Skip to content

K
kernel_linux

OpenHarmony / kernel_linux
上一次同步 大约 4 年

通知 14
Star 8
Fork 2
K
kernel_linux
切换分支/标签
  1. 08 2月, 2010 2 次提交
    • A
      sched: cpuacct: Use bigger percpu counter batch values for stats counters · fa535a77
      由 Anton Blanchard 提交于 
      When CONFIG_VIRT_CPU_ACCOUNTING and CONFIG_CGROUP_CPUACCT are
enabled we can call cpuacct_update_stats with values much larger
than percpu_counter_batch.  This means the call to
percpu_counter_add will always add to the global count which is
protected by a spinlock and we end up with a global spinlock in
the scheduler.

Based on an idea by KOSAKI Motohiro, this patch scales the batch
value by cputime_one_jiffy such that we have the same batch
limit as we would if CONFIG_VIRT_CPU_ACCOUNTING was disabled.
His patch did this once at boot but that initialisation happened
too early on PowerPC (before time_init) and it was never updated
at runtime as a result of a hotplug cpu add/remove.

This patch instead scales percpu_counter_batch by
cputime_one_jiffy at runtime, which keeps the batch correct even
after cpu hotplug operations.  We cap it at INT_MAX in case of
overflow.

For architectures that do not support
CONFIG_VIRT_CPU_ACCOUNTING, cputime_one_jiffy is the constant 1
and gcc is smart enough to optimise min(s32
percpu_counter_batch, INT_MAX) to just percpu_counter_batch at
least on x86 and PowerPC.  So there is no need to add an #ifdef.

On a 64 thread PowerPC box with CONFIG_VIRT_CPU_ACCOUNTING and
CONFIG_CGROUP_CPUACCT enabled, a context switch microbenchmark
is 234x faster and almost matches a CONFIG_CGROUP_CPUACCT
disabled kernel:

 CONFIG_CGROUP_CPUACCT disabled:   16906698 ctx switches/sec
 CONFIG_CGROUP_CPUACCT enabled:       61720 ctx switches/sec
 CONFIG_CGROUP_CPUACCT + patch:	   16663217 ctx switches/sec

Tested with:

 wget http://ozlabs.org/~anton/junkcode/context_switch.c
 make context_switch
 for i in `seq 0 63`; do taskset -c $i ./context_switch & done
 vmstat 1
Signed-off-by: NAnton Blanchard <anton@samba.org>
Reviewed-by: NKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: NBalbir Singh <balbir@linux.vnet.ibm.com>
Tested-by: NBalbir Singh <balbir@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
      fa535a77
    • A
      kernel/sched.c: Suppress unused var warning · 50200df4
      由 Andrew Morton 提交于 
      On UP:

 kernel/sched.c: In function 'wake_up_new_task':
 kernel/sched.c:2631: warning: unused variable 'cpu'
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
      50200df4
  3. 04 2月, 2010 2 次提交
  5. 03 2月, 2010 6 次提交
  7. 02 2月, 2010 2 次提交
  9. 01 2月, 2010 1 次提交
    • J
      softlockup: Add sched_clock_tick() to avoid kernel warning on kgdb resume · d6ad3e28
      由 Jason Wessel 提交于 
      When CONFIG_HAVE_UNSTABLE_SCHED_CLOCK is set, sched_clock() gets
the time from hardware such as the TSC on x86. In this
configuration kgdb will report a softlock warning message on
resuming or detaching from a debug session.

Sequence of events in the problem case:

 1) "cpu sched clock" and "hardware time" are at 100 sec prior
    to a call to kgdb_handle_exception()

 2) Debugger waits in kgdb_handle_exception() for 80 sec and on
    exit the following is called ...  touch_softlockup_watchdog() -->
    __raw_get_cpu_var(touch_timestamp) = 0;

 3) "cpu sched clock" = 100s (it was not updated, because the
    interrupt was disabled in kgdb) but the "hardware time" = 180 sec

 4) The first timer interrupt after resuming from
    kgdb_handle_exception updates the watchdog from the "cpu sched clock"

update_process_times() { ...  run_local_timers() -->
softlockup_tick() --> check (touch_timestamp == 0) (it is "YES"
here, we have set "touch_timestamp = 0" at kgdb) -->
__touch_softlockup_watchdog() ***(A)--> reset "touch_timestamp"
to "get_timestamp()" (Here, the "touch_timestamp" will still be
set to 100s.)  ...

