commit c291ee6221 upstream.
Since the rework of the sparse interrupt code to actually free the
unused interrupt descriptors there exists a race between the /proc
interfaces to the irq subsystem and the code which frees the interrupt
descriptor.
CPU0 CPU1
show_interrupts()
desc = irq_to_desc(X);
free_desc(desc)
remove_from_radix_tree();
kfree(desc);
raw_spinlock_irq(&desc->lock);
/proc/interrupts is the only interface which can actively corrupt
kernel memory via the lock access. /proc/stat can only read from freed
memory. Extremly hard to trigger, but possible.
The interfaces in /proc/irq/N/ are not affected by this because the
removal of the proc file is serialized in procfs against concurrent
readers/writers. The removal happens before the descriptor is freed.
For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue
as the descriptor is never freed. It's merely cleared out with the irq
descriptor lock held. So any concurrent proc access will either see
the old correct value or the cleared out ones.
Protect the lookup and access to the irq descriptor in
show_interrupts() with the sparse_irq_lock.
Provide kstat_irqs_usr() which is protecting the lookup and access
with sparse_irq_lock and switch /proc/stat to use it.
Document the existing kstat_irqs interfaces so it's clear that the
caller needs to take care about protection. The users of these
interfaces are either not affected due to SPARSE_IRQ=n or already
protected against removal.
Fixes: 1f5a5b87f7 "genirq: Implement a sane sparse_irq allocator"
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a5fd9733a3 upstream.
commit 4dbd27711c "tick: export nohz tick idle symbols for module
use" was merged via the thermal tree without an explicit ack from the
relevant maintainers.
The exports are abused by the intel powerclamp driver which implements
a fake idle state from a sched FIFO task. This causes all kinds of
wreckage in the NOHZ core code which rightfully assumes that
tick_nohz_idle_enter/exit() are only called from the idle task itself.
Recent changes in the NOHZ core lead to a failure of the powerclamp
driver and now people try to hack completely broken and backwards
workarounds into the NOHZ core code. This is completely unacceptable
and just papers over the real problem. There are way more subtle
issues lurking around the corner.
The real solution is to fix the powerclamp driver by rewriting it with
a sane concept, but that's beyond the scope of this.
So the only solution for now is to remove the calls into the core NOHZ
code from the powerclamp trainwreck along with the exports.
Fixes: d6d71ee4a1 "PM: Introduce Intel PowerClamp Driver"
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Fengguang Wu <fengguang.wu@intel.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Cc: Pan Jacob jun <jacob.jun.pan@intel.com>
Cc: LKP <lkp@01.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Zhang Rui <rui.zhang@intel.com>
Link: http://lkml.kernel.org/r/alpine.DEB.2.11.1412181110110.17382@nanos
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9fc81d8742 upstream.
We allow PMU driver to change the cpu on which the event
should be installed to. This happened in patch:
e2d37cd213 ("perf: Allow the PMU driver to choose the CPU on which to install events")
This patch also forces all the group members to follow
the currently opened events cpu if the group happened
to be moved.
This and the change of event->cpu in perf_install_in_context()
function introduced in:
0cda4c0231 ("perf: Introduce perf_pmu_migrate_context()")
forces group members to change their event->cpu,
if the currently-opened-event's PMU changed the cpu
and there is a group move.
Above behaviour causes problem for breakpoint events,
which uses event->cpu to touch cpu specific data for
breakpoints accounting. By changing event->cpu, some
breakpoints slots were wrongly accounted for given
cpu.
Vinces's perf fuzzer hit this issue and caused following
WARN on my setup:
WARNING: CPU: 0 PID: 20214 at arch/x86/kernel/hw_breakpoint.c:119 arch_install_hw_breakpoint+0x142/0x150()
Can't find any breakpoint slot
[...]
This patch changes the group moving code to keep the event's
original cpu.
Reported-by: Vince Weaver <vince@deater.net>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Vince Weaver <vince@deater.net>
Cc: Yan, Zheng <zheng.z.yan@intel.com>
Link: http://lkml.kernel.org/r/1418243031-20367-3-git-send-email-jolsa@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 269ad8015a upstream.
The dl_runtime_exceeded() function is supposed to ckeck if
a SCHED_DEADLINE task must be throttled, by checking if its
current runtime is <= 0. However, it also checks if the
scheduling deadline has been missed (the current time is
larger than the current scheduling deadline), further
decreasing the runtime if this happens.
This "double accounting" is wrong:
- In case of partitioned scheduling (or single CPU), this
happens if task_tick_dl() has been called later than expected
(due to small HZ values). In this case, the current runtime is
also negative, and replenish_dl_entity() can take care of the
deadline miss by recharging the current runtime to a value smaller
than dl_runtime
- In case of global scheduling on multiple CPUs, scheduling
deadlines can be missed even if the task did not consume more
runtime than expected, hence penalizing the task is wrong
This patch fix this problem by throttling a SCHED_DEADLINE task
only when its runtime becomes negative, and not modifying the runtime
Signed-off-by: Luca Abeni <luca.abeni@unitn.it>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@gmail.com>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1418813432-20797-3-git-send-email-luca.abeni@unitn.it
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6a503c3be9 upstream.
According to global EDF, tasks should be migrated between runqueues
without checking if their scheduling deadlines and runtimes are valid.
However, SCHED_DEADLINE currently performs such a check:
a migration happens doing:
deactivate_task(rq, next_task, 0);
set_task_cpu(next_task, later_rq->cpu);
activate_task(later_rq, next_task, 0);
which ends up calling dequeue_task_dl(), setting the new CPU, and then
calling enqueue_task_dl().
enqueue_task_dl() then calls enqueue_dl_entity(), which calls
update_dl_entity(), which can modify scheduling deadline and runtime,
breaking global EDF scheduling.
As a result, some of the properties of global EDF are not respected:
for example, a taskset {(30, 80), (40, 80), (120, 170)} scheduled on
two cores can have unbounded response times for the third task even
if 30/80+40/80+120/170 = 1.5809 < 2
This can be fixed by invoking update_dl_entity() only in case of
wakeup, or if this is a new SCHED_DEADLINE task.
Signed-off-by: Luca Abeni <luca.abeni@unitn.it>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@gmail.com>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1418813432-20797-2-git-send-email-luca.abeni@unitn.it
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 24c037ebf5 upstream.
alloc_pid() does get_pid_ns() beforehand but forgets to put_pid_ns() if it
fails because disable_pid_allocation() was called by the exiting
child_reaper.
We could simply move get_pid_ns() down to successful return, but this fix
tries to be as trivial as possible.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Aaron Tomlin <atomlin@redhat.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Serge Hallyn <serge.hallyn@ubuntu.com>
Cc: Sterling Alexander <stalexan@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 041d7b98ff upstream.
A regression was caused by commit 780a7654cee8:
audit: Make testing for a valid loginuid explicit.
(which in turn attempted to fix a regression caused by e1760bd)
When audit_krule_to_data() fills in the rules to get a listing, there was a
missing clause to convert back from AUDIT_LOGINUID_SET to AUDIT_LOGINUID.
This broke userspace by not returning the same information that was sent and
expected.
