Message-ID: <20110514142230.GD23665@htj.dyndns.org>
Date:	Sat, 14 May 2011 16:22:30 +0200
From:	Tejun Heo <tj@...nel.org>
To:	oleg@...hat.com, jan.kratochvil@...hat.com,
	vda.linux@...glemail.com
Cc:	linux-kernel@...r.kernel.org, torvalds@...ux-foundation.org,
	akpm@...ux-foundation.org, indan@....nu, bdonlan@...il.com
Subject: [PATCH UPDATED 8/9] ptrace: move JOBCTL_TRAPPING wait to wait(2)
 and ptrace_check_attach()

Currently, JOBCTL_TRAPPING is used by PTRACE_ATTACH and SEIZE to hide
TASK_STOPPED -> TRACED transition from ptracer.  If tracee is in group
stop, TRAPPING is set, tracee is kicked and tracer waits for the
transition to complete before completing attach.  This prevents tracer
from seeing tracee during transition.

The transition is visible only through wait(2) and following ptrace(2)
requests.  Without TRAPPING, WNOHANG which should succeed right after
attach (when tracer knows tracee was stopped) might fail and likewise
for the following ptrace requests.

TRAPPING will also be used to implement ptrace notification re-traps,
which can be initiated by tasks other than tracer.  To allow this,
this patch moves TRAPPING wait from attach completion path to
operations which are actually affected by the transition - wait(2) and
following ptrace(2) requests.

As reliably checking and modifying TASK_STOPPED/TRACED transition
together with JOBCTL_TRAPPING require siglock and both ptrace and wait
paths are holding tasklist_lock and siglock where TRAPPING check is
needed, ptrace_wait_trapping() assumes both locks to be held on entry
and releases them if it actually had to wait for TRAPPING.

Both wait and ptrace paths are updated to retry the operation after
TRAPPING wait.  Note that wait_task_stopped() now always grabs siglock
for ptrace waits.  This can be avoided with "task_stopped_code() ->
rmb() -> TRAPPING -> rmb() -> task_stopped_code()" sequence but given
that ptrace isn't particularly sensitive to performance or
scalability, choosing simpler implementation seems better.

Both ptrace(2) and wait(2) use restart_syscall() to retry after
waiting for TRAPPING.  This simplifies the implementation and will be
useful when TRAPPING sleep is converted to be interruptible.

Note that, after this change, PTRACE_ATTACH may return before the
transition completes and the ptracer might see the tracee in transient
TASK_RUNNING state via /proc/PID/stat; however, wait(2) and the
following ptrace requests would behave correctly regardless.  This is
userland visible behavior change.

Signed-off-by: Tejun Heo <tj@...nel.org>
---
Forgot to update comment in ptrace_attach().  Nothing changes other
than comment update.  Will update git tree accordingly.

Thanks.

 include/linux/ptrace.h |    1 
 kernel/exit.c          |   18 ++++++++++++--
 kernel/ptrace.c        |   62 ++++++++++++++++++++++++++++++++++++++-----------
 3 files changed, 66 insertions(+), 15 deletions(-)

Index: work/kernel/ptrace.c
===================================================================
--- work.orig/kernel/ptrace.c
+++ work/kernel/ptrace.c
@@ -30,6 +30,47 @@ static int ptrace_trapping_sleep_fn(void
 	return 0;
 }
 
+/**
+ * ptrace_wait_trapping - wait ptracee to finish %TASK_TRACED/STOPPED transition
+ * @child: child to wait for
+ *
+ * There are cases where ptracer needs to ask the ptracee to [re]enter
+ * %TASK_TRACED which involves the tracee going through %TASK_RUNNING
+ * briefly, which could affect operation of ptrace(2) and wait(2).
+ *
+ * %JOBCTL_TRAPPING is used to hide such transitions from the ptracer.
+ * It's set when such transition is initiated by the ptracer and cleared on
+ * completion.  Operations which may be affected should call this function
+ * to make sure no transition is in progress before proceeding.
+ *
+ * This function checks whether @child is TRAPPING and, if so, waits for
+ * the transition to complete.
+ *
+ * CONTEXT:
+ * read_lock(&tasklist_lock) and spin_lock_irq(&child->sighand->siglock).
+ * On %true return, both locks are released and the function might have
+ * slept.
+ *
+ * RETURNS:
+ * %false if @child wasn't trapping and nothing happened.  %true if waited
+ * for trapping transition and released both locks.
+ */
+bool ptrace_wait_trapping(struct task_struct *child)
+	__releases(&child->sighand->siglock)
+	__releases(&tasklist_lock)
+{
+	if (likely(!(child->jobctl & JOBCTL_TRAPPING)))
+		return false;
+
+	spin_unlock_irq(&child->sighand->siglock);
+	get_task_struct(child);
+	read_unlock(&tasklist_lock);
+	wait_on_bit(&child->jobctl, ilog2(JOBCTL_TRAPPING),
+		    ptrace_trapping_sleep_fn, TASK_UNINTERRUPTIBLE);
+	put_task_struct(child);
+	return true;
+}
+
 /*
  * ptrace a task: make the debugger its new parent and
  * move it to the ptrace list.
@@ -140,6 +181,8 @@ int ptrace_check_attach(struct task_stru
 		WARN_ON_ONCE(task_is_stopped(child));
 		if (task_is_traced(child) || ignore_state)
 			ret = 0;
+		else if (ptrace_wait_trapping(child))
+			return restart_syscall();
 		spin_unlock_irq(&child->sighand->siglock);
 	}
 	read_unlock(&tasklist_lock);
@@ -203,7 +246,6 @@ bool ptrace_may_access(struct task_struc
 
