Skip to content

C
clone-Linux

gsplhtlxg / clone-Linux

通知 2
Star 0
Fork 1
C
clone-Linux

体验新版 GitCode，发现更多精彩内容 >>

切换分支/标签
  1. 09 2月, 2016 1 次提交
    • M
      sched/debug: Make schedstats a runtime tunable that is disabled by default · cb251765
      由 Mel Gorman 提交于 
      schedstats is very useful during debugging and performance tuning but it
incurs overhead to calculate the stats. As such, even though it can be
disabled at build time, it is often enabled as the information is useful.

This patch adds a kernel command-line and sysctl tunable to enable or
disable schedstats on demand (when it's built in). It is disabled
by default as someone who knows they need it can also learn to enable
it when necessary.

The benefits are dependent on how scheduler-intensive the workload is.
If it is then the patch reduces the number of cycles spent calculating
the stats with a small benefit from reducing the cache footprint of the
scheduler.

These measurements were taken from a 48-core 2-socket
machine with Xeon(R) E5-2670 v3 cpus although they were also tested on a
single socket machine 8-core machine with Intel i7-3770 processors.

netperf-tcp
                           4.5.0-rc1             4.5.0-rc1
                             vanilla          nostats-v3r1
Hmean    64         560.45 (  0.00%)      575.98 (  2.77%)
Hmean    128        766.66 (  0.00%)      795.79 (  3.80%)
Hmean    256        950.51 (  0.00%)      981.50 (  3.26%)
Hmean    1024      1433.25 (  0.00%)     1466.51 (  2.32%)
Hmean    2048      2810.54 (  0.00%)     2879.75 (  2.46%)
Hmean    3312      4618.18 (  0.00%)     4682.09 (  1.38%)
Hmean    4096      5306.42 (  0.00%)     5346.39 (  0.75%)
Hmean    8192     10581.44 (  0.00%)    10698.15 (  1.10%)
Hmean    16384    18857.70 (  0.00%)    18937.61 (  0.42%)

Small gains here, UDP_STREAM showed nothing intresting and neither did
the TCP_RR tests. The gains on the 8-core machine were very similar.

tbench4
                                 4.5.0-rc1             4.5.0-rc1
                                   vanilla          nostats-v3r1
Hmean    mb/sec-1         500.85 (  0.00%)      522.43 (  4.31%)
Hmean    mb/sec-2         984.66 (  0.00%)     1018.19 (  3.41%)
Hmean    mb/sec-4        1827.91 (  0.00%)     1847.78 (  1.09%)
Hmean    mb/sec-8        3561.36 (  0.00%)     3611.28 (  1.40%)
Hmean    mb/sec-16       5824.52 (  0.00%)     5929.03 (  1.79%)
Hmean    mb/sec-32      10943.10 (  0.00%)    10802.83 ( -1.28%)
Hmean    mb/sec-64      15950.81 (  0.00%)    16211.31 (  1.63%)
Hmean    mb/sec-128     15302.17 (  0.00%)    15445.11 (  0.93%)
Hmean    mb/sec-256     14866.18 (  0.00%)    15088.73 (  1.50%)
Hmean    mb/sec-512     15223.31 (  0.00%)    15373.69 (  0.99%)
Hmean    mb/sec-1024    14574.25 (  0.00%)    14598.02 (  0.16%)
Hmean    mb/sec-2048    13569.02 (  0.00%)    13733.86 (  1.21%)
Hmean    mb/sec-3072    12865.98 (  0.00%)    13209.23 (  2.67%)

Small gains of 2-4% at low thread counts and otherwise flat.  The
gains on the 8-core machine were slightly different

tbench4 on 8-core i7-3770 single socket machine
Hmean    mb/sec-1        442.59 (  0.00%)      448.73 (  1.39%)
Hmean    mb/sec-2        796.68 (  0.00%)      794.39 ( -0.29%)
Hmean    mb/sec-4       1322.52 (  0.00%)     1343.66 (  1.60%)
Hmean    mb/sec-8       2611.65 (  0.00%)     2694.86 (  3.19%)
Hmean    mb/sec-16      2537.07 (  0.00%)     2609.34 (  2.85%)
Hmean    mb/sec-32      2506.02 (  0.00%)     2578.18 (  2.88%)
Hmean    mb/sec-64      2511.06 (  0.00%)     2569.16 (  2.31%)
Hmean    mb/sec-128     2313.38 (  0.00%)     2395.50 (  3.55%)
Hmean    mb/sec-256     2110.04 (  0.00%)     2177.45 (  3.19%)
Hmean    mb/sec-512     2072.51 (  0.00%)     2053.97 ( -0.89%)

In constract, this shows a relatively steady 2-3% gain at higher thread
counts. Due to the nature of the patch and the type of workload, it's
not a surprise that the result will depend on the CPU used.

