In the current implementation of the EPOLLEXCLUSIVE flag (added for
4.5-rc1), if epoll waiters create different POLL* sets and register them
as exclusive against the same target fd, the current implementation will
stop waking any further waiters once it finds the first idle waiter.
This means that waiters could miss wakeups in certain cases.
For example, when we wake up a pipe for reading we do:
wake_up_interruptible_sync_poll(&pipe->wait, POLLIN | POLLRDNORM); So if
one epoll set or epfd is added to pipe p with POLLIN and a second set
epfd2 is added to pipe p with POLLRDNORM, only epfd may receive the
wakeup since the current implementation will stop after it finds any
intersection of events with a waiter that is blocked in epoll_wait().
We could potentially address this by requiring all epoll waiters that
are added to p be required to pass the same set of POLL* events. IE the
first EPOLL_CTL_ADD that passes EPOLLEXCLUSIVE establishes the set POLL*
flags to be used by any other epfds that are added as EPOLLEXCLUSIVE.
However, I think it might be somewhat confusing interface as we would
have to reference count the number of users for that set, and so
userspace would have to keep track of that count, or we would need a
more involved interface. It also adds some shared state that we'd have
store somewhere. I don't think anybody will want to bloat
__wait_queue_head for this.
I think what we could do instead, is to simply restrict EPOLLEXCLUSIVE
such that it can only be specified with EPOLLIN and/or EPOLLOUT. So
that way if the wakeup includes 'POLLIN' and not 'POLLOUT', we can stop
once we hit the first idle waiter that specifies the EPOLLIN bit, since
any remaining waiters that only have 'POLLOUT' set wouldn't need to be
woken. Likewise, we can do the same thing if 'POLLOUT' is in the wakeup
bit set and not 'POLLIN'. If both 'POLLOUT' and 'POLLIN' are set in the
wake bit set (there is at least one example of this I saw in fs/pipe.c),
then we just wake the entire exclusive list. Having both 'POLLOUT' and
'POLLIN' both set should not be on any performance critical path, so I
think that's ok (in fs/pipe.c its in pipe_release()). We also continue
to include EPOLLERR and EPOLLHUP by default in any exclusive set. Thus,
the user can specify EPOLLERR and/or EPOLLHUP but is not required to do
so.
Since epoll waiters may be interested in other events as well besides
EPOLLIN, EPOLLOUT, EPOLLERR and EPOLLHUP, these can still be added by
doing a 'dup' call on the target fd and adding that as one normally
would with EPOLL_CTL_ADD. Since I think that the POLLIN and POLLOUT
events are what we are interest in balancing, I think that the 'dup'
thing could perhaps be added to only one of the waiter threads.
However, I think that EPOLLIN, EPOLLOUT, EPOLLERR and EPOLLHUP should be
sufficient for the majority of use-cases.
Since EPOLLEXCLUSIVE is intended to be used with a target fd shared
among multiple epfds, where between 1 and n of the epfds may receive an
event, it does not satisfy the semantics of EPOLLONESHOT where only 1
epfd would get an event. Thus, it is not allowed to be specified in
conjunction with EPOLLEXCLUSIVE.
EPOLL_CTL_MOD is also not allowed if the fd was previously added as
EPOLLEXCLUSIVE. It seems with the limited number of flags to not be as
interesting, but this could be relaxed at some further point.
Signed-off-by: Jason Baron <jbaron@akamai.com>
Tested-by: Madars Vitolins <m@silodev.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Al Viro <viro@ftp.linux.org.uk>
Cc: Eric Wong <normalperson@yhbt.net>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Hagen Paul Pfeifer <hagen@jauu.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently, epoll file descriptors or epfds (the fd returned from
epoll_create[1]()) that are added to a shared wakeup source are always
added in a non-exclusive manner. This means that when we have multiple
epfds attached to a shared fd source they are all woken up. This creates
thundering herd type behavior.
Introduce a new 'EPOLLEXCLUSIVE' flag that can be passed as part of the
'event' argument during an epoll_ctl() EPOLL_CTL_ADD operation. This new
flag allows for exclusive wakeups when there are multiple epfds attached
to a shared fd event source.
The implementation walks the list of exclusive waiters, and queues an
event to each epfd, until it finds the first waiter that has threads
blocked on it via epoll_wait(). The idea is to search for threads which
are idle and ready to process the wakeup events. Thus, we queue an event
to at least 1 epfd, but may still potentially queue an event to all epfds
that are attached to the shared fd source.
Performance testing was done by Madars Vitolins using a modified version
of Enduro/X. The use of the 'EPOLLEXCLUSIVE' flag reduce the length of
this particular workload from 860s down to 24s.
Sample epoll_clt text:
EPOLLEXCLUSIVE
Sets an exclusive wakeup mode for the epfd file descriptor that is
being attached to the target file descriptor, fd. Thus, when an event
occurs and multiple epfd file descriptors are attached to the same
target file using EPOLLEXCLUSIVE, one or more epfds will receive an
event with epoll_wait(2). The default in this scenario (when
EPOLLEXCLUSIVE is not set) is for all epfds to receive an event.
EPOLLEXCLUSIVE may only be specified with the op EPOLL_CTL_ADD.
Signed-off-by: Jason Baron <jbaron@akamai.com>
Tested-by: Madars Vitolins <m@silodev.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Al Viro <viro@ftp.linux.org.uk>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Eric Wong <normalperson@yhbt.net>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Hagen Paul Pfeifer <hagen@jauu.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
After waking up a task waiting for an event, we explicitly mark it as
TASK_RUNNING (which is necessary as we do the checks for wakeups as
TASK_INTERRUPTIBLE). Once running and dealing with actually delivering
the events, we're obviously not planning on calling schedule, thus we can
relax the implied barrier and simply update the state with
__set_current_state().
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
seq_printf functions shouldn't really check the return value.
