diff options
Diffstat (limited to 'kernel/posix-cpu-timers.c')
| -rw-r--r-- | kernel/posix-cpu-timers.c | 512 |
1 files changed, 260 insertions, 252 deletions
diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index c42a03aef36f..153dcb2639c3 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c | |||
| @@ -7,6 +7,93 @@ | |||
| 7 | #include <linux/errno.h> | 7 | #include <linux/errno.h> |
| 8 | #include <linux/math64.h> | 8 | #include <linux/math64.h> |
| 9 | #include <asm/uaccess.h> | 9 | #include <asm/uaccess.h> |
| 10 | #include <linux/kernel_stat.h> | ||
| 11 | |||
| 12 | /* | ||
| 13 | * Allocate the thread_group_cputime structure appropriately and fill in the | ||
| 14 | * current values of the fields. Called from copy_signal() via | ||
| 15 | * thread_group_cputime_clone_thread() when adding a second or subsequent | ||
| 16 | * thread to a thread group. Assumes interrupts are enabled when called. | ||
| 17 | */ | ||
| 18 | int thread_group_cputime_alloc(struct task_struct *tsk) | ||
| 19 | { | ||
| 20 | struct signal_struct *sig = tsk->signal; | ||
| 21 | struct task_cputime *cputime; | ||
| 22 | |||
| 23 | /* | ||
| 24 | * If we have multiple threads and we don't already have a | ||
| 25 | * per-CPU task_cputime struct (checked in the caller), allocate | ||
| 26 | * one and fill it in with the times accumulated so far. We may | ||
| 27 | * race with another thread so recheck after we pick up the sighand | ||
| 28 | * lock. | ||
| 29 | */ | ||
| 30 | cputime = alloc_percpu(struct task_cputime); | ||
| 31 | if (cputime == NULL) | ||
| 32 | return -ENOMEM; | ||
| 33 | spin_lock_irq(&tsk->sighand->siglock); | ||
| 34 | if (sig->cputime.totals) { | ||
| 35 | spin_unlock_irq(&tsk->sighand->siglock); | ||
| 36 | free_percpu(cputime); | ||
| 37 | return 0; | ||
| 38 | } | ||
| 39 | sig->cputime.totals = cputime; | ||
| 40 | cputime = per_cpu_ptr(sig->cputime.totals, smp_processor_id()); | ||
| 41 | cputime->utime = tsk->utime; | ||
| 42 | cputime->stime = tsk->stime; | ||
| 43 | cputime->sum_exec_runtime = tsk->se.sum_exec_runtime; | ||
| 44 | spin_unlock_irq(&tsk->sighand->siglock); | ||
| 45 | return 0; | ||
| 46 | } | ||
| 47 | |||
| 48 | /** | ||
| 49 | * thread_group_cputime - Sum the thread group time fields across all CPUs. | ||
| 50 | * | ||
| 51 | * @tsk: The task we use to identify the thread group. | ||
| 52 | * @times: task_cputime structure in which we return the summed fields. | ||
| 53 | * | ||
| 54 | * Walk the list of CPUs to sum the per-CPU time fields in the thread group | ||
| 55 | * time structure. | ||
| 56 | */ | ||
| 57 | void thread_group_cputime( | ||
| 58 | struct task_struct *tsk, | ||
| 59 | struct task_cputime *times) | ||
| 60 | { | ||
| 61 | struct signal_struct *sig; | ||
| 62 | int i; | ||
| 63 | struct task_cputime *tot; | ||
| 64 | |||
| 65 | sig = tsk->signal; | ||
| 66 | if (unlikely(!sig) || !sig->cputime.totals) { | ||
| 67 | times->utime = tsk->utime; | ||
| 68 | times->stime = tsk->stime; | ||
| 69 | times->sum_exec_runtime = tsk->se.sum_exec_runtime; | ||
| 70 | return; | ||
| 71 | } | ||
| 72 | times->stime = times->utime = cputime_zero; | ||
| 73 | times->sum_exec_runtime = 0; | ||
| 74 | for_each_possible_cpu(i) { | ||
| 75 | tot = per_cpu_ptr(tsk->signal->cputime.totals, i); | ||
| 76 | times->utime = cputime_add(times->utime, tot->utime); | ||
| 77 | times->stime = cputime_add(times->stime, tot->stime); | ||
| 78 | times->sum_exec_runtime += tot->sum_exec_runtime; | ||
| 79 | } | ||
| 80 | } | ||
| 81 | |||
| 82 | /* | ||
| 83 | * Called after updating RLIMIT_CPU to set timer expiration if necessary. | ||
| 84 | */ | ||
| 85 | void update_rlimit_cpu(unsigned long rlim_new) | ||
| 86 | { | ||
| 87 | cputime_t cputime; | ||
| 88 | |||
| 89 | cputime = secs_to_cputime(rlim_new); | ||
| 90 | if (cputime_eq(current->signal->it_prof_expires, cputime_zero) || | ||
| 91 | cputime_lt(current->signal->it_prof_expires, cputime)) { | ||
| 92 | spin_lock_irq(¤t->sighand->siglock); | ||
| 93 | set_process_cpu_timer(current, CPUCLOCK_PROF, &cputime, NULL); | ||
| 94 | spin_unlock_irq(¤t->sighand->siglock); | ||
