Date: Mon, 31 Jan 2005 13:12:12 +0300 From: Vladimir Grebenschikov <vova@fbsd.ru> To: Kris Kennaway <kris@obsecurity.org> Cc: "current@freebsd.org" <current@freebsd.org> Subject: Re: do_execve() finding vmspace_destroyed set under load Message-ID: <1107166332.1055.1.camel@localhost> In-Reply-To: <20050130025217.GA32612@xor.obsecurity.org> References: <20050130025217.GA32612@xor.obsecurity.org>
index | next in thread | previous in thread | raw e-mail
÷ ÓÂ, 29/01/2005 × 18:52 -0800, Kris Kennaway ÐÉÛÅÔ:
> I'm seeing the following code in do_execve() frequently being
> triggered when scripts are executed on my SMP machine under load:
Probably related, I have 100% panic on exec of single linux static
binary, it happens in do_execve
sorry, no console here, just retype:
Fatal trap 12: page fault in kernel mode
current process = 1638 (tcsh)
db> trace
exec_map_first_page(...) at exec_map_first_page+0x32
do_execve(...) at do_execve+0x2df
kern_execve(...) at kern_execve+0x124
execve(...)
...
(tcsh is parent process)
No problems with other linux or native binaries.
> if (imgp->vmspace_destroyed) {
> /* sorry, no more process anymore. exit gracefully */
> #ifdef MAC
> mac_execve_exit(imgp);
> if (interplabel != NULL)
> mac_vnode_label_free(interplabel);
> #endif
> exit1(td, W_EXITCODE(0, SIGABRT));
> /* NOT REACHED */
> error = 0;
> }
>
> Needless to say, the scripts get pretty unhappy when they're summarily
> aborted. What is the cause of this?
>
> Kris
--
Vladimir B. Grebenchikov
vova@fbsd.ru
help
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1107166332.1055.1.camel>