Checking seq_has_overflowed() occasionally is used instead.
Update vfs documentation.
Link: http://lkml.kernel.org/p/e37e6e7b76acbdcc3bb4ab2a57c8f8ca1ae11b9a.1412031505.git.joe@perches.com
Cc: David S. Miller <davem@davemloft.net>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Joe Perches <joe@perches.com>
[ did a few clean ups ]
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
When calling epoll_ctl with operation EPOLL_CTL_DEL, structure epds is
not initialized but ep_take_care_of_epollwakeup reads its event field.
When this unintialized field has EPOLLWAKEUP bit set, a capability check
is done for CAP_BLOCK_SUSPEND in ep_take_care_of_epollwakeup. This
produces unexpected messages in the audit log, such as (on a system
running SELinux):
type=AVC msg=audit(1408212798.866:410): avc: denied
{ block_suspend } for pid=7754 comm="dbus-daemon" capability=36
scontext=unconfined_u:unconfined_r:unconfined_t
tcontext=unconfined_u:unconfined_r:unconfined_t
tclass=capability2 permissive=1
type=SYSCALL msg=audit(1408212798.866:410): arch=c000003e syscall=233
success=yes exit=0 a0=3 a1=2 a2=9 a3=7fffd4d66ec0 items=0 ppid=1
pid=7754 auid=1000 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0
fsgid=0 tty=(none) ses=3 comm="dbus-daemon"
exe="/usr/bin/dbus-daemon"
subj=unconfined_u:unconfined_r:unconfined_t key=(null)
("arch=c000003e syscall=233 a1=2" means "epoll_ctl(op=EPOLL_CTL_DEL)")
Remove use of epds in epoll_ctl when op == EPOLL_CTL_DEL.
Fixes: 4d7e30d989 ("epoll: Add a flag, EPOLLWAKEUP, to prevent suspend while epoll events are ready")
Signed-off-by: Nicolas Iooss <nicolas.iooss_linux@m4x.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Arve Hjønnevåg <arve@android.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This fixes use-after-free of epi->fllink.next inside list loop macro.
This loop actually releases elements in the body. The list is
rcu-protected but here we cannot hold rcu_read_lock because we need to
lock mutex inside.
The obvious solution is to use list_for_each_entry_safe(). RCU-ness
isn't essential because nobody can change this list under us, it's final
fput for this file.
The bug was introduced by ae10b2b4eb ("epoll: optimize EPOLL_CTL_DEL
using rcu")
Signed-off-by: Konstantin Khlebnikov <koct9i@gmail.com>
Reported-by: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Stable <stable@vger.kernel.org> # 3.13+
Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Jason Baron <jbaron@akamai.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This typedef is unnecessary and should just be removed.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The EPOLL_CTL_DEL path of epoll contains a classic, ab-ba deadlock.
That is, epoll_ctl(a, EPOLL_CTL_DEL, b, x), will deadlock with
epoll_ctl(b, EPOLL_CTL_DEL, a, x). The deadlock was introduced with
commmit 67347fe4e6 ("epoll: do not take global 'epmutex' for simple
topologies").
The acquistion of the ep->mtx for the destination 'ep' was added such
that a concurrent EPOLL_CTL_ADD operation would see the correct state of
the ep (Specifically, the check for '!list_empty(&f.file->f_ep_links')
However, by simply not acquiring the lock, we do not serialize behind
the ep->mtx from the add path, and thus may perform a full path check
when if we had waited a little longer it may not have been necessary.
However, this is a transient state, and performing the full loop
checking in this case is not harmful.
The important point is that we wouldn't miss doing the full loop
checking when required, since EPOLL_CTL_ADD always locks any 'ep's that
its operating upon. The reason we don't need to do lock ordering in the
add path, is that we are already are holding the global 'epmutex'
whenever we do the double lock. Further, the original posting of this
patch, which was tested for the intended performance gains, did not
perform this additional locking.
Signed-off-by: Jason Baron <jbaron@akamai.com>
Cc: Nathan Zimmer <nzimmer@sgi.com>
Cc: Eric Wong <normalperson@yhbt.net>
Cc: Nelson Elhage <nelhage@nelhage.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Drop EPOLLWAKEUP from epoll events mask if CONFIG_PM_SLEEP is disabled.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Merge first patch-bomb from Andrew Morton:
"Quite a lot of other stuff is banked up awaiting further
next->mainline merging, but this batch contains:
- Lots of random misc patches
- OCFS2
- Most of MM
- backlight updates
- lib/ updates
- printk updates
- checkpatch updates
- epoll tweaking
- rtc updates
- hfs
- hfsplus
- documentation
- procfs
- update gcov to gcc-4.7 format
- IPC"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (269 commits)
ipc, msg: fix message length check for negative values
ipc/util.c: remove unnecessary work pending test
devpts: plug the memory leak in kill_sb
./Makefile: export initial ramdisk compression config option
init/Kconfig: add option to disable kernel compression
drivers: w1: make w1_slave::flags long to avoid memory corruption
drivers/w1/masters/ds1wm.cuse dev_get_platdata()
drivers/memstick/core/ms_block.c: fix unreachable state in h_msb_read_page()
drivers/memstick/core/mspro_block.c: fix attributes array allocation
drivers/pps/clients/pps-gpio.c: remove redundant of_match_ptr
kernel/panic.c: reduce 1 byte usage for print tainted buffer
gcov: reuse kbasename helper
kernel/gcov/fs.c: use pr_warn()
kernel/module.c: use pr_foo()
gcov: compile specific gcov implementation based on gcc version
gcov: add support for gcc 4.7 gcov format
gcov: move gcov structs definitions to a gcc version specific file
kernel/taskstats.c: return -ENOMEM when alloc memory fails in add_del_listener()
kernel/taskstats.c: add nla_nest_cancel() for failure processing between nla_nest_start() and nla_nest_end()
kernel/sysctl_binary.c: use scnprintf() instead of snprintf()
...