The rule:
auditctl -a exit,never -F auid=-1
gives:
auditctl -l
LIST_RULES: exit,never f24=0 syscall=all
when it should give:
LIST_RULES: exit,never auid=-1 (0xffffffff) syscall=all
Tag it so that it is reported the same way it was set. Create a new
private flags audit_krule field (pflags) to store it that won't interact with
the public one from the API.
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <pmoore@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 66d2f338ee upstream.
Now that setgroups can be disabled and not reenabled, setting gid_map
without privielge can now be enabled when setgroups is disabled.
This restores most of the functionality that was lost when unprivileged
setting of gid_map was removed. Applications that use this functionality
will need to check to see if they use setgroups or init_groups, and if they
don't they can be fixed by simply disabling setgroups before writing to
gid_map.
Reviewed-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9cc46516dd upstream.
- Expose the knob to user space through a proc file /proc/<pid>/setgroups
A value of "deny" means the setgroups system call is disabled in the
current processes user namespace and can not be enabled in the
future in this user namespace.
A value of "allow" means the segtoups system call is enabled.
- Descendant user namespaces inherit the value of setgroups from
their parents.
- A proc file is used (instead of a sysctl) as sysctls currently do
not allow checking the permissions at open time.
- Writing to the proc file is restricted to before the gid_map
for the user namespace is set.
This ensures that disabling setgroups at a user namespace
level will never remove the ability to call setgroups
from a process that already has that ability.
A process may opt in to the setgroups disable for itself by
creating, entering and configuring a user namespace or by calling
setns on an existing user namespace with setgroups disabled.
Processes without privileges already can not call setgroups so this
is a noop. Prodcess with privilege become processes without
privilege when entering a user namespace and as with any other path
to dropping privilege they would not have the ability to call
setgroups. So this remains within the bounds of what is possible
without a knob to disable setgroups permanently in a user namespace.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f0d62aec93 upstream.
Generalize id_map_mutex so it can be used for more state of a user namespace.
Reviewed-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f95d7918bd upstream.
If you did not create the user namespace and are allowed
to write to uid_map or gid_map you should already have the necessary
privilege in the parent user namespace to establish any mapping
you want so this will not affect userspace in practice.
Limiting unprivileged uid mapping establishment to the creator of the
user namespace makes it easier to verify all credentials obtained with
the uid mapping can be obtained without the uid mapping without
privilege.
Limiting unprivileged gid mapping establishment (which is temporarily
absent) to the creator of the user namespace also ensures that the
combination of uid and gid can already be obtained without privilege.
This is part of the fix for CVE-2014-8989.
Reviewed-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 80dd00a237 upstream.
setresuid allows the euid to be set to any of uid, euid, suid, and
fsuid. Therefor it is safe to allow an unprivileged user to map
their euid and use CAP_SETUID privileged with exactly that uid,
as no new credentials can be obtained.
I can not find a combination of existing system calls that allows setting
uid, euid, suid, and fsuid from the fsuid making the previous use
of fsuid for allowing unprivileged mappings a bug.
This is part of a fix for CVE-2014-8989.
Reviewed-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit be7c6dba23 upstream.
As any gid mapping will allow and must allow for backwards
compatibility dropping groups don't allow any gid mappings to be
established without CAP_SETGID in the parent user namespace.
For a small class of applications this change breaks userspace
and removes useful functionality. This small class of applications
includes tools/testing/selftests/mount/unprivilged-remount-test.c
Most of the removed functionality will be added back with the addition
of a one way knob to disable setgroups. Once setgroups is disabled
setting the gid_map becomes as safe as setting the uid_map.
For more common applications that set the uid_map and the gid_map
with privilege this change will have no affect.
This is part of a fix for CVE-2014-8989.
Reviewed-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 273d2c67c3 upstream.
setgroups is unique in not needing a valid mapping before it can be called,
in the case of setgroups(0, NULL) which drops all supplemental groups.
The design of the user namespace assumes that CAP_SETGID can not actually
be used until a gid mapping is established. Therefore add a helper function
to see if the user namespace gid mapping has been established and call
that function in the setgroups permission check.
This is part of the fix for CVE-2014-8989, being able to drop groups
without privilege using user namespaces.
Reviewed-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 0542f17bf2 upstream.
The rule is simple. Don't allow anything that wouldn't be allowed
without unprivileged mappings.
It was previously overlooked that establishing gid mappings would
allow dropping groups and potentially gaining permission to files and
directories that had lesser permissions for a specific group than for
all other users.
This is the rule needed to fix CVE-2014-8989 and prevent any other
security issues with new_idmap_permitted.
The reason for this rule is that the unix permission model is old and
there are programs out there somewhere that take advantage of every
little corner of it. So allowing a uid or gid mapping to be
established without privielge that would allow anything that would not
be allowed without that mapping will result in expectations from some
code somewhere being violated. Violated expectations about the
behavior of the OS is a long way to say a security issue.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 7ff4d90b4c upstream.
Today there are 3 instances of setgroups and due to an oversight their
permission checking has diverged. Add a common function so that
they may all share the same permission checking code.
This corrects the current oversight in the current permission checks
and adds a helper to avoid this in the future.
A user namespace security fix will update this new helper, shortly.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 82975bc6a6 upstream.
x86 call do_notify_resume on paranoid returns if TIF_UPROBE is set but
not on non-paranoid returns. I suspect that this is a mistake and that
the code only works because int3 is paranoid.
Setting _TIF_NOTIFY_RESUME in the uprobe code was probably a workaround
for the x86 bug. With that bug fixed, we can remove _TIF_NOTIFY_RESUME
from the uprobes code.
Reported-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit b3f207855f upstream.
When running a 32-bit userspace on a 64-bit kernel (eg. i386
application on x86_64 kernel or 32-bit arm userspace on arm64
kernel) some of the perf ioctls must be treated with special
care, as they have a pointer size encoded in the command.
For example, PERF_EVENT_IOC_ID in 32-bit world will be encoded
as 0x80042407, but 64-bit kernel will expect 0x80082407. In
result the ioctl will fail returning -ENOTTY.
This patch solves the problem by adding code fixing up the
size as compat_ioctl file operation.
Reported-by: Drew Richardson <drew.richardson@arm.com>
Signed-off-by: Pawel Moll <pawel.moll@arm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1402671812-9078-1-git-send-email-pawel.moll@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: David Ahern <daahern@cisco.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2aa792e6fa upstream.
The rcu_gp_kthread_wake() function checks for three conditions before
waking up grace period kthreads:
* Is the thread we are trying to wake up the current thread?
* Are the gp_flags zero? (all threads wait on non-zero gp_flags condition)
* Is there no thread created for this flavour, hence nothing to wake up?
If any one of these condition is true, we do not call wake_up().
It was found that there are quite a few avoidable wake ups both during
idle time and under stress induced by rcutorture.
Idle:
Total:66000, unnecessary:66000, case1:61827, case2:66000, case3:0
Total:68000, unnecessary:68000, case1:63696, case2:68000, case3:0
rcutorture:
Total:254000, unnecessary:254000, case1:199913, case2:254000, case3:0
Total:256000, unnecessary:256000, case1:201784, case2:256000, case3:0
Here case{1-3} are the cases listed above. We can avoid these wake
ups by using rcu_gp_kthread_wake() to conditionally wake up the grace
period kthreads.