hackbench-pipes
                         4.5.0-rc1             4.5.0-rc1
                           vanilla          nostats-v3r1
Amean    1        0.0637 (  0.00%)      0.0660 ( -3.59%)
Amean    4        0.1229 (  0.00%)      0.1181 (  3.84%)
Amean    7        0.1921 (  0.00%)      0.1911 (  0.52%)
Amean    12       0.3117 (  0.00%)      0.2923 (  6.23%)
Amean    21       0.4050 (  0.00%)      0.3899 (  3.74%)
Amean    30       0.4586 (  0.00%)      0.4433 (  3.33%)
Amean    48       0.5910 (  0.00%)      0.5694 (  3.65%)
Amean    79       0.8663 (  0.00%)      0.8626 (  0.43%)
Amean    110      1.1543 (  0.00%)      1.1517 (  0.22%)
Amean    141      1.4457 (  0.00%)      1.4290 (  1.16%)
Amean    172      1.7090 (  0.00%)      1.6924 (  0.97%)
Amean    192      1.9126 (  0.00%)      1.9089 (  0.19%)

Some small gains and losses and while the variance data is not included,
it's close to the noise. The UMA machine did not show anything particularly
different

pipetest
                             4.5.0-rc1             4.5.0-rc1
                               vanilla          nostats-v2r2
Min         Time        4.13 (  0.00%)        3.99 (  3.39%)
1st-qrtle   Time        4.38 (  0.00%)        4.27 (  2.51%)
2nd-qrtle   Time        4.46 (  0.00%)        4.39 (  1.57%)
3rd-qrtle   Time        4.56 (  0.00%)        4.51 (  1.10%)
Max-90%     Time        4.67 (  0.00%)        4.60 (  1.50%)
Max-93%     Time        4.71 (  0.00%)        4.65 (  1.27%)
Max-95%     Time        4.74 (  0.00%)        4.71 (  0.63%)
Max-99%     Time        4.88 (  0.00%)        4.79 (  1.84%)
Max         Time        4.93 (  0.00%)        4.83 (  2.03%)
Mean        Time        4.48 (  0.00%)        4.39 (  1.91%)
Best99%Mean Time        4.47 (  0.00%)        4.39 (  1.91%)
Best95%Mean Time        4.46 (  0.00%)        4.38 (  1.93%)
Best90%Mean Time        4.45 (  0.00%)        4.36 (  1.98%)
Best50%Mean Time        4.36 (  0.00%)        4.25 (  2.49%)
Best10%Mean Time        4.23 (  0.00%)        4.10 (  3.13%)
Best5%Mean  Time        4.19 (  0.00%)        4.06 (  3.20%)
Best1%Mean  Time        4.13 (  0.00%)        4.00 (  3.39%)

Small improvement and similar gains were seen on the UMA machine.

The gain is small but it stands to reason that doing less work in the
scheduler is a good thing. The downside is that the lack of schedstats and
tracepoints may be surprising to experts doing performance analysis until
they find the existence of the schedstats= parameter or schedstats sysctl.
It will be automatically activated for latencytop and sleep profiling to
alleviate the problem. For tracepoints, there is a simple warning as it's
not safe to activate schedstats in the context when it's known the tracepoint
may be wanted but is unavailable.
Signed-off-by: NMel Gorman <mgorman@techsingularity.net>
Reviewed-by: NMatt Fleming <matt@codeblueprint.co.uk>
Reviewed-by: NSrikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <mgalbraith@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1454663316-22048-1-git-send-email-mgorman@techsingularity.netSigned-off-by: NIngo Molnar <mingo@kernel.org>
      cb251765
  3. 21 1月, 2016 2 次提交
  5. 20 1月, 2016 1 次提交
    • W
      pipe: limit the per-user amount of pages allocated in pipes · 759c0114
      由 Willy Tarreau 提交于 
      On no-so-small systems, it is possible for a single process to cause an
OOM condition by filling large pipes with data that are never read. A
typical process filling 4000 pipes with 1 MB of data will use 4 GB of
memory. On small systems it may be tricky to set the pipe max size to
prevent this from happening.

This patch makes it possible to enforce a per-user soft limit above
which new pipes will be limited to a single page, effectively limiting
them to 4 kB each, as well as a hard limit above which no new pipes may
be created for this user. This has the effect of protecting the system
against memory abuse without hurting other users, and still allowing
pipes to work correctly though with less data at once.

The limit are controlled by two new sysctls : pipe-user-pages-soft, and
pipe-user-pages-hard. Both may be disabled by setting them to zero. The
default soft limit allows the default number of FDs per process (1024)
to create pipes of the default size (64kB), thus reaching a limit of 64MB
before starting to create only smaller pipes. With 256 processes limited
to 1024 FDs each, this results in 1024*64kB + (256*1024 - 1024) * 4kB =
1084 MB of memory allocated for a user. The hard limit is disabled by
default to avoid breaking existing applications that make intensive use
of pipes (eg: for splicing).

