Skip to content

cloud-kernel
cloud-kernel

openanolis / cloud-kernel
1 年多 前同步成功

通知 160
Star 36
Fork 7
cloud-kernel
cloud-kernel
切换分支/标签
  1. 29 9月, 2017 2 次提交
  3. 15 9月, 2017 2 次提交
    • T
      sched/wait: Introduce wakeup boomark in wake_up_page_bit · 11a19c7b
      由 Tim Chen 提交于 
      Now that we have added breaks in the wait queue scan and allow bookmark
on scan position, we put this logic in the wake_up_page_bit function.

We can have very long page wait list in large system where multiple
pages share the same wait list. We break the wake up walk here to allow
other cpus a chance to access the list, and not to disable the interrupts
when traversing the list for too long.  This reduces the interrupt and
rescheduling latency, and excessive page wait queue lock hold time.

[ v2: Remove bookmark_wake_function ]
Signed-off-by: NTim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      11a19c7b
    • T
      sched/wait: Break up long wake list walk · 2554db91
      由 Tim Chen 提交于 
      We encountered workloads that have very long wake up list on large
systems. A waker takes a long time to traverse the entire wake list and
execute all the wake functions.

We saw page wait list that are up to 3700+ entries long in tests of
large 4 and 8 socket systems. It took 0.8 sec to traverse such list
during wake up. Any other CPU that contends for the list spin lock will
spin for a long time. It is a result of the numa balancing migration of
hot pages that are shared by many threads.

Multiple CPUs waking are queued up behind the lock, and the last one
queued has to wait until all CPUs did all the wakeups.

The page wait list is traversed with interrupt disabled, which caused
various problems. This was the original cause that triggered the NMI
watch dog timer in: https://patchwork.kernel.org/patch/9800303/ . Only
extending the NMI watch dog timer there helped.

This patch bookmarks the waker's scan position in wake list and break
the wake up walk, to allow access to the list before the waker resume
its walk down the rest of the wait list. It lowers the interrupt and
rescheduling latency.

This patch also provides a performance boost when combined with the next
patch to break up page wakeup list walk. We saw 22% improvement in the
will-it-scale file pread2 test on a Xeon Phi system running 256 threads.

[ v2: Merged in Linus' changes to remove the bookmark_wake_function, and
  simply access to flags. ]
Reported-by: NKan Liang <kan.liang@intel.com>
Tested-by: NKan Liang <kan.liang@intel.com>
Signed-off-by: NTim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      2554db91
  5. 12 9月, 2017 4 次提交
  7. 11 9月, 2017 1 次提交
  9. 09 9月, 2017 3 次提交
  11. 07 9月, 2017 2 次提交
  13. 29 8月, 2017 1 次提交
    • Y
      smp: Avoid using two cache lines for struct call_single_data · 966a9671
      由 Ying Huang 提交于 
      struct call_single_data is used in IPIs to transfer information between
CPUs.  Its size is bigger than sizeof(unsigned long) and less than
cache line size.  Currently it is not allocated with any explicit alignment
requirements.  This makes it possible for allocated call_single_data to
cross two cache lines, which results in double the number of the cache lines
that need to be transferred among CPUs.

This can be fixed by requiring call_single_data to be aligned with the
size of call_single_data. Currently the size of call_single_data is the
power of 2.  If we add new fields to call_single_data, we may need to
add padding to make sure the size of new definition is the power of 2
as well.

Fortunately, this is enforced by GCC, which will report bad sizes.

To set alignment requirements of call_single_data to the size of
call_single_data, a struct definition and a typedef is used.

To test the effect of the patch, I used the vm-scalability multiple
thread swap test case (swap-w-seq-mt).  The test will create multiple
threads and each thread will eat memory until all RAM and part of swap
is used, so that huge number of IPIs are triggered when unmapping
memory.  In the test, the throughput of memory writing improves ~5%
compared with misaligned call_single_data, because of faster IPIs.
Suggested-by: NPeter Zijlstra <peterz@infradead.org>
Signed-off-by: NHuang, Ying <ying.huang@intel.com>
[ Add call_single_data_t and align with size of call_single_data. ]
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Aaron Lu <aaron.lu@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/87bmnqd6lz.fsf@yhuang-mobile.sh.intel.comSigned-off-by: NIngo Molnar <mingo@kernel.org>
      966a9671
  15. 28 8月, 2017 1 次提交
    • L
      Minor page waitqueue cleanups · 3510ca20
      由 Linus Torvalds 提交于 
      Tim Chen and Kan Liang have been battling a customer load that shows
extremely long page wakeup lists.  The cause seems to be constant NUMA
migration of a hot page that is shared across a lot of threads, but the
actual root cause for the exact behavior has not been found.