There is a comment about an implied barrier supplied by the wake_up()
logic. This barrier is necessary for the awakened thread to see the
updated ->gp_flags. This flag is always being updated with the root node
lock held. Also, the awakened thread tries to acquire the root node lock
before reading ->gp_flags because of which there is proper ordering.
Hence this commit tries to avoid calling wake_up() whenever we can by
using rcu_gp_kthread_wake() function.
Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>
CC: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Kamal Mostafa <kamal@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 48a7639ce8 upstream.
The rcu_start_gp_advanced() function currently uses irq_work_queue()
to defer wakeups of the RCU grace-period kthread. This deferring
is necessary to avoid RCU-scheduler deadlocks involving the rcu_node
structure's lock, meaning that RCU cannot call any of the scheduler's
wake-up functions while holding one of these locks.
Unfortunately, the second and subsequent calls to irq_work_queue() are
ignored, and the first call will be ignored (aside from queuing the work
item) if the scheduler-clock tick is turned off. This is OK for many
uses, especially those where irq_work_queue() is called from an interrupt
or softirq handler, because in those cases the scheduler-clock-tick state
will be re-evaluated, which will turn the scheduler-clock tick back on.
On the next tick, any deferred work will then be processed.
However, this strategy does not always work for RCU, which can be invoked
at process level from idle CPUs. In this case, the tick might never
be turned back on, indefinitely defering a grace-period start request.
Note that the RCU CPU stall detector cannot see this condition, because
there is no RCU grace period in progress. Therefore, we can (and do!)
see long tens-of-seconds stalls in grace-period handling. In theory,
we could see a full grace-period hang, but rcutorture testing to date
has seen only the tens-of-seconds stalls. Event tracing demonstrates
that irq_work_queue() is being called repeatedly to no effect during
these stalls: The "newreq" event appears repeatedly from a task that is
not one of the grace-period kthreads.
In theory, irq_work_queue() might be fixed to avoid this sort of issue,
but RCU's requirements are unusual and it is quite straightforward to pass
wake-up responsibility up through RCU's call chain, so that the wakeup
happens when the offending locks are released.
This commit therefore makes this change. The rcu_start_gp_advanced(),
rcu_start_future_gp(), rcu_accelerate_cbs(), rcu_advance_cbs(),
__note_gp_changes(), and rcu_start_gp() functions now return a boolean
which indicates when a wake-up is needed. A new rcu_gp_kthread_wake()
does the wakeup when it is necessary and safe to do so: No self-wakes,
no wake-ups if the ->gp_flags field indicates there is no need (as in
someone else did the wake-up before we got around to it), and no wake-ups
before the grace-period kthread has been created.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
[ Pranith: backport to 3.13-stable: just rcu_gp_kthread_wake(),
prereq for 2aa792e "rcu: Use rcu_gp_kthread_wake() to wake up grace
period kthreads" ]
Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>
Signed-off-by: Kamal Mostafa <kamal@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 799b601451 upstream.
Audit rules disappear when an inode they watch is evicted from the cache.
This is likely not what we want.
The guilty commit is "fsnotify: allow marks to not pin inodes in core",
which didn't take into account that audit_tree adds watches with a zero
mask.
Adding any mask should fix this.
Fixes: 90b1e7a578 ("fsnotify: allow marks to not pin inodes in core")
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Paul Moore <pmoore@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 897f1acbb6 upstream.
Add a space between subj= and feature= fields to make them parsable.
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <pmoore@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9ef9151477 upstream.
When an AUDIT_GET_FEATURE message is sent from userspace to the kernel, it
should reply with a message tagged as an AUDIT_GET_FEATURE type with a struct
audit_feature. The current reply is a message tagged as an AUDIT_GET
type with a struct audit_feature.
This appears to have been a cut-and-paste-eo in commit b0fed40.
Reported-by: Steve Grubb <sgrubb@redhat.com>
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f1e3a0932f upstream.
Probability of use-after-free isn't zero in this place.
Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20140922183636.11015.83611.stgit@localhost
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6891c4509c upstream.
If userland creates a timer without specifying a sigevent info, we'll
create one ourself, using a stack local variable. Particularly will we
use the timer ID as sival_int. But as sigev_value is a union containing
a pointer and an int, that assignment will only partially initialize
sigev_value on systems where the size of a pointer is bigger than the
size of an int. On such systems we'll copy the uninitialized stack bytes
from the timer_create() call to userland when the timer actually fires
and we're going to deliver the signal.
Initialize sigev_value with 0 to plug the stack info leak.
Found in the PaX patch, written by the PaX Team.
Fixes: 5a9fa73072 ("posix-timers: kill ->it_sigev_signo and...")
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Brad Spengler <spender@grsecurity.net>
Cc: PaX Team <pageexec@freemail.hu>
Link: http://lkml.kernel.org/r/1412456799-32339-1-git-send-email-minipli@googlemail.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 94fb823fcb upstream.
If a device's dev_pm_ops::freeze callback fails during the QUIESCE
phase, we don't rollback things correctly calling the thaw and complete
callbacks. This could leave some devices in a suspended state in case of
an error during resuming from hibernation.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 5695be142e upstream.
PM freezer relies on having all tasks frozen by the time devices are
getting frozen so that no task will touch them while they are getting
frozen. But OOM killer is allowed to kill an already frozen task in
order to handle OOM situtation. In order to protect from late wake ups
OOM killer is disabled after all tasks are frozen. This, however, still
keeps a window open when a killed task didn't manage to die by the time
freeze_processes finishes.
Reduce the race window by checking all tasks after OOM killer has been
disabled. This is still not race free completely unfortunately because
oom_killer_disable cannot stop an already ongoing OOM killer so a task
might still wake up from the fridge and get killed without
freeze_processes noticing. Full synchronization of OOM and freezer is,
however, too heavy weight for this highly unlikely case.
Introduce and check oom_kills counter which gets incremented early when
the allocator enters __alloc_pages_may_oom path and only check all the
tasks if the counter changes during the freezing attempt. The counter
is updated so early to reduce the race window since allocator checked
oom_killer_disabled which is set by PM-freezing code. A false positive
will push the PM-freezer into a slow path but that is not a big deal.
Changes since v1
- push the re-check loop out of freeze_processes into
check_frozen_processes and invert the condition to make the code more
readable as per Rafael
Fixes: f660daac47 (oom: thaw threads if oom killed thread is frozen before deferring)
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 51fae6da64 upstream.
Since f660daac47 (oom: thaw threads if oom killed thread is frozen
before deferring) OOM killer relies on being able to thaw a frozen task
to handle OOM situation but a3201227f8 (freezer: make freezing() test
freeze conditions in effect instead of TIF_FREEZE) has reorganized the
code and stopped clearing freeze flag in __thaw_task. This means that
the target task only wakes up and goes into the fridge again because the
freezing condition hasn't changed for it. This reintroduces the bug
fixed by f660daac47.
Fix the issue by checking for TIF_MEMDIE thread flag in
freezing_slow_path and exclude the task from freezing completely. If a
task was already frozen it would get woken by __thaw_task from OOM killer
and get out of freezer after rechecking freezing().
