proc_pid_make_inode:
ei->pid = get_pid(task_pid(task));
I think this is not safe. get_pid() can be preempted after checking "pid
!= NULL". Then the task exits, does detach_pid(), and RCU frees the pid.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
task_state() needs tasklist_lock to protect ->parent/->real_parent. However
task->parent points to nowhere only when the actions below happen in order
1) release_task(task)
2) release_task(task->parent)
3) a grace period passed
But 3) implies that the memory ops from 1) should be finished, so pid_alive()
can't be true in such a case.
Otherwise, we don't care if ->parent/->real_parent changes under us.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Drop tasklist_lock. ->siglock protects almost all interesting data
(including sub-threads traversal) except:
->signal->tty
protected by tty_mutex
->real_parent
the task can't be unhashed while we are holding
->siglock, so ->real_parent can change from under us
but we can safely dereference it under rcu_read_lock()
->pgrp/->session
we can get inconsistent numbers if the task does
sys_setsid/daemonize at the same time. I hope this
is acceptable.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
I think it is hardly possible to read the current do_each_task_pid(). The
new version is much simpler and makes the code smaller.
Only the do_each_task_pid change is tested, the do_each_pid_task isn't.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
We don't need to disable irqs to clear current->io_context, it is protected
by ->alloc_lock. Even IF it was possible to submit I/O from IRQ on behalf of
current this irq_disable() can't help: current_io_context() will re-instantiate
->io_context after irq_enable().
We don't need task_lock() or local_irq_disable() to clear ioc->task. This can't
prevent other CPUs from playing with our io_context anyway.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
- It is not possible to have task->mm == &init_mm.
- task_lock() buys nothing for 'if (!p->mm)' check.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Nick Piggin <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
The only one usage of TASK_DEAD outside of last schedule path,
select_bad_process:
for_each_task(p) {
if (!p->mm)
continue;
...
if (p->state == TASK_DEAD)
continue;
...
TASK_DEAD state is set at the end of do_exit(), this means that p->mm
was already set == NULL by exit_mm(), so this task was already rejected
by 'if (!p->mm)' above.
Note also that the caller holds tasklist_lock, this means that p can't
pass exit_notify() and then set TASK_DEAD when p->mm != NULL.
Also, remove open-coded is_init().
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Nick Piggin <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
I am not sure about this patch, I am asking Ingo to take a decision.
task_struct->state == EXIT_DEAD is a very special case, to avoid a confusion
it makes sense to introduce a new state, TASK_DEAD, while EXIT_DEAD should
live only in ->exit_state as documented in sched.h.
Note that this state is not visible to user-space, get_task_state() masks off
unsuitable states.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Ingo Molnar <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
schedule() checks PF_DEAD on every context switch and sets ->state = EXIT_DEAD
to ensure that the exiting task will be deactivated. Note that this EXIT_DEAD
is in fact a "random" value, we can use any bit except normal TASK_XXX values.
It is better to set this state in do_exit() along with PF_DEAD flag and remove
that check in schedule().
We are safe wrt concurrent try_to_wake_up() (for example ptrace, tkill), it
can not change task's ->state: the 'state' argument of try_to_wake_up() can't
have EXIT_DEAD bit. And in case when try_to_wake_up() sees a stale value of
->state == TASK_RUNNING it will do nothing.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Ingo Molnar <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
It is ok to do find_task_by_pid() + get_task_struct() under
rcu_read_lock(), we cand drop tasklist_lock.
Note that testing of ->exit_state is racy with or without tasklist anyway.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Ingo Molnar <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
copy_process:
// holds tasklist_lock + ->siglock
/*
* inherit ioprio
*/
p->ioprio = current->ioprio;
Why? ->ioprio was already copied in dup_task_struct(). I guess this is
needed to ensure that the child can't escape
sys_ioprio_set(IOPRIO_WHO_{PGRP,USER}), yes?
In that case we don't need ->siglock held, and the comment should be
updated.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Jens Axboe <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
I am not sure this patch is correct: I can't understand what the current
code does, and I don't know what it was supposed to do.