Tim has a patch that batches the wait list traversal at wakeup time, so
that we at least don't get long uninterruptible cases where we traverse
and wake up thousands of processes and get nasty latency spikes.  That
is likely 4.14 material, but we're still discussing the page waitqueue
specific parts of it.

In the meantime, I've tried to look at making the page wait queues less
expensive, and failing miserably.  If you have thousands of threads
waiting for the same page, it will be painful.  We'll need to try to
figure out the NUMA balancing issue some day, in addition to avoiding
the excessive spinlock hold times.

That said, having tried to rewrite the page wait queues, I can at least
fix up some of the braindamage in the current situation. In particular:

 (a) we don't want to continue walking the page wait list if the bit
     we're waiting for already got set again (which seems to be one of
     the patterns of the bad load).  That makes no progress and just
     causes pointless cache pollution chasing the pointers.

 (b) we don't want to put the non-locking waiters always on the front of
     the queue, and the locking waiters always on the back.  Not only is
     that unfair, it means that we wake up thousands of reading threads
     that will just end up being blocked by the writer later anyway.

Also add a comment about the layout of 'struct wait_page_key' - there is
an external user of it in the cachefiles code that means that it has to
match the layout of 'struct wait_bit_key' in the two first members.  It
so happens to match, because 'struct page *' and 'unsigned long *' end
up having the same values simply because the page flags are the first
member in struct page.

Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Christopher Lameter <cl@linux.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      3510ca20
  17. 25 8月, 2017 4 次提交
    • P
      sched/debug: Optimize sched_domain sysctl generation · bbdacdfe
      由 Peter Zijlstra 提交于 
      Currently we unconditionally destroy all sysctl bits and regenerate
them after we've rebuild the domains (even if that rebuild is a
no-op).

And since we unconditionally (re)build the sysctl for all possible
CPUs, onlining all CPUs gets us O(n^2) time. Instead change this to
only rebuild the bits for CPUs we've actually installed new domains
on.
Reported-by: NOfer Levi(SW) <oferle@mellanox.com>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: NIngo Molnar <mingo@kernel.org>
      bbdacdfe
    • P
      sched/topology: Avoid pointless rebuild · 09e0dd8e
      由 Peter Zijlstra 提交于 
      Fix partition_sched_domains() to try and preserve the existing machine
wide domain instead of unconditionally destroying it. We do this by
attempting to allocate the new single domain, only when that fails to
we reuse the fallback_doms.

When using fallback_doms we need to first destroy and then recreate
because both the old and new could be backed by it.
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Ofer Levi(SW) <oferle@mellanox.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vineet.Gupta1@synopsys.com <Vineet.Gupta1@synopsys.com>
Cc: rusty@rustcorp.com.au <rusty@rustcorp.com.au>
Signed-off-by: NIngo Molnar <mingo@kernel.org>
      09e0dd8e
    • P
      sched/topology: Improve comments · a090c4f2
      由 Peter Zijlstra 提交于 
      Mike provided a better comment for destroy_sched_domain() ...
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>
Signed-off-by: NIngo Molnar <mingo@kernel.org>
      a090c4f2
    • S
      sched/topology: Fix memory leak in __sdt_alloc() · 213c5a45
      由 Shu Wang 提交于 
      Found this issue by kmemleak: the 'sg' and 'sgc' pointers from
__sdt_alloc() might be leaked as each domain holds many groups' ref,
but in destroy_sched_domain(), it only declined the first group ref.

Onlining and offlining a CPU can trigger this leak, and cause OOM.

Reproducer for my 6 CPUs machine:

  while true
  do
      echo 0 > /sys/devices/system/cpu/cpu5/online;
      echo 1 > /sys/devices/system/cpu/cpu5/online;
  done

