Message-ID: <54CBDBAF.7010309@gmx.de>
Date: Fri, 30 Jan 2015 20:29:51 +0100
From: Heinrich Schuchardt <xypron.glpk@gmx.de>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Icedove/31.3.0
MIME-Version: 1.0
To: submit@bugs.x2go.org
Subject: nx-X11: Size of pointer 'server_support' used instead of size of
 its data
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit

package: nx-libs
version: head


In nx-libs/nx-X11/extras/Mesa/src/glx/x11/glxextensions.c
the length of the zeroed memory is the length of the pointer (4 bytes in
case of a 32bit system) not the length of the data the pointer points to
(8 byte).

static void
__glXProcessServerString( const struct extension_info * ext,
			  const char * server_string,
			  unsigned char * server_support )
{
   unsigned  base;
   unsigned  len;

   (void) memset( server_support, 0, sizeof( server_support ) );

Furthermore the length of the memory area pointed to by server_support
is defined in varying ways in the coding:

#define __GL_EXT_BYTES   ((__NUM_GL_EXTS + 7) / 8)


unsigned char server_support[ __GL_EXT_BYTES ];
unsigned char server_support[8];

Currently __NUM_GL_EXTS = 123, so __GL_EXT_BYTES = 8.

What is expected to happen if __GL_EXT_BYTES > 8 after defining six more
values in the unamed (sic!) enum with the different bits?

This questionable code was identified with cppcheck.
http://cppcheck.sourceforge.net/

Best regards

Heinrich Schuchardt