Reported-by: socketpair@gmail.com
Reported-by: NTetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Mitigates: CVE-2013-4312 (Linux 2.0+)
Suggested-by: NLinus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NWilly Tarreau <w@1wt.eu>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
      759c0114
  7. 11 1月, 2016 1 次提交
  9. 06 1月, 2016 3 次提交
    • J
      sched/core: Move sched_entity::avg into separate cache line · 5a107804
      由 Jiri Olsa 提交于 
      The sched_entity::avg collides with read-mostly sched_entity data.

The perf c2c tool showed many read HITM accesses across
many CPUs for sched_entity's cfs_rq and my_q, while having
at the same time tons of stores for avg.

After placing sched_entity::avg into separate cache line,
the perf bench sched pipe showed around 20 seconds speedup.

NOTE I cut out all perf events except for cycles and
instructions from following output.

Before:
  $ perf stat -r 5 perf bench sched pipe -l 10000000
  # Running 'sched/pipe' benchmark:
  # Executed 10000000 pipe operations between two processes

       Total time: 270.348 [sec]

        27.034805 usecs/op
            36989 ops/sec
   ...

     245,537,074,035      cycles                    #    1.433 GHz
     187,264,548,519      instructions              #    0.77  insns per cycle

       272.653840535 seconds time elapsed           ( +-  1.31% )

After:
  $ perf stat -r 5 perf bench sched pipe -l 10000000
  # Running 'sched/pipe' benchmark:
  # Executed 10000000 pipe operations between two processes

       Total time: 251.076 [sec]

        25.107678 usecs/op
            39828 ops/sec
  ...

     244,573,513,928      cycles                    #    1.572 GHz
     187,409,641,157      instructions              #    0.76  insns per cycle

       251.679315188 seconds time elapsed           ( +-  0.31% )
Signed-off-by: NJiri Olsa <jolsa@kernel.org>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Joe Mario <jmario@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1449606239-28602-1-git-send-email-jolsa@kernel.orgSigned-off-by: NIngo Molnar <mingo@kernel.org>
      5a107804
    • P
      sched/core: Fix unserialized r-m-w scribbling stuff · be958bdc
      由 Peter Zijlstra 提交于 
      Some of the sched bitfieds (notably sched_reset_on_fork) can be set
on other than current, this can cause the r-m-w to race with other
updates.

Since all the sched bits are serialized by scheduler locks, pull them
in a separate word.
Reported-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: akpm@linux-foundation.org
Cc: hannes@cmpxchg.org
Cc: mhocko@kernel.org
Cc: vdavydov@parallels.com
Link: http://lkml.kernel.org/r/20151125150207.GM11639@twins.programming.kicks-ass.netSigned-off-by: NIngo Molnar <mingo@kernel.org>
      be958bdc
    • S
      sched/core: Check tgid in is_global_init() · 570f5241
      由 Sergey Senozhatsky 提交于 
      Our global init task can have sub-threads, so ->pid check is not reliable
enough for is_global_init(), we need to check tgid instead. This has been
spotted by Oleg and a fix was proposed by Richard a long time ago (see the
link below).

Oleg wrote:

  : Because is_global_init() is only true for the main thread of /sbin/init.
  :
  : Just look at oom_unkillable_task(). It tries to not kill init. But, say,
  : select_bad_process() can happily find a sub-thread of is_global_init()
  : and still kill it.

I recently hit the problem in question; re-sending the patch (to the
best of my knowledge it has never been submitted) with updated function
comment. Credit goes to Oleg and Richard.
Suggested-by: NRichard Guy Briggs <rgb@redhat.com>
Reported-by: NOleg Nesterov <oleg@redhat.com>
Signed-off-by: NSergey Senozhatsky <sergey.senozhatsky@gmail.com>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: NSerge Hallyn <serge.hallyn@canonical.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Eric W . Biederman <ebiederm@xmission.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Serge E . Hallyn <serge.hallyn@ubuntu.com>
Cc: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://www.redhat.com/archives/linux-audit/2013-December/msg00086.htmlSigned-off-by: NIngo Molnar <mingo@kernel.org>
      570f5241
  11. 09 12月, 2015 1 次提交
    • T
      watchdog: introduce touch_softlockup_watchdog_sched() · 03e0d461
      由 Tejun Heo 提交于 
      touch_softlockup_watchdog() is used to tell watchdog that scheduler
stall is expected.  One group of usage is from paths where the task
may not be able to yield for a long time such as performing slow PIO
to finicky device and coming out of suspend.  The other is to account
for scheduler and timer going idle.

For scheduler softlockup detection, there's no reason to distinguish
the two cases; however, workqueue lockup detector is planned and it
can use the same signals from the former group while the latter would
spuriously prevent detection.  This patch introduces a new function
touch_softlockup_watchdog_sched() and convert the latter group to call
it instead.  For now, it just calls touch_softlockup_watchdog() and
there's no functional difference.
Signed-off-by: NTejun Heo <tj@kernel.org>
Cc: Ulrich Obergfell <uobergfe@redhat.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
      03e0d461
  13. 04 12月, 2015 2 次提交
  15. 23 11月, 2015 1 次提交
  17. 10 11月, 2015 1 次提交
    • R
      coredump: add DAX filtering for ELF coredumps · 5037835c
      由 Ross Zwisler 提交于 
      Add two new flags to the existing coredump mechanism for ELF files to
allow us to explicitly filter DAX mappings.  This is desirable because
DAX mappings, like hugetlb mappings, have the potential to be very
large.