Pull vfs updates from Al Viro:
"All kinds of stuff this time around; some more notable parts:
- RCU'd vfsmounts handling
- new primitives for coredump handling
- files_lock is gone
- Bruce's delegations handling series
- exportfs fixes
plus misc stuff all over the place"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (101 commits)
ecryptfs: ->f_op is never NULL
locks: break delegations on any attribute modification
locks: break delegations on link
locks: break delegations on rename
locks: helper functions for delegation breaking
locks: break delegations on unlink
namei: minor vfs_unlink cleanup
locks: implement delegations
locks: introduce new FL_DELEG lock flag
vfs: take i_mutex on renamed file
vfs: rename I_MUTEX_QUOTA now that it's not used for quotas
vfs: don't use PARENT/CHILD lock classes for non-directories
vfs: pull ext4's double-i_mutex-locking into common code
exportfs: fix quadratic behavior in filehandle lookup
exportfs: better variable name
exportfs: move most of reconnect_path to helper function
exportfs: eliminate unused "noprogress" counter
exportfs: stop retrying once we race with rename/remove
exportfs: clear DISCONNECTED on all parents sooner
exportfs: more detailed comment for path_reconnect
...
When calling EPOLL_CTL_ADD for an epoll file descriptor that is attached
directly to a wakeup source, we do not need to take the global 'epmutex',
unless the epoll file descriptor is nested. The purpose of taking the
'epmutex' on add is to prevent complex topologies such as loops and deep
wakeup paths from forming in parallel through multiple EPOLL_CTL_ADD
operations. However, for the simple case of an epoll file descriptor
attached directly to a wakeup source (with no nesting), we do not need to
hold the 'epmutex'.
This patch along with 'epoll: optimize EPOLL_CTL_DEL using rcu' improves
scalability on larger systems. Quoting Nathan Zimmer's mail on SPECjbb
performance:
"On the 16 socket run the performance went from 35k jOPS to 125k jOPS. In
addition the benchmark when from scaling well on 10 sockets to scaling
well on just over 40 sockets.
...
Currently the benchmark stops scaling at around 40-44 sockets but it seems like
I found a second unrelated bottleneck."
[akpm@linux-foundation.org: use `bool' for boolean variables, remove unneeded/undesirable cast of void*, add missed ep_scan_ready_list() kerneldoc]
Signed-off-by: Jason Baron <jbaron@akamai.com>
Tested-by: Nathan Zimmer <nzimmer@sgi.com>
Cc: Eric Wong <normalperson@yhbt.net>
Cc: Nelson Elhage <nelhage@nelhage.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nathan Zimmer found that once we get over 10+ cpus, the scalability of
SPECjbb falls over due to the contention on the global 'epmutex', which is
taken in on EPOLL_CTL_ADD and EPOLL_CTL_DEL operations.
Patch #1 removes the 'epmutex' lock completely from the EPOLL_CTL_DEL path
by using rcu to guard against any concurrent traversals.
Patch #2 remove the 'epmutex' lock from EPOLL_CTL_ADD operations for
simple topologies. IE when adding a link from an epoll file descriptor to
a wakeup source, where the epoll file descriptor is not nested.
This patch (of 2):
Optimize EPOLL_CTL_DEL such that it does not require the 'epmutex' by
converting the file->f_ep_links list into an rcu one. In this way, we can
traverse the epoll network on the add path in parallel with deletes.
Since deletes can't create loops or worse wakeup paths, this is safe.
This patch in combination with the patch "epoll: Do not take global 'epmutex'
for simple topologies", shows a dramatic performance improvement in
scalability for SPECjbb.
Signed-off-by: Jason Baron <jbaron@akamai.com>
Tested-by: Nathan Zimmer <nzimmer@sgi.com>
Cc: Eric Wong <normalperson@yhbt.net>
Cc: Nelson Elhage <nelhage@nelhage.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
CC: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This reverts commit 1c441e9212 (epoll: use freezable blocking call)
which is reported to cause user space memory corruption to happen
after suspend to RAM.
Since it appears to be extremely difficult to root cause this
problem, it is best to revert the offending commit and try to address
the original issue in a better way later.
References: https://bugzilla.kernel.org/show_bug.cgi?id=61781
Reported-by: Natrio <natrio@list.ru>
Reported-by: Jeff Pohlmeyer <yetanothergeek@gmail.com>
Bisected-by: Leo Wolf <jclw@ymail.com>
Fixes: 1c441e9212 (epoll: use freezable blocking call)
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: 3.11+ <stable@vger.kernel.org> # 3.11+
ep_free() might iterate on a huge set of epitems and hold cpu too long.
Add two cond_resched() in order to yield cpu to other tasks. This is safe
as we only hold mutexes in this function.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Theodore Ts'o <tytso@mit.edu>
Acked-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge first patch-bomb from Andrew Morton:
- various misc bits
- I'm been patchmonkeying ocfs2 for a while, as Joel and Mark have been
distracted. There has been quite a bit of activity.
- About half the MM queue
- Some backlight bits
- Various lib/ updates
- checkpatch updates
- zillions more little rtc patches
- ptrace
- signals
- exec
- procfs
- rapidio
- nbd
- aoe
- pps
- memstick
- tools/testing/selftests updates
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (445 commits)
tools/testing/selftests: don't assume the x bit is set on scripts
selftests: add .gitignore for kcmp
selftests: fix clean target in kcmp Makefile
selftests: add .gitignore for vm
selftests: add hugetlbfstest
self-test: fix make clean
selftests: exit 1 on failure
kernel/resource.c: remove the unneeded assignment in function __find_resource
aio: fix wrong comment in aio_complete()
drivers/w1/slaves/w1_ds2408.c: add magic sequence to disable P0 test mode
drivers/memstick/host/r592.c: convert to module_pci_driver
drivers/memstick/host/jmb38x_ms: convert to module_pci_driver
pps-gpio: add device-tree binding and support
drivers/pps/clients/pps-gpio.c: convert to module_platform_driver
drivers/pps/clients/pps-gpio.c: convert to devm_* helpers
drivers/parport/share.c: use kzalloc
Documentation/accounting/getdelays.c: avoid strncpy in accounting tool
aoe: update internal version number to v83
aoe: update copyright date
aoe: perform I/O completions in parallel
...
sigprocmask() should die. None of the current callers actually
need this strange interface.
