commit 4fdd595e4f upstream.
A while ago, when the "trace" file was opened, tracing was stopped, and
code was added to stop recording the comms to saved_cmdlines, for mapping
of the pids to the task name.
Code has been added that only records the comm if a trace event occurred,
and there's no reason to not trace it if the trace file is opened.
Cc: stable@vger.kernel.org
Fixes: 7ffbd48d5c ("tracing: Cache comms only after an event occurred")
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 85550c83da upstream.
The saved_cmdlines is used to map pids to the task name, such that the
output of the tracing does not just show pids, but also gives a human
readable name for the task.
If the name is not mapped, the output looks like this:
<...>-1316 [005] ...2 132.044039: ...
Instead of this:
gnome-shell-1316 [005] ...2 132.044039: ...
The names are updated when tracing is running, but are skipped if tracing
is stopped. Unfortunately, this stops the recording of the names if the
top level tracer is stopped, and not if there's other tracers active.
The recording of a name only happens when a new event is written into a
ring buffer, so there is no need to test if tracing is on or not. If
tracing is off, then no event is written and no need to test if tracing is
off or not.
Remove the check, as it hides the names of tasks for events in the
instance buffers.
Cc: stable@vger.kernel.org
Fixes: 7ffbd48d5c ("tracing: Cache comms only after an event occurred")
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 89529d8b8f upstream.
The trace_clock_global() tries to make sure the events between CPUs is
somewhat in order. A global value is used and updated by the latest read
of a clock. If one CPU is ahead by a little, and is read by another CPU, a
lock is taken, and if the timestamp of the other CPU is behind, it will
simply use the other CPUs timestamp.
The lock is also only taken with a "trylock" due to tracing, and strange
recursions can happen. The lock is not taken at all in NMI context.
In the case where the lock is not able to be taken, the non synced
timestamp is returned. But it will not be less than the saved global
timestamp.
The problem arises because when the time goes "backwards" the time
returned is the saved timestamp plus 1. If the lock is not taken, and the
plus one to the timestamp is returned, there's a small race that can cause
the time to go backwards!
CPU0 CPU1
---- ----
trace_clock_global() {
ts = clock() [ 1000 ]
trylock(clock_lock) [ success ]
global_ts = ts; [ 1000 ]
<interrupted by NMI>
trace_clock_global() {
ts = clock() [ 999 ]
if (ts < global_ts)
ts = global_ts + 1 [ 1001 ]
trylock(clock_lock) [ fail ]
return ts [ 1001]
}
unlock(clock_lock);
return ts; [ 1000 ]
}
trace_clock_global() {
ts = clock() [ 1000 ]
if (ts < global_ts) [ false 1000 == 1000 ]
trylock(clock_lock) [ success ]
global_ts = ts; [ 1000 ]
unlock(clock_lock)
return ts; [ 1000 ]
}
The above case shows to reads of trace_clock_global() on the same CPU, but
the second read returns one less than the first read. That is, time when
backwards, and this is not what is allowed by trace_clock_global().
This was triggered by heavy tracing and the ring buffer checker that tests
for the clock going backwards:
Ring buffer clock went backwards: 20613921464 -> 20613921463
------------[ cut here ]------------
WARNING: CPU: 2 PID: 0 at kernel/trace/ring_buffer.c:3412 check_buffer+0x1b9/0x1c0
Modules linked in:
[..]
[CPU: 2]TIME DOES NOT MATCH expected:20620711698 actual:20620711697 delta:6790234 before:20613921463 after:20613921463
[20613915818] PAGE TIME STAMP
[20613915818] delta:0
[20613915819] delta:1
[20613916035] delta:216
[20613916465] delta:430
[20613916575] delta:110
[20613916749] delta:174
[20613917248] delta:499
[20613917333] delta:85
[20613917775] delta:442
[20613917921] delta:146
[20613918321] delta:400
[20613918568] delta:247
[20613918768] delta:200
[20613919306] delta:538
[20613919353] delta:47
[20613919980] delta:627
[20613920296] delta:316
[20613920571] delta:275
[20613920862] delta:291
[20613921152] delta:290
[20613921464] delta:312
[20613921464] delta:0 TIME EXTEND
[20613921464] delta:0
This happened more than once, and always for an off by one result. It also
started happening after commit aafe104aa9 was added.
Cc: stable@vger.kernel.org
Fixes: aafe104aa9 ("tracing: Restructure trace_clock_global() to never block")
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6c14133d2d upstream.
It was reported that a bug on arm64 caused a bad ip address to be used for
updating into a nop in ftrace_init(), but the error path (rightfully)
returned -EINVAL and not -EFAULT, as the bug caused more than one error to
occur. But because -EINVAL was returned, the ftrace_bug() tried to report
what was at the location of the ip address, and read it directly. This
caused the machine to panic, as the ip was not pointing to a valid memory
address.