| 95 | } | ||
| 96 | } | ||
| 10 | 97 | ||
| 11 | static int check_clock(const clockid_t which_clock) | 98 | static int check_clock(const clockid_t which_clock) |
| 12 | { | 99 | { |
| @@ -158,10 +245,6 @@ static inline cputime_t virt_ticks(struct task_struct *p) | |||
| 158 | { | 245 | { |
| 159 | return p->utime; | 246 | return p->utime; |
| 160 | } | 247 | } |
| 161 | static inline unsigned long long sched_ns(struct task_struct *p) | ||
| 162 | { | ||
| 163 | return task_sched_runtime(p); | ||
| 164 | } | ||
| 165 | 248 | ||
| 166 | int posix_cpu_clock_getres(const clockid_t which_clock, struct timespec *tp) | 249 | int posix_cpu_clock_getres(const clockid_t which_clock, struct timespec *tp) |
| 167 | { | 250 | { |
| @@ -211,7 +294,7 @@ static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p, | |||
| 211 | cpu->cpu = virt_ticks(p); | 294 | cpu->cpu = virt_ticks(p); |
| 212 | break; | 295 | break; |
| 213 | case CPUCLOCK_SCHED: | 296 | case CPUCLOCK_SCHED: |
| 214 | cpu->sched = sched_ns(p); | 297 | cpu->sched = p->se.sum_exec_runtime + task_delta_exec(p); |
| 215 | break; | 298 | break; |
| 216 | } | 299 | } |
| 217 | return 0; | 300 | return 0; |
| @@ -220,59 +303,30 @@ static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p, | |||
| 220 | /* | 303 | /* |
| 221 | * Sample a process (thread group) clock for the given group_leader task. | 304 | * Sample a process (thread group) clock for the given group_leader task. |
| 222 | * Must be called with tasklist_lock held for reading. | 305 | * Must be called with tasklist_lock held for reading. |
| 223 | * Must be called with tasklist_lock held for reading, and p->sighand->siglock. | ||
| 224 | */ | 306 | */ |
| 225 | static int cpu_clock_sample_group_locked(unsigned int clock_idx, | 307 | static int cpu_clock_sample_group(const clockid_t which_clock, |
| 226 | struct task_struct *p, | 308 | struct task_struct *p, |
| 227 | union cpu_time_count *cpu) | 309 | union cpu_time_count *cpu) |
| 228 | { | 310 | { |
| 229 | struct task_struct *t = p; | 311 | struct task_cputime cputime; |
| 230 | switch (clock_idx) { | 312 | |
| 313 | thread_group_cputime(p, &cputime); | ||
| 314 | switch (which_clock) { | ||
| 231 | default: | 315 | default: |
| 232 | return -EINVAL; | 316 | return -EINVAL; |
| 233 | case CPUCLOCK_PROF: | 317 | case CPUCLOCK_PROF: |
| 234 | cpu->cpu = cputime_add(p->signal->utime, p->signal->stime); | 318 | cpu->cpu = cputime_add(cputime.utime, cputime.stime); |
| 235 | do { | ||
| 236 | cpu->cpu = cputime_add(cpu->cpu, prof_ticks(t)); | ||
| 237 | t = next_thread(t); | ||
| 238 | } while (t != p); | ||
| 239 | break; | 319 | break; |
| 240 | case CPUCLOCK_VIRT: | 320 | case CPUCLOCK_VIRT: |
| 241 | cpu->cpu = p->signal->utime; | 321 | cpu->cpu = cputime.utime; |
| 242 | do { | ||
| 243 | cpu->cpu = cputime_add(cpu->cpu, virt_ticks(t)); | ||
| 244 | t = next_thread(t); | ||
| 245 | } while (t != p); | ||
| 246 | break; | 322 | break; |
| 247 | case CPUCLOCK_SCHED: | 323 | case CPUCLOCK_SCHED: |
| 248 | cpu->sched = p->signal->sum_sched_runtime; | 324 | cpu->sched = cputime.sum_exec_runtime + task_delta_exec(p); |
| 249 | /* Add in each other live thread. */ | ||
| 250 | while ((t = next_thread(t)) != p) { | ||
| 251 | cpu->sched += t->se.sum_exec_runtime; | ||
| 252 | } | ||
| 253 | cpu->sched += sched_ns(p); | ||
| 254 | break; | 325 | break; |
| 255 | } | 326 | } |
| 256 | return 0; | 327 | return 0; |
| 257 | } | 328 | } |
| 258 | 329 | ||
| 259 | /* | ||
| 260 | * Sample a process (thread group) clock for the given group_leader task. | ||
| 261 | * Must be called with tasklist_lock held for reading. | ||
| 262 | */ | ||
| 263 | static int cpu_clock_sample_group(const clockid_t which_clock, | ||
| 264 | struct task_struct *p, | ||
| 265 | union cpu_time_count *cpu) | ||
| 266 | { | ||
| 267 | int ret; | ||
| 268 | unsigned long flags; | ||
| 269 | spin_lock_irqsave(&p->sighand->siglock, flags); | ||
| 270 | ret = cpu_clock_sample_group_locked(CPUCLOCK_WHICH(which_clock), p, | ||
| 271 | cpu); | ||
| 272 | spin_unlock_irqrestore(&p->sighand->siglock, flags); | ||
