linux/kernel
KaiGai Kohei bb231fe3a5 [PATCH] Fix pacct bug in multithreading case.
I noticed a bug on the process accounting facility.  In multi-threading
process, some data would be recorded incorrectly when the group_leader dies
earlier than one or more threads.  The attached patch fixes this problem.

See below.  'bugacct' is a test program that create a worker thread after 4
seconds sleeping, then the group_leader dies soon.  The worker thread
consume CPU/Memory for 6 seconds, then exit.  We can estimate 10 seconds as
etime and 6 seconds as stime + utime.  This is a sample program which the
group_leader dies earlier than other threads.

The results of same binary execution on different kernel are below.
-- accounted records --------------------
         |   btime  | utime | stime | etime | minflt | majflt |   comm  |
original | 13:16:40 |  0.00 |  0.00 |  6.10 |    171 |      0 | bugacct |
 patched | 13:20:21 |  5.83 |  0.18 | 10.03 |  32776 |      0 | bugacct |
(*) bugacct allocates 128MB memory, thus 128MB / 4KB = 32768 of minflt is
    appropriate.

-- Test results in original kernel ------
$ date; time -p ./bugacct
Tue Mar 28 13:16:36 JST 2006  <- But pacct said btime is 13:16:40
real 10.11                    <- But pacct said etime is 6.10
user 5.96                     <- But pacct said utime is 0.00
sys 0.14                      <- But pacct said stime is 0.00
$
-- Test results in patched kernel -------
$ date; time -p ./bugacct
Tue Mar 28 13:20:21 JST 2006
real 10.04
user 5.83
sys 0.19
$

In the original 2.6.16 kernel, pacct records btime, utime, stime, etime and
minflt incorrectly.  In my opinion, this problem is caused by an assumption
that group_leader dies last.

The following section calculates process running time for etime and btime.
But it means running time of the thread that dies last, not process.  The
start_time of the first thread in the process (group_leader) should be
reduced from uptime to calculate etime and btime correctly.

   ---- do_acct_process() in kernel/acct.c:
   /* calculate run_time in nsec*/
   do_posix_clock_monotonic_gettime(&uptime);
   run_time = (u64)uptime.tv_sec*NSEC_PER_SEC + uptime.tv_nsec;
   run_time -= (u64)current->start_time.tv_sec*NSEC_PER_SEC
                                   + current->start_time.tv_nsec;
   ----

The following section calculates stime and utime of the process.
But it might count the utime and stime of the group_leader duplicatly
and ignore the utime and stime of the thread dies last, when one or
more threads remain after group_leader dead.
The ac_utime should be calculated as the sum of the signal->utime
and utime of the thread dies last. The ac_stime should be done also.

   ---- do_acct_process() in kernel/acct.c:
   jiffies = cputime_to_jiffies(cputime_add(current->group_leader->utime,
                                            current->signal->utime));
   ac.ac_utime = encode_comp_t(jiffies_to_AHZ(jiffies));
   jiffies = cputime_to_jiffies(cputime_add(current->group_leader->stime,
                                            current->signal->stime));
   ac.ac_stime = encode_comp_t(jiffies_to_AHZ(jiffies));
   ----

The part of the minflt/majflt calculation has same problem.
This patch solves those problems, I think.

Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-31 12:18:54 -08:00
..
irq [PATCH] warn if free_irq() is called from IRQ context 2006-03-26 08:56:53 -08:00
power [PATCH] Fix suspend with traced tasks 2006-03-31 12:18:50 -08:00
.gitignore gitignore: ignore more generated files 2006-01-03 11:35:26 +01:00
acct.c [PATCH] Fix pacct bug in multithreading case. 2006-03-31 12:18:54 -08:00
audit.c [PATCH] Don't pass boot parameters to argv_init[] 2006-03-31 12:18:53 -08:00
audit.h [PATCH] sem2mutex: audit_netlink_sem 2006-03-20 14:08:55 -05:00
auditfilter.c [PATCH] sem2mutex: audit_netlink_sem 2006-03-20 14:08:55 -05:00
auditsc.c Merge branch 'audit.b3' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/audit-current 2006-03-25 09:24:53 -08:00
capability.c [PATCH] refactor capable() to one implementation, add __capable() helper 2006-03-25 08:22:56 -08:00
compat.c [PATCH] lightweight robust futexes: compat 2006-03-27 08:44:49 -08:00
configs.c update the email address of Randy Dunlap 2006-01-03 13:37:51 +01:00
cpu.c [PATCH] Notifier chain update: API changes 2006-03-27 08:44:50 -08:00
cpuset.c [PATCH] cpuset: remove useless local variable initialization 2006-03-24 07:33:24 -08:00
dma.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
exec_domain.c [PATCH] Fix module refcount leak in __set_personality() 2006-03-24 07:33:30 -08:00
exit.c [PATCH] cleanup __exit_signal->cleanup_sighand path 2006-03-28 18:36:44 -08:00
extable.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
fork.c [PATCH] cleanup __exit_signal->cleanup_sighand path 2006-03-28 18:36:44 -08:00
futex_compat.c [PATCH] compat_sys_futex() warning fix 2006-03-28 09:16:09 -08:00
futex.c [PATCH] lightweight robust futexes updates 2006-03-27 08:44:49 -08:00
hrtimer.c [PATCH] hrtimers: remove data field 2006-03-26 08:57:03 -08:00
intermodule.c [PATCH] missing license tag in intermodule 2006-02-05 11:06:52 -08:00
itimer.c [PATCH] hrtimers: remove data field 2006-03-26 08:57:03 -08:00
kallsyms.c [PATCH] fix missing includes 2005-10-30 17:37:32 -08:00
Kconfig.hz [PATCH] i386: Selectable Frequency of the Timer Interrupt 2005-06-23 09:45:10 -07:00
Kconfig.preempt [PATCH] sched: voluntary kernel preemption 2005-06-25 16:24:45 -07:00
kexec.c [PATCH] move capable() to capability.h 2006-01-11 18:42:13 -08:00
kfifo.c [PATCH] gfp flags annotations - part 1 2005-10-08 15:00:57 -07:00
kmod.c [PATCH] wait_for_helper: trivial style cleanup 2006-03-28 18:36:41 -08:00
kprobes.c [PATCH] kretprobe instance recycled by parent process 2006-03-26 08:57:04 -08:00
ksysfs.c [PATCH] fix build error if CONFIG_SYSFS=n 2006-03-24 07:33:31 -08:00
kthread.c [PATCH] find_task_by_pid() needs tasklist_lock 2006-03-25 08:22:57 -08:00
Makefile [PATCH] lightweight robust futexes: compat 2006-03-27 08:44:49 -08:00
module.c [PATCH] symversion warning fix 2006-03-28 09:16:02 -08:00
mutex-debug.c [PATCH] fix/simplify mutex debugging code 2006-01-11 08:14:16 -08:00
mutex-debug.h [PATCH] mutex subsystem, debugging code 2006-01-09 15:59:20 -08:00
mutex.c [PATCH] mutex: trivial whitespace cleanups 2006-01-10 14:27:59 -08:00
mutex.h [PATCH] mutex subsystem, core 2006-01-09 15:59:19 -08:00
panic.c [PATCH] Notifier chain update: API changes 2006-03-27 08:44:50 -08:00
params.c [PATCH] Change dash2underscore() return value to char 2006-03-28 09:16:03 -08:00
pid.c [PATCH] pidhash: don't count idle threads 2006-03-28 18:36:41 -08:00
posix-cpu-timers.c [PATCH] hrtimer: switch clock_nanosleep to hrtimer nanosleep API 2006-01-10 08:01:38 -08:00
posix-timers.c [PATCH] hrtimers: remove data field 2006-03-26 08:57:03 -08:00
printk.c [PATCH] console_setup() depends (wrongly?) on CONFIG_PRINTK 2006-03-24 07:33:27 -08:00
profile.c [PATCH] Notifier chain update: API changes 2006-03-27 08:44:50 -08:00
ptrace.c [PATCH] don't use REMOVE_LINKS/SET_LINKS for reparenting 2006-03-28 18:36:41 -08:00
rcupdate.c [PATCH] rcu_process_callbacks: don't cli() while testing ->nxtlist 2006-03-24 07:33:20 -08:00
rcutorture.c [PATCH] for_each_possible_cpu: fixes for generic part 2006-03-28 09:16:05 -08:00
relay.c [PATCH] relay: consolidate sendfile() and read() code 2006-03-23 19:58:45 +01:00
resource.c [PATCH] kernel/resource.c: __check_region(): remove pointless __deprecated 2006-01-10 08:02:02 -08:00
sched.c [PATCH] for_each_possible_cpu: fixes for generic part 2006-03-28 09:16:05 -08:00
seccomp.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
signal.c [PATCH] Fix suspend with traced tasks 2006-03-31 12:18:50 -08:00
softirq.c [PATCH] on_each_cpu(): disable local interrupts 2006-03-22 07:53:59 -08:00
softlockup.c [PATCH] Notifier chain update: API changes 2006-03-27 08:44:50 -08:00
spinlock.c [PATCH] BUILD_LOCK_OPS: cleanup preempt_disable() usage 2006-03-23 07:38:16 -08:00
stop_machine.c [PATCH] Remove set_fs() in stop_machine() 2006-01-10 08:01:25 -08:00
sys_ni.c [PATCH] lightweight robust futexes: core 2006-03-27 08:44:49 -08:00
sys.c [PATCH] sys_times: don't take tasklist_lock 2006-03-28 18:36:43 -08:00
sysctl.c [PATCH] Range checking in do_proc_dointvec_(userhz_)jiffies_conv 2006-03-24 07:33:20 -08:00
time.c [PATCH] hrtimers: remove nsec_t typedef 2006-03-26 08:57:03 -08:00
timer.c [PATCH] __mod_timer: simplify ->base changing 2006-03-31 12:18:53 -08:00
uid16.c [PATCH] move capable() to capability.h 2006-01-11 18:42:13 -08:00
user.c [PATCH] free_uid() locking improvement 2006-03-24 07:33:20 -08:00
wait.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
workqueue.c [SCSI] add execute_in_process_context() API 2006-02-27 23:34:40 -06:00