Instead, read the ip address with copy_from_kernel_nofault() to safely
access the memory, and if it faults, report that the address faulted,
otherwise report what was in that location.
Link: https://lore.kernel.org/lkml/20210607032329.28671-1-mark-pk.tsai@mediatek.com/
Cc: stable@vger.kernel.org
Fixes: 05736a427f ("ftrace: warn on failure to disable mcount callers")
Reported-by: Mark-PK Tsai <mark-pk.tsai@mediatek.com>
Tested-by: Mark-PK Tsai <mark-pk.tsai@mediatek.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6c605f8371 upstream.
KCSAN reports a data race between increment and decrement of pin_count:
write to 0xffff888237c2d4e0 of 4 bytes by task 15740 on cpu 1:
find_get_context kernel/events/core.c:4617
__do_sys_perf_event_open kernel/events/core.c:12097 [inline]
__se_sys_perf_event_open kernel/events/core.c:11933
...
read to 0xffff888237c2d4e0 of 4 bytes by task 15743 on cpu 0:
perf_unpin_context kernel/events/core.c:1525 [inline]
__do_sys_perf_event_open kernel/events/core.c:12328 [inline]
__se_sys_perf_event_open kernel/events/core.c:11933
...
Because neither read-modify-write here is atomic, this can lead to one
of the operations being lost, resulting in an inconsistent pin_count.
Fix it by adding the missing locking in the CPU-event case.
Fixes: fe4b04fa31 ("perf: Cure task_oncpu_function_call() races")
Reported-by: syzbot+142c9018f5962db69c7e@syzkaller.appspotmail.com
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20210527104711.2671610-1-elver@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 940d71c646 ]
If VCPU is suspended (VM suspend) in wq_watchdog_timer_fn() then
once this VCPU resumes it will see the new jiffies value, while it
may take a while before IRQ detects PVCLOCK_GUEST_STOPPED on this
VCPU and updates all the watchdogs via pvclock_touch_watchdogs().
There is a small chance of misreported WQ stalls in the meantime,
because new jiffies is time_after() old 'ts + thresh'.
wq_watchdog_timer_fn()
{
for_each_pool(pool, pi) {
if (time_after(jiffies, ts + thresh)) {
pr_emerg("BUG: workqueue lockup - pool");
}
}
}
Save jiffies at the beginning of this function and use that value
for stall detection. If VM gets suspended then we continue using
"old" jiffies value and old WQ touch timestamps. If IRQ at some
point restarts the stall detection cycle (pvclock_touch_watchdogs())
then old jiffies will always be before new 'ts + thresh'.
Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 45e1ba4083 ]
This patch effectively reverts the commit a3e72739b7 ("cgroup: fix
too early usage of static_branch_disable()"). The commit 6041186a32
("init: initialize jump labels before command line option parsing") has
moved the jump_label_init() before parse_args() which has made the
commit a3e72739b7 unnecessary. On the other hand there are
consequences of disabling the controllers later as there are subsystems
doing the controller checks for different decisions. One such incident
is reported [1] regarding the memory controller and its impact on memory
reclaim code.
[1] https://lore.kernel.org/linux-mm/921e53f3-4b13-aab8-4a9e-e83ff15371e4@nec.com
Signed-off-by: Shakeel Butt <shakeelb@google.com>
Reported-by: NOMURA JUNICHI(野村 淳一) <junichi.nomura@nec.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Tested-by: Jun'ichi Nomura <junichi.nomura@nec.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit dbb5afad10 ]
Suppose we have 2 threads, the group-leader L and a sub-theread T,
both parked in ptrace_stop(). Debugger tries to resume both threads
and does
ptrace(PTRACE_CONT, T);
ptrace(PTRACE_CONT, L);
If the sub-thread T execs in between, the 2nd PTRACE_CONT doesn not
resume the old leader L, it resumes the post-exec thread T which was
actually now stopped in PTHREAD_EVENT_EXEC. In this case the
PTHREAD_EVENT_EXEC event is lost, and the tracer can't know that the
tracee changed its pid.
This patch makes ptrace() fail in this case until debugger does wait()
and consumes PTHREAD_EVENT_EXEC which reports old_pid. This affects all
ptrace requests except the "asynchronous" PTRACE_INTERRUPT/KILL.
The patch doesn't add the new PTRACE_ option to not complicate the API,
and I _hope_ this won't cause any noticeable regression:
- If debugger uses PTRACE_O_TRACEEXEC and the thread did an exec
and the tracer does a ptrace request without having consumed
the exec event, it's 100% sure that the thread the ptracer
thinks it is targeting does not exist anymore, or isn't the
same as the one it thinks it is targeting.