Update the coredump_filter documentation in
Documentation/filesystems/proc.txt so that it addresses the new DAX
coredump flags.  Also update the documented default value of
coredump_filter to be consistent with the core(5) man page.  The
documentation being updated talks about bit 4, Dump ELF headers, which
is enabled if CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS is turned on in the
kernel config.  This kernel config option defaults to "y" if both ELF
binaries and coredump are enabled.
Signed-off-by: NRoss Zwisler <ross.zwisler@linux.intel.com>
Acked-by: NJeff Moyer <jmoyer@redhat.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>
      5037835c
  19. 07 11月, 2015 3 次提交
  21. 06 11月, 2015 3 次提交
  23. 15 10月, 2015 2 次提交
    • J
      posix_cpu_timer: Reduce unnecessary sighand lock contention · c8d75aa4
      由 Jason Low 提交于 
      It was found while running a database workload on large systems that
significant time was spent trying to acquire the sighand lock.

The issue was that whenever an itimer expired, many threads ended up
simultaneously trying to send the signal. Most of the time, nothing
happened after acquiring the sighand lock because another thread
had just already sent the signal and updated the "next expire" time.
The fastpath_timer_check() didn't help much since the "next expire"
time was updated after the threads exit fastpath_timer_check().

This patch addresses this by having the thread_group_cputimer structure
maintain a boolean to signify when a thread in the group is already
checking for process wide timers, and adds extra logic in the fastpath
to check the boolean.
Signed-off-by: NJason Low <jason.low2@hp.com>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Reviewed-by: NGeorge Spelvin <linux@horizon.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: hideaki.kimura@hpe.com
Cc: terry.rudd@hpe.com
Cc: scott.norton@hpe.com
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1444849677-29330-5-git-send-email-jason.low2@hp.comSigned-off-by: NThomas Gleixner <tglx@linutronix.de>
      c8d75aa4
    • J
      posix_cpu_timer: Convert cputimer->running to bool · d5c373eb
      由 Jason Low 提交于 
      In the next patch in this series, a new field 'checking_timer' will
be added to 'struct thread_group_cputimer'. Both this and the
existing 'running' integer field are just used as boolean values. To
save space in the structure, we can make both of these fields booleans.

This is a preparatory patch to convert the existing running integer
field to a boolean.
Suggested-by: NGeorge Spelvin <linux@horizon.com>
Signed-off-by: NJason Low <jason.low2@hp.com>
Reviewed: George Spelvin <linux@horizon.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: hideaki.kimura@hpe.com
Cc: terry.rudd@hpe.com
Cc: scott.norton@hpe.com
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1444849677-29330-4-git-send-email-jason.low2@hp.comSigned-off-by: NThomas Gleixner <tglx@linutronix.de>
      d5c373eb
  25. 13 10月, 2015 1 次提交
    • A
      bpf: charge user for creation of BPF maps and programs · aaac3ba9
      由 Alexei Starovoitov 提交于 
      since eBPF programs and maps use kernel memory consider it 'locked' memory
from user accounting point of view and charge it against RLIMIT_MEMLOCK limit.
This limit is typically set to 64Kbytes by distros, so almost all
bpf+tracing programs would need to increase it, since they use maps,
but kernel charges maximum map size upfront.
For example the hash map of 1024 elements will be charged as 64Kbyte.
It's inconvenient for current users and changes current behavior for root,
but probably worth doing to be consistent root vs non-root.

Similar accounting logic is done by mmap of perf_event.
Signed-off-by: NAlexei Starovoitov <ast@plumgrid.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      aaac3ba9
  27. 06 10月, 2015 2 次提交
    • P
      sched/core: Create preempt_count invariant · 609ca066
      由 Peter Zijlstra 提交于 
      Assuming units of PREEMPT_DISABLE_OFFSET for preempt_count() numbers.

Now that TASK_DEAD no longer results in preempt_count() == 3 during
scheduling, we will always call context_switch() with preempt_count()
== 2.

However, we don't always end up with preempt_count() == 2 in
finish_task_switch() because new tasks get created with
preempt_count() == 1.

Create FORK_PREEMPT_COUNT and set it to 2 and use that in the right
places. Note that we cannot use INIT_PREEMPT_COUNT as that serves
another purpose (boot).

