All of lore.kernel.org
 help / color / mirror / Atom feed
From: ebiederm@xmission.com (Eric W. Biederman)
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andy Lutomirski <luto@kernel.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	linux-arch <linux-arch@vger.kernel.org>,
	Oleg Nesterov <oleg@redhat.com>,
	Al Viro <viro@zeniv.linux.org.uk>,
	Kees Cook <keescook@chromium.org>
Subject: Re: [PATCH 13/20] signal: Implement force_fatal_sig
Date: Mon, 25 Oct 2021 23:57:48 -0500	[thread overview]
Message-ID: <87v91kqt6b.fsf@disp2133> (raw)
In-Reply-To: <CAHk-=whJETM0MHqWQKCVALBkJX-Th5471z5FW3gFJO5c73L6QA@mail.gmail.com> (Linus Torvalds's message of "Mon, 25 Oct 2021 16:15:44 -0700")

Linus Torvalds <torvalds@linux-foundation.org> writes:

> On Mon, Oct 25, 2021 at 3:41 PM Andy Lutomirski <luto@kernel.org> wrote:
>>
>> I'm rather nervous about all this, and I'm also nervous about the
>> existing code.  A quick skim is finding plenty of code paths that assume
>> force_sigsegv (or a do_exit that this series touches) are genuinely
>> unrecoverable.
>
> I was going to say "what are you talking about", because clearly Eric
> kept it all fatal.
>
> But then looked at that patch a bit more before I claimed you were wrong.
>
> And yeah, Eric's force_fatal_sig() is completely broken.
>
> It claims to force a fatal signal, but doesn't actually do that at
> all, and is completely misnamed.
>
> It just uses "force_sig_info_to_task()", which still allows user space
> to catch signals - so it's not "fatal" in the least. It only punches
> through SIG_IGN and blocked signals.

Rereading this I think you might be misreading something.

force_siginfo_to_task takes a sigdfl parameter which I am setting in
force_fatal_signal.

When that sigdfl paramter is set force_siginfo_to_task always changes
the signal handler to SIGDFL, and always unblocks the signal.

Because the siglock remains held over send_signal none of those
properties can change during send_signal.  Which means that as long
as we are not talking about a coredump signal complete_signal is
guaranteed to recognize the signal as fatal immediately.

For coredump signals there is a race where siglock is dropped before
get_signal is called that could result in the signal handler being
changed or the signal being blocked.  Which is why I pointed out the
problem is coredumps.

But assuming userspace does not change something in that narrow window
the signal will most definitely be fatal to the target process.

Just as soon as I know if we can have per signal_struct coredumps
without causing regressions I will close the final race.  I can do it
either way but the code is much less complicated with per signal_struct
coredumps.

Hoisting the current zap_threads from fs/coredump.c into complete_signal
is a pain and a half.  While just the per_signal struct part is already
there, and the code just needs a few tweaks to allow get_signal to act
as the coredump rendezvous location.

Eric

  parent reply	other threads:[~2021-10-26  4:58 UTC|newest]