The comment says:
* can't change policy, except between SCHED_NORMAL
* and SCHED_BATCH:
The code:
if (((policy != SCHED_NORMAL && p->policy != SCHED_BATCH) &&
(policy != SCHED_BATCH && p->policy != SCHED_NORMAL)) &&
But this is equivalent to:
if ( (is_rt_policy(policy) && has_rt_policy(p)) &&
which means something different. We can't _decrease_ the current
->rt_priority with such a check (if rlim[RLIMIT_RTPRIO] == 0).
Probably, it was supposed to be:
if ( !(policy == SCHED_NORMAL && p->policy == SCHED_BATCH) &&
!(policy == SCHED_BATCH && p->policy == SCHED_NORMAL)
this matches the comment, but strange: it doesn't allow to _drop_ the
realtime priority when rlim[RLIMIT_RTPRIO] == 0.
I think the right check would be:
/* can't set/change rt policy */
if (is_rt_policy(policy) &&
policy != p->policy &&
!rlim_rtprio)
return -EPERM;
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Nick Piggin <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
do_each_thread() is rcu-safe, and all tasks which use this ->mm must sleep
in wait_for_completion(&mm->core_done) at this point, so we can use RCU
locks.
Also, remove unneeded INIT_LIST_HEAD(new) before list_add(new, head).
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-By: David Howells <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
do_each_thread() is rcu-safe, and all tasks which use this ->mm must sleep
in wait_for_completion(&mm->core_done) at this point, so we can use RCU
locks.
Also, remove unneeded INIT_LIST_HEAD(new) before list_add(new, head).
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
rcu_do_batch() decrements rdp->qlen with irqs enabled. This is not good,
it can also be modified by call_rcu() from interrupt.
Decrement ->qlen once with irqs disabled, after a main loop.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Dipankar Sarma <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
It is not possible to find a sub-thread in ->children/->ptrace_children
lists, ptrace_attach() does not allow to attach to sub-threads.
Even if it was possible to ptrace the task from the same thread group,
we can't allow to release ->group_leader while there are others (ptracer)
threads in the same group.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
futex_find_get_task:
if (p->state == EXIT_ZOMBIE || p->exit_state == EXIT_ZOMBIE)
return NULL;
I can't understand this. First, p->state can't be EXIT_ZOMBIE. The
->exit_state check looks strange too. Sub-threads or tasks whose ->parent
ignores SIGCHLD go directly to EXIT_DEAD state (I am ignoring a ptrace
case). Why EXIT_DEAD tasks should be ok? Yes, EXIT_ZOMBIE is more
important (a task may stay zombie for a long time), but this doesn't mean
we should explicitely ignore other EXIT_XXX states.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Ingo Molnar <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
sched_setscheduler() looks at ->signal->rlim[]. It is unsafe do
dereference ->signal unless tasklist_lock or ->siglock is held (or p ==
current). We pin the task structure, but this can't prevent from
release_task()->__exit_signal() which sets ->signal = NULL.
Restore tasklist_lock across the setscheduler call.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Greg KH <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
An exiting task or process which didn't do I/O yet have no io context,
elv_unregister() should check it is not NULL.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Jens Axboe <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Obviously, cfq_cic_link() shouldn't free a just allocated cfq_io_context?
The dead key is from __cic, so drop that.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
I know nothing about io scheduler, but I suspect set_task_ioprio() is not safe.
current_io_context() initializes "struct io_context", then sets ->io_context.
set_task_ioprio() running on another cpu may see the changes out of order, so
->set_ioprio(ioc) may use io_context which was not initialized properly.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
From include/linux/sched.h:
* Careful: do_each_thread/while_each_thread is a double loop so
* 'break' will not work as expected - use goto instead.
*/
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
After the previous patch SIGNAL_GROUP_EXIT implies a pending SIGKILL, we
can remove this check from copy_process() because we already checked
!signal_pending().
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Roland McGrath <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch optimizes zap_threads() for the case when there are no ->mm
users except the current's thread group. In that case we can avoid
'for_each_process()' loop.
It also adds a useful invariant: SIGNAL_GROUP_EXIT (if checked under
->siglock) always implies that all threads (except may be current) have
pending SIGKILL.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Roland McGrath <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This is a preparation for the next patch. No functional changes.
