Peter Zijlstra
88ce30fb88
locking/qspinlock, x86: Provide liveness guarantee ...
commit 7aa54be2976550f17c11a1c3e3630002dea39303 upstream.
On x86 we cannot do fetch_or() with a single instruction and thus end up
using a cmpxchg loop, this reduces determinism. Replace the fetch_or()
with a composite operation: tas-pending + load.
Using two instructions of course opens a window we previously did not
have. Consider the scenario:
CPU0 CPU1 CPU2
1) lock
trylock -> (0,0,1)
2) lock
trylock /* fail */
3) unlock -> (0,0,0)
4) lock
trylock -> (0,0,1)
5) tas-pending -> (0,1,1)
load-val <- (0,1,0) from 3
6) clear-pending-set-locked -> (0,0,1)
FAIL: _2_ owners
where 5) is our new composite operation. When we consider each part of
the qspinlock state as a separate variable (as we can when
_Q_PENDING_BITS == 8) then the above is entirely possible, because
tas-pending will only RmW the pending byte, so the later load is able
to observe prior tail and lock state (but not earlier than its own
trylock, which operates on the whole word, due to coherence).
To avoid this we need 2 things:
- the load must come after the tas-pending (obviously, otherwise it
can trivially observe prior state).
- the tas-pending must be a full word RmW instruction, it cannot be an XCHGB for
example, such that we cannot observe other state prior to setting
pending.
On x86 we can realize this by using "LOCK BTS m32, r32" for
tas-pending followed by a regular load.
Note that observing later state is not a problem:
- if we fail to observe a later unlock, we'll simply spin-wait for
that store to become visible.
- if we observe a later xchg_tail(), there is no difference from that
xchg_tail() having taken place before the tas-pending.
Suggested-by: Will Deacon <will.deacon@arm.com>
Reported-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Will Deacon <will.deacon@arm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: andrea.parri@amarulasolutions.com
Cc: longman@redhat.com
Fixes: 59fb586b4a07 ("locking/qspinlock: Remove unbounded cmpxchg() loop from locking slowpath")
Link: https://lkml.kernel.org/r/20181003130957.183726335@infradead.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
[bigeasy: GEN_BINARY_RMWcc macro redo]
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2018-12-21 14:11:34 +01:00
bpf: Prevent memory disambiguation attack
2018-12-08 13:05:10 +01:00
config: android: enable CONFIG_SECCOMP
2016-10-11 15:06:32 -07:00
kdb: use memmove instead of overlapping memcpy
2018-12-08 13:05:05 +01:00
uprobes: Fix handle_swbp() vs. unregister() + register() race once more
2018-12-08 13:05:06 +01:00
gcov: support GCC 7.1
2017-09-02 07:07:53 +02:00
genirq: Fix race on spurious interrupt detection
2018-11-13 11:17:01 -08:00
livepatch/module: make TAINT_LIVEPATCH module-specific
2016-08-26 14:42:08 +02:00
locking/qspinlock, x86: Provide liveness guarantee
2018-12-21 14:11:34 +01:00
PM / sleep: wakeup: Fix build error caused by missing SRCU support
2018-09-09 20:01:23 +02:00
printk: Fix panic caused by passing log_buf_len to command line
2018-11-13 11:17:01 -08:00
rcu: Call touch_nmi_watchdog() while printing stall warnings
2018-05-30 07:50:50 +02:00
sched/core: Allow __sched_setscheduler() in interrupts when PI is not used
2018-12-01 09:44:25 +01:00
timer/debug: Change /proc/timer_list from 0444 to 0400
2018-12-21 14:11:30 +01:00
tracing: Fix memory leak of instance function hash filters
2018-12-21 14:11:31 +01:00
kernel/acct.c: fix the acct->needcheck check in check_free_space()
2018-01-10 09:29:51 +01:00
kernel/async.c: revert "async: simplify lowest_in_progress()"
2018-02-17 13:21:18 +01:00
audit: return on memory error to avoid null pointer dereference
2018-05-30 07:50:49 +02:00
Merge branch 'stable-4.8' of git://git.infradead.org/users/pcmoore/audit
2016-07-29 17:54:17 -07:00
audit: fix use-after-free in audit_add_watch
2018-09-26 08:36:37 +02:00
audit: allow not equal op for audit by executable
2018-08-03 07:55:25 +02:00
audit: allow not equal op for audit by executable
2018-08-03 07:55:25 +02:00
kbuild: fix kernel/bounds.c 'W=1' warning
2018-11-13 11:16:57 -08:00
ptrace: Capture the ptracer's creds not PT_PTRACE_CAP
2017-01-06 10:40:13 +01:00
cgroup: Fix deadlock in cpu hotplug path
2018-10-13 09:18:56 +02:00
cgroup/pids: remove spurious suspicious RCU usage warning
2017-03-26 13:05:58 +02:00
Revert "x86/speculation: Enable cross-hyperthread spectre v2 STIBP mitigation"