After this, preempt_count() is invariant across the context switch,
with exception of PREEMPT_ACTIVE.
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NIngo Molnar <mingo@kernel.org>
      609ca066
    • P
      sched/core: Simplify INIT_PREEMPT_COUNT · 87dcbc06
      由 Peter Zijlstra 提交于 
      As per the following commit:

  d86ee480 ("sched: optimize cond_resched()")

we need PREEMPT_ACTIVE to avoid cond_resched() from working before
the scheduler is set up.

However, keeping preemption disabled should do the same thing
already, making the PREEMPT_ACTIVE part entirely redundant.

The only complication is !PREEMPT_COUNT kernels, where
PREEMPT_DISABLED ends up being 0. Instead we use an unconditional
PREEMPT_OFFSET to set preempt_count() even on !PREEMPT_COUNT
kernels.
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: NSteven Rostedt <rostedt@goodmis.org>
Reviewed-by: NThomas Gleixner <tglx@linutronix.de>
Reviewed-by: NFrederic Weisbecker <fweisbec@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NIngo Molnar <mingo@kernel.org>
      87dcbc06
  29. 23 9月, 2015 1 次提交
  31. 21 9月, 2015 1 次提交
    • P
      rcu: Use single-stage IPI algorithm for RCU expedited grace period · 8203d6d0
      由 Paul E. McKenney 提交于 
      The current preemptible-RCU expedited grace-period algorithm invokes
synchronize_sched_expedited() to enqueue all tasks currently running
in a preemptible-RCU read-side critical section, then waits for all the
->blkd_tasks lists to drain.  This works, but results in both an IPI and
a double context switch even on CPUs that do not happen to be running
in a preemptible RCU read-side critical section.

This commit implements a new algorithm that causes less OS jitter.
This new algorithm IPIs all online CPUs that are not idle (from an
RCU perspective), but refrains from self-IPIs.  If a CPU receiving
this IPI is not in a preemptible RCU read-side critical section (or
is just now exiting one), it pushes quiescence up the rcu_node tree,
otherwise, it sets a flag that will be handled by the upcoming outermost
rcu_read_unlock(), which will then push quiescence up the tree.

The expedited grace period must of course wait on any pre-existing blocked
readers, and newly blocked readers must be queued carefully based on
the state of both the normal and the expedited grace periods.  This
new queueing approach also avoids the need to update boost state,
courtesy of the fact that blocked tasks are no longer ever migrated to
the root rcu_node structure.
Signed-off-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
      8203d6d0
  33. 17 9月, 2015 1 次提交
    • T
      sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem · 1ed13287
      由 Tejun Heo 提交于 
      Note: This commit was originally committed as d59cfc09 but got
      reverted by 0c986253 due to the performance regression from
      the percpu_rwsem write down/up operations added to cgroup task
      migration path.  percpu_rwsem changes which alleviate the
      performance issue are pending for v4.4-rc1 merge window.
      Re-apply.

The cgroup side of threadgroup locking uses signal_struct->group_rwsem
to synchronize against threadgroup changes.  This per-process rwsem
adds small overhead to thread creation, exit and exec paths, forces
cgroup code paths to do lock-verify-unlock-retry dance in a couple
places and makes it impossible to atomically perform operations across
multiple processes.

This patch replaces signal_struct->group_rwsem with a global
percpu_rwsem cgroup_threadgroup_rwsem which is cheaper on the reader
side and contained in cgroups proper.  This patch converts one-to-one.

This does make writer side heavier and lower the granularity; however,
cgroup process migration is a fairly cold path, we do want to optimize
thread operations over it and cgroup migration operations don't take
enough time for the lower granularity to matter.
Signed-off-by: NTejun Heo <tj@kernel.org>
Link: http://lkml.kernel.org/g/55F8097A.7000206@de.ibm.com
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
      1ed13287
  35. 16 9月, 2015 1 次提交
    • T
      Revert "sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem" · 0c986253
      由 Tejun Heo 提交于 
      This reverts commit d59cfc09.

d59cfc09 ("sched, cgroup: replace signal_struct->group_rwsem with
a global percpu_rwsem") and b5ba75b5 ("cgroup: simplify
threadgroup locking") changed how cgroup synchronizes against task
fork and exits so that it uses global percpu_rwsem instead of
per-process rwsem; unfortunately, the write [un]lock paths of
percpu_rwsem always involve synchronize_rcu_expedited() which turned
out to be too expensive.

Improvements for percpu_rwsem are scheduled to be merged in the coming
v4.4-rc1 merge window which alleviates this issue.  For now, revert
the two commits to restore per-process rwsem.  They will be re-applied
for the v4.4-rc1 merge window.
Signed-off-by: NTejun Heo <tj@kernel.org>
Link: http://lkml.kernel.org/g/55F8097A.7000206@de.ibm.comReported-by: NChristian Borntraeger <borntraeger@de.ibm.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: stable@vger.kernel.org # v4.2+
      0c986253
  37. 13 9月, 2015 2 次提交
    • D
      sched/fair: Make utilization tracking CPU scale-invariant · e3279a2e
      由 Dietmar Eggemann 提交于 
      Besides the existing frequency scale-invariance correction factor, apply
CPU scale-invariance correction factor to utilization tracking to
compensate for any differences in compute capacity. This could be due to
micro-architectural differences (i.e. instructions per seconds) between
cpus in HMP systems (e.g. big.LITTLE), and/or differences in the current
maximum frequency supported by individual cpus in SMP systems. In the
existing implementation utilization isn't comparable between cpus as it
is relative to the capacity of each individual CPU.