Basically, this patch moves '->flags & SIGNAL_GROUP_EXIT' check into
zap_threads(), and 'complete(vfork_done)' into coredump_wait outside of
->mmap_sem protected area.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Roland McGrath <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch removes tasklist_lock from zap_threads().
This is safe wrt:
do_exit:
The caller holds mm->mmap_sem. This means that task which
shares the same ->mm can't pass exit_mm(), so it can't be
unhashed from init_task.tasks or ->thread_group lists.
fork:
None of sub-threads can fork after zap_process(leader). All
processes which were created before this point should be
visible to zap_threads() because copy_process() adds the new
process to the tail of init_task.tasks list, and ->siglock
lock/unlock provides a memory barrier.
de_thread:
It does list_replace_rcu(&leader->tasks, ¤t->tasks).
So zap_threads() will see either old or new leader, it does
not matter. However, it can change p->sighand, so we should
use lock_task_sighand() in zap_process().
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
With this patch zap_process() sets SIGNAL_GROUP_EXIT while sending SIGKILL to
the thread group. This means that a TASK_TRACED task
1. Will be awakened by signal_wake_up(1)
2. Can't sleep again via ptrace_notify()
3. Can't go to do_signal_stop() after return
from ptrace_stop() in get_signal_to_deliver()
So we can remove all ptrace related stuff from coredump path.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Cc: Roland McGrath <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
With this patch a thread group is killed atomically under ->siglock. This is
faster because we can use sigaddset() instead of force_sig_info() and this is
used in further patches.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Acked-by: Roland McGrath <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
zap_threads() iterates over all threads to find those ones which share
current->mm. All threads in the thread group share the same ->mm, so we can
skip entire thread group if it has another ->mm.
This patch shifts the killing of thread group into the newly added
zap_process() function. This looks as unnecessary complication, but it is
used in further patches.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Acked-by: Roland McGrath <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
We should keep the value of old_leader->tasks.next in de_thread, otherwise
we can't do for_each_process/do_each_thread without tasklist_lock held.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Try to make next_tid() a bit more readable and deletes unnecessary
"pid_alive(pos)" check.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
first_tid:
/* If nr exceeds the number of threads there is nothing todo */
if (nr) {
if (nr >= get_nr_threads(leader))
goto done;
}
This is not reliable: sub-threads can exit after this check, so the
'for' loop below can overlap and proc_task_readdir() can return an
already filldir'ed dirents.
for (; pos && pid_alive(pos); pos = next_thread(pos)) {
if (--nr > 0)
continue;
Off-by-one error, will return 'leader' when nr == 1.
This patch tries to fix these problems and simplify the code.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
list_splice_init(list, head) does unneeded job if it is known that
list_empty(head) == 1. We can use list_replace_init() instead.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: David S. Miller <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
list_replace() is similar to list_replace_rcu(), but unlike
list_replace_rcu() it
could be used when list_empty(old) == 1
doesn't use barriers
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
arm_timer() checks PF_EXITING to prevent BUG_ON(->exit_state)
in run_posix_cpu_timers().
However, for some reason it does so only for CPUCLOCK_PERTHREAD
case (which is imho wrong).
Also, this check is not reliable, PF_EXITING could be set on
another cpu without any locks/barriers just after the check,
so it can't prevent from attaching the timer to the exiting
task.
The previous patch makes this check unneeded.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
do_exit() clears ->it_##clock##_expires, but nothing prevents
another cpu to attach the timer to exiting process after that.
arm_timer() tries to protect against this race, but the check
is racy.
After exit_notify() does 'write_unlock_irq(&tasklist_lock)' and
before do_exit() calls 'schedule() local timer interrupt can find
tsk->exit_state != 0. If that state was EXIT_DEAD (or another cpu
does sys_wait4) interrupted task has ->signal == NULL.
At this moment exiting task has no pending cpu timers, they were
cleanuped in __exit_signal()->posix_cpu_timers_exit{,_group}(),
so we can just return from irq.