- To some degree this patch adds nothing new. In the scenario
above ptrace(L) can fail with -ESRCH if it is called after the
execing sub-thread wakes the leader up and before it "steals"
the leader's pid.
Test-case:
#include <stdio.h>
#include <unistd.h>
#include <signal.h>
#include <sys/ptrace.h>
#include <sys/wait.h>
#include <errno.h>
#include <pthread.h>
#include <assert.h>
void *tf(void *arg)
{
execve("/usr/bin/true", NULL, NULL);
assert(0);
return NULL;
}
int main(void)
{
int leader = fork();
if (!leader) {
kill(getpid(), SIGSTOP);
pthread_t th;
pthread_create(&th, NULL, tf, NULL);
for (;;)
pause();
return 0;
}
waitpid(leader, NULL, WSTOPPED);
ptrace(PTRACE_SEIZE, leader, 0,
PTRACE_O_TRACECLONE | PTRACE_O_TRACEEXEC);
waitpid(leader, NULL, 0);
ptrace(PTRACE_CONT, leader, 0,0);
waitpid(leader, NULL, 0);
int status, thread = waitpid(-1, &status, 0);
assert(thread > 0 && thread != leader);
assert(status == 0x80137f);
ptrace(PTRACE_CONT, thread, 0,0);
/*
* waitid() because waitpid(leader, &status, WNOWAIT) does not
* report status. Why ????
*
* Why WEXITED? because we have another kernel problem connected
* to mt-exec.
*/
siginfo_t info;
assert(waitid(P_PID, leader, &info, WSTOPPED|WEXITED|WNOWAIT) == 0);
assert(info.si_pid == leader && info.si_status == 0x0405);
/* OK, it sleeps in ptrace(PTRACE_EVENT_EXEC == 0x04) */
assert(ptrace(PTRACE_CONT, leader, 0,0) == -1);
assert(errno == ESRCH);
assert(leader == waitpid(leader, &status, WNOHANG));
assert(status == 0x04057f);
assert(ptrace(PTRACE_CONT, leader, 0,0) == 0);
return 0;
}
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Reported-by: Simon Marchi <simon.marchi@efficios.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Pedro Alves <palves@redhat.com>
Acked-by: Simon Marchi <simon.marchi@efficios.com>
Acked-by: Jan Kratochvil <jan.kratochvil@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit aafe104aa9 upstream.
It was reported that a fix to the ring buffer recursion detection would
cause a hung machine when performing suspend / resume testing. The
following backtrace was extracted from debugging that case:
Call Trace:
trace_clock_global+0x91/0xa0
__rb_reserve_next+0x237/0x460
ring_buffer_lock_reserve+0x12a/0x3f0
trace_buffer_lock_reserve+0x10/0x50
__trace_graph_return+0x1f/0x80
trace_graph_return+0xb7/0xf0
? trace_clock_global+0x91/0xa0
ftrace_return_to_handler+0x8b/0xf0
? pv_hash+0xa0/0xa0
return_to_handler+0x15/0x30
? ftrace_graph_caller+0xa0/0xa0
? trace_clock_global+0x91/0xa0
? __rb_reserve_next+0x237/0x460
? ring_buffer_lock_reserve+0x12a/0x3f0
? trace_event_buffer_lock_reserve+0x3c/0x120
? trace_event_buffer_reserve+0x6b/0xc0
? trace_event_raw_event_device_pm_callback_start+0x125/0x2d0
? dpm_run_callback+0x3b/0xc0
? pm_ops_is_empty+0x50/0x50
? platform_get_irq_byname_optional+0x90/0x90
? trace_device_pm_callback_start+0x82/0xd0
? dpm_run_callback+0x49/0xc0
With the following RIP:
RIP: 0010:native_queued_spin_lock_slowpath+0x69/0x200
Since the fix to the recursion detection would allow a single recursion to
happen while tracing, this lead to the trace_clock_global() taking a spin
lock and then trying to take it again:
ring_buffer_lock_reserve() {
trace_clock_global() {
arch_spin_lock() {
queued_spin_lock_slowpath() {
/* lock taken */
(something else gets traced by function graph tracer)
ring_buffer_lock_reserve() {
trace_clock_global() {
arch_spin_lock() {
queued_spin_lock_slowpath() {
/* DEAD LOCK! */
Tracing should *never* block, as it can lead to strange lockups like the
above.
Restructure the trace_clock_global() code to instead of simply taking a
lock to update the recorded "prev_time" simply use it, as two events
happening on two different CPUs that calls this at the same time, really
doesn't matter which one goes first. Use a trylock to grab the lock for
updating the prev_time, and if it fails, simply try again the next time.