Changes since v1
- put TIF_MEMDIE check into freezing_slowpath rather than in __refrigerator
as per Oleg
- return __thaw_task into oom_scan_process_thread because
oom_kill_process will not wake task in the fridge because it is
sleeping uninterruptible
[mhocko@suse.cz: rewrote the changelog]
Fixes: a3201227f8 (freezer: make freezing() test freeze conditions in effect instead of TIF_FREEZE)
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d3051b489a upstream.
A panic was seen in the following sitation.
There are two threads running on the system. The first thread is a system
monitoring thread that is reading /proc/modules. The second thread is
loading and unloading a module (in this example I'm using my simple
dummy-module.ko). Note, in the "real world" this occurred with the qlogic
driver module.
When doing this, the following panic occurred:
------------[ cut here ]------------
kernel BUG at kernel/module.c:3739!
invalid opcode: 0000 [#1] SMP
Modules linked in: binfmt_misc sg nfsv3 rpcsec_gss_krb5 nfsv4 dns_resolver nfs fscache intel_powerclamp coretemp kvm_intel kvm crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel lrw igb gf128mul glue_helper iTCO_wdt iTCO_vendor_support ablk_helper ptp sb_edac cryptd pps_core edac_core shpchp i2c_i801 pcspkr wmi lpc_ich ioatdma mfd_core dca ipmi_si nfsd ipmi_msghandler auth_rpcgss nfs_acl lockd sunrpc xfs libcrc32c sr_mod cdrom sd_mod crc_t10dif crct10dif_common mgag200 syscopyarea sysfillrect sysimgblt i2c_algo_bit drm_kms_helper ttm isci drm libsas ahci libahci scsi_transport_sas libata i2c_core dm_mirror dm_region_hash dm_log dm_mod [last unloaded: dummy_module]
CPU: 37 PID: 186343 Comm: cat Tainted: GF O-------------- 3.10.0+ #7
Hardware name: Intel Corporation S2600CP/S2600CP, BIOS RMLSDP.86I.00.29.D696.1311111329 11/11/2013
task: ffff8807fd2d8000 ti: ffff88080fa7c000 task.ti: ffff88080fa7c000
RIP: 0010:[<ffffffff810d64c5>] [<ffffffff810d64c5>] module_flags+0xb5/0xc0
RSP: 0018:ffff88080fa7fe18 EFLAGS: 00010246
RAX: 0000000000000003 RBX: ffffffffa03b5200 RCX: 0000000000000000
RDX: 0000000000001000 RSI: ffff88080fa7fe38 RDI: ffffffffa03b5000
RBP: ffff88080fa7fe28 R08: 0000000000000010 R09: 0000000000000000
R10: 0000000000000000 R11: 000000000000000f R12: ffffffffa03b5000
R13: ffffffffa03b5008 R14: ffffffffa03b5200 R15: ffffffffa03b5000
FS: 00007f6ae57ef740(0000) GS:ffff88101e7a0000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000404f70 CR3: 0000000ffed48000 CR4: 00000000001407e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Stack:
ffffffffa03b5200 ffff8810101e4800 ffff88080fa7fe70 ffffffff810d666c
ffff88081e807300 000000002e0f2fbf 0000000000000000 ffff88100f257b00
ffffffffa03b5008 ffff88080fa7ff48 ffff8810101e4800 ffff88080fa7fee0
Call Trace:
[<ffffffff810d666c>] m_show+0x19c/0x1e0
[<ffffffff811e4d7e>] seq_read+0x16e/0x3b0
[<ffffffff812281ed>] proc_reg_read+0x3d/0x80
[<ffffffff811c0f2c>] vfs_read+0x9c/0x170
[<ffffffff811c1a58>] SyS_read+0x58/0xb0
[<ffffffff81605829>] system_call_fastpath+0x16/0x1b
Code: 48 63 c2 83 c2 01 c6 04 03 29 48 63 d2 eb d9 0f 1f 80 00 00 00 00 48 63 d2 c6 04 13 2d 41 8b 0c 24 8d 50 02 83 f9 01 75 b2 eb cb <0f> 0b 66 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 55 48 89 e5 41
RIP [<ffffffff810d64c5>] module_flags+0xb5/0xc0
RSP <ffff88080fa7fe18>
Consider the two processes running on the system.
CPU 0 (/proc/modules reader)
CPU 1 (loading/unloading module)
CPU 0 opens /proc/modules, and starts displaying data for each module by
traversing the modules list via fs/seq_file.c:seq_open() and
fs/seq_file.c:seq_read(). For each module in the modules list, seq_read
does
op->start() <-- this is a pointer to m_start()
op->show() <- this is a pointer to m_show()
op->stop() <-- this is a pointer to m_stop()
The m_start(), m_show(), and m_stop() module functions are defined in
kernel/module.c. The m_start() and m_stop() functions acquire and release
the module_mutex respectively.
ie) When reading /proc/modules, the module_mutex is acquired and released
for each module.
m_show() is called with the module_mutex held. It accesses the module
struct data and attempts to write out module data. It is in this code
path that the above BUG_ON() warning is encountered, specifically m_show()
calls
static char *module_flags(struct module *mod, char *buf)
{
int bx = 0;
BUG_ON(mod->state == MODULE_STATE_UNFORMED);
...
The other thread, CPU 1, in unloading the module calls the syscall
delete_module() defined in kernel/module.c. The module_mutex is acquired
for a short time, and then released. free_module() is called without the
module_mutex. free_module() then sets mod->state = MODULE_STATE_UNFORMED,
also without the module_mutex. Some additional code is called and then the
module_mutex is reacquired to remove the module from the modules list:
/* Now we can delete it from the lists */
mutex_lock(&module_mutex);
stop_machine(__unlink_module, mod, NULL);
mutex_unlock(&module_mutex);
This is the sequence of events that leads to the panic.
CPU 1 is removing dummy_module via delete_module(). It acquires the
module_mutex, and then releases it. CPU 1 has NOT set dummy_module->state to
MODULE_STATE_UNFORMED yet.
CPU 0, which is reading the /proc/modules, acquires the module_mutex and
acquires a pointer to the dummy_module which is still in the modules list.
CPU 0 calls m_show for dummy_module. The check in m_show() for
MODULE_STATE_UNFORMED passed for dummy_module even though it is being
torn down.
Meanwhile CPU 1, which has been continuing to remove dummy_module without
holding the module_mutex, now calls free_module() and sets
dummy_module->state to MODULE_STATE_UNFORMED.
CPU 0 now calls module_flags() with dummy_module and ...
static char *module_flags(struct module *mod, char *buf)
{
int bx = 0;
BUG_ON(mod->state == MODULE_STATE_UNFORMED);
and BOOM.
Acquire and release the module_mutex lock around the setting of
MODULE_STATE_UNFORMED in the teardown path, which should resolve the
problem.
Testing: In the unpatched kernel I can panic the system within 1 minute by
doing
while (true) do insmod dummy_module.ko; rmmod dummy_module.ko; done
and
while (true) do cat /proc/modules; done
in separate terminals.
In the patched kernel I was able to run just over one hour without seeing
any issues. I also verified the output of panic via sysrq-c and the output
of /proc/modules looks correct for all three states for the dummy_module.
dummy_module 12661 0 - Unloading 0xffffffffa03a5000 (OE-)
dummy_module 12661 0 - Live 0xffffffffa03bb000 (OE)
dummy_module 14015 1 - Loading 0xffffffffa03a5000 (OE+)
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 66339c31bc upstream.
dl_bw_of() dereferences rq->rd which has to have RCU read lock held.