John Stultz recently confirmed this bug, see
http://marc.theaimsgroup.com/?l=linux-kernel&m=115015841413687
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
If the local timer interrupt happens just after do_exit() sets PF_EXITING
(and before it clears ->it_xxx_expires) run_posix_cpu_timers() will call
check_process_timers() with tasklist_lock + ->siglock held and
check_process_timers:
t = tsk;
do {
....
do {
t = next_thread(t);
} while (unlikely(t->flags & PF_EXITING));
} while (t != tsk);
the outer loop will never stop.
Actually, the window is bigger. Another process can attach the timer
after ->it_xxx_expires was cleared (see the next commit) and the 'if
(PF_EXITING)' check in arm_timer() is racy (see the one after that).
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Commit e56d090310
[PATCH] RCU signal handling
made this BUG_ON() unsafe. This code runs under ->siglock,
while switch_exec_pids() takes tasklist_lock.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Since base and new_base are of the same type now, we can save one 'if'
branch and simplify the code a bit.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Ingo Molnar <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Commit a4a6198b80:
[PATCH] tvec_bases too large for per-cpu data
introduced "struct tvec_t_base_s boot_tvec_bases" which is visible at
compile time. This means we can kill __init_timer_base and move
timer_base_s's content into tvec_t_base_s.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
send_sigqueue() checks PF_EXITING, then locks p->sighand->siglock. This is
unsafe: 'p' can exit in between and set ->sighand = NULL. The race is
theoretical, the window is tiny and irqs are disabled by the caller, so I
don't think we need the fix for -stable tree.
Convert send_sigqueue() to use lock_task_sighand() helper.
Also, delete 'p->flags & PF_EXITING' re-check, it is unneeded and the
comment is wrong.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
The previous patch has changed callsites of do_notify_parent_cldstop() so that
to_self == (->ptrace & PT_PTRACED) always (as it should be). We can remove
this parameter now.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: john stultz <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Remove an obscure 'stop_count < 0' check in finish_stop(). The previous patch
made this case impossible.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: john stultz <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
do_signal_stop() considers 'thread_group_empty()' as a special case.
This was needed to avoid taking tasklist_lock. Since this lock is
unneeded any longer, we can remove this special case and simplify
the code even more.
Also, before this patch, finish_stop() was called with stop_count == -1
for 'thread_group_empty()' case. This is not strictly wrong, but confusing
and unneeded.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: john stultz <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Move 'tsk->sighand = NULL' from cleanup_sighand() to __exit_signal(). This
makes the exit path more understandable and allows us to do
cleanup_sighand() outside of ->siglock protected section.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Eric W. Biederman wrote:
>
> Ok. SUSV3/Posix is clear, fork is atomic with respect
> to signals. Either a signal comes before or after a
> fork but not during. (See the rationale section).
> http://www.opengroup.org/onlinepubs/000095399/functions/fork.html
>
> The tasklist_lock does not stop forks from adding to a process
> group. The forks stall while the tasklist_lock is held, but a fork
> that began before we grabbed the tasklist_lock simply completes
> afterwards, and the child does not receive the signal.
This also means that SIGSTOP or sig_kernel_coredump() signal can't
be delivered to pgrp/session reliably.
With this patch copy_process() returns -ERESTARTNOINTR when it
detects a pending signal, fork() will be restarted transparently
after handling the signals.
This patch also deletes now unneeded "group_stop_count > 0" check,
copy_process() can no longer succeed while group stop in progress.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-By: Eric Biederman <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch kills PIDTYPE_TGID pid_type thus saving one hash table in
kernel/pid.c and speeding up subthreads create/destroy a bit. It is also a
preparation for the further tref/pids rework.
This patch adds 'struct list_head thread_group' to 'struct task_struct'
instead.
We don't detach group leader from PIDTYPE_PID namespace until another
thread inherits it's ->pid == ->tgid, so we are safe wrt premature
free_pidmap(->tgid) call.
Currently there are no users of find_task_by_pid_type(PIDTYPE_TGID).
