From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <systemtap-return-3126-listarch-systemtap=sources.redhat.com@sourceware.org>
Received: (qmail 7843 invoked by alias); 23 Apr 2006 04:32:58 -0000
Received: (qmail 7819 invoked by uid 48); 23 Apr 2006 04:32:53 -0000
Date: Sun, 23 Apr 2006 04:32:00 -0000
Message-ID: <20060423043253.7818.qmail@sourceware.org>
From: "eteo at redhat dot com" <sourceware-bugzilla@sourceware.org>
To: systemtap@sources.redhat.com
In-Reply-To: <20060112181937.2149.hunt@redhat.com>
References: <20060112181937.2149.hunt@redhat.com>
Reply-To: sourceware-bugzilla@sourceware.org
Subject: [Bug translator/2149] check return value for _stp_map_set_xx()
X-Bugzilla-Reason: AssignedTo
Mailing-List: contact systemtap-help@sourceware.org; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:systemtap-subscribe@sourceware.org>
List-Post: <mailto:systemtap@sourceware.org>
List-Help: <mailto:systemtap-help@sourceware.org>, <http://sourceware.org/lists.html#faqs>
Sender: systemtap-owner@sourceware.org
X-SW-Source: 2006-q2/txt/msg00204.txt.bz2


------- Additional Comments From eteo at redhat dot com  2006-04-23 04:32 -------
Before fix, _stp_map_set_xx do not have their return values checked:

[eteo@eteo src]$ stap -DMAXMAPENTRIES=1 -p3 -u -e 'probe begin { a[2]="hello" ;
a[3]="a"} global a' | grep _stp_map_set
     _stp_map_set_is (global_a, l->__tmp0, (l->__tmp2[0] ? l->__tmp2 : NULL));
     _stp_map_set_is (global_a, l->__tmp4, (l->__tmp6[0] ? l->__tmp6 : NULL));

After applying fix:

[eteo@eteo src]$ stap -DMAXMAPENTRIES=1 -p3 -u -e 'probe begin { a[2]="hello" ;
a[3]="a"} global a' | grep _stp_map_set
      { int rc = _stp_map_set_is (global_a, l->__tmp0, (l->__tmp2[0] ? l->__tmp2
: NULL)); if (unlikely(rc)) c->last_error = "Array overflow, check
MAXMAPENTRIES"; };
      { int rc = _stp_map_set_is (global_a, l->__tmp4, (l->__tmp6[0] ? l->__tmp6
: NULL)); if (unlikely(rc)) c->last_error = "Array overflow, check
MAXMAPENTRIES"; };

[eteo@eteo src]$ stap -DMAXMAPENTRIES=1 -u -e 'probe begin { a[2]="hello" ;
a[3]="a"} global a'
ERROR: Array overflow, check MAXMAPENTRIES near identifier 'a' at <input>:1:30
WARNING: Number of errors: 1, skipped probes: 0

Here's the patch. Please feedback.

Index: ChangeLog
===================================================================
RCS file: /cvs/systemtap/src/ChangeLog,v
retrieving revision 1.360
diff -u -3 -r1.360 ChangeLog
--- ChangeLog   23 Apr 2006 03:28:45 -0000      1.360
+++ ChangeLog   23 Apr 2006 04:24:42 -0000
@@ -1,5 +1,11 @@
 2006-04-23  Eugene Teo  <eteo@redhat.com>

+       PR 2149
+       * translate.cxx (mapvar::set): Test _stp_map_set_xx() for
+       array overflows.
+
+2006-04-23  Eugene Teo  <eteo@redhat.com>
+
        * small_demos/ansi_colors.stp: Add an example of using octal
        escape sequences to display all possible ansi colors.

Index: translate.cxx
===================================================================
RCS file: /cvs/systemtap/src/translate.cxx,v
retrieving revision 1.114
diff -u -3 -r1.114 translate.cxx
--- translate.cxx       22 Apr 2006 06:46:00 -0000      1.114
+++ translate.cxx       23 Apr 2006 04:24:46 -0000
@@ -581,14 +581,20 @@

   string set (vector<tmpvar> const & indices, tmpvar const & val) const
   {
+    string res = "{ int rc = ";
+
     // impedance matching: empty strings -> NULL
     if (type() == pe_string)
-      return (call_prefix("set", indices)
+      res += (call_prefix("set", indices)
              + ", (" + val.qname() + "[0] ? " + val.qname() + " : NULL))");
     else if (type() == pe_long)
-      return (call_prefix("set", indices) + ", " + val.qname() + ")");
+      res += (call_prefix("set", indices) + ", " + val.qname() + ")");
     else
       throw semantic_error("setting a value of an unsupported map type");
+
+    res += "; if (unlikely(rc)) c->last_error = \"Array overflow, check
MAXMAPENTRIES\"; }";
+
+    return res;
   }

   string hist() const

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |WAITING


http://sourceware.org/bugzilla/show_bug.cgi?id=2149

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.