/linux-6.12.1/Documentation/translations/zh_CN/scheduler/ |
D | sched-nice-design.rst | 4 :Original: Documentation/scheduler/sched-nice-design.rst 11 调度器nice值设计 14 本文档解释了新的Linux调度器中修改和精简后的nice级别的实现思路。 16 Linux的nice级别总是非常脆弱,人们持续不断地缠着我们,让nice +19的任务占用 20 nice级别的支持在历史上是与时间片长度耦合的,而时间片单位是由HZ滴答驱动的, 23 在O(1)调度器中(2003年),我们改变了负的nice级别,使它们比2.4内核更强 24 (人们对这一变化很满意),而且我们还故意校正了线性时间片准则,使得nice +19 39 -*----------------------------------*-----> [nice level] 51 nice +19级别运行数量颇多的应用程序)。 53 因此,对于HZ=1000,我们将nice +19改为5毫秒,因为这感觉像是正确的最小 [all …]
|
D | sched-design-CFS.rst | 46 CFS剩下的其它设计,一般脱离了这个简单的概念,附加的设计包括nice级别,多处理,以及各种 92 CFS调度器处理nice级别和SCHED_BATCH的能力比之前的原始调度器更强:两种类型的工作负载 110 - SCHED_IDLE:它比nice 19更弱,不过它不是真正的idle定时器调度器,因为要避免给机器
|
D | sched-energy.rst | 248 b. 所有任务,不管它们的nice值是多大,都应该被提供了足够多的CPU算力。 254 又回到了在CPU密集的条件下基于负载的算法做负载均衡。这更好地尊重了任务的nice值。
|
D | index.rst | 29 sched-nice-design
|
/linux-6.12.1/Documentation/scheduler/ |
D | sched-nice-design.rst | 6 nice-levels implementation in the new Linux scheduler. 9 pestered us to make nice +19 tasks use up much less CPU time. 12 scheduler, (otherwise we'd have done it long ago) because nice level 16 In the O(1) scheduler (in 2003) we changed negative nice levels to be 19 rule so that nice +19 level would be _exactly_ 1 jiffy. To better 34 -*----------------------------------*-----> [nice level] 49 people were running number crunching apps at nice +19.) 51 So for HZ=1000 we changed nice +19 to 5msecs, because that felt like the 53 But the fundamental HZ-sensitive property for nice+19 still remained, 54 and we never got a single complaint about nice +19 being too _weak_ in [all …]
|
D | index.rst | 21 sched-nice-design
|
/linux-6.12.1/kernel/sched/ |
D | autogroup.c | 230 int proc_sched_autogroup_set_nice(struct task_struct *p, int nice) in proc_sched_autogroup_set_nice() argument 237 if (nice < MIN_NICE || nice > MAX_NICE) in proc_sched_autogroup_set_nice() 240 err = security_task_setnice(current, nice); in proc_sched_autogroup_set_nice() 244 if (nice < 0 && !can_nice(current, nice)) in proc_sched_autogroup_set_nice() 254 idx = array_index_nospec(nice + 20, 40); in proc_sched_autogroup_set_nice() 260 ag->nice = nice; in proc_sched_autogroup_set_nice() 276 seq_printf(m, "/autogroup-%ld nice %d\n", ag->id, ag->nice); in proc_sched_autogroup_show_task()
|
D | syscalls.c | 19 static inline int __normal_prio(int policy, int rt_prio, int nice) in __normal_prio() argument 28 prio = NICE_TO_PRIO(nice); in __normal_prio() 65 void set_user_nice(struct task_struct *p, long nice) in set_user_nice() argument 71 if (task_nice(p) == nice || nice < MIN_NICE || nice > MAX_NICE) in set_user_nice() 89 p->static_prio = NICE_TO_PRIO(nice); in set_user_nice() 100 p->static_prio = NICE_TO_PRIO(nice); in set_user_nice() 126 static bool is_nice_reduction(const struct task_struct *p, const int nice) in is_nice_reduction() argument 129 int nice_rlim = nice_to_rlimit(nice); in is_nice_reduction() 139 int can_nice(const struct task_struct *p, const int nice) in can_nice() argument 141 return is_nice_reduction(p, nice) || capable(CAP_SYS_NICE); in can_nice() [all …]
