mbox series

[0/12] ptrace: cleaning up ptrace_stop

Message ID 87k0b7v9yk.fsf_-_@email.froward.int.ebiederm.org
Headers show
Series ptrace: cleaning up ptrace_stop | expand

Message

Eric W. Biederman April 29, 2022, 9:46 p.m. UTC 
The states TASK_STOPPED and TASK_TRACE are special in they can not
handle spurious wake-ups.  This plus actively depending upon and
changing the value of tsk->__state causes problems for PREEMPT_RT and
Peter's freezer rewrite.

There are a lot of details we have to get right to sort out the
technical challenges and this is my parred back version of the changes
that contains just those problems I see good solutions to that I believe
are ready.

In particular I don't have a solution that is ready for the challenges
presented by wait_task_inactive.

I hope we can review these changes and then have a firm foundation
for the rest of the challenges.

There are cleanups to the ptrace support for xtensa, um, and
ia64.

I have sucked in the first patch of Peter's freezer change as
with minor modifications I believe it is ready to go.

Eric W. Biederman (12):
      signal: Rename send_signal send_signal_locked
      signal: Replace __group_send_sig_info with send_signal_locked
      ptrace/um: Replace PT_DTRACE with TIF_SINGLESTEP
      ptrace/xtensa: Replace PT_SINGLESTEP with TIF_SINGLESTEP
      signal: Use lockdep_assert_held instead of assert_spin_locked
      ptrace: Reimplement PTRACE_KILL by always sending SIGKILL
      ptrace: Don't change __state
      ptrace: Remove arch_ptrace_attach
      ptrace: Always take siglock in ptrace_resume
      ptrace: Only return signr from ptrace_stop if it was provided
      ptrace: Always call schedule in ptrace_stop
      sched,signal,ptrace: Rework TASK_TRACED, TASK_STOPPED state

 arch/ia64/include/asm/ptrace.h    |   4 --
 arch/ia64/kernel/ptrace.c         |  57 ----------------
 arch/um/include/asm/thread_info.h |   2 +
 arch/um/kernel/exec.c             |   2 +-
 arch/um/kernel/process.c          |   2 +-
 arch/um/kernel/ptrace.c           |   8 +--
 arch/um/kernel/signal.c           |   4 +-
 arch/xtensa/kernel/ptrace.c       |   4 +-
 arch/xtensa/kernel/signal.c       |   4 +-
 drivers/tty/tty_jobctrl.c         |   4 +-
 include/linux/ptrace.h            |   7 --
 include/linux/sched.h             |  10 ++-
 include/linux/sched/jobctl.h      |  10 +++
 include/linux/sched/signal.h      |  23 ++++++-
 include/linux/signal.h            |   3 +-
 kernel/ptrace.c                   |  88 +++++++++----------------
 kernel/signal.c                   | 135 +++++++++++++++++---------------------
 kernel/time/posix-cpu-timers.c    |   6 +-
 18 files changed, 145 insertions(+), 228 deletions(-)

Eric

Comments

Sebastian Andrzej Siewior May 2, 2022, 1:38 p.m. UTC | #1 
On 2022-04-29 16:46:59 [-0500], Eric W. Biederman wrote:
> 
> The states TASK_STOPPED and TASK_TRACE are special in they can not
> handle spurious wake-ups.  This plus actively depending upon and
> changing the value of tsk->__state causes problems for PREEMPT_RT and
> Peter's freezer rewrite.

PREEMPT_RT wise, I had to duct tape wait_task_inactive() and remove the
preempt-disable section in ptrace_stop() (like previously). This reduces
the amount of __state + saved_state checks and looks otherwise stable in
light testing.

Sebastian