From owner-freebsd-hackers Wed Dec 21 00:58:19 1994 Return-Path: hackers-owner Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id AAA00140 for hackers-outgoing; Wed, 21 Dec 1994 00:58:19 GMT Received: from sovcom.kiae.su (sovcom.kiae.su [144.206.136.1]) by freefall.cdrom.com (8.6.9/8.6.6) with SMTP id AAA00129 for ; Wed, 21 Dec 1994 00:58:10 GMT Received: by sovcom.kiae.su id AA19815 (5.65.kiae-2 for FreeBSD-hackers@freebsd.org); Tue, 20 Dec 1994 15:19:31 +0300 Received: by elvisti.kiev.ua (uumail/ache) id AA18019; Tue, 20 Dec 1994 13:10:45 +0200 Xref: store comp.os.386bsd.bugs:796 comp.os.386bsd.questions:6868 Newsgroups: comp.os.386bsd.bugs,comp.os.386bsd.questions Path: store!office.elvisti.kiev.ua!stesin From: stesin@elvisti.kiev.ua (Andrew V. Stesin) Subject: [?!] So, is this a bug in the kernel execve() code? X-Newsreader: TIN [version 1.2 PL2] Nntp-Posting-Host: office.elvisti.kiev.ua Organization: Electronni Visti InformAgency (ElVisti) Message-Id: Date: Tue, 20 Dec 1994 01:20:06 GMT Apparently-To: FreeBSD-hackers@freebsd.org Sender: hackers-owner@freebsd.org Precedence: bulk Hi BSDers, as I wrote a few days ago, there was a strange thing noticed with FreeBSD-1.1.5R. Here is an (incorrect, I know! ;) program in C: int main (int ac, char **av) { return (execve("/bin/ls", av[1], 0)); } anyway, given some command line parameters, this reboots the system immediately on our machine. Your mileage may vary -- this is the shortest example from the set. I can describe this as: broken arguments to execve() system call, like nonterminated argv[] array, or random arguments, -- may _often_ cause the immediate and silent OS reboot on our system. This effect was discovered by Natalie Vinokurova, nata@bitmcnit.bryansk.su. We tried to reproduce her example and (after a few attempts with "Bus error" and segfaults) we found it... for a pity... :( I tried to dig into that piece of kernel code, but sorry -- I'm not a wisard, I'm not a kernel hacker, I'm not even on a mailing list. I personally have no idea about how legal this behavior is and how to fix it. I don't even know -- was this reported yet or not? So, questions. 1. How valid the described behavior is? If it isn't considered valid, how one can fix it? 2. Is there a set of "official" patches for 1.1.5? Where? We're running vanilla 1.1.5.0 kernel, I only replaced some user-area utilities. Maybe some bugfix patches are needed? Thanks for your attention; comments and suggestions are welcome! -- With best wishes -- Andrew Stesin, system administrator.