  unreferenced object 0xffff88007d772a80 (size 64):
    comm "cpuhp/5", pid 39, jiffies 4294719962 (age 35.251s)
    hex dump (first 32 bytes):
      c0 22 77 7d 00 88 ff ff 02 00 00 00 01 00 00 00  ."w}............
      40 2a 77 7d 00 88 ff ff 00 00 00 00 00 00 00 00  @*w}............
    backtrace:
      [<ffffffff8176525a>] kmemleak_alloc+0x4a/0xa0
      [<ffffffff8121efe1>] __kmalloc_node+0xf1/0x280
      [<ffffffff810d94a8>] build_sched_domains+0x1e8/0xf20
      [<ffffffff810da674>] partition_sched_domains+0x304/0x360
      [<ffffffff81139557>] cpuset_update_active_cpus+0x17/0x40
      [<ffffffff810bdb2e>] sched_cpu_activate+0xae/0xc0
      [<ffffffff810900e0>] cpuhp_invoke_callback+0x90/0x400
      [<ffffffff81090597>] cpuhp_up_callbacks+0x37/0xb0
      [<ffffffff81090887>] cpuhp_thread_fun+0xd7/0xf0
      [<ffffffff810b37e0>] smpboot_thread_fn+0x110/0x160
      [<ffffffff810af5d9>] kthread+0x109/0x140
      [<ffffffff81770e45>] ret_from_fork+0x25/0x30
      [<ffffffffffffffff>] 0xffffffffffffffff

  unreferenced object 0xffff88007d772a40 (size 64):
    comm "cpuhp/5", pid 39, jiffies 4294719962 (age 35.251s)
    hex dump (first 32 bytes):
      03 00 00 00 00 00 00 00 00 04 00 00 00 00 00 00  ................
      00 04 00 00 00 00 00 00 4f 3c fc ff 00 00 00 00  ........O<......
    backtrace:
      [<ffffffff8176525a>] kmemleak_alloc+0x4a/0xa0
      [<ffffffff8121efe1>] __kmalloc_node+0xf1/0x280
      [<ffffffff810da16d>] build_sched_domains+0xead/0xf20
      [<ffffffff810da674>] partition_sched_domains+0x304/0x360
      [<ffffffff81139557>] cpuset_update_active_cpus+0x17/0x40
      [<ffffffff810bdb2e>] sched_cpu_activate+0xae/0xc0
      [<ffffffff810900e0>] cpuhp_invoke_callback+0x90/0x400
      [<ffffffff81090597>] cpuhp_up_callbacks+0x37/0xb0
      [<ffffffff81090887>] cpuhp_thread_fun+0xd7/0xf0
      [<ffffffff810b37e0>] smpboot_thread_fn+0x110/0x160
      [<ffffffff810af5d9>] kthread+0x109/0x140
      [<ffffffff81770e45>] ret_from_fork+0x25/0x30
      [<ffffffffffffffff>] 0xffffffffffffffff
Reported-by: NChunyu Hu <chuhu@redhat.com>
Signed-off-by: NShu Wang <shuwang@redhat.com>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: NChunyu Hu <chuhu@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: liwang@redhat.com
Link: http://lkml.kernel.org/r/1502351536-9108-1-git-send-email-shuwang@redhat.comSigned-off-by: NIngo Molnar <mingo@kernel.org>
      213c5a45
  19. 18 8月, 2017 2 次提交
    • V
      cpufreq: schedutil: Always process remote callback with slow switching · c49cbc19
      由 Viresh Kumar 提交于 
      The frequency update from the utilization update handlers can be divided
into two parts:

(A) Finding the next frequency
(B) Updating the frequency

While any CPU can do (A), (B) can be restricted to a group of CPUs only,
depending on the current platform.

For platforms where fast cpufreq switching is possible, both (A) and (B)
are always done from the same CPU and that CPU should be capable of
changing the frequency of the target CPU.

But for platforms where fast cpufreq switching isn't possible, after
doing (A) we wake up a kthread which will eventually do (B). This
kthread is already bound to the right set of CPUs, i.e. only those which
can change the frequency of CPUs of a cpufreq policy. And so any CPU
can actually do (A) in this case, as the frequency is updated from the
right set of CPUs only.

Check cpufreq_can_do_remote_dvfs() only for the fast switching case.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
      c49cbc19
    • V
      cpufreq: schedutil: Don't restrict kthread to related_cpus unnecessarily · e2cabe48
      由 Viresh Kumar 提交于 
      Utilization update callbacks are now processed remotely, even on the
CPUs that don't share cpufreq policy with the target CPU (if
dvfs_possible_from_any_cpu flag is set).

But in non-fast switch paths, the frequency is changed only from one of
policy->related_cpus. This happens because the kthread which does the
actual update is bound to a subset of CPUs (i.e. related_cpus).