Each segment of the sched_avg.util_sum geometric series is now scaled
by the CPU performance factor too so the sched_avg.util_avg of each
sched entity will be invariant from the particular CPU of the HMP/SMP
system on which the sched entity is scheduled.

With this patch, the utilization of a CPU stays relative to the max CPU
performance of the fastest CPU in the system.

In contrast to utilization (sched_avg.util_sum), load
(sched_avg.load_sum) should not be scaled by compute capacity. The
utilization metric is based on running time which only makes sense when
cpus are _not_ fully utilized (utilization cannot go beyond 100% even if
more tasks are added), where load is runnable time which isn't limited
by the capacity of the CPU and therefore is a better metric for
overloaded scenarios. If we run two nice-0 busy loops on two cpus with
different compute capacity their load should be similar since their
compute demands are the same. We have to assume that the compute demand
of any task running on a fully utilized CPU (no spare cycles = 100%
utilization) is high and the same no matter of the compute capacity of
its current CPU, hence we shouldn't scale load by CPU capacity.
Signed-off-by: NDietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: NMorten Rasmussen <morten.rasmussen@arm.com>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/55CE7409.1000700@arm.comSigned-off-by: NIngo Molnar <mingo@kernel.org>
      e3279a2e
    • D
      sched/fair: Make load tracking frequency scale-invariant · e0f5f3af
      由 Dietmar Eggemann 提交于 
      Apply frequency scaling correction factor to per-entity load tracking to
make it frequency invariant. Currently, load appears bigger when the CPU
is running slower which affects load-balancing decisions.

Each segment of the sched_avg.load_sum geometric series is now scaled by
the current frequency so that the sched_avg.load_avg of each sched entity
will be invariant from frequency scaling.

Moreover, cfs_rq.runnable_load_sum is scaled by the current frequency as
well.
Signed-off-by: NDietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: NMorten Rasmussen <morten.rasmussen@arm.com>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: NVincent Guittot <vincent.guittot@linaro.org>
Cc: Dietmar Eggemann <Dietmar.Eggemann@arm.com>
Cc: Juri Lelli <Juri.Lelli@arm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: daniel.lezcano@linaro.org
Cc: mturquette@baylibre.com
Cc: pang.xunlei@zte.com.cn
Cc: rjw@rjwysocki.net
Cc: sgurrappadi@nvidia.com
Cc: yuyang.du@intel.com
Link: http://lkml.kernel.org/r/1439569394-11974-2-git-send-email-morten.rasmussen@arm.comSigned-off-by: NIngo Molnar <mingo@kernel.org>
      e0f5f3af
  39. 05 9月, 2015 2 次提交
    • M
      mm: defer flush of writable TLB entries · d950c947
      由 Mel Gorman 提交于 
      If a PTE is unmapped and it's dirty then it was writable recently.  Due to
deferred TLB flushing, it's best to assume a writable TLB cache entry
exists.  With that assumption, the TLB must be flushed before any IO can
start or the page is freed to avoid lost writes or data corruption.  This
patch defers flushing of potentially writable TLBs as long as possible.
Signed-off-by: NMel Gorman <mgorman@suse.de>
Reviewed-by: NRik van Riel <riel@redhat.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Acked-by: NIngo Molnar <mingo@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      d950c947
    • M
      mm: send one IPI per CPU to TLB flush all entries after unmapping pages · 72b252ae
      由 Mel Gorman 提交于 
      An IPI is sent to flush remote TLBs when a page is unmapped that was
potentially accesssed by other CPUs.  There are many circumstances where
this happens but the obvious one is kswapd reclaiming pages belonging to a
running process as kswapd and the task are likely running on separate
CPUs.

On small machines, this is not a significant problem but as machine gets
larger with more cores and more memory, the cost of these IPIs can be
high.  This patch uses a simple structure that tracks CPUs that
potentially have TLB entries for pages being unmapped.  When the unmapping
is complete, the full TLB is flushed on the assumption that a refill cost
is lower than flushing individual entries.

Architectures wishing to do this must give the following guarantee.

        If a clean page is unmapped and not immediately flushed, the
        architecture must guarantee that a write to that linear address
        from a CPU with a cached TLB entry will trap a page fault.

This is essentially what the kernel already depends on but the window is
much larger with this patch applied and is worth highlighting.  The
architecture should consider whether the cost of the full TLB flush is
higher than sending an IPI to flush each individual entry.  An additional
architecture helper called flush_tlb_local is required.  It's a trivial
wrapper with some accounting in the x86 case.

