linux-stable.git/kernel/sched, branch v5.0.17

sched/deadline: Correctly handle active 0-lag timers

2019-05-02T08:02:49+00:00

commit 1b02cd6a2d7f3e2a6a5262887d2cb2912083e42f upstream.

syzbot reported the following warning:

   [ ] WARNING: CPU: 4 PID: 17089 at kernel/sched/deadline.c:255 task_non_contending+0xae0/0x1950

line 255 of deadline.c is:

	WARN_ON(hrtimer_active(&dl_se->inactive_timer));

in task_non_contending().

Unfortunately, in some cases (for example, a deadline task
continuosly blocking and waking immediately) it can happen that
a task blocks (and task_non_contending() is called) while the
0-lag timer is still active.

In this case, the safest thing to do is to immediately decrease
the running bandwidth of the task, without trying to re-arm the 0-lag timer.

Signed-off-by: luca abeni 
Signed-off-by: Peter Zijlstra (Intel) 
Acked-by: Juri Lelli 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: chengjian (D) 
Link: https://lkml.kernel.org/r/20190325131530.34706-1-luca.abeni@santannapisa.it
Signed-off-by: Ingo Molnar 
Signed-off-by: Greg Kroah-Hartman

sched/numa: Fix a possible divide-by-zero

2019-05-02T08:02:45+00:00

commit a860fa7b96e1a1c974556327aa1aee852d434c21 upstream.

sched_clock_cpu() may not be consistent between CPUs. If a task
migrates to another CPU, then se.exec_start is set to that CPU's
rq_clock_task() by update_stats_curr_start(). Specifically, the new
value might be before the old value due to clock skew.

So then if in numa_get_avg_runtime() the expression:

  'now - p->last_task_numa_placement'

ends up as -1, then the divider '*period + 1' in task_numa_placement()
is 0 and things go bang. Similar to update_curr(), check if time goes
backwards to avoid this.

[ peterz: Wrote new changelog. ]
[ mingo: Tweaked the code comment. ]

Signed-off-by: Xie XiuQi 
Signed-off-by: Peter Zijlstra (Intel) 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: cj.chengjian@huawei.com
Cc: 
Link: http://lkml.kernel.org/r/20190425080016.GX11158@hirez.programming.kicks-ass.net
Signed-off-by: Ingo Molnar 
Signed-off-by: Greg Kroah-Hartman

sched/fair: Limit sched_cfs_period_timer() loop to avoid hard lockup

2019-04-27T07:37:44+00:00

[ Upstream commit 2e8e19226398db8265a8e675fcc0118b9e80c9e8 ]

With extremely short cfs_period_us setting on a parent task group with a large
number of children the for loop in sched_cfs_period_timer() can run until the
watchdog fires. There is no guarantee that the call to hrtimer_forward_now()
will ever return 0.  The large number of children can make
do_sched_cfs_period_timer() take longer than the period.

 NMI watchdog: Watchdog detected hard LOCKUP on cpu 24
 RIP: 0010:tg_nop+0x0/0x10
  
  walk_tg_tree_from+0x29/0xb0
  unthrottle_cfs_rq+0xe0/0x1a0
  distribute_cfs_runtime+0xd3/0xf0
  sched_cfs_period_timer+0xcb/0x160
  ? sched_cfs_slack_timer+0xd0/0xd0
  __hrtimer_run_queues+0xfb/0x270
  hrtimer_interrupt+0x122/0x270
  smp_apic_timer_interrupt+0x6a/0x140
  apic_timer_interrupt+0xf/0x20
  

To prevent this we add protection to the loop that detects when the loop has run
too many times and scales the period and quota up, proportionally, so that the timer
can complete before then next period expires.  This preserves the relative runtime
quota while preventing the hard lockup.

A warning is issued reporting this state and the new values.