| 273 | return ret; | ||
| 274 | } | ||
| 275 | |||
| 276 | 330 | ||
| 277 | int posix_cpu_clock_get(const clockid_t which_clock, struct timespec *tp) | 331 | int posix_cpu_clock_get(const clockid_t which_clock, struct timespec *tp) |
| 278 | { | 332 | { |
| @@ -471,80 +525,11 @@ void posix_cpu_timers_exit(struct task_struct *tsk) | |||
| 471 | } | 525 | } |
| 472 | void posix_cpu_timers_exit_group(struct task_struct *tsk) | 526 | void posix_cpu_timers_exit_group(struct task_struct *tsk) |
| 473 | { | 527 | { |
| 474 | cleanup_timers(tsk->signal->cpu_timers, | 528 | struct task_cputime cputime; |
| 475 | cputime_add(tsk->utime, tsk->signal->utime), | ||
| 476 | cputime_add(tsk->stime, tsk->signal->stime), | ||
| 477 | tsk->se.sum_exec_runtime + tsk->signal->sum_sched_runtime); | ||
| 478 | } | ||
| 479 | 529 | ||
| 480 | 530 | thread_group_cputime(tsk, &cputime); | |
| 481 | /* | 531 | cleanup_timers(tsk->signal->cpu_timers, |
| 482 | * Set the expiry times of all the threads in the process so one of them | 532 | cputime.utime, cputime.stime, cputime.sum_exec_runtime); |
| 483 | * will go off before the process cumulative expiry total is reached. | ||
| 484 | */ | ||
| 485 | static void process_timer_rebalance(struct task_struct *p, | ||
| 486 | unsigned int clock_idx, | ||
| 487 | union cpu_time_count expires, | ||
| 488 | union cpu_time_count val) | ||
| 489 | { | ||
| 490 | cputime_t ticks, left; | ||
| 491 | unsigned long long ns, nsleft; | ||
| 492 | struct task_struct *t = p; | ||
| 493 | unsigned int nthreads = atomic_read(&p->signal->live); | ||
| 494 | |||
| 495 | if (!nthreads) | ||
| 496 | return; | ||
| 497 | |||
| 498 | switch (clock_idx) { | ||
| 499 | default: | ||
| 500 | BUG(); | ||
| 501 | break; | ||
| 502 | case CPUCLOCK_PROF: | ||
| 503 | left = cputime_div_non_zero(cputime_sub(expires.cpu, val.cpu), | ||
| 504 | nthreads); | ||
| 505 | do { | ||
| 506 | if (likely(!(t->flags & PF_EXITING))) { | ||
| 507 | ticks = cputime_add(prof_ticks(t), left); | ||
| 508 | if (cputime_eq(t->it_prof_expires, | ||
| 509 | cputime_zero) || | ||
| 510 | cputime_gt(t->it_prof_expires, ticks)) { | ||
| 511 | t->it_prof_expires = ticks; | ||
| 512 | } | ||
| 513 | } | ||
| 514 | t = next_thread(t); | ||
| 515 | } while (t != p); | ||
| 516 | break; | ||
| 517 | case CPUCLOCK_VIRT: | ||
| 518 | left = cputime_div_non_zero(cputime_sub(expires.cpu, val.cpu), | ||
| 519 | nthreads); | ||
| 520 | do { | ||
| 521 | if (likely(!(t->flags & PF_EXITING))) { | ||
| 522 | ticks = cputime_add(virt_ticks(t), left); | ||
| 523 | if (cputime_eq(t->it_virt_expires, | ||
| 524 | cputime_zero) || | ||
| 525 | cputime_gt(t->it_virt_expires, ticks)) { | ||
| 526 | t->it_virt_expires = ticks; | ||
| 527 | } | ||
| 528 | } | ||
| 529 | t = next_thread(t); | ||
| 530 | } while (t != p); | ||
| 531 | break; | ||
| 532 | case CPUCLOCK_SCHED: | ||
| 533 | nsleft = expires.sched - val.sched; | ||
| 534 | do_div(nsleft, nthreads); | ||
| 535 | nsleft = max_t(unsigned long long, nsleft, 1); | ||
| 536 | do { | ||
| 537 | if (likely(!(t->flags & PF_EXITING))) { | ||
| 538 | ns = t->se.sum_exec_runtime + nsleft; | ||
| 539 | if (t->it_sched_expires == 0 || | ||
| 540 | t->it_sched_expires > ns) { | ||
| 541 | t->it_sched_expires = ns; | ||
| 542 | } | ||
| 543 | } | ||
| 544 | t = next_thread(t); | ||
| 545 | } while (t != p); | ||
| 546 | break; | ||
| 547 | } | ||
| 548 | } | 533 | } |
| 549 | 534 | ||
| 550 | static void clear_dead_task(struct k_itimer *timer, union cpu_time_count now) | 535 | static void clear_dead_task(struct k_itimer *timer, union cpu_time_count now) |
| @@ -608,29 +593,32 @@ static void arm_timer(struct k_itimer *timer, union cpu_time_count now) | |||
| 608 | default: | 593 | default: |
| 609 | BUG(); | 594 | BUG(); |
| 610 | case CPUCLOCK_PROF: | 595 | case CPUCLOCK_PROF: |
| 611 | if (cputime_eq(p->it_prof_expires, | 596 | if (cputime_eq(p->cputime_expires.prof_exp, |
| 612 | cputime_zero) || | 597 | cputime_zero) || |
| 613 | cputime_gt(p->it_prof_expires, | 598 | cputime_gt(p->cputime_expires.prof_exp, |