Allow frequency to be remotely updated as well (i.e. call
__cpufreq_driver_target()) if dvfs_possible_from_any_cpu flag is set.
Reported-by: NPavan Kondeti <pkondeti@codeaurora.org>
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
      e2cabe48
  21. 17 8月, 2017 3 次提交
    • P
      completion: Replace spin_unlock_wait() with lock/unlock pair · dec13c42
      由 Paul E. McKenney 提交于 
      There is no agreed-upon definition of spin_unlock_wait()'s semantics,
and it appears that all callers could do just as well with a lock/unlock
pair.  This commit therefore replaces the spin_unlock_wait() call in
completion_done() with spin_lock() followed immediately by spin_unlock().
This should be safe from a performance perspective because the lock
will be held only the wakeup happens really quickly.
Signed-off-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Andrea Parri <parri.andrea@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Reviewed-by: NSteven Rostedt (VMware) <rostedt@goodmis.org>
      dec13c42
    • M
      membarrier: Provide expedited private command · 22e4ebb9
      由 Mathieu Desnoyers 提交于 
      Implement MEMBARRIER_CMD_PRIVATE_EXPEDITED with IPIs using cpumask built
from all runqueues for which current thread's mm is the same as the
thread calling sys_membarrier. It executes faster than the non-expedited
variant (no blocking). It also works on NOHZ_FULL configurations.

Scheduler-wise, it requires a memory barrier before and after context
switching between processes (which have different mm). The memory
barrier before context switch is already present. For the barrier after
context switch:

* Our TSO archs can do RELEASE without being a full barrier. Look at
  x86 spin_unlock() being a regular STORE for example.  But for those
  archs, all atomics imply smp_mb and all of them have atomic ops in
  switch_mm() for mm_cpumask(), and on x86 the CR3 load acts as a full
  barrier.

* From all weakly ordered machines, only ARM64 and PPC can do RELEASE,
  the rest does indeed do smp_mb(), so there the spin_unlock() is a full
  barrier and we're good.

* ARM64 has a very heavy barrier in switch_to(), which suffices.

* PPC just removed its barrier from switch_to(), but appears to be
  talking about adding something to switch_mm(). So add a
  smp_mb__after_unlock_lock() for now, until this is settled on the PPC
  side.

Changes since v3:
- Properly document the memory barriers provided by each architecture.

Changes since v2:
- Address comments from Peter Zijlstra,
- Add smp_mb__after_unlock_lock() after finish_lock_switch() in
  finish_task_switch() to add the memory barrier we need after storing
  to rq->curr. This is much simpler than the previous approach relying
  on atomic_dec_and_test() in mmdrop(), which actually added a memory
  barrier in the common case of switching between userspace processes.
- Return -EINVAL when MEMBARRIER_CMD_SHARED is used on a nohz_full
  kernel, rather than having the whole membarrier system call returning
  -ENOSYS. Indeed, CMD_PRIVATE_EXPEDITED is compatible with nohz_full.
  Adapt the CMD_QUERY mask accordingly.

Changes since v1:
- move membarrier code under kernel/sched/ because it uses the
  scheduler runqueue,
- only add the barrier when we switch from a kernel thread. The case
  where we switch from a user-space thread is already handled by
  the atomic_dec_and_test() in mmdrop().
- add a comment to mmdrop() documenting the requirement on the implicit
  memory barrier.

CC: Peter Zijlstra <peterz@infradead.org>
CC: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
CC: Boqun Feng <boqun.feng@gmail.com>
CC: Andrew Hunter <ahh@google.com>
CC: Maged Michael <maged.michael@gmail.com>
CC: gromer@google.com
CC: Avi Kivity <avi@scylladb.com>
CC: Benjamin Herrenschmidt <benh@kernel.crashing.org>
CC: Paul Mackerras <paulus@samba.org>
CC: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NMathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: NDave Watson <davejwatson@fb.com>
      22e4ebb9
    • S
      sched/completion: Document that reinit_completion() must be called after complete_all() · 9c878320
      由 Steven Rostedt 提交于 
      The complete_all() function modifies the completion's "done" variable to
UINT_MAX, and no other caller (wait_for_completion(), etc) will modify
it back to zero. That means that any call to complete_all() must have a
reinit_completion() before that completion can be used again.

Document this fact by the complete_all() function.

Also document that completion_done() will always return true if
complete_all() is called.
Signed-off-by: NSteven Rostedt (VMware) <rostedt@goodmis.org>
Acked-by: NLinus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20170816131202.195c2f4b@gandalf.local.homeSigned-off-by: NIngo Molnar <mingo@kernel.org>
      9c878320
  23. 12 8月, 2017 1 次提交
  25. 11 8月, 2017 2 次提交
  27. 10 8月, 2017 12 次提交