If it failed to be taken, that means something else is already updating
it.
Link: https://lkml.kernel.org/r/20210430121758.650b6e8a@gandalf.local.home
Cc: stable@vger.kernel.org
Tested-by: Konstantin Kharlamov <hi-angel@yandex.ru>
Tested-by: Todd Brandt <todd.e.brandt@linux.intel.com>
Fixes: b02414c8f0 ("ring-buffer: Fix recursion protection transitions between interrupt context") # started showing the problem
Fixes: 14131f2f98 ("tracing: implement trace_clock_*() APIs") # where the bug happened
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=212761
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 785e3c0a3a upstream.
The default max PID is set by PID_MAX_DEFAULT, and the tracing
infrastructure uses this number to map PIDs to the comm names of the
tasks, such output of the trace can show names from the recorded PIDs in
the ring buffer. This mapping is also exported to user space via the
"saved_cmdlines" file in the tracefs directory.
But currently the mapping expects the PIDs to be less than
PID_MAX_DEFAULT, which is the default maximum and not the real maximum.
Recently, systemd will increases the maximum value of a PID on the system,
and when tasks are traced that have a PID higher than PID_MAX_DEFAULT, its
comm is not recorded. This leads to the entire trace to have "<...>" as
the comm name, which is pretty useless.
Instead, keep the array mapping the size of PID_MAX_DEFAULT, but instead
of just mapping the index to the comm, map a mask of the PID
(PID_MAX_DEFAULT - 1) to the comm, and find the full PID from the
map_cmdline_to_pid array (that already exists).
This bug goes back to the beginning of ftrace, but hasn't been an issue
until user space started increasing the maximum value of PIDs.
Link: https://lkml.kernel.org/r/20210427113207.3c601884@gandalf.local.home
Cc: stable@vger.kernel.org
Fixes: bc0c38d139 ("ftrace: latency tracer infrastructure")
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8c9af478c0 upstream.
# echo switch_mm:traceoff > /sys/kernel/tracing/set_ftrace_filter
will cause switch_mm to stop tracing by the traceoff command.
# echo -n switch_mm:traceoff > /sys/kernel/tracing/set_ftrace_filter
does nothing.
The reason is that the parsing in the write function only processes
commands if it finished parsing (there is white space written after the
command). That's to handle:
write(fd, "switch_mm:", 10);
write(fd, "traceoff", 8);
cases, where the command is broken over multiple writes.
The problem is if the file descriptor is closed, then the write call is
not processed, and the command needs to be processed in the release code.
The release code can handle matching of functions, but does not handle
commands.
Cc: stable@vger.kernel.org
Fixes: eda1e32855 ("tracing: handle broken names in ftrace filter")
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 4fbf5d6837 upstream.
The FUTEX_WAIT operand has historically a relative timeout which means that
the clock id is irrelevant as relative timeouts on CLOCK_REALTIME are not
subject to wall clock changes and therefore are mapped by the kernel to
CLOCK_MONOTONIC for simplicity.
If a caller would set FUTEX_CLOCK_REALTIME for FUTEX_WAIT the timeout is
still treated relative vs. CLOCK_MONOTONIC and then the wait arms that
timeout based on CLOCK_REALTIME which is broken and obviously has never
been used or even tested.
Reject any attempt to use FUTEX_CLOCK_REALTIME with FUTEX_WAIT again.
The desired functionality can be achieved with FUTEX_WAIT_BITSET and a
FUTEX_BITSET_MATCH_ANY argument.
Fixes: 337f13046f ("futex: Allow FUTEX_CLOCK_REALTIME with FUTEX_WAIT op")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20210422194704.834797921@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 0687c66b5f ]
The debug_work_activate() is called on the premise that
the work can be inserted, because if wq be in WQ_DRAINING
status, insert work may be failed.
Fixes: e41e704bc4 ("workqueue: improve destroy_workqueue() debuggability")
Signed-off-by: Zqiang <qiang.zhang@windriver.com>
Reviewed-by: Lai Jiangshan <jiangshanlai@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 3054d06719 upstream.
If audit_list_rules_send() fails when trying to create a new thread
to send the rules it also fails to cleanup properly, leaking a
reference to a net structure. This patch fixes the error patch and
renames audit_send_list() to audit_send_list_thread() to better
match its cousin, audit_send_reply_thread().
Reported-by: teroincn@gmail.com
Reviewed-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Cc: <stable@vger.kernel.org> # 4.9.x
Signed-off-by: Wen Yang <wenyang@linux.alibaba.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a48b284b40 upstream.