The impact of this patch depends on the workload as measuring any benefit
requires both mapped pages co-located on the LRU and memory pressure.  The
case with the biggest impact is multiple processes reading mapped pages
taken from the vm-scalability test suite.  The test case uses NR_CPU
readers of mapped files that consume 10*RAM.

Linear mapped reader on a 4-node machine with 64G RAM and 48 CPUs

                                           4.2.0-rc1          4.2.0-rc1
                                             vanilla       flushfull-v7
Ops lru-file-mmap-read-elapsed      159.62 (  0.00%)   120.68 ( 24.40%)
Ops lru-file-mmap-read-time_range    30.59 (  0.00%)     2.80 ( 90.85%)
Ops lru-file-mmap-read-time_stddv     6.70 (  0.00%)     0.64 ( 90.38%)

           4.2.0-rc1    4.2.0-rc1
             vanilla flushfull-v7
User          581.00       611.43
System       5804.93      4111.76
Elapsed       161.03       122.12

This is showing that the readers completed 24.40% faster with 29% less
system CPU time.  From vmstats, it is known that the vanilla kernel was
interrupted roughly 900K times per second during the steady phase of the
test and the patched kernel was interrupts 180K times per second.

The impact is lower on a single socket machine.

                                           4.2.0-rc1          4.2.0-rc1
                                             vanilla       flushfull-v7
Ops lru-file-mmap-read-elapsed       25.33 (  0.00%)    20.38 ( 19.54%)
Ops lru-file-mmap-read-time_range     0.91 (  0.00%)     1.44 (-58.24%)
Ops lru-file-mmap-read-time_stddv     0.28 (  0.00%)     0.47 (-65.34%)

           4.2.0-rc1    4.2.0-rc1
             vanilla flushfull-v7
User           58.09        57.64
System        111.82        76.56
Elapsed        27.29        22.55

It's still a noticeable improvement with vmstat showing interrupts went
from roughly 500K per second to 45K per second.

The patch will have no impact on workloads with no memory pressure or have
relatively few mapped pages.  It will have an unpredictable impact on the
workload running on the CPU being flushed as it'll depend on how many TLB
entries need to be refilled and how long that takes.  Worst case, the TLB
will be completely cleared of active entries when the target PFNs were not
resident at all.

[sasha.levin@oracle.com: trace tlb flush after disabling preemption in try_to_unmap_flush]
Signed-off-by: NMel Gorman <mgorman@suse.de>
Reviewed-by: NRik van Riel <riel@redhat.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Acked-by: NIngo Molnar <mingo@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NSasha Levin <sasha.levin@oracle.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      72b252ae
  41. 12 8月, 2015 1 次提交
  43. 03 8月, 2015 4 次提交
    • Y
      sched/fair: Rewrite runnable load and utilization average tracking · 9d89c257
      由 Yuyang Du 提交于 
      The idea of runnable load average (let runnable time contribute to weight)
was proposed by Paul Turner and Ben Segall, and it is still followed by
this rewrite. This rewrite aims to solve the following issues:

1. cfs_rq's load average (namely runnable_load_avg and blocked_load_avg) is
   updated at the granularity of an entity at a time, which results in the
   cfs_rq's load average is stale or partially updated: at any time, only
   one entity is up to date, all other entities are effectively lagging
   behind. This is undesirable.

   To illustrate, if we have n runnable entities in the cfs_rq, as time
   elapses, they certainly become outdated:

     t0: cfs_rq { e1_old, e2_old, ..., en_old }

   and when we update:

     t1: update e1, then we have cfs_rq { e1_new, e2_old, ..., en_old }

     t2: update e2, then we have cfs_rq { e1_old, e2_new, ..., en_old }

     ...

   We solve this by combining all runnable entities' load averages together
   in cfs_rq's avg, and update the cfs_rq's avg as a whole. This is based
   on the fact that if we regard the update as a function, then:

   w * update(e) = update(w * e) and

   update(e1) + update(e2) = update(e1 + e2), then

   w1 * update(e1) + w2 * update(e2) = update(w1 * e1 + w2 * e2)

   therefore, by this rewrite, we have an entirely updated cfs_rq at the
   time we update it:

     t1: update cfs_rq { e1_new, e2_new, ..., en_new }

     t2: update cfs_rq { e1_new, e2_new, ..., en_new }

     ...

2. cfs_rq's load average is different between top rq->cfs_rq and other
   task_group's per CPU cfs_rqs in whether or not blocked_load_average
   contributes to the load.

   The basic idea behind runnable load average (the same for utilization)
   is that the blocked state is taken into account as opposed to only
   accounting for the currently runnable state. Therefore, the average
   should include both the runnable/running and blocked load averages.
   This rewrite does that.

   In addition, we also combine runnable/running and blocked averages
   of all entities into the cfs_rq's average, and update it together at
   once. This is based on the fact that:

     update(runnable) + update(blocked) = update(runnable + blocked)

   This significantly reduces the code as we don't need to separately
   maintain/update runnable/running load and blocked load.

