[PATCH] Fix ether_line

[PATCH] Fix ether_line

[Jakub Jelinek]

Hi!

ether_line ("0:c0:f0:46:5f:97 host.ether.com \t# comment", ...)
would return 0:0:0:6:f:7 with -fwriteable-strings or segfault
otherwise. The following patch fixes both bugs.

2002-05-23  Jakub Jelinek  <jakub@redhat.com>

	* inet/ether_line.c (ether_line): Fix a typo causing only
	lower 4 bits of each ethernet address byte being assigned.
	Don't modify what line points to.
	* inet/tst-ether_aton.c (main): Add ether_line tests.

--- libc/inet/ether_line.c.jj	Thu Aug 23 18:47:43 2001
+++ libc/inet/ether_line.c	Thu May 23 18:38:03 2002
@@ -1,4 +1,4 @@
-/* Copyright (C) 1996, 1999 Free Software Foundation, Inc.
+/* Copyright (C) 1996, 1999, 2002 Free Software Foundation, Inc.
    This file is part of the GNU C Library.

    The GNU C Library is free software; you can redistribute it and/or
@@ -46,7 +46,7 @@ ether_line (const char *line, struct eth
 	  if ((ch < '0' || ch > '9') && (ch < 'a' || ch > 'f'))
 	    return -1;
 	  number <<= 4;
-	  number = isdigit (ch) ? (ch - '0') : (ch - 'a' + 10);
+	  number += isdigit (ch) ? (ch - '0') : (ch - 'a' + 10);
 
 	  ch = *line;
 	  if (cnt < 5 && ch != ':')
@@ -65,15 +65,15 @@ ether_line (const char *line, struct eth
   cp = __strchrnul (line, '#');
   while (cp > line && isspace (cp[-1]))
     --cp;
-  *cp = '\0';
 
-  if (*line == '\0')
+  if (cp == line)
     /* No hostname.  */
     return -1;
 
   /* XXX This can cause trouble because the hostname might be too long
      but we have no possibility to check it here.  */
-  strcpy (hostname, line);
+  memcpy (hostname, line, cp - line);
+  hostname [cp - line] = '\0';
 
   return 0;
 }
--- libc/inet/tst-ether_aton.c.jj	Fri May  1 12:39:33 1998
+++ libc/inet/tst-ether_aton.c	Thu May 23 18:44:14 2002
@@ -1,29 +1,57 @@
 #include <stdio.h>
+#include <string.h>
 #include <netinet/ether.h>
 
 int
 main (int argc, char *argv[])
 {
-  struct ether_addr *val;
-  int result;
+  struct ether_addr *valp, val;
+  int result, r;
+  char hostname[32], buf[64], *p;
 
-  val = ether_aton ("12:34:56:78:9a:bc");
+  valp = ether_aton ("12:34:56:78:9a:bc");
 
   printf ("ether_aton (\"12:34:56:78:9a:bc\") = %hhx:%hhx:%hhx:%hhx:%hhx:%hhx\n",
-	  val->ether_addr_octet[0],
-	  val->ether_addr_octet[1],
-	  val->ether_addr_octet[2],
-	  val->ether_addr_octet[3],
-	  val->ether_addr_octet[4],
-	  val->ether_addr_octet[5]);
-
-
-  result = (val->ether_addr_octet[0] != 0x12
-	    || val->ether_addr_octet[1] != 0x34
-	    || val->ether_addr_octet[2] != 0x56
-	    || val->ether_addr_octet[3] != 0x78
-	    || val->ether_addr_octet[4] != 0x9a
-	    || val->ether_addr_octet[5] != 0xbc);
+	  valp->ether_addr_octet[0],
+	  valp->ether_addr_octet[1],
+	  valp->ether_addr_octet[2],
+	  valp->ether_addr_octet[3],
+	  valp->ether_addr_octet[4],
+	  valp->ether_addr_octet[5]);
+
+  result = (valp->ether_addr_octet[0] != 0x12
+	    || valp->ether_addr_octet[1] != 0x34
+	    || valp->ether_addr_octet[2] != 0x56
+	    || valp->ether_addr_octet[3] != 0x78
+	    || valp->ether_addr_octet[4] != 0x9a
+	    || valp->ether_addr_octet[5] != 0xbc);
+
+  if ((r = ether_line ("0:c0:f0:46:5f:97 host.ether.com \t# comment",
+		       &val, hostname)) == 0)
+    {
+      ether_ntoa_r (&val, buf);
+      p = strchr (buf, '\0');
+      *p++ = ' ';
+      strcpy (p, hostname);
+
+      printf ("ether_line (\"0:c0:f0:46:5f:97 host.ether.com\") = \"%s\"\n",
+	      buf);
+
+      result |= strcmp ("0:c0:f0:46:5f:97 host.ether.com", buf) != 0;
+    }
+  else
+    {
+      printf ("ether_line (\"0:c0:f0:46:5f:97 host.ether.com\") = %d\n", r);
+      result |= 1;
+    }
+
+  r = ether_line ("0:c0:2:d0 foo.bar   ", &val, hostname);
+  printf ("ether_line (\"0:c0:2:d0 foo.bar   \") = %d\n", r);
+  result |= r != -1;
+
+  r = ether_line ("0:c0:2:d0:1a:2a  ", &val, hostname);
+  printf ("ether_line (\"0:c0:2:d0:1a:2a  \") = %d\n", r);
+  result |= r != -1;
 
   return result;
 }

	Jakub

Re: [PATCH] Fix ether_line

[Ulrich Drepper]

[-- Attachment #1: Type: text/plain, Size: 572 bytes --]

On Thu, 2002-05-23 at 09:59, Jakub Jelinek wrote:

> ether_line ("0:c0:f0:46:5f:97 host.ether.com \t# comment", ...)
> would return 0:0:0:6:f:7 with -fwriteable-strings or segfault
> otherwise. The following patch fixes both bugs.

Right.  Seems to be a very popular function.  It's been around forever. 
I've applied the patch.  Thanks,

-- 
---------------.                          ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Red Hat          `--' drepper at redhat.com   `------------------------

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 232 bytes --]