (intptr_t ip)
+{
+  free ((void *)ip); /* { dg-message "first 'free'" } */
+  free ((void *)ip); /* { dg-warning "double-'free'" } */
+}


More information about the Gcc-cvs mailing list