Change fs/eventpoll.c to use set_current_blocked(). This also
means we should not worry about SIGKILL/SIGSTOP.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Eric Wong <normalperson@yhbt.net>
Cc: Jason Baron <jbaron@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Avoid waking up every thread sleeping in an epoll_wait call during
suspend and resume by calling a freezable blocking call. Previous
patches modified the freezer to avoid sending wakeups to threads
that are blocked in freezable blocking calls.
This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Pull compat cleanup from Al Viro:
"Mostly about syscall wrappers this time; there will be another pile
with patches in the same general area from various people, but I'd
rather push those after both that and vfs.git pile are in."
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/signal:
syscalls.h: slightly reduce the jungles of macros
get rid of union semop in sys_semctl(2) arguments
make do_mremap() static
sparc: no need to sign-extend in sync_file_range() wrapper
ppc compat wrappers for add_key(2) and request_key(2) are pointless
x86: trim sys_ia32.h
x86: sys32_kill and sys32_mprotect are pointless
get rid of compat_sys_semctl() and friends in case of ARCH_WANT_OLD_COMPAT_IPC
merge compat sys_ipc instances
consolidate compat lookup_dcookie()
convert vmsplice to COMPAT_SYSCALL_DEFINE
switch getrusage() to COMPAT_SYSCALL_DEFINE
switch epoll_pwait to COMPAT_SYSCALL_DEFINE
convert sendfile{,64} to COMPAT_SYSCALL_DEFINE
switch signalfd{,4}() to COMPAT_SYSCALL_DEFINE
make SYSCALL_DEFINE<n>-generated wrappers do asmlinkage_protect
make HAVE_SYSCALL_WRAPPERS unconditional
consolidate cond_syscall and SYSCALL_ALIAS declarations
teach SYSCALL_DEFINE<n> how to deal with long long/unsigned long long
get rid of duplicate logics in __SC_....[1-6] definitions
It is always safe to use RCU_INIT_POINTER to NULL a pointer. This results
in slightly smaller/faster code.
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This reduces the amount of code inside the ready list iteration loops for
better readability IMHO.
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Technically we do not need to hold ep->mtx during ep_free since we are
certain there are no other users of ep at that point. However, lockdep
complains with a "suspicious rcu_dereference_check() usage!" message; so
lock the mutex before ep_remove to silence the warning.
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Arve Hjønnevåg <arve@android.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: NeilBrown <neilb@suse.de>,
Cc: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This prevents wakeup_source destruction when a user hits the item with
EPOLL_CTL_MOD while ep_poll_callback is running.
Tested with CONFIG_SPARSE_RCU_POINTER=y and "make fs/eventpoll.o C=2"
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Arve Hjønnevåg <arve@android.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: NeilBrown <neilb@suse.de>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It is common for epoll users to have thousands of epitems, so saving a
cache line on every allocation leads to large memory savings.
Since epitem allocations are cache-aligned, reducing sizeof(struct
epitem) from 136 bytes to 128 bytes will allow it to squeeze under a
cache line boundary on x86_64.
Via /sys/kernel/slab/eventpoll_epi, I see the following changes on my
x86_64 Core2 Duo (which has 64-byte cache alignment):
object_size : 192 => 128
objs_per_slab: 21 => 32
Also, add a BUILD_BUG_ON() to check for future accidental breakage.
[akpm@linux-foundation.org: use __packed, for all architectures]
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
EPOLL_CTL_MOD sets the interest mask before calling f_op->poll() to
ensure events are not missed. Since the modifications to the interest
mask are not protected by the same lock as ep_poll_callback, we need to
ensure the change is visible to other CPUs calling ep_poll_callback.
We also need to ensure f_op->poll() has an up-to-date view of past
events which occured before we modified the interest mask. So this
barrier also pairs with the barrier in wq_has_sleeper().
This should guarantee either ep_poll_callback or f_op->poll() (or both)
will notice the readiness of a recently-ready/modified item.
This issue was encountered by Andreas Voellmy and Junchang(Jason) Wang in:
http://thread.gmane.org/gmane.linux.kernel/1408782/
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Cc: Hans Verkuil <hans.verkuil@cisco.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Mauro Carvalho Chehab <mchehab@infradead.org>
Cc: David Miller <davem@davemloft.net>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andreas Voellmy <andreas.voellmy@yale.edu>
Tested-by: "Junchang(Jason) Wang" <junchang.wang@yale.edu>
Cc: netdev@vger.kernel.org
Cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Revert commit 03a7beb55b ("epoll: support for disabling items, and a
self-test app") pending resolution of the issues identified by Michael
Kerrisk, copied below.
We'll revisit this for 3.8.
: I've taken a look at this patch as it currently stands in 3.7-rc1, and
: done a bit of testing. (By the way, the test program
: tools/testing/selftests/epoll/test_epoll.c does not compile...)
:
: There are one or two places where the behavior seems a little strange,
: so I have a question or two at the end of this mail. But other than
: that, I want to check my understanding so that the interface can be
: correctly documented.
:
: Just to go though my understanding, the problem is the following
: scenario in a multithreaded application:
:
: 1. Multiple threads are performing epoll_wait() operations,
: and maintaining a user-space cache that contains information
: corresponding to each file descriptor being monitored by
: epoll_wait().
