public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH 1/2] analyzer: consider that realloc could shrink the buffer [PR106539]
@ 2022-08-09 21:19 Tim Lange
  2022-08-09 21:19 ` [PATCH 2/2] analyzer: out-of-bounds checker [PR106000] Tim Lange
                   ` (2 more replies)
  0 siblings, 3 replies; 8+ messages in thread
From: Tim Lange @ 2022-08-09 21:19 UTC (permalink / raw)
  To: gcc-patches; +Cc: dmalcolm, Tim Lange

This patch adds the "shrinks buffer" case to the success_with_move
modelling of realloc.

2022-08-09  Tim Lange  <mail@tim-lange.me>

gcc/analyzer/ChangeLog:

	PR analyzer/106539
	* region-model-impl-calls.cc (region_model::impl_call_realloc):
	Add get_copied_size function and pass the result as the size of the
	new sized_region.

---
 gcc/analyzer/region-model-impl-calls.cc | 37 ++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)

diff --git a/gcc/analyzer/region-model-impl-calls.cc b/gcc/analyzer/region-model-impl-calls.cc
index 8c38e9206fa..50a19a52a21 100644
--- a/gcc/analyzer/region-model-impl-calls.cc
+++ b/gcc/analyzer/region-model-impl-calls.cc
@@ -737,9 +737,11 @@ region_model::impl_call_realloc (const call_details &cd)
 						  old_size_sval);
 	      const svalue *buffer_content_sval
 		= model->get_store_value (sized_old_reg, cd.get_ctxt ());
+	      const svalue *copied_size_sval
+		= get_copied_size (old_size_sval, new_size_sval);
 	      const region *sized_new_reg
 		= model->m_mgr->get_sized_region (new_reg, NULL,
-						  old_size_sval);
+						  copied_size_sval);
 	      model->set_value (sized_new_reg, buffer_content_sval,
 				cd.get_ctxt ());
 	    }
@@ -774,6 +776,39 @@ region_model::impl_call_realloc (const call_details &cd)
       else
 	return true;
     }
+
+  private:
+    /* Return the size svalue for the new region allocated by realloc.  */
+    const svalue *get_copied_size (const svalue *old_size_sval,
+				   const svalue *new_size_sval) const
+    {
+      tree old_size_cst = old_size_sval->maybe_get_constant ();
+      tree new_size_cst = new_size_sval->maybe_get_constant ();
+
+      if (old_size_cst && new_size_cst)
+	{
+	  /* Both are constants and comparable.  */
+	  tree cmp = fold_binary (LT_EXPR, boolean_type_node,
+				  old_size_cst, new_size_cst);
+
+	  if (cmp == boolean_true_node)
+	    return old_size_sval;
+	  else
+	    return new_size_sval;
+	}
+      else if (new_size_cst)
+	{
+	  /* OLD_SIZE_SVAL is symbolic, so return that.  */
+	  return old_size_sval;
+	}
+      else
+	{
+	  /* NEW_SIZE_SVAL is symbolic or both are symbolic.
+	     Return NEW_SIZE_SVAL, because implementations of realloc
+	     probably only moves the buffer if the new size is larger.  */
+	  return new_size_sval;
+	}
+    }
   };
 
   /* Body of region_model::impl_call_realloc.  */
-- 
2.37.1


^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2022-08-11 19:30 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-08-09 21:19 [PATCH 1/2] analyzer: consider that realloc could shrink the buffer [PR106539] Tim Lange
2022-08-09 21:19 ` [PATCH 2/2] analyzer: out-of-bounds checker [PR106000] Tim Lange
2022-08-09 22:44   ` David Malcolm
2022-08-09 22:02 ` [PATCH 1/2] analyzer: consider that realloc could shrink the buffer [PR106539] David Malcolm
2022-08-11 17:24 ` [PATCH 1/2 v2] " Tim Lange
2022-08-11 17:24   ` [PATCH 2/2 v2] analyzer: out-of-bounds checker [PR106000] Tim Lange
2022-08-11 19:30     ` David Malcolm
2022-08-11 19:25   ` [PATCH 1/2 v2] analyzer: consider that realloc could shrink the buffer [PR106539] David Malcolm

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).