Thread overview: 110+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-20 17:32 [PATCH 00/20] exit cleanups Eric W. Biederman
2021-10-20 17:32 ` [OpenRISC] " Eric W. Biederman
2021-10-20 17:32 ` Eric W. Biederman
2021-10-20 17:43 ` [PATCH 01/20] exit/doublefault: Remove apparently bogus comment about rewind_stack_do_exit Eric W. Biederman
2021-10-21 16:02   ` Kees Cook
2021-10-20 17:43 ` [PATCH 02/20] exit: Remove calls of do_exit after noreturn versions of die Eric W. Biederman
2021-10-20 17:43   ` [OpenRISC] " Eric W. Biederman
2021-10-21 16:02   ` Kees Cook
2021-10-21 16:02     ` [OpenRISC] " Kees Cook
2021-10-21 16:25     ` Eric W. Biederman
2021-10-21 16:25       ` [OpenRISC] " Eric W. Biederman
2021-10-20 17:43 ` [PATCH 03/20] reboot: Remove the unreachable panic after do_exit in reboot(2) Eric W. Biederman
2021-10-21 16:05   ` Kees Cook
2021-10-20 17:43 ` [PATCH 04/20] signal/sparc32: Remove unreachable do_exit in do_sparc_fault Eric W. Biederman
2021-10-21 16:05   ` Kees Cook
2021-10-20 17:43 ` [PATCH 05/20] signal/mips: Update (_save|_restore)_fp_context to fail with -EFAULT Eric W. Biederman
2021-10-21 16:06   ` Kees Cook
2021-10-24  4:24   ` Maciej W. Rozycki
2021-10-25 20:55     ` Eric W. Biederman
2021-10-24 15:27   ` Thomas Bogendoerfer
2021-10-20 17:43 ` [PATCH 06/20] signal/sh: Use force_sig(SIGKILL) instead of do_group_exit(SIGKILL) Eric W. Biederman
2021-10-20 19:57   ` Linus Torvalds
2021-10-27 14:24     ` Rich Felker
2021-10-21 16:08   ` Kees Cook
2021-10-20 17:43 ` [PATCH 07/20] signal/powerpc: On swapcontext failure force SIGSEGV Eric W. Biederman
2021-10-20 17:43   ` Eric W. Biederman
2021-10-21 16:09   ` Kees Cook
2021-10-21 16:09     ` Kees Cook
2021-10-20 17:43 ` [PATCH 08/20] signal/sparc: In setup_tsb_params convert open coded BUG into BUG Eric W. Biederman
2021-10-21 16:12   ` Kees Cook
2021-10-20 17:43 ` [PATCH 09/20] signal/vm86_32: Replace open coded BUG_ON with an actual BUG_ON Eric W. Biederman
2021-10-21 16:15   ` Kees Cook
2021-11-12 15:40   ` Eric W. Biederman
2021-11-12 17:51     ` Brian Gerst
2021-11-12 19:57       ` Eric W. Biederman
2021-11-12 20:40         ` Linus Torvalds
2021-11-12 21:03           ` Eric W. Biederman
2021-11-12 21:23             ` Linus Torvalds
2021-11-12 21:24               ` Linus Torvalds
2021-11-12 21:37                 ` [GIT PULL ] signal/vm86_32: Remove pointless test in BUG_ON Eric W. Biederman
2021-11-13 19:15                   ` pr-tracker-bot
2021-11-12 21:43                 ` [PATCH 09/20] signal/vm86_32: Replace open coded BUG_ON with an actual BUG_ON Eric W. Biederman
2021-10-20 17:43 ` [PATCH 10/20] signal/vm86_32: Properly send SIGSEGV when the vm86 state cannot be saved Eric W. Biederman
2021-10-21 16:16   ` Kees Cook
2021-10-21 17:02     ` Eric W. Biederman
2021-10-21 20:33       ` Kees Cook
2021-10-21 23:08   ` Andy Lutomirski
2021-10-24 16:06     ` Eric W. Biederman
     [not found]   ` <875ytkygfj.fsf_-_@disp2133>
2021-10-25 21:12     ` [PATCH v2 10/32] " Linus Torvalds
2021-10-25 21:28       ` Eric W. Biederman
2021-10-25 22:25     ` Andy Lutomirski
2021-10-25 23:45       ` Linus Torvalds
2021-10-26  0:21         ` Andy Lutomirski
2021-10-20 17:43 ` [PATCH 11/20] signal/s390: Use force_sigsegv in default_trap_handler Eric W. Biederman
