linux-kernel - Re: Q: ptrace_signal() && PTRACE_SETSIGINFO (Was: SIGSTOP && ptrace)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20090424063159.BCEDAFC3C7@magilla.sf.frob.com>
Date:	Thu, 23 Apr 2009 23:31:59 -0700 (PDT)
From:	Roland McGrath <roland@...hat.com>
To:	Oleg Nesterov <oleg@...hat.com>
Cc:	linux-kernel@...r.kernel.org,
	Jan Kratochvil <jan.kratochvil@...hat.com>,
	Denys Vlasenko <dvlasenk@...hat.com>
Subject: Re: Q: ptrace_signal() && PTRACE_SETSIGINFO (Was: SIGSTOP && ptrace)

> Yes. PTRACE_SETSIGINFO can change *info if debugger wants something
> special. But then we do:
> 
> 	if (signr != info->si_signo) {
> 		info->si_signo = signr;
[...]
> Why? If the tracer changes ->exit_code it should know what it does.

If it uses PTRACE_SETSIGINFO it should know what it does, and update
the siginfo_t to match the signal it passes to PTRACE_CONT et al.

> Why do we reset *info?

PTRACE_SETSIGINFO did not always exist, and even now might not be used by a
simple-minded application.  If the user is sophisticated, it calls
PTRACE_SETSIGINFO and then passes the signal number to match.  If not, it
never calls PTRACE_SETSIGINFO at all, but expects the signal number it
chose to pass in PTRACE_CONT to behave "normally" in the tracee.  

We reset the siginfo_t the tracee will see to match what a kill() from the
debugger would have looked like.  Otherwise the tracee could be confused by
the siginfo_t values that don't make sense for the signal number delivered.
(The simple-minded debugger's ptrace code could even predate SA_SIGINFO
handlers and tracees that could see the siginfo_t.)

> But the real question, how can PTRACE_SETSIGINFO change ->si_signo
> (for example, for do_signal_stop(si_signo)) if this in fact is not
> allowed?

It's allowed.  You just have to pass the same value you set in si_signo as
the argument to PTRACE_CONT after you do PTRACE_SETSIGINFO.

Thanks,
Roland
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/