2018-11-23 12:57:41 +01:00
sched/cpuset/pm: Fix cpuset vs. suspend-resume bugs
2017-10-12 11:51:25 +02:00
cred: Reject inodes with invalid ids in set_create_file_as()
2016-06-30 18:05:09 -05:00
kernel/exit.c: avoid undefined behaviour when calling wait4()
2018-05-19 10:27:01 +02:00
kernel/extable.c: mark core_kernel_text notrace
2017-07-21 07:42:21 +02:00
kthread: Fix use-after-free if kthread fork fails
2018-09-19 22:47:10 +02:00
freezer, oom: check TIF_MEMDIE on the correct task
2016-07-28 16:07:41 -07:00
futex: futex_wake_op, do not fail on invalid op
2018-11-10 07:42:54 -08:00
kernel: make groups_sort calling a responsibility group_info allocators
2018-01-10 09:29:52 +01:00
hung_task: allow hung_task_panic when hung_task_warnings is 0
2016-10-11 15:06:33 -07:00
jump_label: Invoke jump_label_test() via early_initcall()
2017-12-14 09:28:24 +01:00
kcov: ensure irq code sees a valid area
2018-08-03 07:55:12 +02:00
kexec: allow architectures to override boot mapping
2016-08-02 19:35:27 -04:00
objtool, x86: Add several functions and files to the objtool whitelist
2018-06-05 10:28:57 +02:00
kexec: fix double-free when failing to relocate the purgatory
2016-09-01 17:52:01 -07:00
kprobes: Return error if we fail to reuse kprobe instead of BUG_ON()
2018-11-13 11:16:50 -08:00
kexec: add a kexec_crash_loaded() function
2016-08-02 19:35:30 -04:00
kthread, tracing: Don't expose half-written comm when creating kthreads
2018-08-03 07:55:12 +02:00
kernel/watchdog.c: move hardlockup detector to separate file
2017-06-17 06:41:57 +02:00
Fix: Disable sys_membarrier when nohz_full is enabled
2017-03-12 06:41:45 +01:00
mm: Fix devm_memremap_pages() collision handling
2018-02-28 10:18:34 +01:00
module: exclude SHN_UNDEF symbols from kallsyms api
2018-10-03 17:01:48 -07:00
padata: free correct variable
2017-05-20 14:28:40 +02:00
kernel/panic.c: add missing \n
2017-07-05 14:40:24 +02:00
pidns: disable pid allocation if pid_ns_prepare_proc() is failed in alloc_pid()
2018-04-13 19:47:53 +02:00
pid_ns: Sleep in TASK_INTERRUPTIBLE in zap_pid_ns_processes
2017-05-25 15:44:38 +02:00
profile: Convert to hotplug state machine
2016-07-15 10:41:42 +02:00
ptrace: Properly initialize ptracer_cred on fork
2017-06-14 15:05:54 +02:00
kernel/relay.c: limit kmalloc size to KMALLOC_MAX_SIZE
2018-05-30 07:50:29 +02:00
resource: fix integer overflow at reallocation
2018-04-24 09:34:09 +02:00
seccomp: Move speculation migitation control to arch code
2018-05-22 16:58:02 +02:00
signal: Introduce COMPAT_SIGMINSTKSZ for use in compat_sys_sigaltstack
2018-12-21 14:11:29 +01:00
cpu/hotplug: Fix SMT supported evaluation
2018-08-15 18:14:53 +02:00
kthread/smpboot: do not park in kthread_create_on_cpu()
2016-10-11 15:06:33 -07:00
Mark HI and TASKLET softirq synchronous
2018-08-15 18:14:42 +02:00
stacktrace, lockdep: Fix address, newline ugliness
2017-02-14 15:25:42 -08:00
stop_machine: Use raw spinlocks
2018-08-03 07:55:24 +02:00
sys: don't hold uts_sem while accessing userspace memory
2018-09-09 20:01:24 +02:00
x86/pkeys: Fix pkeys build breakage for some non-x86 arches
2016-09-13 14:41:36 +02:00
namei: allow restricted O_CREAT of FIFOs and regular files
2018-12-01 09:44:25 +01:00
kernel/sysctl_binary.c: use generic UUID library
2016-05-20 17:58:30 -07:00
task_work: use READ_ONCE/lockless_dereference, avoid pi_lock if !task_works
2016-08-02 19:35:02 -04:00
taskstats: fix the length of cgroupstats_cmd_get_policy
2016-11-03 16:55:58 -04:00
torture: Convert torture_shutdown() to hrtimer
2016-08-22 10:01:49 -07:00
tracepoint: Do not warn on ENOMEM
2018-05-09 09:50:20 +02:00
kernel/ucount.c: mark user_header with kmemleak_ignore()
2017-06-17 06:41:51 +02:00
kernel: make groups_sort calling a responsibility group_info allocators
2018-01-10 09:29:52 +01:00
smp: Add function to execute a function synchronously on a CPU
2016-09-05 13:52:39 +02:00
userns: move user access out of the mutex
2018-09-09 20:01:24 +02:00
Merge branch 'nsfs-ioctls' into HEAD
2016-09-22 20:00:36 -05:00
sys: don't hold uts_sem while accessing userspace memory
2018-09-09 20:01:24 +02:00
kernel/watchdog: prevent false hardlockup on overloaded system
2017-06-17 06:41:57 +02:00
kernel/watchdog: prevent false hardlockup on overloaded system
2017-06-17 06:41:57 +02:00
workqueue: use put_device() instead of kfree()
2018-05-30 07:50:36 +02:00
workqueue: Fix NULL pointer dereference
2017-11-15 15:53:17 +01:00
88ce30fb88