 static int ptrace_attach(struct task_struct *task)
 {
-	bool wait_trap = false;
 	int retval;
 
 	audit_ptrace(task);
@@ -249,15 +291,13 @@ static int ptrace_attach(struct task_str
 	 * If the task is already STOPPED, set JOBCTL_STOP_PENDING and
 	 * TRAPPING, and kick it so that it transits to TRACED.  TRAPPING
 	 * will be cleared if the child completes the transition or any
-	 * event which clears the group stop states happens.  We'll wait
-	 * for the transition to complete before returning from this
-	 * function.
+	 * event which clears the group stop states happens.
 	 *
-	 * This hides STOPPED -> RUNNING -> TRACED transition from the
-	 * attaching thread but a different thread in the same group can
-	 * still observe the transient RUNNING state.  IOW, if another
-	 * thread's WNOHANG wait(2) on the stopped tracee races against
-	 * ATTACH, the wait(2) may fail due to the transient RUNNING.
+	 * This is to hide STOPPED -> RUNNING -> TRACED transition from
+	 * wait(2) and ptrace(2).  If called before the transition is
+	 * complete, both will wait for TRAPPING to be cleared and retry,
+	 * thus hiding the transition from userland; however, the transient
+	 * RUNNING state is still visible through /proc.
 	 *
 	 * The following task_is_stopped() test is safe as both transitions
 	 * in and out of STOPPED are protected by siglock.
@@ -265,7 +305,6 @@ static int ptrace_attach(struct task_str
 	if (task_is_stopped(task)) {
 		task->jobctl |= JOBCTL_STOP_PENDING | JOBCTL_TRAPPING;
 		signal_wake_up(task, 1);
-		wait_trap = true;
 	}
 
 	spin_unlock(&task->sighand->siglock);
@@ -276,9 +315,6 @@ unlock_tasklist:
 unlock_creds:
 	mutex_unlock(&task->signal->cred_guard_mutex);
 out:
-	if (wait_trap)
-		wait_on_bit(&task->jobctl, ilog2(JOBCTL_TRAPPING),
-			    ptrace_trapping_sleep_fn, TASK_UNINTERRUPTIBLE);
 	return retval;
 }
 
Index: work/kernel/exit.c
===================================================================
--- work.orig/kernel/exit.c
+++ work/kernel/exit.c
@@ -1409,15 +1409,29 @@ static int wait_task_stopped(struct wait
 	if (!ptrace && !(wo->wo_flags & WUNTRACED))
 		return 0;
 
-	if (!task_stopped_code(p, ptrace))
+	/*
+	 * For ptrace waits, we can't reliably check whether wait condition
+	 * exists without grabbing siglock due to JOBCTL_TRAPPING
+	 * transitions.  A task might be temporarily in TASK_RUNNING while
+	 * trapping which should be transparent to the ptracer.
+	 *
+	 * Note that we can avoid unconditionally grabbing siglock by
+	 * wrapping TRAPPING test with two rmb's; however, let's stick with
+	 * simpler implementation for now.
+	 */
+	if (!ptrace && !(p->signal->flags & SIGNAL_STOP_STOPPED))
 		return 0;
 
 	exit_code = 0;
 	spin_lock_irq(&p->sighand->siglock);
 
 	p_code = task_stopped_code(p, ptrace);
-	if (unlikely(!p_code))
+	if (unlikely(!p_code)) {
+		/* if trapping, wait for it and restart the whole process */
+		if (ptrace && ptrace_wait_trapping(p))
+			return restart_syscall();
 		goto unlock_sig;
+	}
 
 	exit_code = *p_code;
 	if (!exit_code)
Index: work/include/linux/ptrace.h
===================================================================
--- work.orig/include/linux/ptrace.h
+++ work/include/linux/ptrace.h
@@ -105,6 +105,7 @@ extern long arch_ptrace(struct task_stru
 extern int ptrace_readdata(struct task_struct *tsk, unsigned long src, char __user *dst, int len);
 extern int ptrace_writedata(struct task_struct *tsk, char __user *src, unsigned long dst, int len);
 extern void ptrace_disable(struct task_struct *);
+extern bool ptrace_wait_trapping(struct task_struct *child);
 extern int ptrace_check_attach(struct task_struct *task, bool ignore_state);
 extern int ptrace_request(struct task_struct *child, long request,
 			  unsigned long addr, unsigned long data);
--
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/

Powered by blists - more mailing lists