Signed-off-by: Phil Auld 
Signed-off-by: Peter Zijlstra (Intel) 
Cc: 
Cc: Anton Blanchard 
Cc: Ben Segall 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: https://lkml.kernel.org/r/20190319130005.25492-1-pauld@redhat.com
Signed-off-by: Ingo Molnar 
Signed-off-by: Sasha Levin

sched/core: Fix buffer overflow in cgroup2 property cpu.max

2019-04-20T07:16:50+00:00

[ Upstream commit 4c47acd824aaaa8fc6dc519fb4e08d1522105b7a ]

Add limit into sscanf format string for on-stack buffer.

Signed-off-by: Konstantin Khlebnikov 
Signed-off-by: Peter Zijlstra (Intel) 
Acked-by: Tejun Heo 
Cc: Andy Lutomirski 
Cc: Borislav Petkov 
Cc: Dave Hansen 
Cc: H. Peter Anvin 
Cc: Johannes Weiner 
Cc: Li Zefan 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Rik van Riel 
Cc: Thomas Gleixner 
Fixes: 0d5936344f30 ("sched: Implement interface for cgroup unified hierarchy")
Link: https://lkml.kernel.org/r/155189230232.2620.13120481613524200065.stgit@buzz
Signed-off-by: Ingo Molnar 
Signed-off-by: Sasha Levin

sched/cpufreq: Fix 32-bit math overflow

2019-04-20T07:16:49+00:00

[ Upstream commit a23314e9d88d89d49e69db08f60b7caa470f04e1 ]

Vincent Wang reported that get_next_freq() has a mult overflow bug on
32-bit platforms in the IOWAIT boost case, since in that case {util,max}
are in freq units instead of capacity units.

Solve this by moving the IOWAIT boost to capacity units. And since this
means @max is constant; simplify the code.

Reported-by: Vincent Wang 
Tested-by: Vincent Wang 
Signed-off-by: Peter Zijlstra (Intel) 
Acked-by: Rafael J. Wysocki 
Cc: Andy Lutomirski 
Cc: Borislav Petkov 
Cc: Chunyan Zhang 
Cc: Dave Hansen 
Cc: H. Peter Anvin 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Quentin Perret 
Cc: Rafael J. Wysocki 
Cc: Rik van Riel 
Cc: Thomas Gleixner 
Link: https://lkml.kernel.org/r/20190305083202.GU32494@hirez.programming.kicks-ass.net
Signed-off-by: Ingo Molnar 
Signed-off-by: Sasha Levin

sched/fair: Do not re-read ->h_load_next during hierarchical load calculation

2019-04-17T06:39:52+00:00

commit 0e9f02450da07fc7b1346c8c32c771555173e397 upstream.

A NULL pointer dereference bug was reported on a distribution kernel but
the same issue should be present on mainline kernel. It occured on s390
but should not be arch-specific.  A partial oops looks like:

  Unable to handle kernel pointer dereference in virtual kernel address space
  ...
  Call Trace:
    ...
    try_to_wake_up+0xfc/0x450
    vhost_poll_wakeup+0x3a/0x50 [vhost]
    __wake_up_common+0xbc/0x178
    __wake_up_common_lock+0x9e/0x160
    __wake_up_sync_key+0x4e/0x60
    sock_def_readable+0x5e/0x98

The bug hits any time between 1 hour to 3 days. The dereference occurs
in update_cfs_rq_h_load when accumulating h_load. The problem is that
cfq_rq->h_load_next is not protected by any locking and can be updated
by parallel calls to task_h_load. Depending on the compiler, code may be
generated that re-reads cfq_rq->h_load_next after the check for NULL and
then oops when reading se->avg.load_avg. The dissassembly showed that it
was possible to reread h_load_next after the check for NULL.

While this does not appear to be an issue for later compilers, it's still
an accident if the correct code is generated. Full locking in this path
would have high overhead so this patch uses READ_ONCE to read h_load_next
only once and check for NULL before dereferencing. It was confirmed that
there were no further oops after 10 days of testing.

As Peter pointed out, it is also necessary to use WRITE_ONCE() to avoid any
potential problems with store tearing.