| 614 | nt->expires.cpu)) | 599 | nt->expires.cpu)) |
| 615 | p->it_prof_expires = nt->expires.cpu; | 600 | p->cputime_expires.prof_exp = |
| 601 | nt->expires.cpu; | ||
| 616 | break; | 602 | break; |
| 617 | case CPUCLOCK_VIRT: | 603 | case CPUCLOCK_VIRT: |
| 618 | if (cputime_eq(p->it_virt_expires, | 604 | if (cputime_eq(p->cputime_expires.virt_exp, |
| 619 | cputime_zero) || | 605 | cputime_zero) || |
| 620 | cputime_gt(p->it_virt_expires, | 606 | cputime_gt(p->cputime_expires.virt_exp, |
| 621 | nt->expires.cpu)) | 607 | nt->expires.cpu)) |
| 622 | p->it_virt_expires = nt->expires.cpu; | 608 | p->cputime_expires.virt_exp = |
| 609 | nt->expires.cpu; | ||
| 623 | break; | 610 | break; |
| 624 | case CPUCLOCK_SCHED: | 611 | case CPUCLOCK_SCHED: |
| 625 | if (p->it_sched_expires == 0 || | 612 | if (p->cputime_expires.sched_exp == 0 || |
| 626 | p->it_sched_expires > nt->expires.sched) | 613 | p->cputime_expires.sched_exp > |
| 627 | p->it_sched_expires = nt->expires.sched; | 614 | nt->expires.sched) |
| 615 | p->cputime_expires.sched_exp = | ||
| 616 | nt->expires.sched; | ||
| 628 | break; | 617 | break; |
| 629 | } | 618 | } |
| 630 | } else { | 619 | } else { |
| 631 | /* | 620 | /* |
| 632 | * For a process timer, we must balance | 621 | * For a process timer, set the cached expiration time. |
| 633 | * all the live threads' expirations. | ||
| 634 | */ | 622 | */ |
| 635 | switch (CPUCLOCK_WHICH(timer->it_clock)) { | 623 | switch (CPUCLOCK_WHICH(timer->it_clock)) { |
| 636 | default: | 624 | default: |
| @@ -641,7 +629,9 @@ static void arm_timer(struct k_itimer *timer, union cpu_time_count now) | |||
| 641 | cputime_lt(p->signal->it_virt_expires, | 629 | cputime_lt(p->signal->it_virt_expires, |
| 642 | timer->it.cpu.expires.cpu)) | 630 | timer->it.cpu.expires.cpu)) |
| 643 | break; | 631 | break; |
| 644 | goto rebalance; | 632 | p->signal->cputime_expires.virt_exp = |
| 633 | timer->it.cpu.expires.cpu; | ||
| 634 | break; | ||
| 645 | case CPUCLOCK_PROF: | 635 | case CPUCLOCK_PROF: |
| 646 | if (!cputime_eq(p->signal->it_prof_expires, | 636 | if (!cputime_eq(p->signal->it_prof_expires, |
| 647 | cputime_zero) && | 637 | cputime_zero) && |
| @@ -652,13 +642,12 @@ static void arm_timer(struct k_itimer *timer, union cpu_time_count now) | |||
| 652 | if (i != RLIM_INFINITY && | 642 | if (i != RLIM_INFINITY && |
| 653 | i <= cputime_to_secs(timer->it.cpu.expires.cpu)) | 643 | i <= cputime_to_secs(timer->it.cpu.expires.cpu)) |
| 654 | break; | 644 | break; |
| 655 | goto rebalance; | 645 | p->signal->cputime_expires.prof_exp = |
| 646 | timer->it.cpu.expires.cpu; | ||
| 647 | break; | ||
| 656 | case CPUCLOCK_SCHED: | 648 | case CPUCLOCK_SCHED: |
| 657 | rebalance: | 649 | p->signal->cputime_expires.sched_exp = |
| 658 | process_timer_rebalance( | 650 | timer->it.cpu.expires.sched; |
| 659 | timer->it.cpu.task, | ||
| 660 | CPUCLOCK_WHICH(timer->it_clock), | ||
| 661 | timer->it.cpu.expires, now); | ||
| 662 | break; | 651 | break; |
| 663 | } | 652 | } |
| 664 | } | 653 | } |
| @@ -969,13 +958,13 @@ static void check_thread_timers(struct task_struct *tsk, | |||
| 969 | struct signal_struct *const sig = tsk->signal; | 958 | struct signal_struct *const sig = tsk->signal; |
| 970 | 959 | ||
| 971 | maxfire = 20; | 960 | maxfire = 20; |
| 972 | tsk->it_prof_expires = cputime_zero; | 961 | tsk->cputime_expires.prof_exp = cputime_zero; |
| 973 | while (!list_empty(timers)) { | 962 | while (!list_empty(timers)) { |
| 974 | struct cpu_timer_list *t = list_first_entry(timers, | 963 | struct cpu_timer_list *t = list_first_entry(timers, |
| 975 | struct cpu_timer_list, | 964 | struct cpu_timer_list, |
| 976 | entry); | 965 | entry); |
| 977 | if (!--maxfire || cputime_lt(prof_ticks(tsk), t->expires.cpu)) { | 966 | if (!--maxfire || cputime_lt(prof_ticks(tsk), t->expires.cpu)) { |
| 978 | tsk->it_prof_expires = t->expires.cpu; | 967 | tsk->cputime_expires.prof_exp = t->expires.cpu; |
| 979 | break; | 968 | break; |
| 980 | } | 969 | } |
| 981 | t->firing = 1; | 970 | t->firing = 1; |
| @@ -984,13 +973,13 @@ static void check_thread_timers(struct task_struct *tsk, | |||