|
D | autogroup.h | 17 int nice; member
|
/linux-6.12.1/include/linux/sched/ |
D | prio.h | 27 #define NICE_TO_PRIO(nice) ((nice) + DEFAULT_PRIO) argument 33 static inline long nice_to_rlimit(long nice) in nice_to_rlimit() argument 35 return (MAX_NICE - nice + 1); in nice_to_rlimit()
|
D | autogroup.h | 18 extern int proc_sched_autogroup_set_nice(struct task_struct *p, int nice);
|
/linux-6.12.1/Documentation/translations/zh_CN/core-api/ |
D | workqueue.rst | 577 pool[00] ref= 1 nice= 0 idle/workers= 4/ 4 cpu= 0 578 pool[01] ref= 1 nice=-20 idle/workers= 2/ 2 cpu= 0 579 pool[02] ref= 1 nice= 0 idle/workers= 4/ 4 cpu= 1 580 pool[03] ref= 1 nice=-20 idle/workers= 2/ 2 cpu= 1 581 pool[04] ref= 1 nice= 0 idle/workers= 4/ 4 cpu= 2 582 pool[05] ref= 1 nice=-20 idle/workers= 2/ 2 cpu= 2 583 pool[06] ref= 1 nice= 0 idle/workers= 3/ 3 cpu= 3 584 pool[07] ref= 1 nice=-20 idle/workers= 2/ 2 cpu= 3 585 pool[08] ref=42 nice= 0 idle/workers= 6/ 6 cpus=0000000f 586 pool[09] ref=28 nice= 0 idle/workers= 3/ 3 cpus=00000003 [all …]
|
/linux-6.12.1/fs/proc/ |
D | stat.c | 85 u64 user, nice, system, idle, iowait, irq, softirq, steal; in show_stat() local 92 user = nice = system = idle = iowait = in show_stat() 106 nice += cpustat[CPUTIME_NICE]; in show_stat() 128 seq_put_decimal_ull(p, " ", nsec_to_clock_t(nice)); in show_stat() 147 nice = cpustat[CPUTIME_NICE]; in show_stat() 158 seq_put_decimal_ull(p, " ", nsec_to_clock_t(nice)); in show_stat()
|
/linux-6.12.1/drivers/gpu/drm/omapdrm/ |
D | omap_dmm_tiler.c | 1023 static void text_map(char **map, int xdiv, char *nice, int yd, int x0, int x1) in text_map() argument 1026 int w = (map_width(xdiv, x0, x1) - strlen(nice)) / 2; in text_map() 1029 while (*nice) in text_map() 1030 *p++ = *nice++; in text_map() 1034 static void map_1d_info(char **map, int xdiv, int ydiv, char *nice, in map_1d_info() argument 1037 sprintf(nice, "%dK", tcm_sizeof(*a) * 4); in map_1d_info() 1039 text_map(map, xdiv, nice, (a->p0.y + a->p1.y) / 2 / ydiv, 0, in map_1d_info() 1042 if (strlen(nice) < map_width(xdiv, a->p0.x, 256 - 1)) in map_1d_info() 1043 text_map(map, xdiv, nice, a->p0.y / ydiv, in map_1d_info() 1045 else if (strlen(nice) < map_width(xdiv, 0, a->p1.x)) in map_1d_info() [all …]
|
/linux-6.12.1/Documentation/admin-guide/namespaces/ |
D | resource-control.rst | 9 users programs to play nice this problems becomes more acute. 14 memory user's they don't trust to play nice can use.