Should the need arise, we can use find_task_by_pid()->group_leader.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-By: Eric Biederman <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
do_sigaction() does not need tasklist_lock anymore, we can simplify the code.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
do_signal_stop() does not need tasklist_lock anymore. So it does not need to
do misc re-checks, and we can simplify the code.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
handle_stop_signal() does not need tasklist_lock for SIG_KERNEL_STOP_MASK
signals anymore.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch moves __unhash_process() call from realease_task() to
__exit_signal(), so __detach_pid() is called with ->siglock held.
This means we don't need tasklist_lock to iterate over thread group anymore:
copy_process() was already changed to do attach_pid()
under ->siglock.
Eric's "pidhash-kill-switch_exec_pids.patch" from -mm
changed de_thread() so it doesn't touch PIDTYPE_TGID.
NOTE: de_thread() still needs some attention. It still changes task->pid
lockless. Taking ->sighand.siglock here allows to do more tasklist_lock
removals.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch reverts 'CONFIG_SMP && thread_group_empty()' optimization in
sys_times(). The reason is that the next patch breaks memory ordering which
is needed for that optimization.
tasklist_lock in sys_times() will be eliminated completely by further patch.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
__exit_signal() is private to release_task() now. I think it is better to
make it static in kernel/exit.c and export flush_sigqueue() instead - this
function is much more simple and straightforward.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Cosmetic, rename __exit_sighand to cleanup_sighand and move it close to
copy_sighand().
This matches copy_signal/cleanup_signal naming, and I think it is easier to
follow.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Acked-by: "Paul E. McKenney" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
__exit_signal() does important cleanups atomically under ->siglock. It is
also called from copy_process's error path. This is not good, for example we
can't move __unhash_process() under ->siglock for that reason.
We should not mix these 2 paths, just look at ugly 'if (p->sighand)' under
'bad_fork_cleanup_sighand:' label. For copy_process() case it is sufficient
to just backout copy_signal(), nothing more.
Again, nobody can see this task yet. For CLONE_THREAD case we just decrement
signal->count, otherwise nobody can see this ->signal and we can free it
lockless.
This patch assumes it is safe to do exit_thread_group_keys() without
tasklist_lock.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Acked-by: David Howells <[email protected]>
Signed-off-by: Adrian Bunk <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
The only caller of exit_sighand(tsk) is copy_process's error path. We can
call __exit_sighand() directly and kill exit_sighand().
This 'tsk' was not yet registered in pid_hash[] or init_task.tasks, it has no
external references, nobody can see it, and
IF (clone_flags & CLONE_SIGHAND)
At least 'current' has a reference to ->sighand, this
means atomic_dec_and_test(sighand->count) can't be true.
ELSE
Nobody can see this ->sighand, this means we can free it
without any locking.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Acked-by: "Paul E. McKenney" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Add lock_task_sighand() helper and converts group_send_sig_info() to use
it. Hopefully we will have more users soon.
This patch also removes '!sighand->count' and '!p->usage' checks, I think
they both are bogus, racy and unneeded (but probably it makes sense to
restore them as BUG_ON()s).
->sighand is cleared and it's ->count is decremented in release_task() with
sighand->siglock held, so it is a bug to have '!p->usage || !->count' after
we already locked and verified it is the same. On the other hand, an
already dead task without ->sighand can have a non-zero ->usage due to
ptrace, for example.
If we read the stale value of ->sighand we must see the change after
spin_lock(), because that change was done while holding that same old
->sighand.siglock.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch borrows a clever Hugh's 'struct anon_vma' trick.
Without tasklist_lock held we can't trust task->sighand until we locked it
and re-checked that it is still the same.
But this means we don't need to defer 'kmem_cache_free(sighand)'. We can
return the memory to slab immediately, all we need is to be sure that
sighand->siglock can't dissapear inside rcu protected section.
To do so we need to initialize ->siglock inside ctor function,
SLAB_DESTROY_BY_RCU does the rest.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Use remove_parent/add_parent instead of open coding.
No changes in kernel/exit.o
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
daemonize() calls set_special_pids(1,1), while init and kernel threads spawned
from init/main.c:init() run with 0,0 special pids. This patch changes
INIT_SIGNALS() so that that they run with ->pgrp == ->session == 1 also. This
patch relies on fact that swapper's pid == 1.
Now we have no hashed zero pids in pid_hash[].
User-space visibible change is that now /sbin/init runs with (1,1) special
pids and becomes a session leader.