    scheduler_tick() ***(B)--> sched_clock_tick() (update "cpu sched
    clock" to "hardware time" = 180s) ...  }

 5) The Second timer interrupt handler appears to have a large
    jump and trips the softlockup warning.

update_process_times() { ...  run_local_timers() -->
softlockup_tick() --> "cpu sched clock" - "touch_timestamp" =
180s-100s > 60s --> printk "soft lockup error messages" ...  }

note: ***(A) reset "touch_timestamp" to
"get_timestamp(this_cpu)"

Why is "touch_timestamp" 100 sec, instead of 180 sec?

When CONFIG_HAVE_UNSTABLE_SCHED_CLOCK is set, the call trace of
get_timestamp() is:

get_timestamp(this_cpu)
 -->cpu_clock(this_cpu)
 -->sched_clock_cpu(this_cpu)
 -->__update_sched_clock(sched_clock_data, now)

The __update_sched_clock() function uses the GTOD tick value to
create a window to normalize the "now" values.  So if "now"
value is too big for sched_clock_data, it will be ignored.

The fix is to invoke sched_clock_tick() to update "cpu sched
clock" in order to recover from this state.  This is done by
introducing the function touch_softlockup_watchdog_sync(). This
allows kgdb to request that the sched clock is updated when the
watchdog thread runs the first time after a resume from kgdb.

[yong.zhang0@gmail.com: Use per cpu instead of an array]
Signed-off-by: NJason Wessel <jason.wessel@windriver.com>
Signed-off-by: NDongdong Deng <Dongdong.Deng@windriver.com>
Cc: kgdb-bugreport@lists.sourceforge.net
Cc: peterz@infradead.org
LKML-Reference: <1264631124-4837-2-git-send-email-jason.wessel@windriver.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
      d6ad3e28
  11. 30 1月, 2010 2 次提交
    • J
      perf, hw_breakpoint, kgdb: Do not take mutex for kernel debugger · 5352ae63
      由 Jason Wessel 提交于 
      This patch fixes the regression in functionality where the
kernel debugger and the perf API do not nicely share hw
breakpoint reservations.

The kernel debugger cannot use any mutex_lock() calls because it
can start the kernel running from an invalid context.

A mutex free version of the reservation API needed to get
created for the kernel debugger to safely update hw breakpoint
reservations.

The possibility for a breakpoint reservation to be concurrently
processed at the time that kgdb interrupts the system is
improbable. Should this corner case occur the end user is
warned, and the kernel debugger will prohibit updating the
hardware breakpoint reservations.

Any time the kernel debugger reserves a hardware breakpoint it
will be a system wide reservation.
Signed-off-by: NJason Wessel <jason.wessel@windriver.com>
Acked-by: NFrederic Weisbecker <fweisbec@gmail.com>
Cc: kgdb-bugreport@lists.sourceforge.net
Cc: K.Prasad <prasad@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: torvalds@linux-foundation.org
LKML-Reference: <1264719883-7285-3-git-send-email-jason.wessel@windriver.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
      5352ae63
    • J
      x86, hw_breakpoints, kgdb: Fix kgdb to use hw_breakpoint API · cc096749
      由 Jason Wessel 提交于 
      In the 2.6.33 kernel, the hw_breakpoint API is now used for the
performance event counters.  The hw_breakpoint_handler() now
consumes the hw breakpoints that were previously set by kgdb
arch specific code.  In order for kgdb to work in conjunction
with this core API change, kgdb must use some of the low level
functions of the hw_breakpoint API to install, uninstall, and
deal with hw breakpoint reservations.

The kgdb core required a change to call kgdb_disable_hw_debug
anytime a slave cpu enters kgdb_wait() in order to keep all the
hw breakpoints in sync as well as to prevent hitting a hw
breakpoint while kgdb is active.

During the architecture specific initialization of kgdb, it will
pre-allocate 4 disabled (struct perf event **) structures.  Kgdb
will use these to manage the capabilities for the 4 hw
breakpoint registers, per cpu.  Right now the hw_breakpoint API
does not have a way to ask how many breakpoints are available,
on each CPU so it is possible that the install of a breakpoint
might fail when kgdb restores the system to the run state.  The
intent of this patch is to first get the basic functionality of
hw breakpoints working and leave it to the person debugging the
kernel to understand what hw breakpoints are in use and what
restrictions have been imposed as a result.  Breakpoint
constraints will be dealt with in a future patch.