:
: 2. At some point, a thread wants to delete (EPOLL_CTL_DEL)
: a file descriptor from the epoll interest list, and
: delete the corresponding record from the user-space cache.
:
: 3. The problem with (2) is that some other thread may have
: previously done an epoll_wait() that retrieved information
: about the fd in question, and may be in the middle of using
: information in the cache that relates to that fd. Thus,
: there is a potential race.
:
: 4. The race can't solved purely in user space, because doing
: so would require applying a mutex across the epoll_wait()
: call, which would of course blow thread concurrency.
:
: Right?
:
: Your solution is the EPOLL_CTL_DISABLE operation. I want to
: confirm my understanding about how to use this flag, since
: the description that has accompanied the patches so far
: has been a bit sparse
:
: 0. In the scenario you're concerned about, deleting a file
: descriptor means (safely) doing the following:
: (a) Deleting the file descriptor from the epoll interest list
: using EPOLL_CTL_DEL
: (b) Deleting the corresponding record in the user-space cache
:
: 1. It's only meaningful to use this EPOLL_CTL_DISABLE in
: conjunction with EPOLLONESHOT.
:
: 2. Using EPOLL_CTL_DISABLE without using EPOLLONESHOT in
: conjunction is a logical error.
:
: 3. The correct way to code multithreaded applications using
: EPOLL_CTL_DISABLE and EPOLLONESHOT is as follows:
:
: a. All EPOLL_CTL_ADD and EPOLL_CTL_MOD operations should
: should EPOLLONESHOT.
:
: b. When a thread wants to delete a file descriptor, it
: should do the following:
:
: [1] Call epoll_ctl(EPOLL_CTL_DISABLE)
: [2] If the return status from epoll_ctl(EPOLL_CTL_DISABLE)
: was zero, then the file descriptor can be safely
: deleted by the thread that made this call.
: [3] If the epoll_ctl(EPOLL_CTL_DISABLE) fails with EBUSY,
: then the descriptor is in use. In this case, the calling
: thread should set a flag in the user-space cache to
: indicate that the thread that is using the descriptor
: should perform the deletion operation.
:
: Is all of the above correct?
:
: The implementation depends on checking on whether
: (events & ~EP_PRIVATE_BITS) == 0
: This replies on the fact that EPOLL_CTL_AD and EPOLL_CTL_MOD always
: set EPOLLHUP and EPOLLERR in the 'events' mask, and EPOLLONESHOT
: causes those flags (as well as all others in ~EP_PRIVATE_BITS) to be
: cleared.
:
: A corollary to the previous paragraph is that using EPOLL_CTL_DISABLE
: is only useful in conjunction with EPOLLONESHOT. However, as things
: stand, one can use EPOLL_CTL_DISABLE on a file descriptor that does
: not have EPOLLONESHOT set in 'events' This results in the following
: (slightly surprising) behavior:
:
: (a) The first call to epoll_ctl(EPOLL_CTL_DISABLE) returns 0
: (the indicator that the file descriptor can be safely deleted).
: (b) The next call to epoll_ctl(EPOLL_CTL_DISABLE) fails with EBUSY.
:
: This doesn't seem particularly useful, and in fact is probably an
: indication that the user made a logic error: they should only be using
: epoll_ctl(EPOLL_CTL_DISABLE) on a file descriptor for which
: EPOLLONESHOT was set in 'events'. If that is correct, then would it
: not make sense to return an error to user space for this case?
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: "Paton J. Lewis" <palewis@adobe.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Enhanced epoll_ctl to support EPOLL_CTL_DISABLE, which disables an epoll
item. If epoll_ctl doesn't return -EBUSY in this case, it is then safe to
delete the epoll item in a multi-threaded environment. Also added a new
test_epoll self- test app to both demonstrate the need for this feature
and test it.
Signed-off-by: Paton J. Lewis <palewis@adobe.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Paul Holland <pholland@adobe.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
As soon as we'd installed the file into descriptor table, it can
get closed by another thread. Freeing ep in process...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
As discussed in
http://thread.gmane.org/gmane.linux.kernel/1249726/focus=1288990,
the capability introduced in 4d7e30d989
to govern EPOLLWAKEUP seems misnamed: this capability is about governing
the ability to suspend the system, not using a particular API flag
(EPOLLWAKEUP). We should make the name of the capability more general
to encourage reuse in related cases. (Whether or not this capability
should also be used to govern the use of /sys/power/wake_lock is a
question that needs to be separately resolved.)
This patch renames the capability to CAP_BLOCK_SUSPEND. In order to ensure
that the old capability name doesn't make it out into the wild, could you
please apply and push up the tree to ensure that it is incorporated
for the 3.5 release.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Everyone either defines it in arch thread_info.h or has TIF_RESTORE_SIGMASK
and picks default set_restore_sigmask() in linux/thread_info.h. Kill the
ifdefs, slap #error in linux/thread_info.h to catch breakage when new ones
get merged.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Commit 4d7e30d (epoll: Add a flag, EPOLLWAKEUP, to prevent
suspend while epoll events are ready) caused some applications to
malfunction, because they set the bit corresponding to the new
EPOLLWAKEUP flag in their eventpoll flags and they don't have the
new CAP_EPOLLWAKEUP capability.
To prevent that from happening, change epoll_ctl() to clear
EPOLLWAKEUP in epds.events if the caller doesn't have the
CAP_EPOLLWAKEUP capability instead of failing and returning an
error code, which allows the affected applications to function
normally.
Reported-and-tested-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
When an epoll_event, that has the EPOLLWAKEUP flag set, is ready, a
wakeup_source will be active to prevent suspend. This can be used to
handle wakeup events from a driver that support poll, e.g. input, if
that driver wakes up the waitqueue passed to epoll before allowing
suspend.