Signed-off-by: Mel Gorman 
Signed-off-by: Peter Zijlstra (Intel) 
Reviewed-by: Valentin Schneider 
Cc: Linus Torvalds 
Cc: Mike Galbraith 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: 
Fixes: 685207963be9 ("sched: Move h_load calculation to task_h_load()")
Link: https://lkml.kernel.org/r/20190319123610.nsivgf3mjbjjesxb@techsingularity.net
Signed-off-by: Ingo Molnar 
Signed-off-by: Greg Kroah-Hartman

sched/core: Use READ_ONCE()/WRITE_ONCE() in move_queued_task()/task_rq_lock()

2019-04-05T20:34:43+00:00

[ Upstream commit c546951d9c9300065bad253ecdf1ac59ce9d06c8 ]

move_queued_task() synchronizes with task_rq_lock() as follows:

	move_queued_task()		task_rq_lock()

	[S] ->on_rq = MIGRATING		[L] rq = task_rq()
	WMB (__set_task_cpu())		ACQUIRE (rq->lock);
	[S] ->cpu = new_cpu		[L] ->on_rq

where "[L] rq = task_rq()" is ordered before "ACQUIRE (rq->lock)" by an
address dependency and, in turn, "ACQUIRE (rq->lock)" is ordered before
"[L] ->on_rq" by the ACQUIRE itself.

Use READ_ONCE() to load ->cpu in task_rq() (c.f., task_cpu()) to honor
this address dependency.  Also, mark the accesses to ->cpu and ->on_rq
with READ_ONCE()/WRITE_ONCE() to comply with the LKMM.

Signed-off-by: Andrea Parri 
Signed-off-by: Peter Zijlstra (Intel) 
Cc: Alan Stern 
Cc: Linus Torvalds 
Cc: Mike Galbraith 
Cc: Paul E. McKenney 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: Will Deacon 
Link: https://lkml.kernel.org/r/20190121155240.27173-1-andrea.parri@amarulasolutions.com
Signed-off-by: Ingo Molnar 
Signed-off-by: Sasha Levin

sched/debug: Initialize sd_sysctl_cpus if !CONFIG_CPUMASK_OFFSTACK

2019-04-05T20:34:43+00:00

[ Upstream commit 1ca4fa3ab604734e38e2a3000c9abf788512ffa7 ]

register_sched_domain_sysctl() copies the cpu_possible_mask into
sd_sysctl_cpus, but only if sd_sysctl_cpus hasn't already been
allocated (ie, CONFIG_CPUMASK_OFFSTACK is set).  However, when
CONFIG_CPUMASK_OFFSTACK is not set, sd_sysctl_cpus is left
uninitialized (all zeroes) and the kernel may fail to initialize
sched_domain sysctl entries for all possible CPUs.

This is visible to the user if the kernel is booted with maxcpus=n, or
if ACPI tables have been modified to leave CPUs offline, and then
checking for missing /proc/sys/kernel/sched_domain/cpu* entries.

Fix this by separating the allocation and initialization, and adding a
flag to initialize the possible CPU entries while system booting only.

Tested-by: Syuuichirou Ishii 
Tested-by: Tarumizu, Kohei 
Signed-off-by: Hidetoshi Seto 
Signed-off-by: Peter Zijlstra (Intel) 
Reviewed-by: Masayoshi Mizuma 
Acked-by: Joe Lawrence 
Cc: Linus Torvalds 
Cc: Masayoshi Mizuma 
Cc: Mike Galbraith 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: https://lkml.kernel.org/r/20190129151245.5073-1-msys.mizuma@gmail.com
Signed-off-by: Ingo Molnar 
Signed-off-by: Sasha Levin

sched/topology: Fix percpu data types in struct sd_data & struct s_data

2019-04-05T20:34:40+00:00

[ Upstream commit 99687cdbb3f6c8e32bcc7f37496e811f30460e48 ]

The percpu members of struct sd_data and s_data are declared as:

	struct ... ** __percpu member;

So their type is:

	__percpu pointer to pointer to struct ...

But looking at how they're used, their type should be:

	pointer to __percpu pointer to struct ...

and they should thus be declared as:

	struct ... * __percpu *member;

So fix the placement of '__percpu' in the definition of these
structures.