Probability of use-after-free isn't zero here.
Also add lockdep assert into dl_bw_cpus().
Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20140922183624.11015.71558.stgit@localhost
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 086ba77a6d upstream.
ARM has some private syscalls (for example, set_tls(2)) which lie
outside the range of NR_syscalls. If any of these are called while
syscall tracing is being performed, out-of-bounds array access will
occur in the ftrace and perf sys_{enter,exit} handlers.
# trace-cmd record -e raw_syscalls:* true && trace-cmd report
...
true-653 [000] 384.675777: sys_enter: NR 192 (0, 1000, 3, 4000022, ffffffff, 0)
true-653 [000] 384.675812: sys_exit: NR 192 = 1995915264
true-653 [000] 384.675971: sys_enter: NR 983045 (76f74480, 76f74000, 76f74b28, 76f74480, 76f76f74, 1)
true-653 [000] 384.675988: sys_exit: NR 983045 = 0
...
# trace-cmd record -e syscalls:* true
[ 17.289329] Unable to handle kernel paging request at virtual address aaaaaace
[ 17.289590] pgd = 9e71c000
[ 17.289696] [aaaaaace] *pgd=00000000
[ 17.289985] Internal error: Oops: 5 [#1] PREEMPT SMP ARM
[ 17.290169] Modules linked in:
[ 17.290391] CPU: 0 PID: 704 Comm: true Not tainted 3.18.0-rc2+ #21
[ 17.290585] task: 9f4dab00 ti: 9e710000 task.ti: 9e710000
[ 17.290747] PC is at ftrace_syscall_enter+0x48/0x1f8
[ 17.290866] LR is at syscall_trace_enter+0x124/0x184
Fix this by ignoring out-of-NR_syscalls-bounds syscall numbers.
Commit cd0980fc8a "tracing: Check invalid syscall nr while tracing syscalls"
added the check for less than zero, but it should have also checked
for greater than NR_syscalls.
Link: http://lkml.kernel.org/p/1414620418-29472-1-git-send-email-rabin@rab.in
Fixes: cd0980fc8a "tracing: Check invalid syscall nr while tracing syscalls"
Signed-off-by: Rabin Vincent <rabin@rab.in>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 76835b0ebf upstream.
Commit b0c29f79ec (futexes: Avoid taking the hb->lock if there's
nothing to wake up) changes the futex code to avoid taking a lock when
there are no waiters. This code has been subsequently fixed in commit
11d4616bd0 (futex: revert back to the explicit waiter counting code).
Both the original commit and the fix-up rely on get_futex_key_refs() to
always imply a barrier.
However, for private futexes, none of the cases in the switch statement
of get_futex_key_refs() would be hit and the function completes without
a memory barrier as required before checking the "waiters" in
futex_wake() -> hb_waiters_pending(). The consequence is a race with a
thread waiting on a futex on another CPU, allowing the waker thread to
read "waiters == 0" while the waiter thread to have read "futex_val ==
locked" (in kernel).
Without this fix, the problem (user space deadlocks) can be seen with
Android bionic's mutex implementation on an arm64 multi-cluster system.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Reported-by: Matteo Franchin <Matteo.Franchin@arm.com>
Fixes: b0c29f79ec (futexes: Avoid taking the hb->lock if there's nothing to wake up)
Acked-by: Davidlohr Bueso <dave@stgolabs.net>
Tested-by: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Darren Hart <dvhart@linux.intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 615d6e8756 upstream.
This patch is a continuation of efforts trying to optimize find_vma(),
avoiding potentially expensive rbtree walks to locate a vma upon faults.
The original approach (https://lkml.org/lkml/2013/11/1/410), where the
largest vma was also cached, ended up being too specific and random,
thus further comparison with other approaches were needed. There are
two things to consider when dealing with this, the cache hit rate and
the latency of find_vma(). Improving the hit-rate does not necessarily
translate in finding the vma any faster, as the overhead of any fancy
caching schemes can be too high to consider.
We currently cache the last used vma for the whole address space, which
provides a nice optimization, reducing the total cycles in find_vma() by
up to 250%, for workloads with good locality. On the other hand, this
simple scheme is pretty much useless for workloads with poor locality.
Analyzing ebizzy runs shows that, no matter how many threads are
running, the mmap_cache hit rate is less than 2%, and in many situations
below 1%.
The proposed approach is to replace this scheme with a small per-thread
cache, maximizing hit rates at a very low maintenance cost.
Invalidations are performed by simply bumping up a 32-bit sequence
number. The only expensive operation is in the rare case of a seq
number overflow, where all caches that share the same address space are
flushed. Upon a miss, the proposed replacement policy is based on the
page number that contains the virtual address in question. Concretely,
the following results are seen on an 80 core, 8 socket x86-64 box:
1) System bootup: Most programs are single threaded, so the per-thread
scheme does improve ~50% hit rate by just adding a few more slots to
the cache.
+----------------+----------+------------------+
| caching scheme | hit-rate | cycles (billion) |
+----------------+----------+------------------+
| baseline | 50.61% | 19.90 |
| patched | 73.45% | 13.58 |
+----------------+----------+------------------+
2) Kernel build: This one is already pretty good with the current
approach as we're dealing with good locality.
+----------------+----------+------------------+
| caching scheme | hit-rate | cycles (billion) |
+----------------+----------+------------------+
| baseline | 75.28% | 11.03 |
| patched | 88.09% | 9.31 |
+----------------+----------+------------------+
3) Oracle 11g Data Mining (4k pages): Similar to the kernel build workload.
+----------------+----------+------------------+
| caching scheme | hit-rate | cycles (billion) |
+----------------+----------+------------------+
| baseline | 70.66% | 17.14 |
| patched | 91.15% | 12.57 |
+----------------+----------+------------------+
4) Ebizzy: There's a fair amount of variation from run to run, but this
approach always shows nearly perfect hit rates, while baseline is just
about non-existent. The amounts of cycles can fluctuate between
anywhere from ~60 to ~116 for the baseline scheme, but this approach
reduces it considerably. For instance, with 80 threads:
+----------------+----------+------------------+
| caching scheme | hit-rate | cycles (billion) |
+----------------+----------+------------------+
| baseline | 1.06% | 91.54 |
| patched | 99.97% | 14.18 |
+----------------+----------+------------------+
[akpm@linux-foundation.org: fix nommu build, per Davidlohr]
[akpm@linux-foundation.org: document vmacache_valid() logic]
[akpm@linux-foundation.org: attempt to untangle header files]
[akpm@linux-foundation.org: add vmacache_find() BUG_ON]
[hughd@google.com: add vmacache_valid_mm() (from Oleg)]
[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: adjust and enhance comments]
Signed-off-by: Davidlohr Bueso <davidlohr@hp.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Reviewed-by: Michel Lespinasse <walken@google.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Tested-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d26914d117 upstream.
Since put_mems_allowed() is strictly optional, its a seqcount retry, we
don't need to evaluate the function if the allocation was in fact
successful, saving a smp_rmb some loads and comparisons on some relative
fast-paths.