If audit_send_reply() fails when trying to create a new thread to
send the reply it also fails to cleanup properly, leaking a reference
to a net structure. This patch fixes the error path and makes a
handful of other cleanups that came up while fixing the code.
Reported-by: teroincn@gmail.com
Reviewed-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Cc: <stable@vger.kernel.org> # 4.9.x
Signed-off-by: Wen Yang <wenyang@linux.alibaba.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9deb193af6 upstream.
Commit cbc3b92ce0 fixed an issue to modify the macros of the stack trace
event so that user space could parse it properly. Originally the stack
trace format to user space showed that the called stack was a dynamic
array. But it is not actually a dynamic array, in the way that other
dynamic event arrays worked, and this broke user space parsing for it. The
update was to make the array look to have 8 entries in it. Helper
functions were added to make it parse it correctly, as the stack was
dynamic, but was determined by the size of the event stored.
Although this fixed user space on how it read the event, it changed the
internal structure used for the stack trace event. It changed the array
size from [0] to [8] (added 8 entries). This increased the size of the
stack trace event by 8 words. The size reserved on the ring buffer was the
size of the stack trace event plus the number of stack entries found in
the stack trace. That commit caused the amount to be 8 more than what was
needed because it did not expect the caller field to have any size. This
produced 8 entries of garbage (and reading random data) from the stack
trace event:
<idle>-0 [002] d... 1976396.837549: <stack trace>
=> trace_event_raw_event_sched_switch
=> __traceiter_sched_switch
=> __schedule
=> schedule_idle
=> do_idle
=> cpu_startup_entry
=> secondary_startup_64_no_verify
=> 0xc8c5e150ffff93de
=> 0xffff93de
=> 0
=> 0
=> 0xc8c5e17800000000
=> 0x1f30affff93de
=> 0x00000004
=> 0x200000000
Instead, subtract the size of the caller field from the size of the event
to make sure that only the amount needed to store the stack trace is
reserved.
Link: https://lore.kernel.org/lkml/your-ad-here.call-01617191565-ext-9692@work.hours/
Cc: stable@vger.kernel.org
Fixes: cbc3b92ce0 ("tracing: Set kernel_stack's caller size properly")
Reported-by: Vasily Gorbik <gor@linux.ibm.com>
Tested-by: Vasily Gorbik <gor@linux.ibm.com>
Acked-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9f5d1c336a upstream.
Gratian managed to trigger the BUG_ON(!newowner) in fixup_pi_state_owner().
This is one possible chain of events leading to this:
Task Prio Operation
T1 120 lock(F)
T2 120 lock(F) -> blocks (top waiter)
T3 50 (RT) lock(F) -> boosts T1 and blocks (new top waiter)
XX timeout/ -> wakes T2
signal
T1 50 unlock(F) -> wakes T3 (rtmutex->owner == NULL, waiter bit is set)
T2 120 cleanup -> try_to_take_mutex() fails because T3 is the top waiter
and the lower priority T2 cannot steal the lock.
-> fixup_pi_state_owner() sees newowner == NULL -> BUG_ON()
The comment states that this is invalid and rt_mutex_real_owner() must
return a non NULL owner when the trylock failed, but in case of a queued
and woken up waiter rt_mutex_real_owner() == NULL is a valid transient
state. The higher priority waiter has simply not yet managed to take over
the rtmutex.
The BUG_ON() is therefore wrong and this is just another retry condition in
fixup_pi_state_owner().
Drop the locks, so that T3 can make progress, and then try the fixup again.
Gratian provided a great analysis, traces and a reproducer. The analysis is
to the point, but it confused the hell out of that tglx dude who had to
page in all the futex horrors again. Condensed version is above.
[ tglx: Wrote comment and changelog ]
Fixes: c1e2f0eaf0 ("futex: Avoid violating the 10th rule of futex")
Reported-by: Gratian Crisan <gratian.crisan@ni.com>
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/87a6w6x7bb.fsf@ni.com
Link: https://lore.kernel.org/r/87sg9pkvf7.fsf@nanos.tec.linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 921c7ebd13 upstream.
If should_futex_fail() returns true in futex_wake_pi(), then the 'ret'
variable is set to -EFAULT and then immediately overwritten. So the failure
injection is non-functional.
Fix it by actually leaving the function and returning -EFAULT.
The Fixes tag is kinda blury because the initial commit which introduced
failure injection was already sloppy, but the below mentioned commit broke
it completely.
[ tglx: Massaged changelog ]
Fixes: 6b4f4bc9cb ("locking/futex: Allow low-level atomic operations to return -EAGAIN")
Signed-off-by: Mateusz Nosek <mateusznosek0@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20200927000858.24219-1-mateusznosek0@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ca16d5bee5 upstream.