| 984 | 973 | ||
| 985 | ++timers; | 974 | ++timers; |
| 986 | maxfire = 20; | 975 | maxfire = 20; |
| 987 | tsk->it_virt_expires = cputime_zero; | 976 | tsk->cputime_expires.virt_exp = cputime_zero; |
| 988 | while (!list_empty(timers)) { | 977 | while (!list_empty(timers)) { |
| 989 | struct cpu_timer_list *t = list_first_entry(timers, | 978 | struct cpu_timer_list *t = list_first_entry(timers, |
| 990 | struct cpu_timer_list, | 979 | struct cpu_timer_list, |
| 991 | entry); | 980 | entry); |
| 992 | if (!--maxfire || cputime_lt(virt_ticks(tsk), t->expires.cpu)) { | 981 | if (!--maxfire || cputime_lt(virt_ticks(tsk), t->expires.cpu)) { |
| 993 | tsk->it_virt_expires = t->expires.cpu; | 982 | tsk->cputime_expires.virt_exp = t->expires.cpu; |
| 994 | break; | 983 | break; |
| 995 | } | 984 | } |
| 996 | t->firing = 1; | 985 | t->firing = 1; |
| @@ -999,13 +988,13 @@ static void check_thread_timers(struct task_struct *tsk, | |||
| 999 | 988 | ||
| 1000 | ++timers; | 989 | ++timers; |
| 1001 | maxfire = 20; | 990 | maxfire = 20; |
| 1002 | tsk->it_sched_expires = 0; | 991 | tsk->cputime_expires.sched_exp = 0; |
| 1003 | while (!list_empty(timers)) { | 992 | while (!list_empty(timers)) { |
| 1004 | struct cpu_timer_list *t = list_first_entry(timers, | 993 | struct cpu_timer_list *t = list_first_entry(timers, |
| 1005 | struct cpu_timer_list, | 994 | struct cpu_timer_list, |
| 1006 | entry); | 995 | entry); |
| 1007 | if (!--maxfire || tsk->se.sum_exec_runtime < t->expires.sched) { | 996 | if (!--maxfire || tsk->se.sum_exec_runtime < t->expires.sched) { |
| 1008 | tsk->it_sched_expires = t->expires.sched; | 997 | tsk->cputime_expires.sched_exp = t->expires.sched; |
| 1009 | break; | 998 | break; |
| 1010 | } | 999 | } |
| 1011 | t->firing = 1; | 1000 | t->firing = 1; |
| @@ -1055,10 +1044,10 @@ static void check_process_timers(struct task_struct *tsk, | |||
| 1055 | { | 1044 | { |
| 1056 | int maxfire; | 1045 | int maxfire; |
| 1057 | struct signal_struct *const sig = tsk->signal; | 1046 | struct signal_struct *const sig = tsk->signal; |
| 1058 | cputime_t utime, stime, ptime, virt_expires, prof_expires; | 1047 | cputime_t utime, ptime, virt_expires, prof_expires; |
| 1059 | unsigned long long sum_sched_runtime, sched_expires; | 1048 | unsigned long long sum_sched_runtime, sched_expires; |
| 1060 | struct task_struct *t; | ||
| 1061 | struct list_head *timers = sig->cpu_timers; | 1049 | struct list_head *timers = sig->cpu_timers; |
| 1050 | struct task_cputime cputime; | ||
| 1062 | 1051 | ||
| 1063 | /* | 1052 | /* |
| 1064 | * Don't sample the current process CPU clocks if there are no timers. | 1053 | * Don't sample the current process CPU clocks if there are no timers. |
| @@ -1074,18 +1063,10 @@ static void check_process_timers(struct task_struct *tsk, | |||
| 1074 | /* | 1063 | /* |
| 1075 | * Collect the current process totals. | 1064 | * Collect the current process totals. |
| 1076 | */ | 1065 | */ |
| 1077 | utime = sig->utime; | 1066 | thread_group_cputime(tsk, &cputime); |
| 1078 | stime = sig->stime; | 1067 | utime = cputime.utime; |
| 1079 | sum_sched_runtime = sig->sum_sched_runtime; | 1068 | ptime = cputime_add(utime, cputime.stime); |
| 1080 | t = tsk; | 1069 | sum_sched_runtime = cputime.sum_exec_runtime; |
| 1081 | do { | ||
| 1082 | utime = cputime_add(utime, t->utime); | ||
| 1083 | stime = cputime_add(stime, t->stime); | ||
| 1084 | sum_sched_runtime += t->se.sum_exec_runtime; | ||
| 1085 | t = next_thread(t); | ||
| 1086 | } while (t != tsk); | ||
| 1087 | ptime = cputime_add(utime, stime); | ||
| 1088 | |||
| 1089 | maxfire = 20; | 1070 | maxfire = 20; |
| 1090 | prof_expires = cputime_zero; | 1071 | prof_expires = cputime_zero; |
| 1091 | while (!list_empty(timers)) { | 1072 | while (!list_empty(timers)) { |
| @@ -1193,60 +1174,18 @@ static void check_process_timers(struct task_struct *tsk, | |||
| 1193 | } | 1174 | } |
| 1194 | } | 1175 | } |
| 1195 | 1176 | ||
| 1196 | if (!cputime_eq(prof_expires, cputime_zero) || | 1177 | if (!cputime_eq(prof_expires, cputime_zero) && |