Signed-off-by: Arve Hjønnevåg <arve@android.com>
Reviewed-by: NeilBrown <neilb@suse.de>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
An epoll_ctl(,EPOLL_CTL_ADD,,) operation can return '-ELOOP' to prevent
circular epoll dependencies from being created. However, in that case we
do not properly clear the 'tfile_check_list'. Thus, add a call to
clear_tfile_check_list() for the -ELOOP case.
Signed-off-by: Jason Baron <jbaron@redhat.com>
Reported-by: Yurij M. Plotnikov <Yurij.Plotnikov@oktetlabs.ru>
Cc: Nelson Elhage <nelhage@nelhage.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Tested-by: Alexandra N. Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)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=G9mT
-----END PGP SIGNATURE-----
Merge tag 'split-asm_system_h-for-linus-20120328' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-asm_system
Pull "Disintegrate and delete asm/system.h" from David Howells:
"Here are a bunch of patches to disintegrate asm/system.h into a set of
separate bits to relieve the problem of circular inclusion
dependencies.
I've built all the working defconfigs from all the arches that I can
and made sure that they don't break.
The reason for these patches is that I recently encountered a circular
dependency problem that came about when I produced some patches to
optimise get_order() by rewriting it to use ilog2().
This uses bitops - and on the SH arch asm/bitops.h drags in
asm-generic/get_order.h by a circuituous route involving asm/system.h.
The main difficulty seems to be asm/system.h. It holds a number of
low level bits with no/few dependencies that are commonly used (eg.
memory barriers) and a number of bits with more dependencies that
aren't used in many places (eg. switch_to()).
These patches break asm/system.h up into the following core pieces:
(1) asm/barrier.h
Move memory barriers here. This already done for MIPS and Alpha.
(2) asm/switch_to.h
Move switch_to() and related stuff here.
(3) asm/exec.h
Move arch_align_stack() here. Other process execution related bits
could perhaps go here from asm/processor.h.
(4) asm/cmpxchg.h
Move xchg() and cmpxchg() here as they're full word atomic ops and
frequently used by atomic_xchg() and atomic_cmpxchg().
(5) asm/bug.h
Move die() and related bits.
(6) asm/auxvec.h
Move AT_VECTOR_SIZE_ARCH here.
Other arch headers are created as needed on a per-arch basis."
Fixed up some conflicts from other header file cleanups and moving code
around that has happened in the meantime, so David's testing is somewhat
weakened by that. We'll find out anything that got broken and fix it..
* tag 'split-asm_system_h-for-linus-20120328' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-asm_system: (38 commits)
Delete all instances of asm/system.h
Remove all #inclusions of asm/system.h
Add #includes needed to permit the removal of asm/system.h
Move all declarations of free_initmem() to linux/mm.h
Disintegrate asm/system.h for OpenRISC
Split arch_align_stack() out from asm-generic/system.h
Split the switch_to() wrapper out of asm-generic/system.h
Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h
Create asm-generic/barrier.h
Make asm-generic/cmpxchg.h #include asm-generic/cmpxchg-local.h
Disintegrate asm/system.h for Xtensa
Disintegrate asm/system.h for Unicore32 [based on ver #3, changed by gxt]
Disintegrate asm/system.h for Tile
Disintegrate asm/system.h for Sparc
Disintegrate asm/system.h for SH
Disintegrate asm/system.h for Score
Disintegrate asm/system.h for S390
Disintegrate asm/system.h for PowerPC
Disintegrate asm/system.h for PA-RISC
Disintegrate asm/system.h for MN10300
...
Remove all #inclusions of asm/system.h preparatory to splitting and killing
it. Performed with the following command:
perl -p -i -e 's!^#\s*include\s*<asm/system[.]h>.*\n!!' `grep -Irl '^#\s*include\s*<asm/system[.]h>' *`
Signed-off-by: David Howells <dhowells@redhat.com>
We never use the length variable.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Jason Baron <jbaron@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Looking for a bug in -rt, I stumbled across this code here from: commit
2dfa4eeab0 ("epoll keyed wakeups: teach epoll about hints coming with
the wakeup key"), specifically:
#ifdef CONFIG_DEBUG_LOCK_ALLOC
static inline void ep_wake_up_nested(wait_queue_head_t *wqueue,
unsigned long events, int subclass)
{
unsigned long flags;
spin_lock_irqsave_nested(&wqueue->lock, flags, subclass);
wake_up_locked_poll(wqueue, events);
spin_unlock_irqrestore(&wqueue->lock, flags);
}
#else
static inline void ep_wake_up_nested(wait_queue_head_t *wqueue,
unsigned long events, int subclass)
{
wake_up_poll(wqueue, events);
}
#endif
You change the function of ep_wake_up_nested() depending on whether
CONFIG_DEBUG_LOCK_ALLOC is set or not. This looks awfully suspicious,
and there's no comment to explain why. I initially thought that this
was trying to fool lockdep, and hiding a real bug.
Investigating it, I found the creation of wake_up_nested() (which no
longer exists) but was created for the sole purpose of epoll and its
strange wake ups, as explained in commit 0ccf831cbe ("lockdep:
annotate epoll")
Although the commit message says "annotate epoll" the change log is much
better at explaining what is happening than what is in the actual code.
Thus a comment is really necessary here. And to save the time of other
developers from having to go trudging through the git logs trying to
figure out why this code exists.
I took parts of the change log and placed it into a comment above the
affected code. This will make the description of what is happening more
visible to new developers that have to look at this code for the first
time.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In some cases the poll() implementation in a driver has to do different
things depending on the events the caller wants to poll for. An example
is when a driver needs to start a DMA engine if the caller polls for
POLLIN, but doesn't want to do that if POLLIN is not requested but instead
only POLLOUT or POLLPRI is requested. This is something that can happen
in the video4linux subsystem among others.