|
/linux-6.12.1/Documentation/block/ |
D | ioprio.rst | 9 The io priority feature enables users to io nice processes or process groups, 32 to the cpu nice levels just more coarsely implemented. 0 is the highest 33 BE prio level, 7 is the lowest. The mapping between cpu nice level and io 34 nice level is determined as: io_nice = (cpu_nice + 20) / 5.
|
/linux-6.12.1/drivers/crypto/intel/qat/qat_common/ |
D | icp_qat_hw_20_comp.h | 70 __u16 nice; member 105 QAT_FIELD_SET(val32, csr.nice, in ICP_QAT_FW_COMP_20_BUILD_CONFIG_UPPER()
|
D | adf_gen4_dc.c | 41 hw_comp_upper_csr.nice = ICP_QAT_HW_COMP_20_CONFIG_CSR_NICE_PARAM_DEFAULT_VAL; in qat_comp_build_deflate()
|
/linux-6.12.1/Documentation/fb/ |
D | pvr2fb.rst | 10 * It provides a nice large console (128 cols + 48 lines with 1024x768) 30 SUN12x22 font which is very nice at high resolutions.
|
/linux-6.12.1/Documentation/core-api/ |
D | workqueue.rst | 216 served by worker threads with elevated nice level. 650 pool[00] ref= 1 nice= 0 idle/workers= 4/ 4 cpu= 0 651 pool[01] ref= 1 nice=-20 idle/workers= 2/ 2 cpu= 0 652 pool[02] ref= 1 nice= 0 idle/workers= 4/ 4 cpu= 1 653 pool[03] ref= 1 nice=-20 idle/workers= 2/ 2 cpu= 1 654 pool[04] ref= 1 nice= 0 idle/workers= 4/ 4 cpu= 2 655 pool[05] ref= 1 nice=-20 idle/workers= 2/ 2 cpu= 2 656 pool[06] ref= 1 nice= 0 idle/workers= 3/ 3 cpu= 3 657 pool[07] ref= 1 nice=-20 idle/workers= 2/ 2 cpu= 3 658 pool[08] ref=42 nice= 0 idle/workers= 6/ 6 cpus=0000000f [all …]
|
/linux-6.12.1/drivers/parport/ |
D | TODO-parport | 9 a) ECP support would be nice. This can only work if both the port and
|
/linux-6.12.1/drivers/leds/ |
D | TODO | 67 It would be also nice to have useful listing mode -- name, type, 74 ethernet interface would be nice.
|
/linux-6.12.1/kernel/ |
D | workqueue.c | 1205 int high = pool->attrs->nice == HIGHPRI_NICE_LEVEL ? 1 : 0; in bh_pool_irq_work() 1220 if (pool->attrs->nice == HIGHPRI_NICE_LEVEL) in kick_bh_pool() 2747 pool->attrs->nice < 0 ? "H" : ""); in format_worker_id() 2806 set_user_nice(worker->task, pool->attrs->nice); in create_worker() 3637 if (pool->attrs->nice == HIGHPRI_NICE_LEVEL) in drain_dead_softirq_workfn() 3669 if (pool->attrs->nice == HIGHPRI_NICE_LEVEL) in workqueue_softirq_dead() 4626 to->nice = from->nice; in copy_workqueue_attrs() 4657 hash = jhash_1word(attrs->nice, hash); in wqattrs_hash() 4671 if (a->nice != b->nice) in wqattrs_equal() 6140 pool->attrs->nice == HIGHPRI_NICE_LEVEL ? "-hi" : ""); in pr_cont_pool_info() [all …]
|
/linux-6.12.1/Documentation/cpu-freq/ |
D | index.rst | 10 fly. This is a nice method to save battery power, because the lower
|
/linux-6.12.1/Documentation/admin-guide/aoe/ |
D | todo.rst | 9 not been observed, but it would be nice to eliminate any potential for
|