| 1197 | !cputime_eq(virt_expires, cputime_zero) || | 1178 | (cputime_eq(sig->cputime_expires.prof_exp, cputime_zero) || |
| 1198 | sched_expires != 0) { | 1179 | cputime_gt(sig->cputime_expires.prof_exp, prof_expires))) |
| 1199 | /* | 1180 | sig->cputime_expires.prof_exp = prof_expires; |
| 1200 | * Rebalance the threads' expiry times for the remaining | 1181 | if (!cputime_eq(virt_expires, cputime_zero) && |
| 1201 | * process CPU timers. | 1182 | (cputime_eq(sig->cputime_expires.virt_exp, cputime_zero) || |
| 1202 | */ | 1183 | cputime_gt(sig->cputime_expires.virt_exp, virt_expires))) |
| 1203 | 1184 | sig->cputime_expires.virt_exp = virt_expires; | |
| 1204 | cputime_t prof_left, virt_left, ticks; | 1185 | if (sched_expires != 0 && |
| 1205 | unsigned long long sched_left, sched; | 1186 | (sig->cputime_expires.sched_exp == 0 || |
| 1206 | const unsigned int nthreads = atomic_read(&sig->live); | 1187 | sig->cputime_expires.sched_exp > sched_expires)) |
| 1207 | 1188 | sig->cputime_expires.sched_exp = sched_expires; | |
| 1208 | if (!nthreads) | ||
| 1209 | return; | ||
| 1210 | |||
| 1211 | prof_left = cputime_sub(prof_expires, utime); | ||
| 1212 | prof_left = cputime_sub(prof_left, stime); | ||
| 1213 | prof_left = cputime_div_non_zero(prof_left, nthreads); | ||
| 1214 | virt_left = cputime_sub(virt_expires, utime); | ||
| 1215 | virt_left = cputime_div_non_zero(virt_left, nthreads); | ||
| 1216 | if (sched_expires) { | ||
| 1217 | sched_left = sched_expires - sum_sched_runtime; | ||
| 1218 | do_div(sched_left, nthreads); | ||
| 1219 | sched_left = max_t(unsigned long long, sched_left, 1); | ||
| 1220 | } else { | ||
| 1221 | sched_left = 0; | ||
| 1222 | } | ||
| 1223 | t = tsk; | ||
| 1224 | do { | ||
| 1225 | if (unlikely(t->flags & PF_EXITING)) | ||
| 1226 | continue; | ||
| 1227 | |||
| 1228 | ticks = cputime_add(cputime_add(t->utime, t->stime), | ||
| 1229 | prof_left); | ||
| 1230 | if (!cputime_eq(prof_expires, cputime_zero) && | ||
| 1231 | (cputime_eq(t->it_prof_expires, cputime_zero) || | ||
| 1232 | cputime_gt(t->it_prof_expires, ticks))) { | ||
| 1233 | t->it_prof_expires = ticks; | ||
| 1234 | } | ||
| 1235 | |||
| 1236 | ticks = cputime_add(t->utime, virt_left); | ||
| 1237 | if (!cputime_eq(virt_expires, cputime_zero) && | ||
| 1238 | (cputime_eq(t->it_virt_expires, cputime_zero) || | ||
| 1239 | cputime_gt(t->it_virt_expires, ticks))) { | ||
| 1240 | t->it_virt_expires = ticks; | ||
| 1241 | } | ||
| 1242 | |||
| 1243 | sched = t->se.sum_exec_runtime + sched_left; | ||
| 1244 | if (sched_expires && (t->it_sched_expires == 0 || | ||
| 1245 | t->it_sched_expires > sched)) { | ||
| 1246 | t->it_sched_expires = sched; | ||
| 1247 | } | ||
| 1248 | } while ((t = next_thread(t)) != tsk); | ||
| 1249 | } | ||
| 1250 | } | 1189 | } |
| 1251 | 1190 | ||
| 1252 | /* | 1191 | /* |
| @@ -1314,6 +1253,86 @@ out: | |||
| 1314 | ++timer->it_requeue_pending; | 1253 | ++timer->it_requeue_pending; |
| 1315 | } | 1254 | } |
| 1316 | 1255 | ||
| 1256 | /** | ||
| 1257 | * task_cputime_zero - Check a task_cputime struct for all zero fields. | ||
| 1258 | * | ||
| 1259 | * @cputime: The struct to compare. | ||
| 1260 | * | ||
| 1261 | * Checks @cputime to see if all fields are zero. Returns true if all fields | ||
| 1262 | * are zero, false if any field is nonzero. | ||
| 1263 | */ | ||
| 1264 | static inline int task_cputime_zero(const struct task_cputime *cputime) | ||
| 1265 | { | ||
| 1266 | if (cputime_eq(cputime->utime, cputime_zero) && | ||
| 1267 | cputime_eq(cputime->stime, cputime_zero) && | ||
| 1268 | cputime->sum_exec_runtime == 0) | ||
| 1269 | return 1; | ||
| 1270 | return 0; | ||
| 1271 | } | ||
| 1272 | |||
| 1273 | /** | ||
| 1274 | * task_cputime_expired - Compare two task_cputime entities. | ||
| 1275 | * | ||
| 1276 | * @sample: The task_cputime structure to be checked for expiration. | ||
| 1277 | * @expires: Expiration times, against which @sample will be checked. | ||
| 1278 | * | ||
| 1279 | * Checks @sample against @expires to see if any field of @sample has expired. | ||