Since the naming, get/put_mems_allowed() does suggest a mandatory
pairing, rename the interface, as suggested by Mel, to resemble the
seqcount interface.
This gives us: read_mems_allowed_begin() and read_mems_allowed_retry(),
where it is important to note that the return value of the latter call
is inverted from its previous incarnation.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d78c9300c5 upstream.
timeval_to_jiffies tried to round a timeval up to an integral number
of jiffies, but the logic for doing so was incorrect: intervals
corresponding to exactly N jiffies would become N+1. This manifested
itself particularly repeatedly stopping/starting an itimer:
setitimer(ITIMER_PROF, &val, NULL);
setitimer(ITIMER_PROF, NULL, &val);
would add a full tick to val, _even if it was exactly representable in
terms of jiffies_ (say, the result of a previous rounding.) Doing
this repeatedly would cause unbounded growth in val. So fix the math.
Here's what was wrong with the conversion: we essentially computed
(eliding seconds)
jiffies = usec * (NSEC_PER_USEC/TICK_NSEC)
by using scaling arithmetic, which took the best approximation of
NSEC_PER_USEC/TICK_NSEC with denominator of 2^USEC_JIFFIE_SC =
x/(2^USEC_JIFFIE_SC), and computed:
jiffies = (usec * x) >> USEC_JIFFIE_SC
and rounded this calculation up in the intermediate form (since we
can't necessarily exactly represent TICK_NSEC in usec.) But the
scaling arithmetic is a (very slight) *over*approximation of the true
value; that is, instead of dividing by (1 usec/ 1 jiffie), we
effectively divided by (1 usec/1 jiffie)-epsilon (rounding
down). This would normally be fine, but we want to round timeouts up,
and we did so by adding 2^USEC_JIFFIE_SC - 1 before the shift; this
would be fine if our division was exact, but dividing this by the
slightly smaller factor was equivalent to adding just _over_ 1 to the
final result (instead of just _under_ 1, as desired.)
In particular, with HZ=1000, we consistently computed that 10000 usec
was 11 jiffies; the same was true for any exact multiple of
TICK_NSEC.
We could possibly still round in the intermediate form, adding
something less than 2^USEC_JIFFIE_SC - 1, but easier still is to
convert usec->nsec, round in nanoseconds, and then convert using
time*spec*_to_jiffies. This adds one constant multiplication, and is
not observably slower in microbenchmarks on recent x86 hardware.
Tested: the following program:
int main() {
struct itimerval zero = {{0, 0}, {0, 0}};
/* Initially set to 10 ms. */
struct itimerval initial = zero;
initial.it_interval.tv_usec = 10000;
setitimer(ITIMER_PROF, &initial, NULL);
/* Save and restore several times. */
for (size_t i = 0; i < 10; ++i) {
struct itimerval prev;
setitimer(ITIMER_PROF, &zero, &prev);
/* on old kernels, this goes up by TICK_USEC every iteration */
printf("previous value: %ld %ld %ld %ld\n",
prev.it_interval.tv_sec, prev.it_interval.tv_usec,
prev.it_value.tv_sec, prev.it_value.tv_usec);
setitimer(ITIMER_PROF, &prev, NULL);
}
return 0;
}
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Paul Turner <pjt@google.com>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Reviewed-by: Paul Turner <pjt@google.com>
Reported-by: Aaron Jacobs <jacobsa@google.com>
Signed-off-by: Andrew Hunter <ahh@google.com>
[jstultz: Tweaked to apply to 3.17-rc]
Signed-off-by: John Stultz <john.stultz@linaro.org>
[bwh: Backported to 3.16: adjust filename]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 24607f114f upstream.
Commit 651e22f270 "ring-buffer: Always reset iterator to reader page"
fixed one bug but in the process caused another one. The reset is to
update the header page, but that fix also changed the way the cached
reads were updated. The cache reads are used to test if an iterator
needs to be updated or not.
A ring buffer iterator, when created, disables writes to the ring buffer
but does not stop other readers or consuming reads from happening.
Although all readers are synchronized via a lock, they are only
synchronized when in the ring buffer functions. Those functions may
be called by any number of readers. The iterator continues down when
its not interrupted by a consuming reader. If a consuming read
occurs, the iterator starts from the beginning of the buffer.
The way the iterator sees that a consuming read has happened since
its last read is by checking the reader "cache". The cache holds the
last counts of the read and the reader page itself.
Commit 651e22f270 changed what was saved by the cache_read when
the rb_iter_reset() occurred, making the iterator never match the cache.
Then if the iterator calls rb_iter_reset(), it will go into an
infinite loop by checking if the cache doesn't match, doing the reset
and retrying, just to see that the cache still doesn't match! Which
should never happen as the reset is suppose to set the cache to the
current value and there's locks that keep a consuming reader from
having access to the data.
Fixes: 651e22f270 "ring-buffer: Always reset iterator to reader page"
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6c72e3501d upstream.
Oleg noticed that a cleanup by Sylvain actually uncovered a bug; by
calling perf_event_free_task() when failing sched_fork() we will not yet
have done the memset() on ->perf_event_ctxp[] and will therefore try and
'free' the inherited contexts, which are still in use by the parent
process. This is bad..
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Reported-by: Oleg Nesterov <oleg@redhat.com>
Reported-by: Sylvain 'ythier' Hitier <sylvain.hitier@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 43e8317b0b upstream.
Use the observation that, for platform-dependent sleep states
(PM_SUSPEND_STANDBY, PM_SUSPEND_MEM), a given state is either
always supported or always unsupported and store that information
in pm_states[] instead of calling valid_state() every time we
need to check it.
Also do not use valid_state() for PM_SUSPEND_FREEZE, which is always
valid, and move the pm_test_level validity check for PM_SUSPEND_FREEZE
directly into enter_state().
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 27ddcc6596 upstream.
To allow sleep states corresponding to the "mem", "standby" and
"freeze" lables to be different from the pm_states[] indexes of
those strings, introduce struct pm_sleep_state, consisting of
a string label and a state number, and turn pm_states[] into an
array of objects of that type.
This modification should not lead to any functional changes.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3577af70a2 upstream.
We saw a kernel soft lockup in perf_remove_from_context(),
it looks like the `perf` process, when exiting, could not go
out of the retry loop. Meanwhile, the target process was forking
a child. So either the target process should execute the smp
function call to deactive the event (if it was running) or it should
do a context switch which deactives the event.
It seems we optimize out a context switch in perf_event_context_sched_out(),
and what's more important, we still test an obsolete task pointer when
retrying, so no one actually would deactive that event in this situation.
Fix it directly by reloading the task pointer in perf_remove_from_context().
This should cure the above soft lockup.
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1409696840-843-1-git-send-email-xiyou.wangcong@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 474e941bed upstream.
Locks the k_itimer's it_lock member when handling the alarm timer's
expiry callback.
The regular posix timers defined in posix-timers.c have this lock held
during timout processing because their callbacks are routed through
posix_timer_fn(). The alarm timers follow a different path, so they
ought to grab the lock somewhere else.
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Sharvil Nanavati <sharvil@google.com>
Signed-off-by: Richard Larocque <rlarocque@google.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 265b81d23a upstream.