Robust futexes utilize the robust_list mechanism to allow the kernel to
release futexes which are held when a task exits. The exit can be voluntary
or caused by a signal or fault. This prevents that waiters block forever.
The futex operations in user space store a pointer to the futex they are
either locking or unlocking in the op_pending member of the per task robust
list.
After a lock operation has succeeded the futex is queued in the robust list
linked list and the op_pending pointer is cleared.
After an unlock operation has succeeded the futex is removed from the
robust list linked list and the op_pending pointer is cleared.
The robust list exit code checks for the pending operation and any futex
which is queued in the linked list. It carefully checks whether the futex
value is the TID of the exiting task. If so, it sets the OWNER_DIED bit and
tries to wake up a potential waiter.
This is race free for the lock operation but unlock has two race scenarios
where waiters might not be woken up. These issues can be observed with
regular robust pthread mutexes. PI aware pthread mutexes are not affected.
(1) Unlocking task is killed after unlocking the futex value in user space
before being able to wake a waiter.
pthread_mutex_unlock()
|
V
atomic_exchange_rel (&mutex->__data.__lock, 0)
<------------------------killed
lll_futex_wake () |
|
|(__lock = 0)
|(enter kernel)
|
V
do_exit()
exit_mm()
mm_release()
exit_robust_list()
handle_futex_death()
|
|(__lock = 0)
|(uval = 0)
|
V
if ((uval & FUTEX_TID_MASK) != task_pid_vnr(curr))
return 0;
The sanity check which ensures that the user space futex is owned by
the exiting task prevents the wakeup of waiters which in consequence
block infinitely.
(2) Waiting task is killed after a wakeup and before it can acquire the
futex in user space.
OWNER WAITER
futex_wait()
pthread_mutex_unlock() |
| |
|(__lock = 0) |
| |
V |
futex_wake() ------------> wakeup()
|
|(return to userspace)
|(__lock = 0)
|
V
oldval = mutex->__data.__lock
<-----------------killed
atomic_compare_and_exchange_val_acq (&mutex->__data.__lock, |
id | assume_other_futex_waiters, 0) |
|
|
(enter kernel)|
|
V
do_exit()
|
|
V
handle_futex_death()
|
|(__lock = 0)
|(uval = 0)
|
V
if ((uval & FUTEX_TID_MASK) != task_pid_vnr(curr))
return 0;
The sanity check which ensures that the user space futex is owned
by the exiting task prevents the wakeup of waiters, which seems to
be correct as the exiting task does not own the futex value, but
the consequence is that other waiters wont be woken up and block
infinitely.
In both scenarios the following conditions are true:
- task->robust_list->list_op_pending != NULL
- user space futex value == 0
- Regular futex (not PI)
If these conditions are met then it is reasonably safe to wake up a
potential waiter in order to prevent the above problems.
As this might be a false positive it can cause spurious wakeups, but the
waiter side has to handle other types of unrelated wakeups, e.g. signals
gracefully anyway. So such a spurious wakeup will not affect the
correctness of these operations.
This workaround must not touch the user space futex value and cannot set
the OWNER_DIED bit because the lock value is 0, i.e. uncontended. Setting
OWNER_DIED in this case would result in inconsistent state and subsequently
in malfunction of the owner died handling in user space.
The rest of the user space state is still consistent as no other task can
observe the list_op_pending entry in the exiting tasks robust list.
The eventually woken up waiter will observe the uncontended lock value and
take it over.
[ tglx: Massaged changelog and comment. Made the return explicit and not
depend on the subsequent check and added constants to hand into
handle_futex_death() instead of plain numbers. Fixed a few coding
style issues. ]
Fixes: 0771dfefc9 ("[PATCH] lightweight robust futexes: core")
Signed-off-by: Yang Tao <yang.tao172@zte.com.cn>
Signed-off-by: Yi Wang <wang.yi59@zte.com.cn>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/1573010582-35297-1-git-send-email-wang.yi59@zte.com.cn
Link: https://lkml.kernel.org/r/20191106224555.943191378@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6b4f4bc9cb upstream.
Some futex() operations, including FUTEX_WAKE_OP, require the kernel to
perform an atomic read-modify-write of the futex word via the userspace
mapping. These operations are implemented by each architecture in
arch_futex_atomic_op_inuser() and futex_atomic_cmpxchg_inatomic(), which
are called in atomic context with the relevant hash bucket locks held.
Although these routines may return -EFAULT in response to a page fault
generated when accessing userspace, they are expected to succeed (i.e.
return 0) in all other cases. This poses a problem for architectures
that do not provide bounded forward progress guarantees or fairness of
contended atomic operations and can lead to starvation in some cases.