| 1280 | * Returns true if any field of the former is greater than the corresponding | ||
| 1281 | * field of the latter if the latter field is set. Otherwise returns false. | ||
| 1282 | */ | ||
| 1283 | static inline int task_cputime_expired(const struct task_cputime *sample, | ||
| 1284 | const struct task_cputime *expires) | ||
| 1285 | { | ||
| 1286 | if (!cputime_eq(expires->utime, cputime_zero) && | ||
| 1287 | cputime_ge(sample->utime, expires->utime)) | ||
| 1288 | return 1; | ||
| 1289 | if (!cputime_eq(expires->stime, cputime_zero) && | ||
| 1290 | cputime_ge(cputime_add(sample->utime, sample->stime), | ||
| 1291 | expires->stime)) | ||
| 1292 | return 1; | ||
| 1293 | if (expires->sum_exec_runtime != 0 && | ||
| 1294 | sample->sum_exec_runtime >= expires->sum_exec_runtime) | ||
| 1295 | return 1; | ||
| 1296 | return 0; | ||
| 1297 | } | ||
| 1298 | |||
| 1299 | /** | ||
| 1300 | * fastpath_timer_check - POSIX CPU timers fast path. | ||
| 1301 | * | ||
| 1302 | * @tsk: The task (thread) being checked. | ||
| 1303 | * | ||
| 1304 | * Check the task and thread group timers. If both are zero (there are no | ||
| 1305 | * timers set) return false. Otherwise snapshot the task and thread group | ||
| 1306 | * timers and compare them with the corresponding expiration times. Return | ||
| 1307 | * true if a timer has expired, else return false. | ||
| 1308 | */ | ||
| 1309 | static inline int fastpath_timer_check(struct task_struct *tsk) | ||
| 1310 | { | ||
| 1311 | struct signal_struct *sig = tsk->signal; | ||
| 1312 | |||
| 1313 | if (unlikely(!sig)) | ||
| 1314 | return 0; | ||
| 1315 | |||
| 1316 | if (!task_cputime_zero(&tsk->cputime_expires)) { | ||
| 1317 | struct task_cputime task_sample = { | ||
| 1318 | .utime = tsk->utime, | ||
| 1319 | .stime = tsk->stime, | ||
| 1320 | .sum_exec_runtime = tsk->se.sum_exec_runtime | ||
| 1321 | }; | ||
| 1322 | |||
| 1323 | if (task_cputime_expired(&task_sample, &tsk->cputime_expires)) | ||
| 1324 | return 1; | ||
| 1325 | } | ||
| 1326 | if (!task_cputime_zero(&sig->cputime_expires)) { | ||
| 1327 | struct task_cputime group_sample; | ||
| 1328 | |||
| 1329 | thread_group_cputime(tsk, &group_sample); | ||
| 1330 | if (task_cputime_expired(&group_sample, &sig->cputime_expires)) | ||
| 1331 | return 1; | ||
| 1332 | } | ||
| 1333 | return 0; | ||
| 1334 | } | ||
| 1335 | |||
| 1317 | /* | 1336 | /* |
| 1318 | * This is called from the timer interrupt handler. The irq handler has | 1337 | * This is called from the timer interrupt handler. The irq handler has |
| 1319 | * already updated our counts. We need to check if any timers fire now. | 1338 | * already updated our counts. We need to check if any timers fire now. |
| @@ -1326,42 +1345,31 @@ void run_posix_cpu_timers(struct task_struct *tsk) | |||
| 1326 | 1345 | ||
| 1327 | BUG_ON(!irqs_disabled()); | 1346 | BUG_ON(!irqs_disabled()); |
| 1328 | 1347 | ||
| 1329 | #define UNEXPIRED(clock) \ | 1348 | /* |
| 1330 | (cputime_eq(tsk->it_##clock##_expires, cputime_zero) || \ | 1349 | * The fast path checks that there are no expired thread or thread |
| 1331 | cputime_lt(clock##_ticks(tsk), tsk->it_##clock##_expires)) | 1350 | * group timers. If that's so, just return. |
| 1332 | 1351 | */ | |
| 1333 | if (UNEXPIRED(prof) && UNEXPIRED(virt) && | 1352 | if (!fastpath_timer_check(tsk)) |
| 1334 | (tsk->it_sched_expires == 0 || | ||
| 1335 | tsk->se.sum_exec_runtime < tsk->it_sched_expires)) | ||
| 1336 | return; | 1353 | return; |
| 1337 | 1354 | ||
| 1338 | #undef UNEXPIRED | 1355 | spin_lock(&tsk->sighand->siglock); |
| 1339 | |||
| 1340 | /* | 1356 | /* |
| 1341 | * Double-check with locks held. | 1357 | * Here we take off tsk->signal->cpu_timers[N] and |
| 1358 | * tsk->cpu_timers[N] all the timers that are firing, and | ||
| 1359 | * put them on the firing list. | ||
| 1342 | */ | 1360 | */ |
| 1343 | read_lock(&tasklist_lock); | 1361 | check_thread_timers(tsk, &firing); |
| 1344 | if (likely(tsk->signal != NULL)) { | 1362 | check_process_timers(tsk, &firing); |
| 1345 | spin_lock(&tsk->sighand->siglock); | ||
| 1346 | 1363 | ||
| 1347 | /* | 1364 | /* |