Avoids sending a signal to alarm timers created with sigev_notify set to
SIGEV_NONE by checking for that special case in the timeout callback.
The regular posix timers avoid sending signals to SIGEV_NONE timers by
not scheduling any callbacks for them in the first place. Although it
would be possible to do something similar for alarm timers, it's simpler
to handle this as a special case in the timeout.
Prior to this patch, the alarm timer would ignore the sigev_notify value
and try to deliver signals to the process anyway. Even worse, the
sanity check for the value of sigev_signo is skipped when SIGEV_NONE was
specified, so the signal number could be bogus. If sigev_signo was an
unitialized value (as it often would be if SIGEV_NONE is used), then
it's hard to predict which signal will be sent.
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Sharvil Nanavati <sharvil@google.com>
Signed-off-by: Richard Larocque <rlarocque@google.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e86fea7649 upstream.
Returns the time remaining for an alarm timer, rather than the time at
which it is scheduled to expire. If the timer has already expired or it
is not currently scheduled, the it_value's members are set to zero.
This new behavior matches that of the other posix-timers and the POSIX
specifications.
This is a change in user-visible behavior, and may break existing
applications. Hopefully, few users rely on the old incorrect behavior.
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Sharvil Nanavati <sharvil@google.com>
Signed-off-by: Richard Larocque <rlarocque@google.com>
[jstultz: minor style tweak]
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit acbbe6fbb2 upstream.
The C operator <= defines a perfectly fine total ordering on the set of
values representable in a long. However, unlike its namesake in the
integers, it is not translation invariant, meaning that we do not have
"b <= c" iff "a+b <= a+c" for all a,b,c.
This means that it is always wrong to try to boil down the relationship
between two longs to a question about the sign of their difference,
because the resulting relation [a LEQ b iff a-b <= 0] is neither
anti-symmetric or transitive. The former is due to -LONG_MIN==LONG_MIN
(take any two a,b with a-b = LONG_MIN; then a LEQ b and b LEQ a, but a !=
b). The latter can either be seen observing that x LEQ x+1 for all x,
implying x LEQ x+1 LEQ x+2 ... LEQ x-1 LEQ x; or more directly with the
simple example a=LONG_MIN, b=0, c=1, for which a-b < 0, b-c < 0, but a-c >
0.
Note that it makes absolutely no difference that a transmogrying bijection
has been applied before the comparison is done. In fact, had the
obfuscation not been done, one could probably not observe the bug
(assuming all values being compared always lie in one half of the address
space, the mathematical value of a-b is always representable in a long).
As it stands, one can easily obtain three file descriptors exhibiting the
non-transitivity of kcmp().
Side note 1: I can't see that ensuring the MSB of the multiplier is
set serves any purpose other than obfuscating the obfuscating code.
Side note 2:
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <fcntl.h>
#include <unistd.h>
#include <assert.h>
#include <sys/syscall.h>
enum kcmp_type {
KCMP_FILE,
KCMP_VM,
KCMP_FILES,
KCMP_FS,
KCMP_SIGHAND,
KCMP_IO,
KCMP_SYSVSEM,
KCMP_TYPES,
};
pid_t pid;
int kcmp(pid_t pid1, pid_t pid2, int type,
unsigned long idx1, unsigned long idx2)
{
return syscall(SYS_kcmp, pid1, pid2, type, idx1, idx2);
}
int cmp_fd(int fd1, int fd2)
{
int c = kcmp(pid, pid, KCMP_FILE, fd1, fd2);
if (c < 0) {
perror("kcmp");
exit(1);
}
assert(0 <= c && c < 3);
return c;
}
int cmp_fdp(const void *a, const void *b)
{
static const int normalize[] = {0, -1, 1};
return normalize[cmp_fd(*(int*)a, *(int*)b)];
}
#define MAX 100 /* This is plenty; I've seen it trigger for MAX==3 */
int main(int argc, char *argv[])
{
int r, s, count = 0;
int REL[3] = {0,0,0};
int fd[MAX];
pid = getpid();
while (count < MAX) {
r = open("/dev/null", O_RDONLY);
if (r < 0)
break;
fd[count++] = r;
}
printf("opened %d file descriptors\n", count);
for (r = 0; r < count; ++r) {
for (s = r+1; s < count; ++s) {
REL[cmp_fd(fd[r], fd[s])]++;
}
}
printf("== %d\t< %d\t> %d\n", REL[0], REL[1], REL[2]);
qsort(fd, count, sizeof(fd[0]), cmp_fdp);
memset(REL, 0, sizeof(REL));
for (r = 0; r < count; ++r) {
for (s = r+1; s < count; ++s) {
REL[cmp_fd(fd[r], fd[s])]++;
}
}
printf("== %d\t< %d\t> %d\n", REL[0], REL[1], REL[2]);
return (REL[0] + REL[2] != 0);
}
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Reviewed-by: Cyrill Gorcunov <gorcunov@openvz.org>
"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 13c42c2f43 upstream.
futex_wait_requeue_pi() calls futex_wait_setup(). If
futex_wait_setup() succeeds it returns with hb->lock held and
preemption disabled. Now the sanity check after this does:
if (match_futex(&q.key, &key2)) {
ret = -EINVAL;
goto out_put_keys;
}
which releases the keys but does not release hb->lock.
So we happily return to user space with hb->lock held and therefor
preemption disabled.
Unlock hb->lock before taking the exit route.
Reported-by: Dave "Trinity" Jones <davej@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/alpine.DEB.2.10.1409112318500.4178@nanos
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit eb4aec84d6 upstream.
cgroup_pidlist_start() holds cgrp->pidlist_mutex and then calls
pidlist_array_load(), and cgroup_pidlist_stop() releases the mutex.
It is wrong that we release the mutex in the failure path in
pidlist_array_load(), because cgroup_pidlist_stop() will be called
no matter if cgroup_pidlist_start() returns errno or not.
Fixes: 4bac00d16a
Signed-off-by: Zefan Li <lizefan@huawei.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 4ce97dbf50 upstream.
Epoll on trace_pipe can sometimes hang in a weird case. If the ring buffer is
empty when we set waiters_pending but an event shows up exactly at that moment
we can miss being woken up by the ring buffers irq work. Since
ring_buffer_empty() is inherently racey we will sometimes think that the buffer
is not empty. So we don't get woken up and we don't think there are any events
even though there were some ready when we added the watch, which makes us hang.
This patch fixes this by making sure that we are actually on the wait list
before we set waiters_pending, and add a memory barrier to make sure
ring_buffer_empty() is going to be correct.
Link: http://lkml.kernel.org/p/1408989581-23727-1-git-send-email-jbacik@fb.com
Cc: Martin Lau <kafai@fb.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 021de3d904 upstream.
After writting a test to try to trigger the bug that caused the
ring buffer iterator to become corrupted, I hit another bug:
WARNING: CPU: 1 PID: 5281 at kernel/trace/ring_buffer.c:3766 rb_iter_peek+0x113/0x238()
Modules linked in: ipt_MASQUERADE sunrpc [...]