While atomic, the x86 specific kgdb code will call
arch_uninstall_hw_breakpoint() and arch_install_hw_breakpoint()
to manage the cpu specific hw breakpoints.

The net result of these changes allow kgdb to use the same pool
of hw_breakpoints that are used by the perf event API, but
neither knows about future reservations for the available hw
breakpoint slots.
Signed-off-by: NJason Wessel <jason.wessel@windriver.com>
Acked-by: NFrederic Weisbecker <fweisbec@gmail.com>
Cc: kgdb-bugreport@lists.sourceforge.net
Cc: K.Prasad <prasad@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: torvalds@linux-foundation.org
LKML-Reference: <1264719883-7285-2-git-send-email-jason.wessel@windriver.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
      cc096749
  13. 28 1月, 2010 3 次提交
  15. 27 1月, 2010 4 次提交
    • O
      lockdep: Fix check_usage_backwards() error message · 48d50674
      由 Oleg Nesterov 提交于 
      Lockdep has found the real bug, but the output doesn't look right to me:

> =========================================================
> [ INFO: possible irq lock inversion dependency detected ]
> 2.6.33-rc5 #77
> ---------------------------------------------------------
> emacs/1609 just changed the state of lock:
>  (&(&tty->ctrl_lock)->rlock){+.....}, at: [<ffffffff8127c648>] tty_fasync+0xe8/0x190
> but this lock took another, HARDIRQ-unsafe lock in the past:
>  (&(&sighand->siglock)->rlock){-.....}

"HARDIRQ-unsafe" and "this lock took another" looks wrong, afaics.

>   ... key      at: [<ffffffff81c054a4>] __key.46539+0x0/0x8
>   ... acquired at:
>    [<ffffffff81089af6>] __lock_acquire+0x1056/0x15a0
>    [<ffffffff8108a0df>] lock_acquire+0x9f/0x120
>    [<ffffffff81423012>] _raw_spin_lock_irqsave+0x52/0x90
>    [<ffffffff8127c1be>] __proc_set_tty+0x3e/0x150
>    [<ffffffff8127e01d>] tty_open+0x51d/0x5e0

The stack-trace shows that this lock (ctrl_lock) was taken under
->siglock (which is hopefully irq-safe).

This is a clear typo in check_usage_backwards() where we tell the print a
fancy routine we're forwards.
Signed-off-by: NOleg Nesterov <oleg@redhat.com>
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <20100126181641.GA10460@redhat.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
      48d50674
    • M
      tracing/documentation: Cover new frame pointer semantics · 03688970
      由 Mike Frysinger 提交于 
      Update the graph tracer examples to cover the new frame pointer semantics
(in terms of passing it along).  Move the HAVE_FUNCTION_GRAPH_FP_TEST docs
out of the Kconfig, into the right place, and expand on the details.
Signed-off-by: NMike Frysinger <vapier@gentoo.org>
LKML-Reference: <1264165967-18938-1-git-send-email-vapier@gentoo.org>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>
      03688970
    • S
      ring-buffer: Check for end of page in iterator · 3c05d748
      由 Steven Rostedt 提交于 
      If the iterator comes to an empty page for some reason, or if
the page is emptied by a consuming read. The iterator code currently
does not check if the iterator is pass the contents, and may
return a false entry.

This patch adds a check to the ring buffer iterator to test if the
current page has been completely read and sets the iterator to the
next page if necessary.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>
      3c05d748
    • S
      ring-buffer: Check if ring buffer iterator has stale data · 492a74f4
      由 Steven Rostedt 提交于 
      Usually reads of the ring buffer is performed by a single task.
There are two types of reads from the ring buffer.

One is a consuming read which will consume the entry that was read
and the next read will be the entry that follows.

The other is an iterator that will let the user read the contents of
the ring buffer without modifying it. When an iterator is allocated,
writes to the ring buffer are disabled to protect the iterator.

