| /linux/fs/ubifs/ |
| A D | shrinker.c | 58 static int shrink_tnc(struct ubifs_info *c, int nr, int age, int *contention) in shrink_tnc() argument
108 *contention = 1; in shrink_tnc()
144 static int shrink_tnc_trees(int nr, int age, int *contention) in shrink_tnc_trees() argument
167 *contention = 1; in shrink_tnc_trees()
177 *contention = 1; in shrink_tnc_trees()
187 freed += shrink_tnc(c, nr, age, contention); in shrink_tnc_trees()
283 int contention = 0; in ubifs_shrink_scan() local
299 freed = shrink_tnc_trees(nr, OLD_ZNODE_AGE, &contention); in ubifs_shrink_scan()
304 freed += shrink_tnc_trees(nr - freed, YOUNG_ZNODE_AGE, &contention); in ubifs_shrink_scan()
309 freed += shrink_tnc_trees(nr - freed, 0, &contention); in ubifs_shrink_scan()
[all …]
|
| /linux/Documentation/locking/ |
| A D | lockstat.rst | 14 Because things like lock contention can severely impact performance.
51 - number of lock contention that involved x-cpu data
80 It also tracks 4 contention points per class. A contention point is a call site
143 short separator (line 08, 13) from the contention points.
145 Lines 09-12 show the first 4 recorded contention points (the code
151 short separator. The contention points don't match the column descriptors,
152 they have two: contentions and [<IP>] symbol. The second set of contention
|
| A D | locktorture.rst | 19 different critical region behaviors. The amount of contention on the lock
|
| A D | ww-mutex-design.rst | 352 contention for a new lock and hence a true chance of deadlock. In that
360 In general, not much contention is expected. The locks are typically used to
|
| A D | rt-mutex-design.rst | 428 If there is contention on the lock, we go about the slow path
442 contention).
518 CMPXCHG. Since the taking of a mutex on contention always sets the
|
| A D | robust-futexes.rst | 16 field. If userspace notices contention (the lock is already owned and
|
| /linux/tools/perf/scripts/python/bin/ |
| A D | futex-contention-report | 4 perf script $@ -s "$PERF_EXEC_PATH"/scripts/python/futex-contention.py
|
| /linux/kernel/rcu/ |
| A D | Kconfig | 148 against lock contention. Systems that synchronize their
151 lock contention levels acceptably low. Very large systems
157 skew them, which reduces lock contention enough that large
160 lock contention on the leaf-level rcu_node structures unless
167 kernel boot parameter to avoid contention on the rcu_node
|
| /linux/kernel/ |
| A D | Kconfig.hz | 14 contention and cacheline bounces as a result of timer interrupts.
|
| /linux/Documentation/trace/ |
| A D | events-kmem.rst | 72 contention on the lruvec->lru_lock.
82 for order-0 pages, reduces contention on the zone->lock and reduces the
|
| /linux/Documentation/scheduler/ |
| A D | schedutil.txt | 34 two metrics are the same, but once there is contention for the CPU 'running'
36 while 'runnable' will increase to reflect the amount of contention.
|
| A D | sched-bwc.rst | 167 variable. This is a performance tweak that helps prevent added contention on
|
| /linux/Documentation/vm/ |
| A D | split_page_table_lock.rst | 9 multi-threaded applications due high contention on the lock. To improve
|
| /linux/Documentation/networking/ |
| A D | scaling.rst | 195 during CPU contention by dropping packets from large flows slightly
214 and cache contention) and toggled per CPU by setting the relevant bit
409 provides two benefits. First, contention on the device queue lock is
411 (contention can be eliminated completely if each CPU has its own
489 experience no contention. If there are fewer queues than CPUs, then the
|
| /linux/Documentation/devicetree/bindings/sound/ |
| A D | cs35l36.txt | 34 port to prevent bus contention on the output signal
|
| /linux/drivers/gpu/drm/ |
| A D | Kconfig | 104 bool "Enable backtrace history for lock contention"
112 contention. A history of each drm modeset lock path hitting -EDEADLK
|
| /linux/Documentation/RCU/Design/Data-Structures/ |
| A D | Data-Structures.rst | 67 | leaf ``rcu_node`` structures have fanout of 64, the contention on |
76 | if you are using such a system and running into contention problems |
111 contention remains roughly constant up the tree. No matter how many CPUs
114 thus ensuring that the lock contention on that root ``rcu_node``
118 lock contention under control at all tree levels regardless of the level
636 excessive contention for the leaf ``rcu_node`` structures' ``->lock``
|
| /linux/Documentation/accounting/ |
| A D | psi.rst | 13 Without an accurate measure of such contention, users are forced to
|
| /linux/Documentation/devicetree/bindings/bus/ |
| A D | qcom,ebi2.txt | 78 drive the data bus after OE is de-asserted, in order to avoid contention on
|
| /linux/Documentation/arm/ |
| A D | vlocks.rst | 148 reducing bus contention in contended cases.
|
| /linux/fs/squashfs/ |
| A D | Kconfig | 52 it eliminates a memcpy and it also removes the lock contention
|
| /linux/drivers/block/paride/ |
| A D | Transition-notes | 113 that the only possible contention is between scheduling ps_tq followed by
|
| /linux/Documentation/filesystems/ |
| A D | xfs-delayed-logging-design.rst | 652 items, it will get held for a significant time and so spin contention is a
668 possible that this lock will become a contention point, but given the short
669 hold time once per transaction I think that contention is unlikely.
689 much contention on the CIL lock, or too many context switches as a result of
691 given separate wait lists to reduce lock contention and the number of processes
|
| /linux/tools/perf/Documentation/ |
| A D | perf-c2c.txt | 32 for cachelines with highest contention - highest number of HITM accesses.
|
| /linux/lib/ |
| A D | Kconfig.kcsan | 108 due to reducing cache-line contention. The chosen default is a
|