This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[PATCH] Fix uninitialized use of variables.


Compiling gdb head with gcc head reveals two uninitialized uses of
variables. Analysis of the uninitialized uses reveals that gcc is
correct. The following patch fixes both uninitiailized uses of
variables.

In remote.c (unpack_nibble) the variable *val is not set if ishex fails
to find hex digits. Callers of unpack_nibble expect *val to be set. The
solution is to check the return of ishex, and call error appropriately.
In the case that *val is not set, the uninitialized use is never reached
because we call error.

In symtab.c (find_line_common) the variable *exact_match is not set if
no match is found. Callers of find_line_common expect *exact_match to be
set. The solution is to initialize *exact_match to zero, assuming an
inexact match. In the case that we don't find a match in
find_line_common, the statement `(best_index < 0 || !exact)' in
symtab.c:2267 is true, instead of undefined. The comment is adjusted to
indicate that one must look at another symtab if we failed to find a
match `best_index < 0' or we found an inexact match `!exact.'

No regressions on i686-pc-linux-gnu. 

OK to checkin?

Cheers,
Carlos.
-- 
Carlos O'Donell
CodeSourcery
carlos@codesourcery.com
(650) 331-3385 x716

gdb/

2007-10-18  Carlos O'Donell  <carlos@codesourcery.com>

	* remote.c (unpack_nibble): error if buffer is not valid hex.
	* symtab.c (find_line_symtab): `no match' case is important.
	(find_line_common): Default *exact_match = 0 before search.

Index: gdb/remote.c
===================================================================
RCS file: /cvs/src/src/gdb/remote.c,v
retrieving revision 1.271
diff -u -p -r1.271 remote.c
--- gdb/remote.c	8 Oct 2007 12:55:09 -0000	1.271
+++ gdb/remote.c	18 Oct 2007 16:34:05 -0000
@@ -1343,7 +1343,8 @@ unpack_varlen_hex (char *buff,	/* packet
 static char *
 unpack_nibble (char *buf, int *val)
 {
-  ishex (*buf++, val);
+  if (!ishex (*buf++, val))
+    error (_("Unpacked nibble does not contain hex characters."));
   return buf;
 }
 
Index: gdb/symtab.c
===================================================================
RCS file: /cvs/src/src/gdb/symtab.c,v
retrieving revision 1.166
diff -u -p -r1.166 symtab.c
--- gdb/symtab.c	9 Oct 2007 06:59:27 -0000	1.166
+++ gdb/symtab.c	18 Oct 2007 16:34:05 -0000
@@ -2267,10 +2267,10 @@ find_line_symtab (struct symtab *symtab,
   best_index = find_line_common (best_linetable, line, &exact);
   if (best_index < 0 || !exact)
     {
-      /* Didn't find an exact match.  So we better keep looking for
-         another symtab with the same name.  In the case of xcoff,
-         multiple csects for one source file (produced by IBM's FORTRAN
-         compiler) produce multiple symtabs (this is unavoidable
+      /* Didn't find an exact match, or any match.  So we better keep 
+         looking for another symtab with the same name.  In the case of 
+         xcoff, multiple csects for one source file (produced by IBM's 
+         FORTRAN compiler) produce multiple symtabs (this is unavoidable
          assuming csects can be at arbitrary places in memory and that
          the GLOBAL_BLOCK of a symtab has a begin and end address).  */
 
@@ -2411,6 +2411,9 @@ find_line_common (struct linetable *l, i
   int best_index = -1;
   int best = 0;
 
+  /* Assume an inexact match.  */
+  *exact_match = 0;
+
   if (lineno <= 0)
     return -1;
   if (l == 0)
@@ -2436,8 +2439,6 @@ find_line_common (struct linetable *l, i
     }
 
   /* If we got here, we didn't get an exact match.  */
-
-  *exact_match = 0;
   return best_index;
 }
 


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]