The problem exists when consuming reads happen while an iterator is
allocated. Specifically, the kind of read that swaps out an entire
page (used by splice) and replaces it with a new read. If the iterator
is on the page that is swapped out, then the next read may read
from this swapped out page and return garbage.

This patch adds a check when reading the iterator to make sure that
the iterator contents are still valid. If a consuming read has taken
place, the iterator is reset.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>
      492a74f4
  17. 26 1月, 2010 2 次提交
    • T
      clocksource: Prevent potential kgdb dead lock · 7b7422a5
      由 Thomas Gleixner 提交于 
      commit 0f8e8ef7 (clocksource: Simplify clocksource watchdog resume
logic) introduced a potential kgdb dead lock. When the kernel is
stopped by kgdb inside code which holds watchdog_lock then kgdb dead
locks in clocksource_resume_watchdog().

clocksource_resume_watchdog() is called from kbdg via
clocksource_touch_watchdog() to avoid that the clock source watchdog
marks TSC unstable after the kernel has been stopped.

Solve this by replacing spin_lock with a spin_trylock and just return
in case the lock is held. Not resetting the watchdog might result in
TSC becoming marked unstable, but that's an acceptable penalty for
using kgdb.

The timekeeping is anyway easily screwed up by kgdb when the system
uses either jiffies or a clock source which wraps in short intervals
(e.g. pm_timer wraps about every 4.6s), so we really do not have to
worry about that occasional TSC marked unstable side effect.

The second caller of clocksource_resume_watchdog() is
clocksource_resume(). The trylock is safe here as well because the
system is UP at this point, interrupts are disabled and nothing else
can hold watchdog_lock().
Reported-by: NJason Wessel <jason.wessel@windriver.com>
LKML-Reference: <1264480000-6997-4-git-send-email-jason.wessel@windriver.com>
Cc: kgdb-bugreport@lists.sourceforge.net
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: John Stultz <johnstul@us.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
      7b7422a5
    • S
      tracing: Prevent kernel oops with corrupted buffer · 74bf4076
      由 Steven Rostedt 提交于 
      If the contents of the ftrace ring buffer gets corrupted and the trace
file is read, it could create a kernel oops (usualy just killing the user
task thread). This is caused by the checking of the pid in the buffer.
If the pid is negative, it still references the cmdline cache array,
which could point to an invalid address.

The simple fix is to test for negative PIDs.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>
      74bf4076
  19. 23 1月, 2010 3 次提交
  21. 22 1月, 2010 1 次提交
    • P
      sched: Fix fork vs hotplug vs cpuset namespaces · fabf318e
      由 Peter Zijlstra 提交于 
      There are a number of issues:

1) TASK_WAKING vs cgroup_clone (cpusets)

copy_process():

  sched_fork()
    child->state = TASK_WAKING; /* waiting for wake_up_new_task() */
  if (current->nsproxy != p->nsproxy)
     ns_cgroup_clone()
       cgroup_clone()
         mutex_lock(inode->i_mutex)
         mutex_lock(cgroup_mutex)
         cgroup_attach_task()
	   ss->can_attach()
           ss->attach() [ -> cpuset_attach() ]
             cpuset_attach_task()
               set_cpus_allowed_ptr();
                 while (child->state == TASK_WAKING)
                   cpu_relax();
will deadlock the system.


2) cgroup_clone (cpusets) vs copy_process

So even if the above would work we still have:

copy_process():

  if (current->nsproxy != p->nsproxy)
     ns_cgroup_clone()
       cgroup_clone()
         mutex_lock(inode->i_mutex)
         mutex_lock(cgroup_mutex)
         cgroup_attach_task()
	   ss->can_attach()
           ss->attach() [ -> cpuset_attach() ]
             cpuset_attach_task()
               set_cpus_allowed_ptr();
  ...

  p->cpus_allowed = current->cpus_allowed

over-writing the modified cpus_allowed.


3) fork() vs hotplug

  if we unplug the child's cpu after the sanity check when the child
  gets attached to the task_list but before wake_up_new_task() shit
  will meet with fan.

Solve all these issues by moving fork cpu selection into
wake_up_new_task().
Reported-by: NSerge E. Hallyn <serue@us.ibm.com>
Tested-by: NSerge E. Hallyn <serue@us.ibm.com>
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1264106190.4283.1314.camel@laptop>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
      fabf318e
  23. 21 1月, 2010 12 次提交