Quoting Eric W. Biederman:
>
> daemonize consuming pids (1,1) then consumes pgrp 1. So that when
> /sbin/init calls setsid() it thinks /sbin/init is a process group
> leader and setsid() fails. So /sbin/init wants pgrp 1 session 1
> but doesn't get it. I am pretty certain daemonize did not exist so
> /sbin/init got pgrp 1 session 1 in 2.4.
>
> That is the bug that is being fixed.
>
> This patch takes things one step farther and essentially calls
> setsid() for pid == 1 before init is execed. That is new behavior
> but it cleans up the kernel as we now do not need to support the
> case of a process without a process group or a session.
>
> The only process that could have possibly cared was /sbin/init
> and it already calls setsid() because it doesn't want that.
>
> If this was going to break anything noticeable the change in behavior
> from 2.4 to 2.6 would have already done that.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
fork_idle() does unhash_process() just after copy_process(). Contrary,
boot_cpu's idle thread explicitely registers itself for each pid_type with nr
= 0.
copy_process() already checks p->pid != 0 before process_counts++, I think we
can just skip attach_pid() calls and job control inits for idle threads and
kill unhash_process(). We don't need to cleanup ->proc_dentry in fork_idle()
because with this patch idle threads are never hashed in
kernel/pid.c:pid_hash[].
We don't need to hash pid == 0 in pidmap_init(). free_pidmap() is never
called with pid == 0 arg, so it will never be reused. So it is still possible
to use pid == 0 in any PIDTYPE_xxx namespace from kernel/pid.c's POV.
However with this patch we don't hash pid == 0 for PIDTYPE_PID case. We still
have have PIDTYPE_PGID/PIDTYPE_SID entries with pid == 0: /sbin/init and
kernel threads which don't call daemonize().
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Both SET_LINKS() and SET_LINKS/REMOVE_LINKS() have exactly one caller, and
these callers already check thread_group_leader().
This patch kills theese macros, they mix two different things: setting
process's parent and registering it in init_task.tasks list. Callers are
updated to do these actions by hand.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
There are places where kernel uses REMOVE_LINKS/SET_LINKS while changing
process's ->parent. Use add_parent/remove_parent instead, they don't abuse
of global process list.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
add_parent(p, parent) is always called with parent == p->parent, and it makes
no sense to do it differently. This patch removes this argument.
No changes in affected .o files.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
'child_reaper' arg is not used in choose_new_parent().
"->exit_state >= EXIT_ZOMBIE" check is a leftover, was
valid when EXIT_ZOMBIE lived in ->state var.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Eric Biederman <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
I think it is enough to take tasklist_lock for reading while changing
child_reaper:
Reparenting needs write_lock(tasklist_lock)
Only one thread in a thread group can do exec()
sighand->siglock garantees that get_signal_to_deliver()
will not see a stale value of child_reaper.
This means that we can change child_reaper earlier, without calling
zap_other_threads() twice.
"child_reaper = current" is a NOOP when init does exec from main thread, we
don't care.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
__rcu_process_callbacks() disables interrupts to protect itself from
call_rcu() which adds new entries to ->nxtlist.
However we can check "->nxtlist != NULL" with interrupts enabled, we can't
get "false positives" because call_rcu() can only change this condition
from 0 to 1.
Tested with rcutorture.ko.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Dipankar Sarma <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch changes the code from:
preempt_disable();
for (;;) {
...
preempt_disable();
}
to:
for (;;) {
preempt_disable();
...
}
which seems more clean to me and saves a couple of bytes for
each function.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Ingo Molnar <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
If get_next_ra_size() does not grow fast enough, ->prev_page can overrun
the ahead window. This means the caller will read the pages from
->ahead_start + ->ahead_size to ->prev_page synchronously.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Steven Pratt <[email protected]>
Cc: Ram Pai <[email protected]>
Cc: Trond Myklebust <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
sys_unshare() does mmput(new_mm). This is not enough if we have
mm->core_waiters.
This patch is a temporary fix for soon to be released 2.6.16.