3. How task_group entities' share is calculated is complex and imprecise.

   We reduce the complexity in this rewrite to allow a very simple rule:
   the task_group's load_avg is aggregated from its per CPU cfs_rqs's
   load_avgs. Then group entity's weight is simply proportional to its
   own cfs_rq's load_avg / task_group's load_avg. To illustrate,

   if a task_group has { cfs_rq1, cfs_rq2, ..., cfs_rqn }, then,

   task_group_avg = cfs_rq1_avg + cfs_rq2_avg + ... + cfs_rqn_avg, then

   cfs_rqx's entity's share = cfs_rqx_avg / task_group_avg * task_group's share

To sum up, this rewrite in principle is equivalent to the current one, but
fixes the issues described above. Turns out, it significantly reduces the
code complexity and hence increases clarity and efficiency. In addition,
the new averages are more smooth/continuous (no spurious spikes and valleys)
and updated more consistently and quickly to reflect the load dynamics.

As a result, we have less load tracking overhead, better performance,
and especially better power efficiency due to more balanced load.
Signed-off-by: NYuyang Du <yuyang.du@intel.com>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: arjan@linux.intel.com
Cc: bsegall@google.com
Cc: dietmar.eggemann@arm.com
Cc: fengguang.wu@intel.com
Cc: len.brown@intel.com
Cc: morten.rasmussen@arm.com
Cc: pjt@google.com
Cc: rafael.j.wysocki@intel.com
Cc: umgwanakikbuti@gmail.com
Cc: vincent.guittot@linaro.org
Link: http://lkml.kernel.org/r/1436918682-4971-3-git-send-email-yuyang.du@intel.comSigned-off-by: NIngo Molnar <mingo@kernel.org>
      9d89c257
    • K
      sched/preempt: Fix cond_resched_lock() and cond_resched_softirq() · fe32d3cd
      由 Konstantin Khlebnikov 提交于 
      These functions check should_resched() before unlocking spinlock/bh-enable:
preempt_count always non-zero => should_resched() always returns false.
cond_resched_lock() worked iff spin_needbreak is set.

This patch adds argument "preempt_offset" to should_resched().

preempt_count offset constants for that:

  PREEMPT_DISABLE_OFFSET  - offset after preempt_disable()
  PREEMPT_LOCK_OFFSET     - offset after spin_lock()
  SOFTIRQ_DISABLE_OFFSET  - offset after local_bh_distable()
  SOFTIRQ_LOCK_OFFSET     - offset after spin_lock_bh()
Signed-off-by: NKonstantin Khlebnikov <khlebnikov@yandex-team.ru>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Graf <agraf@suse.de>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: bdb43806 ("sched: Extract the basic add/sub preempt_count modifiers")
Link: http://lkml.kernel.org/r/20150715095204.12246.98268.stgit@buzzSigned-off-by: NIngo Molnar <mingo@kernel.org>
      fe32d3cd
    • M
      sched/fair: Beef up wake_wide() · 63b0e9ed
      由 Mike Galbraith 提交于 
      Josef Bacik reported that Facebook sees better performance with their
1:N load (1 dispatch/node, N workers/node) when carrying an old patch
to try very hard to wake to an idle CPU.  While looking at wake_wide(),
I noticed that it doesn't pay attention to the wakeup of a many partner
waker, returning 1 only when waking one of its many partners.

Correct that, letting explicit domain flags override the heuristic.

While at it, adjust task_struct bits, we don't need a 64-bit counter.
Tested-by: NJosef Bacik <jbacik@fb.com>
Signed-off-by: NMike Galbraith <umgwanakikbuti@gmail.com>
[ Tidy things up. ]
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: kernel-team<Kernel-team@fb.com>
Cc: morten.rasmussen@arm.com
Cc: riel@redhat.com
Link: http://lkml.kernel.org/r/1436888390.7983.49.camel@gmail.comSigned-off-by: NIngo Molnar <mingo@kernel.org>
      63b0e9ed
    • P
      sched/cputime: Guarantee stime + utime == rtime · 9d7fb042
      由 Peter Zijlstra 提交于 
      While the current code guarantees monotonicity for stime and utime
independently of one another, it does not guarantee that the sum of
both is equal to the total time we started out with.

This confuses things (and peoples) who look at this sum, like top, and
will report >100% usage followed by a matching period of 0%.

Rework the code to provide both individual monotonicity and a coherent
sum.
Suggested-by: NFredrik Markstrom <fredrik.markstrom@gmail.com>
Reported-by: NFredrik Markstrom <fredrik.markstrom@gmail.com>
Tested-by: NFredrik Markstrom <fredrik.markstrom@gmail.com>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: jason.low2@hp.com
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NIngo Molnar <mingo@kernel.org>
      9d7fb042
  45. 18 7月, 2015 2 次提交
  47. 04 7月, 2015 1 次提交