From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263839AbTDDQlp (for ); Fri, 4 Apr 2003 11:41:45 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263834AbTDDQk6 (for ); Fri, 4 Apr 2003 11:40:58 -0500 Received: from siaag2ab.compuserve.com ([149.174.40.132]:55446 "EHLO siaag2ab.compuserve.com") by vger.kernel.org with ESMTP id S263836AbTDDQhU (for ); Fri, 4 Apr 2003 11:37:20 -0500 Date: Fri, 4 Apr 2003 11:46:02 -0500 From: Chuck Ebbert <76306.1226@compuserve.com> Subject: [BUG?][2.5.66] was: Re: How to fix the ptrace flaw without rebooting To: Alan Cox Cc: linux-kernel Message-ID: <200304041148_MC3-1-32FC-A9A9@compuserve.com> MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Alan wrote: .>> # echo 'x'>/proc/sys/kernel/modprobe .>> bash: /proc/sys/kernel/modprobe: No such file or directory .>Thats not a sufficient fix except for people blindly running the .>example exploit I know that -- it's why I wrote pt_fix after seeing the problems people were having coming up with a stopgap. BTW putting a "ud2a" instruction at the entry point to sys_ptrace causes problems on SMP 2.5.66. On uniprocessor, all attempts to use strace fail with SIGSEGV as expected, but on SMP only the first time does. Further attempts to use strace result in a zombie strace child process hung up at the end of do_exit() (hand copied from sysrq-t): do_exit+0x2f7/0x310 die+0x7f/0x90 do_invalid_op+0x0/0xa0 do_invalid_op+0x87/0xa0 sys_ptrace+0x0/0x6c0 do_page_fault+0x120/0x426 error_code+0x2d/0x38 sys_ptrace+0x0/0x6c0 syscall_call+0x7/0xb 2.4.21pre6 handles this OK. -- Chuck