In these problematic scenarios, we must return back to the core futex
code so that we can drop the hash bucket locks and reschedule if
necessary, much like we do in the case of a page fault.
Allow architectures to return -EAGAIN from their implementations of
arch_futex_atomic_op_inuser() and futex_atomic_cmpxchg_inatomic(), which
will cause the core futex code to reschedule if necessary and return
back to the architecture code later on.
Cc: <stable@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[bwh: Backported to 4.9: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit b061c38bef upstream.
We must not rely on wake_q_add() to delay the wakeup; in particular
commit:
1d0dcb3ad9 ("futex: Implement lockless wakeups")
moved wake_q_add() before smp_store_release(&q->lock_ptr, NULL), which
could result in futex_wait() waking before observing ->lock_ptr ==
NULL and going back to sleep again.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 1d0dcb3ad9 ("futex: Implement lockless wakeups")
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1a1fb985f2 upstream.
commit 56222b212e ("futex: Drop hb->lock before enqueueing on the
rtmutex") changed the locking rules in the futex code so that the hash
bucket lock is not longer held while the waiter is enqueued into the
rtmutex wait list. This made the lock and the unlock path symmetric, but
unfortunately the possible early exit from __rt_mutex_proxy_start() due to
a detected deadlock was not updated accordingly. That allows a concurrent
unlocker to observe inconsitent state which triggers the warning in the
unlock path.
futex_lock_pi() futex_unlock_pi()
lock(hb->lock)
queue(hb_waiter) lock(hb->lock)
lock(rtmutex->wait_lock)
unlock(hb->lock)
// acquired hb->lock
hb_waiter = futex_top_waiter()
lock(rtmutex->wait_lock)
__rt_mutex_proxy_start()
---> fail
remove(rtmutex_waiter);
---> returns -EDEADLOCK
unlock(rtmutex->wait_lock)
// acquired wait_lock
wake_futex_pi()
rt_mutex_next_owner()
--> returns NULL
--> WARN
lock(hb->lock)
unqueue(hb_waiter)
The problem is caused by the remove(rtmutex_waiter) in the failure case of
__rt_mutex_proxy_start() as this lets the unlocker observe a waiter in the
hash bucket but no waiter on the rtmutex, i.e. inconsistent state.
The original commit handles this correctly for the other early return cases
(timeout, signal) by delaying the removal of the rtmutex waiter until the
returning task reacquired the hash bucket lock.
Treat the failure case of __rt_mutex_proxy_start() in the same way and let
the existing cleanup code handle the eventual handover of the rtmutex
gracefully. The regular rt_mutex_proxy_start() gains the rtmutex waiter
removal for the failure case, so that the other callsites are still
operating correctly.
Add proper comments to the code so all these details are fully documented.
Thanks to Peter for helping with the analysis and writing the really
valuable code comments.
Fixes: 56222b212e ("futex: Drop hb->lock before enqueueing on the rtmutex")
Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Co-developed-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: linux-s390@vger.kernel.org
Cc: Stefan Liebler <stli@linux.ibm.com>
Cc: Sebastian Sewior <bigeasy@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1901292311410.1950@nanos.tec.linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 04dc1b2fff upstream.
Markus reported that the glibc/nptl/tst-robustpi8 test was failing after
commit:
cfafcd117d ("futex: Rework futex_lock_pi() to use rt_mutex_*_proxy_lock()")
The following trace shows the problem:
ld-linux-x86-64-2161 [019] .... 410.760971: SyS_futex: 00007ffbeb76b028: 80000875 op=FUTEX_LOCK_PI
ld-linux-x86-64-2161 [019] ...1 410.760972: lock_pi_update_atomic: 00007ffbeb76b028: curval=80000875 uval=80000875 newval=80000875 ret=0
ld-linux-x86-64-2165 [011] .... 410.760978: SyS_futex: 00007ffbeb76b028: 80000875 op=FUTEX_UNLOCK_PI
ld-linux-x86-64-2165 [011] d..1 410.760979: do_futex: 00007ffbeb76b028: curval=80000875 uval=80000875 newval=80000871 ret=0
ld-linux-x86-64-2165 [011] .... 410.760980: SyS_futex: 00007ffbeb76b028: 80000871 ret=0000
ld-linux-x86-64-2161 [019] .... 410.760980: SyS_futex: 00007ffbeb76b028: 80000871 ret=ETIMEDOUT
Task 2165 does an UNLOCK_PI, assigning the lock to the waiter task 2161
which then returns with -ETIMEDOUT. That wrecks the lock state, because now
the owner isn't aware it acquired the lock and removes the pending robust
list entry.