Unfortunately, the current epoll/poll/select implementation doesn't
provide that information reliably. The poll_table_struct does have it: it
has a key field with the event mask. But once a poll() call matches one
or more bits of that mask any following poll() calls are passed a NULL
poll_table pointer.
Also, the eventpoll implementation always left the key field at ~0 instead
of using the requested events mask.
This was changed in eventpoll.c so the key field now contains the actual
events that should be polled for as set by the caller.
The solution to the NULL poll_table pointer is to set the qproc field to
NULL in poll_table once poll() matches the events, not the poll_table
pointer itself. That way drivers can obtain the mask through a new
poll_requested_events inline.
The poll_table_struct can still be NULL since some kernel code calls it
internally (netfs_state_poll() in ./drivers/staging/pohmelfs/netfs.h). In
that case poll_requested_events() returns ~0 (i.e. all events).
Very rarely drivers might want to know whether poll_wait will actually
wait. If another earlier file descriptor in the set already matched the
events the caller wanted to wait for, then the kernel will return from the
select() call without waiting. This might be useful information in order
to avoid doing expensive work.
A new helper function poll_does_not_wait() is added that drivers can use
to detect this situation. This is now used in sock_poll_wait() in
include/net/sock.h. This was the only place in the kernel that needed
this information.
Drivers should no longer access any of the poll_table internals, but use
the poll_requested_events() and poll_does_not_wait() access functions
instead. In order to enforce that the poll_table fields are now prepended
with an underscore and a comment was added warning against using them
directly.
This required a change in unix_dgram_poll() in unix/af_unix.c which used
the key field to get the requested events. It's been replaced by a call
to poll_requested_events().
For qproc it was especially important to change its name since the
behavior of that field changes with this patch since this function pointer
can now be NULL when that wasn't possible in the past.
Any driver accessing the qproc or key fields directly will now fail to compile.
Some notes regarding the correctness of this patch: the driver's poll()
function is called with a 'struct poll_table_struct *wait' argument. This
pointer may or may not be NULL, drivers can never rely on it being one or
the other as that depends on whether or not an earlier file descriptor in
the select()'s fdset matched the requested events.
There are only three things a driver can do with the wait argument:
1) obtain the key field:
events = wait ? wait->key : ~0;
This will still work although it should be replaced with the new
poll_requested_events() function (which does exactly the same).
This will now even work better, since wait is no longer set to NULL
unnecessarily.
2) use the qproc callback. This could be deadly since qproc can now be
NULL. Renaming qproc should prevent this from happening. There are no
kernel drivers that actually access this callback directly, BTW.
3) test whether wait == NULL to determine whether poll would return without
waiting. This is no longer sufficient as the correct test is now
wait == NULL || wait->_qproc == NULL.
However, the worst that can happen here is a slight performance hit in
the case where wait != NULL and wait->_qproc == NULL. In that case the
driver will assume that poll_wait() will actually add the fd to the set
of waiting file descriptors. Of course, poll_wait() will not do that
since it tests for wait->_qproc. This will not break anything, though.
There is only one place in the whole kernel where this happens
(sock_poll_wait() in include/net/sock.h) and that code will be replaced
by a call to poll_does_not_wait() in the next patch.
Note that even if wait->_qproc != NULL drivers cannot rely on poll_wait()
actually waiting. The next file descriptor from the set might match the
event mask and thus any possible waits will never happen.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reviewed-by: Jonathan Corbet <corbet@lwn.net>
Reviewed-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: Davide Libenzi <davidel@xmailserver.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Cc: Mauro Carvalho Chehab <mchehab@infradead.org>
Cc: David Miller <davem@davemloft.net>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 28d82dc1c4 ("epoll: limit paths") that I did to limit the
number of possible wakeup paths in epoll is causing a few applications
to longer work (dovecot for one).
The original patch is really about limiting the amount of epoll nesting
(since epoll fds can be attached to other fds). Thus, we probably can
allow an unlimited number of paths of depth 1. My current patch limits
it at 1000. And enforce the limits on paths that have a greater depth.
This is captured in: https://bugzilla.redhat.com/show_bug.cgi?id=681578
Signed-off-by: Jason Baron <jbaron@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
signalfd_cleanup() ensures that ->signalfd_wqh is not used, but
this is not enough. eppoll_entry->whead still points to the memory
we are going to free, ep_unregister_pollwait()->remove_wait_queue()
is obviously unsafe.
Change ep_poll_callback(POLLFREE) to set eppoll_entry->whead = NULL,
change ep_unregister_pollwait() to check pwq->whead != NULL under
rcu_read_lock() before remove_wait_queue(). We add the new helper,
ep_remove_wait_queue(), for this.
This works because sighand_cachep is SLAB_DESTROY_BY_RCU and because
->signalfd_wqh is initialized in sighand_ctor(), not in copy_sighand.
ep_unregister_pollwait()->remove_wait_queue() can play with already
freed and potentially reused ->sighand, but this is fine. This memory
must have the valid ->signalfd_wqh until rcu_read_unlock().
Reported-by: Maxime Bizon <mbizon@freebox.fr>
Cc: <stable@kernel.org>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch is intentionally incomplete to simplify the review.
It ignores ep_unregister_pollwait() which plays with the same wqh.
See the next change.
epoll assumes that the EPOLL_CTL_ADD'ed file controls everything
f_op->poll() needs. In particular it assumes that the wait queue
can't go away until eventpoll_release(). This is not true in case
of signalfd, the task which does EPOLL_CTL_ADD uses its ->sighand
which is not connected to the file.
This patch adds the special event, POLLFREE, currently only for
epoll. It expects that init_poll_funcptr()'ed hook should do the
necessary cleanup. Perhaps it should be defined as EPOLLFREE in
eventpoll.
