| 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
| ||
|
Message-ID: <202106252038.389B963B6F@keescook>
Date: Fri, 25 Jun 2021 20:42:06 -0700
From: Kees Cook <keescook@...omium.org>
To: "Eric W. Biederman" <ebiederm@...ssion.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
Al Viro <viro@...iv.linux.org.uk>,
Michael Schmitz <schmitzmic@...il.com>,
linux-arch <linux-arch@...r.kernel.org>,
Jens Axboe <axboe@...nel.dk>, Oleg Nesterov <oleg@...hat.com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Richard Henderson <rth@...ddle.net>,
Ivan Kokshaysky <ink@...assic.park.msu.ru>,
Matt Turner <mattst88@...il.com>,
alpha <linux-alpha@...r.kernel.org>,
Geert Uytterhoeven <geert@...ux-m68k.org>,
linux-m68k <linux-m68k@...ts.linux-m68k.org>,
Arnd Bergmann <arnd@...nel.org>, Tejun Heo <tj@...nel.org>
Subject: Re: [PATCH 6/9] signal: Fold do_group_exit into get_signal fixing
io_uring threads
On Thu, Jun 24, 2021 at 02:02:16PM -0500, Eric W. Biederman wrote:
>
> Forld do_group_exit into get_signal as it is the last caller.
>
> Move the group_exit logic above the PF_IO_WORKER exit, ensuring
> that if an PF_IO_WORKER catches SIGKILL every thread in
> the thread group will exit not just the the PF_IO_WORKER.
>
> Now that the information is easily available only set PF_SIGNALED
> when it was a signal that caused the exit.
>
> Signed-off-by: "Eric W. Biederman" <ebiederm@...ssion.com>
> ---
> include/linux/sched/task.h | 1 -
> kernel/exit.c | 31 -------------------------------
> kernel/signal.c | 35 +++++++++++++++++++++++++----------
> 3 files changed, 25 insertions(+), 42 deletions(-)
>
> diff --git a/include/linux/sched/task.h b/include/linux/sched/task.h
> index ef02be869cf2..45525512e3d0 100644
> --- a/include/linux/sched/task.h
> +++ b/include/linux/sched/task.h
> @@ -77,7 +77,6 @@ static inline void exit_thread(struct task_struct *tsk)
> {
> }
> #endif
> -extern void do_group_exit(int);
>
> extern void exit_files(struct task_struct *);
> extern void exit_itimers(struct signal_struct *);
> diff --git a/kernel/exit.c b/kernel/exit.c
> index 921519d80b56..635f434122b7 100644
> --- a/kernel/exit.c
> +++ b/kernel/exit.c
> @@ -892,37 +892,6 @@ SYSCALL_DEFINE1(exit, int, error_code)
> do_exit((error_code&0xff)<<8);
> }
>
> -/*
> - * Take down every thread in the group. This is called by fatal signals
> - * as well as by sys_exit_group (below).
> - */
> -void
> -do_group_exit(int exit_code)
> -{
> - struct signal_struct *sig = current->signal;
> -
> - BUG_ON(exit_code & 0x80); /* core dumps don't get here */
> -
> - if (signal_group_exit(sig))
> - exit_code = sig->group_exit_code;
> - else if (!thread_group_empty(current)) {
> - struct sighand_struct *const sighand = current->sighand;
> -
> - spin_lock_irq(&sighand->siglock);
> - if (signal_group_exit(sig))
> - /* Another thread got here before we took the lock. */
> - exit_code = sig->group_exit_code;
> - else {
> - sig->group_exit_code = exit_code;
> - sig->flags = SIGNAL_GROUP_EXIT;
> - zap_other_threads(current);
Oh, now I see it: the "new code" in start_group_exit() is an open-coded
zap_other_threads()? That wasn't clear to me, but makes sense now.
> - }
> - spin_unlock_irq(&sighand->siglock);
> - }
> -
> - do_exit(exit_code);
> - /* NOTREACHED */
> -}
>
> /*
> * this kills every thread in the thread group. Note that any externally
> diff --git a/kernel/signal.c b/kernel/signal.c
> index c79c010ca5f3..95a076af600a 100644
> --- a/kernel/signal.c
> +++ b/kernel/signal.c
> @@ -2646,6 +2646,7 @@ bool get_signal(struct ksignal *ksig)
> {
> struct sighand_struct *sighand = current->sighand;
> struct signal_struct *signal = current->signal;
> + int exit_code;
> int signr;
>
> if (unlikely(current->task_works))
> @@ -2848,8 +2849,6 @@ bool get_signal(struct ksignal *ksig)
> /*
> * Anything else is fatal, maybe with a core dump.
> */
> - current->flags |= PF_SIGNALED;
> -
> if (sig_kernel_coredump(signr)) {
> if (print_fatal_signals)
> print_fatal_signal(ksig->info.si_signo);
> @@ -2857,14 +2856,33 @@ bool get_signal(struct ksignal *ksig)
> /*
> * If it was able to dump core, this kills all
> * other threads in the group and synchronizes with
> - * their demise. If we lost the race with another
> - * thread getting here, it set group_exit_code
> - * first and our do_group_exit call below will use
> - * that value and ignore the one we pass it.
> + * their demise. If another thread makes it
> + * to do_coredump first, it will set group_exit_code
> + * which will be passed to do_exit.
> */
> do_coredump(&ksig->info);
> }
>
> + /*
> + * Death signals, no core dump.
> + */
> + exit_code = signr;
> + if (signal_group_exit(signal)) {
> + exit_code = signal->group_exit_code;
> + } else {
> + spin_lock_irq(&sighand->siglock);
> + if (signal_group_exit(signal)) {
> + /* Another thread got here before we took the lock. */
> + exit_code = signal->group_exit_code;
> + } else {
> + start_group_exit_locked(signal, exit_code);
And here's the "if we didn't already do start_group_exit(), do it here".
And that state is entirely captured via the SIGNAL_GROUP_EXIT flag.
Cool.
> + }
> + spin_unlock_irq(&sighand->siglock);
> + }
> +
> + if (exit_code & 0x7f)
> + current->flags |= PF_SIGNALED;
> +
> /*
> * PF_IO_WORKER threads will catch and exit on fatal signals
> * themselves. They have cleanup that must be performed, so
> @@ -2873,10 +2891,7 @@ bool get_signal(struct ksignal *ksig)
> if (current->flags & PF_IO_WORKER)
> goto out;
>
> - /*
> - * Death signals, no core dump.
> - */
> - do_group_exit(ksig->info.si_signo);
> + do_exit(exit_code);
> /* NOTREACHED */
> }
> spin_unlock_irq(&sighand->siglock);
> --
> 2.20.1
>
--
Kees Cook
Powered by blists - more mailing lists