This addresses a bunch of Sparse's warnings like:

	warning: incorrect type in initializer (different address spaces)
	  expected void const [noderef]  *__vpp_verify
	  got struct sched_domain **

Signed-off-by: Luc Van Oostenryck 
Signed-off-by: Peter Zijlstra (Intel) 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: https://lkml.kernel.org/r/20190118144936.79158-1-luc.vanoostenryck@gmail.com
Signed-off-by: Ingo Molnar 
Signed-off-by: Sasha Levin

psi: avoid divide-by-zero crash inside virtual machines

2019-02-21T17:01:00+00:00

We've been seeing hard-to-trigger psi crashes when running inside VM
instances:

    divide error: 0000 [#1] SMP PTI
    Modules linked in: [...]
    CPU: 0 PID: 212 Comm: kworker/0:2 Not tainted 4.16.18-119_fbk9_3817_gfe944c98d695 #119
    Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 0.0.0 02/06/2015
    Workqueue: events psi_clock
    RIP: 0010:psi_update_stats+0x270/0x490
    RSP: 0018:ffffc90001117e10 EFLAGS: 00010246
    RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffff8800a35a13f8
    RDX: 0000000000000000 RSI: ffff8800a35a1340 RDI: 0000000000000000
    RBP: 0000000000000658 R08: ffff8800a35a1470 R09: 0000000000000000
    R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
    R13: 0000000000000000 R14: 0000000000000000 R15: 00000000000f8502
    FS:  0000000000000000(0000) GS:ffff88023fc00000(0000) knlGS:0000000000000000
    CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
    CR2: 00007fbe370fa000 CR3: 00000000b1e3a000 CR4: 00000000000006f0
    DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
    DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
    Call Trace:
     psi_clock+0x12/0x50
     process_one_work+0x1e0/0x390
     worker_thread+0x2b/0x3c0
     ? rescuer_thread+0x330/0x330
     kthread+0x113/0x130
     ? kthread_create_worker_on_cpu+0x40/0x40
     ? SyS_exit_group+0x10/0x10
     ret_from_fork+0x35/0x40
    Code: 48 0f 47 c7 48 01 c2 45 85 e4 48 89 16 0f 85 e6 00 00 00 4c 8b 49 10 4c 8b 51 08 49 69 d9 f2 07 00 00 48 6b c0 64 4c 8b 29 31 d2 <48> f7 f7 49 69 d5 8d 06 00 00 48 89 c5 4c 69 f0 00 98 0b 00 48

The Code-line points to `period` being 0 inside update_stats(), and we
divide by that when calculating that period's pressure percentage.

The elapsed period should never be 0.  The reason this can happen is due
to an off-by-one in the idle time / missing period calculation combined
with a coarse sched_clock() in the virtual machine.

The target time for aggregation is advanced into the future on a fixed
grid to prevent clock drift.  So when an aggregation runs after some idle
period, we can not just set it to "now + psi_period", but have to
calculate the downtime and advance the target time relative to itself.

However, if the aggregator was disabled exactly one psi_period (ns), we
drop one idle period in the calculation due to a > when we should do >=.
In that case, next_update will be advanced from 'now - psi_period' to
'now' when it should be moved to 'now + psi_period'.  The run finishes
with last_update == next_update == sched_clock().

With hardware clocks, this exact nanosecond match isn't likely in the
first place; but if it does happen, the clock will still have moved on and
the period non-zero by the time the worker runs.  A pointlessly short
period, but besides the extra work, no harm no foul.  However, a slow
sched_clock() like we have on VMs might not have advanced either by the
time the worker runs again.  And when we calculate the elapsed period, the
result, our pressure divisor, will be 0.  Ouch.

Fix this by correctly handling the situation when the elapsed time between
aggregation runs is precisely two periods, and advance the expiration
timestamp correctly to period into the future.

Link: http://lkml.kernel.org/r/20190214193157.15788-1-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner 
Reported-by: Łukasz Siudut 
Cc: Peter Zijlstra 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds