Now that ignored posix timer signals are requeued and the timers are
rearmed on signal delivery the workaround to keep such timers alive and
self rearm them is not longer required.
Remove the relevant hacks and the not longer required return values from
the related functions. The alarm timer workarounds will be cleaned up in a
separate step.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241105064214.187239060@linutronix.de
Queue posixtimers which have their signal ignored on the ignored list:
1) When the timer fires and the signal has SIG_IGN set
2) When SIG_IGN is installed via sigaction() and a timer signal
is already queued
This only happens when the signal is for a valid timer, which delivered the
signal in periodic mode. One-shot timer signals are correctly dropped.
Due to the lock order constraints (sighand::siglock nests inside
timer::lock) the signal code cannot access any of the timer fields which
are relevant to make this decision, e.g. timer::it_status.
This is addressed by establishing a protection scheme which requires to
lock both locks on the timer side for modifying decision fields in the
timer struct and therefore makes it possible for the signal delivery to
evaluate with only sighand:siglock being held:
1) Move the NULLification of timer->it_signal into the sighand::siglock
protected section of timer_delete() and check timer::it_signal in the
code path which determines whether the signal is dropped or queued on
the ignore list.
This ensures that a deleted timer cannot be moved onto the ignore
list, which would prevent it from being freed on exit() as it is not
longer in the process' posix timer list.
If the timer got moved to the ignored list before deletion then it is
removed from the ignored list under sighand lock in timer_delete().
2) Provide a new timer::it_sig_periodic flag, which gets set in the
signal queue path with both timer and sighand locks held if the timer
is actually in periodic mode at expiry time.
The ignore list code checks this flag under sighand::siglock and drops
the signal when it is not set.
If it is set, then the signal is moved to the ignored list independent
of the actual state of the timer.
When the signal is un-ignored later then the signal is moved back to
the signal queue. On signal delivery the posix timer side decides
about dropping the signal if the timer was re-armed, dis-armed or
deleted based on the signal sequence counter check.
If the thread/process exits then not yet delivered signals are
discarded which means the reference of the timer containing the
sigqueue is dropped and frees the timer.
This is way cheaper than requiring all code paths to lock
sighand::siglock of the target thread/process on any modification of
timer::it_status or going all the way and removing pending signals
from the signal queues on every rearm, disarm or delete operation.
So the protection scheme here is that on the timer side both timer::lock
and sighand::siglock have to be held for modifying
timer::it_signal
timer::it_sig_periodic
which means that on the signal side holding sighand::siglock is enough to
evaluate these fields.
In posixtimer_deliver_signal() holding timer::lock is sufficient to do the
sequence validation against timer::it_signal_seq because a concurrent
expiry is waiting on timer::lock to be released.
This completes the SIG_IGN handling and such timers are not longer self
rearmed which avoids pointless wakeups.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241105064214.120756416@linutronix.de
The posix timer signal handling uses siginfo::si_sys_private for handling
the sequence counter check. That indirection is not longer required and the
sequence count value at signal queueing time can be stored in struct
k_itimer itself.
This removes the requirement of treating siginfo::si_sys_private special as
it's now always zero as the kernel does not touch it anymore.
Suggested-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Link: https://lore.kernel.org/all/20241105064213.852619866@linutronix.de
To cure the SIG_IGN handling for posix interval timers, the preallocated
sigqueue needs to be embedded into struct k_itimer to prevent life time
races of all sorts.
Now that the prerequisites are in place, embed the sigqueue into struct
k_itimer and fixup the relevant usage sites.
Aside of preparing for proper SIG_IGN handling, this spares an extra
allocation.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241105064213.719695194@linutronix.de
To handle posix timers which have their signal ignored via SIG_IGN properly
it is required to requeue a ignored signal for delivery when SIG_IGN is
lifted so the timer gets rearmed.
Split the required code out of send_sigqueue() so it can be reused in
context of sigaction().
While at it rename send_sigqueue() to posixtimer_send_sigqueue() so its
clear what this is about.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241105064213.586453412@linutronix.de
To cure the SIG_IGN handling for posix interval timers, the preallocated
sigqueue needs to be embedded into struct k_itimer to prevent life time
races of all sorts.
To make that work correctly it needs reference counting so that timer
deletion does not free the timer prematuraly when there is a signal queued
or delivered concurrently.
Add a rcuref to the posix timer part.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241105064213.304756440@linutronix.de
The handling of the timer overrun in the signal code is inconsistent as it
takes previous overruns into account. This is just wrong as after the
reprogramming of a timer the overrun count starts over from a clean state,
i.e. 0.
Don't touch info::si_overrun in send_sigqueue() and only store the overrun
value at signal delivery time, which is computed from the timer itself
relative to the expiry time.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241105064213.106738193@linutronix.de
Signals of timers which are reprogammed, disarmed or deleted can deliver
signals related to the past. The POSIX spec is blury about this:
- "The effect of disarming or resetting a timer with pending expiration
notifications is unspecified."
- "The disposition of pending signals for the deleted timer is
unspecified."
In both cases it is reasonable to expect that pending signals are
discarded. Especially in the reprogramming case it does not make sense to
account for previous overruns or to deliver a signal for a timer which has
been disarmed. This makes the behaviour consistent and understandable.
Remove the si_sys_private check from the signal delivery code and invoke
posix_timer_deliver_signal() unconditionally for posix timer related
signals.
Change posix_timer_deliver_signal() so it controls the actual signal
delivery via the return value. It now instructs the signal code to drop the
signal when:
1) The timer does not longer exist in the hash table
2) The timer signal_seq value is not the same as the si_sys_private value
which was set when the signal was queued.
This is also a preparatory change to embed the sigqueue into the k_itimer
structure, which in turn allows to remove the si_sys_private magic.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/all/20241105064213.040348644@linutronix.de
Right now the state tracking is done by two struct members:
- it_active:
A boolean which tracks armed/disarmed state
- it_signal_seq:
A sequence counter which is used to invalidate settings
and prevent rearming
Replace it_active with it_status and keep properly track about the states
in one place.
This allows to reuse it_signal_seq to track reprogramming, disarm and
delete operations in order to drop signals which are related to the state
previous of those operations.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241001083835.670337048@linutronix.de
No point in delivering a signal from the past. POSIX does not specify the
behaviour here:
- "The effect of disarming or resetting a timer with pending expiration
notifications is unspecified."
- "The disposition of pending signals for the deleted timer is unspecified."
In both cases it is reasonable to expect that pending signals are
discarded. Especially in the reprogramming case it does not make sense to
account for previous overruns or to deliver a signal for a timer which has
been disarmed.
Drop the signal as that is conistent and understandable behaviour.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241001083835.553646280@linutronix.de
In case that a timer was reprogrammed or deleted an already pending signal
is obsolete. Right now such signals are kept around and eventually
delivered. While POSIX is blury about this:
- "The effect of disarming or resetting a timer with pending expiration
notifications is unspecified."
- "The disposition of pending signals for the deleted timer is
unspecified."
it is reasonable in both cases to expect that pending signals are discarded
as they have no meaning anymore.
Prepare the signal code to allow dropping posix timer signals.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241001083835.494416923@linutronix.de
The si_sys_private member of the siginfo which is embedded in the
preallocated sigqueue is used by the posix timer code to decide whether a
timer must be reprogrammed on signal delivery.
The handling of this is racy as a long standing comment in that code
documents. It is modified with the timer lock held, but without sighand
lock being held. The actual signal delivery code checks for it under
sighand lock without holding the timer lock.
Hand the new value to send_sigqueue() as argument and store it with sighand
lock held. This is an intermediate change to address this issue.
The arguments to this function will be cleanup in subsequent changes.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/all/20241001083835.434338954@linutronix.de
Rename posix_timer_event() to posix_timer_queue_signal() as this is what
the function is about.
Consolidate the requeue pending and deactivation updates into that function
as there is no point in doing this in all incarnations of posix timers.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
hrtimer based and CPU timers have their own way to install the new interval
and to reset overrun and signal handling related data.
Create a helper function and do the same operation for all variants.
This also makes the handling of the interval consistent. It's only stored
when the timer is actually armed, i.e. timer->it_value != 0. Before that it
was stored unconditionally for posix CPU timers and conditionally for the
other posix timers.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Keeping the overrun count of the previous setup around is just wrong. The
new setting has nothing to do with the previous one and has to start from a
clean slate.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Commit c78f261e5dcb ("posix-timers: Clarify posix_timer_fn() comments")
turns an ifdef CONFIG_HIGH_RES_TIMERS into an conditional on
"IS_ENABLED(CONFIG_HIGHRES_TIMERS)"; note that the new conditional refers
to "HIGHRES_TIMERS" not "HIGH_RES_TIMERS" as before.
Fix this typo introduced in that refactoring.
Fixes: c78f261e5dcb ("posix-timers: Clarify posix_timer_fn() comments")
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20230609094643.26253-1-lukas.bulwahn@gmail.com
release_posix_timers() is called for cleaning up both hashed and unhashed
timers. The cases are differentiated by an argument and the usage is
hideous.
Seperate the actual free path out and use it for unhashed timers. Provide a
function for hashed timers.
No functional change.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230425183313.301432503@linutronix.de
Technically it's not required to set k_itimer::it_signal to NULL on exit()
because there is no other thread anymore which could lookup the timer
concurrently.
Set it to NULL for consistency sake and add a comment to that effect.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230425183313.196462644@linutronix.de
posix_timer_add() tries to allocate a posix timer ID by starting from the
cached ID which was stored by the last successful allocation.
This is done in a loop searching the ID space for a free slot one by
one. The loop has to terminate when the search wrapped around to the
starting point.
But that's racy vs. establishing the starting point. That is read out
lockless, which leads to the following problem:
CPU0 CPU1
posix_timer_add()
start = sig->posix_timer_id;
lock(hash_lock);
... posix_timer_add()
if (++sig->posix_timer_id < 0)
start = sig->posix_timer_id;
sig->posix_timer_id = 0;
So CPU1 can observe a negative start value, i.e. -1, and the loop break
never happens because the condition can never be true:
if (sig->posix_timer_id == start)
break;
While this is unlikely to ever turn into an endless loop as the ID space is
huge (INT_MAX), the racy read of the start value caught the attention of
KCSAN and Dmitry unearthed that incorrectness.
Rewrite it so that all id operations are under the hash lock.
Reported-by: syzbot+5c54bd3eb218bb595aa9@syzkaller.appspotmail.com
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/87bkhzdn6g.ffs@tglx
itimer_delete() has a retry loop when the timer is concurrently expired. On
non-RT kernels this just spin-waits until the timer callback has completed,
except for posix CPU timers which have HAVE_POSIX_CPU_TIMERS_TASK_WORK
enabled.
In that case and on RT kernels the existing task could live lock when
preempting the task which does the timer delivery.
Replace spin_unlock() with an invocation of timer_wait_running() to handle
it the same way as the other retry loops in the posix timer code.
Fixes: ec8f954a40 ("posix-timers: Use a callback for cancel synchronization on PREEMPT_RT")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/87v8g7c50d.ffs@tglx
For some unknown reason the introduction of the timer_wait_running callback
missed to fixup posix CPU timers, which went unnoticed for almost four years.
Marco reported recently that the WARN_ON() in timer_wait_running()
triggers with a posix CPU timer test case.
Posix CPU timers have two execution models for expiring timers depending on
CONFIG_POSIX_CPU_TIMERS_TASK_WORK:
1) If not enabled, the expiry happens in hard interrupt context so
spin waiting on the remote CPU is reasonably time bound.
Implement an empty stub function for that case.
2) If enabled, the expiry happens in task work before returning to user
space or guest mode. The expired timers are marked as firing and moved
from the timer queue to a local list head with sighand lock held. Once
the timers are moved, sighand lock is dropped and the expiry happens in
fully preemptible context. That means the expiring task can be scheduled
out, migrated, interrupted etc. So spin waiting on it is more than
suboptimal.
The timer wheel has a timer_wait_running() mechanism for RT, which uses
a per CPU timer-base expiry lock which is held by the expiry code and the
task waiting for the timer function to complete blocks on that lock.
This does not work in the same way for posix CPU timers as there is no
timer base and expiry for process wide timers can run on any task
belonging to that process, but the concept of waiting on an expiry lock
can be used too in a slightly different way:
- Add a mutex to struct posix_cputimers_work. This struct is per task
and used to schedule the expiry task work from the timer interrupt.
- Add a task_struct pointer to struct cpu_timer which is used to store
a the task which runs the expiry. That's filled in when the task
moves the expired timers to the local expiry list. That's not
affecting the size of the k_itimer union as there are bigger union
members already
- Let the task take the expiry mutex around the expiry function
- Let the waiter acquire a task reference with rcu_read_lock() held and
block on the expiry mutex
This avoids spin-waiting on a task which might not even be on a CPU and
works nicely for RT too.
Fixes: ec8f954a40 ("posix-timers: Use a callback for cancel synchronization on PREEMPT_RT")
Reported-by: Marco Elver <elver@google.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Marco Elver <elver@google.com>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/87zg764ojw.ffs@tglx
The nanosleep syscalls use the restart_block mechanism, with a quirk:
The `type` and `rmtp`/`compat_rmtp` fields are set up unconditionally on
syscall entry, while the rest of the restart_block is only set up in the
unlikely case that the syscall is actually interrupted by a signal (or
pseudo-signal) that doesn't have a signal handler.
If the restart_block was set up by a previous syscall (futex(...,
FUTEX_WAIT, ...) or poll()) and hasn't been invalidated somehow since then,
this will clobber some of the union fields used by futex_wait_restart() and
do_restart_poll().
If userspace afterwards wrongly calls the restart_syscall syscall,
futex_wait_restart()/do_restart_poll() will read struct fields that have
been clobbered.
This doesn't actually lead to anything particularly interesting because
none of the union fields contain trusted kernel data, and
futex(..., FUTEX_WAIT, ...) and poll() aren't syscalls where it makes much
sense to apply seccomp filters to their arguments.
So the current consequences are just of the "if userspace does bad stuff,
it can damage itself, and that's not a problem" flavor.
But still, it seems like a hazard for future developers, so invalidate the
restart_block when partly setting it up in the nanosleep syscalls.
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20230105134403.754986-1-jannh@google.com
As Chris explains, the comment above exit_itimers() is not correct,
we can race with proc_timers_seq_ops. Change exit_itimers() to clear
signal->posix_timers with ->siglock held.
Cc: <stable@vger.kernel.org>
Reported-by: chris@accessvector.net
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge misc updates from Andrew Morton:
"173 patches.
Subsystems affected by this series: ia64, ocfs2, block, and mm (debug,
pagecache, gup, swap, shmem, memcg, selftests, pagemap, mremap,
bootmem, sparsemem, vmalloc, kasan, pagealloc, memory-failure,
hugetlb, userfaultfd, vmscan, compaction, mempolicy, memblock,
oom-kill, migration, ksm, percpu, vmstat, and madvise)"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (173 commits)
mm/madvise: add MADV_WILLNEED to process_madvise()
mm/vmstat: remove unneeded return value
mm/vmstat: simplify the array size calculation
mm/vmstat: correct some wrong comments
mm/percpu,c: remove obsolete comments of pcpu_chunk_populated()
selftests: vm: add COW time test for KSM pages
selftests: vm: add KSM merging time test
mm: KSM: fix data type
selftests: vm: add KSM merging across nodes test
selftests: vm: add KSM zero page merging test
selftests: vm: add KSM unmerge test
selftests: vm: add KSM merge test
mm/migrate: correct kernel-doc notation
mm: wire up syscall process_mrelease
mm: introduce process_mrelease system call
memblock: make memblock_find_in_range method private
mm/mempolicy.c: use in_task() in mempolicy_slab_node()
mm/mempolicy: unify the create() func for bind/interleave/prefer-many policies
mm/mempolicy: advertise new MPOL_PREFERRED_MANY
mm/hugetlb: add support for mempolicy MPOL_PREFERRED_MANY
...
