From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiederm@xmission.com (Eric W. Biederman) Subject: [PATCH 8/9] signal/task_exit: Use start_task_exit in place of do_exit Date: Thu, 24 Jun 2021 14:03:25 -0500 Message-ID: <87pmwb5blu.fsf_-_@disp2133> References: <87sg1lwhvm.fsf@disp2133> <6e47eff8-d0a4-8390-1222-e975bfbf3a65@gmail.com> <924ec53c-2fd9-2e1c-bbb1-3fda49809be4@gmail.com> <87eed4v2dc.fsf@disp2133> <5929e116-fa61-b211-342a-c706dcb834ca@gmail.com> <87fsxjorgs.fsf@disp2133> <87a6njf0ia.fsf@disp2133> <87tulpbp19.fsf@disp2133> <87zgvgabw1.fsf@disp2133> <875yy3850g.fsf_-_@disp2133> Mime-Version: 1.0 Return-path: In-Reply-To: <875yy3850g.fsf_-_@disp2133> (Eric W. Biederman's message of "Thu, 24 Jun 2021 13:57:35 -0500") List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Linus Torvalds Cc: Al Viro , Michael Schmitz , linux-arch , Jens Axboe , Oleg Nesterov , Linux Kernel Mailing List , Richard Henderson , Ivan Kokshaysky , Matt Turner , alpha , Geert Uytterhoeven , linux-m68k , Arnd Bergmann , Ley Foon Tan , Tejun Heo , Kees Cook Reuse start_task_exit_locked to implement start_task_exit. Simplify the exit logic by having all exits go through get_signal. This simplifies the analysis of syncrhonization during exit and gurantees a full set of registers will be available for ptrace to examine at PTRACE_EVENT_EXIT. Signed-off-by: "Eric W. Biederman" --- include/linux/sched/signal.h | 1 + kernel/exit.c | 4 +++- kernel/seccomp.c | 2 +- kernel/signal.c | 12 ++++++++++++ 4 files changed, 17 insertions(+), 2 deletions(-) diff --git a/include/linux/sched/signal.h b/include/linux/sched/signal.h index a958381ba4a9..3f4e69c019b7 100644 --- a/include/linux/sched/signal.h +++ b/include/linux/sched/signal.h @@ -430,6 +430,7 @@ static inline void ptrace_signal_wake_up(struct task_struct *t, bool resume) void start_group_exit(int exit_code); void start_task_exit_locked(struct task_struct *task, int exit_code); +void start_task_exit(int exit_code); void task_join_group_stop(struct task_struct *task); diff --git a/kernel/exit.c b/kernel/exit.c index 635f434122b7..51e0c82b3f7d 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -889,7 +889,9 @@ EXPORT_SYMBOL(complete_and_exit); SYSCALL_DEFINE1(exit, int, error_code) { - do_exit((error_code&0xff)<<8); + start_task_exit((error_code&0xff)<<8); + /* get_signal will call do_exit */ + return 0; } diff --git a/kernel/seccomp.c b/kernel/seccomp.c index b1c06fd1b205..e0c4c123a8bf 100644 --- a/kernel/seccomp.c +++ b/kernel/seccomp.c @@ -1248,7 +1248,7 @@ static int __seccomp_filter(int this_syscall, const struct seccomp_data *sd, force_sig_seccomp(this_syscall, data, true); } else { if (action == SECCOMP_RET_KILL_THREAD) - do_exit(SIGSYS); + start_task_exit(SIGSYS); else start_group_exit(SIGSYS); } diff --git a/kernel/signal.c b/kernel/signal.c index afbc001220dd..63fda9b6bbf9 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1424,6 +1424,18 @@ void start_task_exit_locked(struct task_struct *task, int exit_code) } } +void start_task_exit(int exit_code) +{ + struct task_struct *task = current; + if (!fatal_signal_pending(task)) { + struct sighand_struct *const sighand = task->sighand; + spin_lock_irq(&sighand->siglock); + if (!fatal_signal_pending(current)) + start_task_exit_locked(task, exit_code); + spin_unlock_irq(&sighand->siglock); + } +} + struct sighand_struct *__lock_task_sighand(struct task_struct *tsk, unsigned long *flags) { -- 2.20.1