2021-10-21 16:17   ` Kees Cook
2021-10-26  9:38   ` Christian Borntraeger
2021-10-28 15:56     ` Eric W. Biederman
2021-10-29 19:32       ` Eric W. Biederman
2021-10-20 17:43 ` [PATCH 12/20] exit/kthread: Have kernel threads return instead of calling do_exit Eric W. Biederman
2021-10-21 11:12   ` Christoph Hellwig
2021-10-21 15:11     ` Eric W. Biederman
2021-10-21 16:21   ` Kees Cook
2021-10-20 17:43 ` [PATCH 13/20] signal: Implement force_fatal_sig Eric W. Biederman
2021-10-20 20:05   ` Linus Torvalds
2021-10-20 21:25     ` Eric W. Biederman
2021-10-25 22:41     ` Andy Lutomirski
2021-10-25 23:15       ` Linus Torvalds
2021-10-26  4:45         ` Eric W. Biederman
2021-10-26  4:57         ` Eric W. Biederman [this message]
2021-10-26 16:15           ` Linus Torvalds
2021-10-28 16:33             ` Eric W. Biederman
2021-10-21 16:24   ` Kees Cook
2021-10-21 16:33     ` Eric W. Biederman
2021-10-21 16:39       ` Kees Cook
2021-10-20 17:44 ` [PATCH 14/20] exit/syscall_user_dispatch: Send ordinary signals on failure Eric W. Biederman
2021-10-21 16:25   ` Kees Cook
2021-10-21 16:37     ` Eric W. Biederman
2021-10-21 16:40       ` Kees Cook
2021-10-21 17:05         ` Eric W. Biederman
2021-10-25 22:32     ` Andy Lutomirski
2021-10-21 16:35   ` Gabriel Krisman Bertazi
2021-10-20 17:44 ` [PATCH 15/20] signal/sparc32: Exit with a fatal signal when try_to_clear_window_buffer fails Eric W. Biederman
2021-10-21 16:34   ` Kees Cook
2021-10-21 16:56     ` Eric W. Biederman
2021-10-20 17:44 ` [PATCH 16/20] signal/sparc32: In setup_rt_frame and setup_fram use force_fatal_sig Eric W. Biederman
2021-10-21 16:34   ` Kees Cook
2021-10-20 17:44 ` [PATCH 17/20] signal/x86: In emulate_vsyscall force a signal instead of calling do_exit Eric W. Biederman
2021-10-21 16:36   ` Kees Cook
2021-10-20 17:44 ` [PATCH 18/20] exit/rtl8723bs: Replace the macro thread_exit with a simple return 0 Eric W. Biederman
2021-10-21  7:06   ` Greg KH
2021-10-21 15:06     ` Eric W. Biederman
2021-10-21 16:37   ` Kees Cook
2021-10-20 17:44 ` [PATCH 19/20] exit/rtl8712: " Eric W. Biederman
2021-10-21  7:07   ` Greg KH
2021-10-21 16:37   ` Kees Cook
2021-10-20 17:44 ` [PATCH 20/20] exit/r8188eu: " Eric W. Biederman
2021-10-21  7:07   ` Greg KH
2021-10-21 16:37   ` Kees Cook
2021-10-20 21:51 ` [PATCH 21/20] signal: Replace force_sigsegv(SIGSEGV) with force_fatal_sig(SIGSEGV) Eric W. Biederman
2021-10-20 21:51   ` [OpenRISC] " Eric W. Biederman
2021-10-20 21:51   ` Eric W. Biederman
2021-10-21  8:09   ` Geert Uytterhoeven
2021-10-21  8:09     ` [OpenRISC] " Geert Uytterhoeven
2021-10-21  8:09     ` Geert Uytterhoeven
2021-10-21 13:33     ` Eric W. Biederman
2021-10-21 13:33       ` [OpenRISC] " Eric W. Biederman
2021-10-21 13:33       ` Eric W. Biederman
2021-10-21  8:32   ` Philippe Mathieu-Daudé
2021-10-21  8:32     ` [OpenRISC] " Philippe =?unknown-8bit?q?Mathieu-Daud=C3=A9?=
2021-10-21  8:32     ` Philippe Mathieu-Daudé

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87v91kqt6b.fsf@disp2133 \
    --to=ebiederm@xmission.com \
    --cc=keescook@chromium.org \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luto@kernel.org \
    --cc=oleg@redhat.com \
    --cc=torvalds@linux-foundation.org \
    --cc=viro@zeniv.linux.org.uk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.