If 2161 is killed, the robust list will not clear out this futex and the
subsequent acquire on this futex will then (correctly) result in -ESRCH
which is unexpected by glibc, triggers an internal assertion and dies.
Task 2161 Task 2165
rt_mutex_wait_proxy_lock()
timeout();
/* T2161 is still queued in the waiter list */
return -ETIMEDOUT;
futex_unlock_pi()
spin_lock(hb->lock);
rtmutex_unlock()
remove_rtmutex_waiter(T2161);
mark_lock_available();
/* Make the next waiter owner of the user space side */
futex_uval = 2161;
spin_unlock(hb->lock);
spin_lock(hb->lock);
rt_mutex_cleanup_proxy_lock()
if (rtmutex_owner() !== current)
...
return FAIL;
....
return -ETIMEOUT;
This means that rt_mutex_cleanup_proxy_lock() needs to call
try_to_take_rt_mutex() so it can take over the rtmutex correctly which was
assigned by the waker. If the rtmutex is owned by some other task then this
call is harmless and just confirmes that the waiter is not able to acquire
it.
While there, fix what looks like a merge error which resulted in
rt_mutex_cleanup_proxy_lock() having two calls to
fixup_rt_mutex_waiters() and rt_mutex_wait_proxy_lock() not having any.
Both should have one, since both potentially touch the waiter list.
Fixes: 38d589f2fd ("futex,rt_mutex: Restructure rt_mutex_finish_proxy_lock()")
Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Bug-Spotted-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Florian Weimer <fweimer@redhat.com>
Cc: Darren Hart <dvhart@infradead.org>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Markus Trippelsdorf <markus@trippelsdorf.de>
Link: http://lkml.kernel.org/r/20170519154850.mlomgdsd26drq5j6@hirez.programming.kicks-ass.net
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 97181f9bd5 upstream.
Alexander reported a hrtimer debug_object splat:
ODEBUG: free active (active state 0) object type: hrtimer hint: hrtimer_wakeup (kernel/time/hrtimer.c:1423)
debug_object_free (lib/debugobjects.c:603)
destroy_hrtimer_on_stack (kernel/time/hrtimer.c:427)
futex_lock_pi (kernel/futex.c:2740)
do_futex (kernel/futex.c:3399)
SyS_futex (kernel/futex.c:3447 kernel/futex.c:3415)
do_syscall_64 (arch/x86/entry/common.c:284)
entry_SYSCALL64_slow_path (arch/x86/entry/entry_64.S:249)
Which was caused by commit:
cfafcd117d ("futex: Rework futex_lock_pi() to use rt_mutex_*_proxy_lock()")
... losing the hrtimer_cancel() in the shuffle. Where previously the
hrtimer_cancel() was done by rt_mutex_slowlock() we now need to do it
manually.
Reported-by: Alexander Levin <alexander.levin@verizon.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Fixes: cfafcd117d ("futex: Rework futex_lock_pi() to use rt_mutex_*_proxy_lock()")
Link: http://lkml.kernel.org/r/alpine.DEB.2.20.1704101802370.2906@nanos
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 56222b212e upstream.
When PREEMPT_RT_FULL does the spinlock -> rt_mutex substitution the PI
chain code will (falsely) report a deadlock and BUG.
The problem is that it hold hb->lock (now an rt_mutex) while doing
task_blocks_on_rt_mutex on the futex's pi_state::rtmutex. This, when
interleaved just right with futex_unlock_pi() leads it to believe to see an
AB-BA deadlock.
Task1 (holds rt_mutex, Task2 (does FUTEX_LOCK_PI)
does FUTEX_UNLOCK_PI)
lock hb->lock
lock rt_mutex (as per start_proxy)
lock hb->lock
Which is a trivial AB-BA.
It is not an actual deadlock, because it won't be holding hb->lock by the
time it actually blocks on the rt_mutex, but the chainwalk code doesn't
know that and it would be a nightmare to handle this gracefully.
To avoid this problem, do the same as in futex_unlock_pi() and drop
hb->lock after acquiring wait_lock. This still fully serializes against
futex_unlock_pi(), since adding to the wait_list does the very same lock
dance, and removing it holds both locks.
Aside of solving the RT problem this makes the lock and unlock mechanism
symetric and reduces the hb->lock held time.
Reported-and-tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: juri.lelli@arm.com
Cc: xlpang@redhat.com
Cc: rostedt@goodmis.org
Cc: mathieu.desnoyers@efficios.com
Cc: jdesfossez@efficios.com
Cc: dvhart@infradead.org
Cc: bristot@redhat.com
Link: http://lkml.kernel.org/r/20170322104152.161341537@infradead.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