CPU: 1 PID: 5281 Comm: grep Tainted: G W 3.16.0-rc3-test+ #143
Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./To be filled by O.E.M., BIOS SDBLI944.86P 05/08/2007
0000000000000000 ffffffff81809a80 ffffffff81503fb0 0000000000000000
ffffffff81040ca1 ffff8800796d6010 ffffffff810c138d ffff8800796d6010
ffff880077438c80 ffff8800796d6010 ffff88007abbe600 0000000000000003
Call Trace:
[<ffffffff81503fb0>] ? dump_stack+0x4a/0x75
[<ffffffff81040ca1>] ? warn_slowpath_common+0x7e/0x97
[<ffffffff810c138d>] ? rb_iter_peek+0x113/0x238
[<ffffffff810c138d>] ? rb_iter_peek+0x113/0x238
[<ffffffff810c14df>] ? ring_buffer_iter_peek+0x2d/0x5c
[<ffffffff810c6f73>] ? tracing_iter_reset+0x6e/0x96
[<ffffffff810c74a3>] ? s_start+0xd7/0x17b
[<ffffffff8112b13e>] ? kmem_cache_alloc_trace+0xda/0xea
[<ffffffff8114cf94>] ? seq_read+0x148/0x361
[<ffffffff81132d98>] ? vfs_read+0x93/0xf1
[<ffffffff81132f1b>] ? SyS_read+0x60/0x8e
[<ffffffff8150bf9f>] ? tracesys+0xdd/0xe2
Debugging this bug, which triggers when the rb_iter_peek() loops too
many times (more than 2 times), I discovered there's a case that can
cause that function to legitimately loop 3 times!
rb_iter_peek() is different than rb_buffer_peek() as the rb_buffer_peek()
only deals with the reader page (it's for consuming reads). The
rb_iter_peek() is for traversing the buffer without consuming it, and as
such, it can loop for one more reason. That is, if we hit the end of
the reader page or any page, it will go to the next page and try again.
That is, we have this:
1. iter->head > iter->head_page->page->commit
(rb_inc_iter() which moves the iter to the next page)
try again
2. event = rb_iter_head_event()
event->type_len == RINGBUF_TYPE_TIME_EXTEND
rb_advance_iter()
try again
3. read the event.
But we never get to 3, because the count is greater than 2 and we
cause the WARNING and return NULL.
Up the counter to 3.
Fixes: 69d1b839f7 "ring-buffer: Bind time extend and data events together"
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 651e22f270 upstream.
When performing a consuming read, the ring buffer swaps out a
page from the ring buffer with a empty page and this page that
was swapped out becomes the new reader page. The reader page
is owned by the reader and since it was swapped out of the ring
buffer, writers do not have access to it (there's an exception
to that rule, but it's out of scope for this commit).
When reading the "trace" file, it is a non consuming read, which
means that the data in the ring buffer will not be modified.
When the trace file is opened, a ring buffer iterator is allocated
and writes to the ring buffer are disabled, such that the iterator
will not have issues iterating over the data.
Although the ring buffer disabled writes, it does not disable other
reads, or even consuming reads. If a consuming read happens, then
the iterator is reset and starts reading from the beginning again.
My tests would sometimes trigger this bug on my i386 box:
WARNING: CPU: 0 PID: 5175 at kernel/trace/trace.c:1527 __trace_find_cmdline+0x66/0xaa()
Modules linked in:
CPU: 0 PID: 5175 Comm: grep Not tainted 3.16.0-rc3-test+ #8
Hardware name: /DG965MQ, BIOS MQ96510J.86A.0372.2006.0605.1717 06/05/2006
00000000 00000000 f09c9e1c c18796b3 c1b5d74c f09c9e4c c103a0e3 c1b5154b
f09c9e78 00001437 c1b5d74c 000005f7 c10bd85a c10bd85a c1cac57c f09c9eb0
ed0e0000 f09c9e64 c103a185 00000009 f09c9e5c c1b5154b f09c9e78 f09c9e80^M
Call Trace:
[<c18796b3>] dump_stack+0x4b/0x75
[<c103a0e3>] warn_slowpath_common+0x7e/0x95
[<c10bd85a>] ? __trace_find_cmdline+0x66/0xaa
[<c10bd85a>] ? __trace_find_cmdline+0x66/0xaa
[<c103a185>] warn_slowpath_fmt+0x33/0x35
[<c10bd85a>] __trace_find_cmdline+0x66/0xaa^M
[<c10bed04>] trace_find_cmdline+0x40/0x64
[<c10c3c16>] trace_print_context+0x27/0xec
[<c10c4360>] ? trace_seq_printf+0x37/0x5b
[<c10c0b15>] print_trace_line+0x319/0x39b
[<c10ba3fb>] ? ring_buffer_read+0x47/0x50
[<c10c13b1>] s_show+0x192/0x1ab
[<c10bfd9a>] ? s_next+0x5a/0x7c
[<c112e76e>] seq_read+0x267/0x34c
[<c1115a25>] vfs_read+0x8c/0xef
[<c112e507>] ? seq_lseek+0x154/0x154
[<c1115ba2>] SyS_read+0x54/0x7f
[<c188488e>] syscall_call+0x7/0xb
---[ end trace 3f507febd6b4cc83 ]---
>>>> ##### CPU 1 buffer started ####
Which was the __trace_find_cmdline() function complaining about the pid
in the event record being negative.
After adding more test cases, this would trigger more often. Strangely
enough, it would never trigger on a single test, but instead would trigger
only when running all the tests. I believe that was the case because it
required one of the tests to be shutting down via delayed instances while
a new test started up.
After spending several days debugging this, I found that it was caused by
the iterator becoming corrupted. Debugging further, I found out why
the iterator became corrupted. It happened with the rb_iter_reset().
As consuming reads may not read the full reader page, and only part
of it, there's a "read" field to know where the last read took place.
The iterator, must also start at the read position. In the rb_iter_reset()
code, if the reader page was disconnected from the ring buffer, the iterator
would start at the head page within the ring buffer (where writes still
happen). But the mistake there was that it still used the "read" field
to start the iterator on the head page, where it should always start
at zero because readers never read from within the ring buffer where
writes occur.
I originally wrote a patch to have it set the iter->head to 0 instead
of iter->head_page->read, but then I questioned why it wasn't always
setting the iter to point to the reader page, as the reader page is
still valid. The list_empty(reader_page->list) just means that it was
successful in swapping out. But the reader_page may still have data.
There was a bug report a long time ago that was not reproducible that
had something about trace_pipe (consuming read) not matching trace
(iterator read). This may explain why that happened.
Anyway, the correct answer to this bug is to always use the reader page
an not reset the iterator to inside the writable ring buffer.
Fixes: d769041f86 "ring_buffer: implement new locking"
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Thomas Gleixner
Jiri Olsa
Luca Abeni
Oleg Nesterov
Richard Guy Briggs
Eric W. Biederman
Andy Lutomirski
Pawel Moll
Pranith Kumar
Paul E. McKenney
Miklos Szeredi
Kirill Tkhai
Mathias Krause
Imre Deak
Michal Hocko
Cong Wang
Prarit Bhargava
Rabin Vincent
Catalin Marinas
Davidlohr Bueso
Mel Gorman
Andrew Hunter
Steven Rostedt (Red Hat)
Peter Zijlstra
Rafael J. Wysocki
Cong Wang
Richard Larocque
Rasmus Villemoes
Zefan Li
Josef Bacik
Sasha Levin