__cleanup_sighand() is changed to do wake_up_poll(POLLFREE) if
->signalfd_wqh is not empty, we add the new signalfd_cleanup()
helper.
ep_poll_callback(POLLFREE) simply does list_del_init(task_list).
This make this poll entry inconsistent, but we don't care. If you
share epoll fd which contains our sigfd with another process you
should blame yourself. signalfd is "really special". I simply do
not know how we can define the "right" semantics if it used with
epoll.
The main problem is, epoll calls signalfd_poll() once to establish
the connection with the wait queue, after that signalfd_poll(NULL)
returns the different/inconsistent results depending on who does
EPOLL_CTL_MOD/signalfd_read/etc. IOW: apart from sigmask, signalfd
has nothing to do with the file, it works with the current thread.
In short: this patch is the hack which tries to fix the symptoms.
It also assumes that nobody can take tasklist_lock under epoll
locks, this seems to be true.
Note:
- we do not have wake_up_all_poll() but wake_up_poll()
is fine, poll/epoll doesn't use WQ_FLAG_EXCLUSIVE.
- signalfd_cleanup() uses POLLHUP along with POLLFREE,
we need a couple of simple changes in eventpoll.c to
make sure it can't be "lost".
Reported-by: Maxime Bizon <mbizon@freebox.fr>
Cc: <stable@kernel.org>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The current epoll code can be tickled to run basically indefinitely in
both loop detection path check (on ep_insert()), and in the wakeup paths.
The programs that tickle this behavior set up deeply linked networks of
epoll file descriptors that cause the epoll algorithms to traverse them
indefinitely. A couple of these sample programs have been previously
posted in this thread: https://lkml.org/lkml/2011/2/25/297.
To fix the loop detection path check algorithms, I simply keep track of
the epoll nodes that have been already visited. Thus, the loop detection
becomes proportional to the number of epoll file descriptor and links.
This dramatically decreases the run-time of the loop check algorithm. In
one diabolical case I tried it reduced the run-time from 15 mintues (all
in kernel time) to .3 seconds.
Fixing the wakeup paths could be done at wakeup time in a similar manner
by keeping track of nodes that have already been visited, but the
complexity is harder, since there can be multiple wakeups on different
cpus...Thus, I've opted to limit the number of possible wakeup paths when
the paths are created.
This is accomplished, by noting that the end file descriptor points that
are found during the loop detection pass (from the newly added link), are
actually the sources for wakeup events. I keep a list of these file
descriptors and limit the number and length of these paths that emanate
from these 'source file descriptors'. In the current implemetation I
allow 1000 paths of length 1, 500 of length 2, 100 of length 3, 50 of
length 4 and 10 of length 5. Note that it is sufficient to check the
'source file descriptors' reachable from the newly added link, since no
other 'source file descriptors' will have newly added links. This allows
us to check only the wakeup paths that may have gotten too long, and not
re-check all possible wakeup paths on the system.
In terms of the path limit selection, I think its first worth noting that
the most common case for epoll, is probably the model where you have 1
epoll file descriptor that is monitoring n number of 'source file
descriptors'. In this case, each 'source file descriptor' has a 1 path of
length 1. Thus, I believe that the limits I'm proposing are quite
reasonable and in fact may be too generous. Thus, I'm hoping that the
proposed limits will not prevent any workloads that currently work to
fail.
In terms of locking, I have extended the use of the 'epmutex' to all
epoll_ctl add and remove operations. Currently its only used in a subset
of the add paths. I need to hold the epmutex, so that we can correctly
traverse a coherent graph, to check the number of paths. I believe that
this additional locking is probably ok, since its in the setup/teardown
paths, and doesn't affect the running paths, but it certainly is going to
add some extra overhead. Also, worth noting is that the epmuex was
recently added to the ep_ctl add operations in the initial path loop
detection code using the argument that it was not on a critical path.
Another thing to note here, is the length of epoll chains that is allowed.
Currently, eventpoll.c defines:
/* Maximum number of nesting allowed inside epoll sets */
#define EP_MAX_NESTS 4
This basically means that I am limited to a graph depth of 5 (EP_MAX_NESTS
+ 1). However, this limit is currently only enforced during the loop
check detection code, and only when the epoll file descriptors are added
in a certain order. Thus, this limit is currently easily bypassed. The
newly added check for wakeup paths, stricly limits the wakeup paths to a
length of 5, regardless of the order in which ep's are linked together.
Thus, a side-effect of the new code is a more consistent enforcement of
the graph depth.
Thus far, I've tested this, using the sample programs previously
mentioned, which now either return quickly or return -EINVAL. I've also
testing using the piptest.c epoll tester, which showed no difference in
performance. I've also created a number of different epoll networks and
tested that they behave as expectded.
I believe this solves the original diabolical test cases, while still
preserving the sane epoll nesting.
Signed-off-by: Jason Baron <jbaron@redhat.com>
Cc: Nelson Elhage <nelhage@ksplice.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
epoll can acquire recursively acquire ep->mtx on multiple "struct
eventpoll"s at once in the case where one epoll fd is monitoring another
epoll fd. This is perfectly OK, since we're careful about the lock
ordering, but it causes spurious lockdep warnings. Annotate the recursion
using mutex_lock_nested, and add a comment explaining the nesting rules
for good measure.
Recent versions of systemd are triggering this, and it can also be
demonstrated with the following trivial test program:
--------------------8<--------------------
int main(void) {
int e1, e2;
struct epoll_event evt = {
.events = EPOLLIN
};
e1 = epoll_create1(0);
e2 = epoll_create1(0);
epoll_ctl(e1, EPOLL_CTL_ADD, e2, &evt);
return 0;
}
--------------------8<--------------------
Reported-by: Paul Bolle <pebolle@tiscali.nl>
Tested-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: Nelson Elhage <nelhage@nelhage.com>
Acked-by: Jason Baron <jbaron@redhat.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: <stable@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jason Baron