Signed-off-by: Oleg Nesterov <[email protected]>
[ Checked with Uli: "I'm not planning to use unshare(CLONE_VM). It's
not needed for any functionality planned so far. What we (as in Red
Hat) need unshare() for now is the filesystem side." ]
Signed-off-by: Linus Torvalds <[email protected]>
1. The tracee can go from ptrace_stop() to do_signal_stop()
after __ptrace_unlink(p).
2. It is unsafe to __ptrace_unlink(p) while p->parent may wait
for tasklist_lock in ptrace_detach().
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Roland McGrath <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Eric W. Biederman <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
copy_process:
attach_pid(p, PIDTYPE_PID, p->pid);
attach_pid(p, PIDTYPE_TGID, p->tgid);
What if kill_proc_info(p->pid) happens in between?
copy_process() holds current->sighand.siglock, so we are safe
in CLONE_THREAD case, because current->sighand == p->sighand.
Otherwise, p->sighand is unlocked, the new process is already
visible to the find_task_by_pid(), but have a copy of parent's
'struct pid' in ->pids[PIDTYPE_TGID].
This means that __group_complete_signal() may hang while doing
do ... while (next_thread() != p)
We can solve this problem if we reverse these 2 attach_pid()s:
attach_pid() does wmb()
group_send_sig_info() calls spin_lock(), which
provides a read barrier. // Yes ?
I don't think we can hit this race in practice, but still.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Roland McGrath <[email protected]>
Cc: Ingo Molnar <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
There is a window after copy_process() unlocks ->sighand.siglock
and before it adds the new thread to the thread list.
In that window __group_complete_signal(SIGKILL) will not see the
new thread yet, so this thread will start running while the whole
thread group was supposed to exit.
I beleive we have another good reason to place attach_pid(PID/TGID)
under ->sighand.siglock. We can do the same for
release_task()->__unhash_process()
de_thread()->switch_exec_pids()
After that we don't need tasklist_lock to iterate over the thread
list, and we can simplify things, see for example do_sigaction()
or sys_times().
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Roland McGrath <[email protected]>
Cc: Ingo Molnar <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Pointed out by Linus Torvalds.
sys_signal() forgets to initialize ->sa_mask.
( I suspect arch/ia64/ia32/ia32_signal.c:sys32_signal()
also needs this fix )
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Pointed out by Srivatsa Vaddagiri <[email protected]>.
rcu_do_batch() stops after processing maxbatch callbacks
on ->donelist leaving rcu_tasklet in TASKLET_STATE_SCHED
state.
If CPU_DEAD event happens remaining ->donelist entries are
lost, rcu_offline_cpu() kills this tasklet.
With this patch ->donelist migrates along with ->curlist
and ->nxtlist to the current cpu.
Compile tested.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Paul E. McKenney <[email protected]>
Cc: Srivatsa Vaddagiri <[email protected]>
Cc: Dipankar Sarma <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch moves rcu_state into the rcu_ctrlblk. I think there
are no reasons why we should have 2 different variables to control
rcu state. Every user of rcu_state has also "rcu_ctrlblk *rcp" in
the parameter list.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Paul E. McKenney <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
I think it is better to set ->next_pending in the caller, when
it is needed. This saves one parameter, and this coincides with
cpu_quiet() beahaviour, which sets ->completed = ->cur itself.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Paul E. McKenney <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
__rcu_pending() is rather fat and called twice from rcu_pending().
rcu_pending() has multiple callers, and not that small too.
This patch uninlines both of them.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Paul E. McKenney <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Factor out common code for different RUSAGE_xxx cases.
Don't take ->sighand->siglock in RUSAGE_SELF case, suggested by Ravikiran G
Thirumalai <[email protected]>.
Signed-off-by: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
sys_setpgid() allows to change ->pgrp of ptraced childs.
'man setpgid' does not tell anything about that, so I consider
this behaviour is a bug.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Oren Laadan <[email protected]>
Cc: Roland McGrath <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
setpgid(0, pgid) or setpgid(forked_child_pid, pgid) does not work unless
the calling process is a thread_group_leader().
'man setpgid' does not tell anything about that, so I consider this
behaviour is a bug.
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Oren Laadan <[email protected]>
Cc: Roland McGrath <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Oleg Nesterov