| 1348 | * Here we take off tsk->cpu_timers[N] and tsk->signal->cpu_timers[N] | 1365 | * We must release these locks before taking any timer's lock. |
| 1349 | * all the timers that are firing, and put them on the firing list. | 1366 | * There is a potential race with timer deletion here, as the |
| 1350 | */ | 1367 | * siglock now protects our private firing list. We have set |
| 1351 | check_thread_timers(tsk, &firing); | 1368 | * the firing flag in each timer, so that a deletion attempt |
| 1352 | check_process_timers(tsk, &firing); | 1369 | * that gets the timer lock before we do will give it up and |
| 1353 | 1370 | * spin until we've taken care of that timer below. | |
| 1354 | /* | 1371 | */ |
| 1355 | * We must release these locks before taking any timer's lock. | 1372 | spin_unlock(&tsk->sighand->siglock); |
| 1356 | * There is a potential race with timer deletion here, as the | ||
| 1357 | * siglock now protects our private firing list. We have set | ||
| 1358 | * the firing flag in each timer, so that a deletion attempt | ||
| 1359 | * that gets the timer lock before we do will give it up and | ||
| 1360 | * spin until we've taken care of that timer below. | ||
| 1361 | */ | ||
| 1362 | spin_unlock(&tsk->sighand->siglock); | ||
| 1363 | } | ||
| 1364 | read_unlock(&tasklist_lock); | ||
| 1365 | 1373 | ||
| 1366 | /* | 1374 | /* |
| 1367 | * Now that all the timers on our list have the firing flag, | 1375 | * Now that all the timers on our list have the firing flag, |
| @@ -1389,10 +1397,9 @@ void run_posix_cpu_timers(struct task_struct *tsk) | |||
| 1389 | 1397 | ||
| 1390 | /* | 1398 | /* |
| 1391 | * Set one of the process-wide special case CPU timers. | 1399 | * Set one of the process-wide special case CPU timers. |
| 1392 | * The tasklist_lock and tsk->sighand->siglock must be held by the caller. | 1400 | * The tsk->sighand->siglock must be held by the caller. |
| 1393 | * The oldval argument is null for the RLIMIT_CPU timer, where *newval is | 1401 | * The *newval argument is relative and we update it to be absolute, *oldval |
| 1394 | * absolute; non-null for ITIMER_*, where *newval is relative and we update | 1402 | * is absolute and we update it to be relative. |
| 1395 | * it to be absolute, *oldval is absolute and we update it to be relative. | ||
| 1396 | */ | 1403 | */ |
| 1397 | void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx, | 1404 | void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx, |
| 1398 | cputime_t *newval, cputime_t *oldval) | 1405 | cputime_t *newval, cputime_t *oldval) |
| @@ -1401,7 +1408,7 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx, | |||
| 1401 | struct list_head *head; | 1408 | struct list_head *head; |
| 1402 | 1409 | ||
| 1403 | BUG_ON(clock_idx == CPUCLOCK_SCHED); | 1410 | BUG_ON(clock_idx == CPUCLOCK_SCHED); |
| 1404 | cpu_clock_sample_group_locked(clock_idx, tsk, &now); | 1411 | cpu_clock_sample_group(clock_idx, tsk, &now); |
| 1405 | 1412 | ||
| 1406 | if (oldval) { | 1413 | if (oldval) { |
| 1407 | if (!cputime_eq(*oldval, cputime_zero)) { | 1414 | if (!cputime_eq(*oldval, cputime_zero)) { |
| @@ -1435,13 +1442,14 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx, | |||
| 1435 | cputime_ge(list_first_entry(head, | 1442 | cputime_ge(list_first_entry(head, |
| 1436 | struct cpu_timer_list, entry)->expires.cpu, | 1443 | struct cpu_timer_list, entry)->expires.cpu, |
| 1437 | *newval)) { | 1444 | *newval)) { |
| 1438 | /* | 1445 | switch (clock_idx) { |
| 1439 | * Rejigger each thread's expiry time so that one will | 1446 | case CPUCLOCK_PROF: |
| 1440 | * notice before we hit the process-cumulative expiry time. | 1447 | tsk->signal->cputime_expires.prof_exp = *newval; |
| 1441 | */ | 1448 | break; |
| 1442 | union cpu_time_count expires = { .sched = 0 }; | 1449 | case CPUCLOCK_VIRT: |
| 1443 | expires.cpu = *newval; | 1450 | tsk->signal->cputime_expires.virt_exp = *newval; |
| 1444 | process_timer_rebalance(tsk, clock_idx, expires, now); | 1451 | break; |
| 1452 | } | ||
| 1445 | } | 1453 | } |
| 1446 | } | 1454 | } |
| 1447 | 1455 | ||
