The memcg OOM handling is incredibly fragile and can deadlock. When a
task fails to charge memory, it invokes the OOM killer and loops right
there in the charge code until it succeeds. Comparably, any other task
that enters the charge path at this point will go to a waitqueue right
then and there and sleep until the OOM situation is resolved. The problem
is that these tasks may hold filesystem locks and the mmap_sem; locks that
the selected OOM victim may need to exit.
For example, in one reported case, the task invoking the OOM killer was
about to charge a page cache page during a write(), which holds the
i_mutex. The OOM killer selected a task that was just entering truncate()
and trying to acquire the i_mutex:
OOM invoking task:
mem_cgroup_handle_oom+0x241/0x3b0
mem_cgroup_cache_charge+0xbe/0xe0
add_to_page_cache_locked+0x4c/0x140
add_to_page_cache_lru+0x22/0x50
grab_cache_page_write_begin+0x8b/0xe0
ext3_write_begin+0x88/0x270
generic_file_buffered_write+0x116/0x290
__generic_file_aio_write+0x27c/0x480
generic_file_aio_write+0x76/0xf0 # takes ->i_mutex
do_sync_write+0xea/0x130
vfs_write+0xf3/0x1f0
sys_write+0x51/0x90
system_call_fastpath+0x18/0x1d
OOM kill victim:
do_truncate+0x58/0xa0 # takes i_mutex
do_last+0x250/0xa30
path_openat+0xd7/0x440
do_filp_open+0x49/0xa0
do_sys_open+0x106/0x240
sys_open+0x20/0x30
system_call_fastpath+0x18/0x1d
The OOM handling task will retry the charge indefinitely while the OOM
killed task is not releasing any resources.
A similar scenario can happen when the kernel OOM killer for a memcg is
disabled and a userspace task is in charge of resolving OOM situations.
In this case, ALL tasks that enter the OOM path will be made to sleep on
the OOM waitqueue and wait for userspace to free resources or increase
the group's limit. But a userspace OOM handler is prone to deadlock
itself on the locks held by the waiting tasks. For example one of the
sleeping tasks may be stuck in a brk() call with the mmap_sem held for
writing but the userspace handler, in order to pick an optimal victim,
may need to read files from /proc/<pid>, which tries to acquire the same
mmap_sem for reading and deadlocks.
This patch changes the way tasks behave after detecting a memcg OOM and
makes sure nobody loops or sleeps with locks held:
1. When OOMing in a user fault, invoke the OOM killer and restart the
fault instead of looping on the charge attempt. This way, the OOM
victim can not get stuck on locks the looping task may hold.
2. When OOMing in a user fault but somebody else is handling it
(either the kernel OOM killer or a userspace handler), don't go to
sleep in the charge context. Instead, remember the OOMing memcg in
the task struct and then fully unwind the page fault stack with
-ENOMEM. pagefault_out_of_memory() will then call back into the
memcg code to check if the -ENOMEM came from the memcg, and then
either put the task to sleep on the memcg's OOM waitqueue or just
restart the fault. The OOM victim can no longer get stuck on any
lock a sleeping task may hold.
Debugged by Michal Hocko.
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: azurIt <azurit@pobox.sk>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull KVM updates from Gleb Natapov:
"The highlights of the release are nested EPT and pv-ticketlocks
support (hypervisor part, guest part, which is most of the code, goes
through tip tree). Apart of that there are many fixes for all arches"
Fix up semantic conflicts as discussed in the pull request thread..
* 'next' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (88 commits)
ARM: KVM: Add newlines to panic strings
ARM: KVM: Work around older compiler bug
ARM: KVM: Simplify tracepoint text
ARM: KVM: Fix kvm_set_pte assignment
ARM: KVM: vgic: Bump VGIC_NR_IRQS to 256
ARM: KVM: Bugfix: vgic_bytemap_get_reg per cpu regs
ARM: KVM: vgic: fix GICD_ICFGRn access
ARM: KVM: vgic: simplify vgic_get_target_reg
KVM: MMU: remove unused parameter
KVM: PPC: Book3S PR: Rework kvmppc_mmu_book3s_64_xlate()
KVM: PPC: Book3S PR: Make instruction fetch fallback work for system calls
KVM: PPC: Book3S PR: Don't corrupt guest state when kernel uses VMX
KVM: x86: update masterclock when kvmclock_offset is calculated (v2)
KVM: PPC: Book3S: Fix compile error in XICS emulation
KVM: PPC: Book3S PR: return appropriate error when allocation fails
arch: powerpc: kvm: add signed type cast for comparation
KVM: x86: add comments where MMIO does not return to the emulator
KVM: vmx: count exits to userspace during invalid guest emulation
KVM: rename __kvm_io_bus_sort_cmp to kvm_io_bus_cmp
kvm: optimize away THP checks in kvm_is_mmio_pfn()
...
This reverts commit df54d6fa54.
The commit isn't necessarily wrong, but because it recalculates the
random mmap_base every time, it seems to confuse user memory allocators
that expect contiguous mmap allocations even when the mmap address isn't
specified.
In particular, the MATLAB Java runtime seems to be unhappy. See
https://bugzilla.kernel.org/show_bug.cgi?id=60774
So we'll want to apply the random offset only once, and Radu has a patch
for that. Revert this older commit in order to apply the other one.
Reported-by: Jeff Shorey <shoreyjeff@gmail.com>
Cc: Radu Caragea <sinaelgl@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge a bunch of fixes from Andrew Morton.
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
fs/proc/task_mmu.c: fix buffer overflow in add_page_map()
arch: *: Kconfig: add "kernel/Kconfig.freezer" to "arch/*/Kconfig"
ocfs2: fix null pointer dereference in ocfs2_dir_foreach_blk_id()
x86 get_unmapped_area(): use proper mmap base for bottom-up direction
ocfs2: fix NULL pointer dereference in ocfs2_duplicate_clusters_by_page
ocfs2: Revert 40bd62e to avoid regression in extended allocation
drivers/rtc/rtc-stmp3xxx.c: provide timeout for potentially endless loop polling a HW bit
hugetlb: fix lockdep splat caused by pmd sharing
aoe: adjust ref of head for compound page tails
microblaze: fix clone syscall
mm: save soft-dirty bits on file pages
mm: save soft-dirty bits on swapped pages
memcg: don't initialize kmem-cache destroying work for root caches
Pull scheduler fixes from Ingo Molnar:
"Docbook fixes that make 99% of the diffstat, plus a oneliner fix"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched: Ensure update_cfs_shares() is called for parents of continuously-running tasks
sched: Fix some kernel-doc warnings
Calling freeze_processes sets a global flag that will cause any
process that calls try_to_freeze to enter the refrigerator. It
skips sending a signal to the current task, but if the current
task ever hits try_to_freeze, all threads will be frozen and the
system will deadlock.
Set a new flag, PF_SUSPEND_TASK, on the task that calls
freeze_processes. The flag notifies the freezer that the thread
is involved in suspend and should not be frozen. Also add a
WARN_ON in thaw_processes if the caller does not have the
PF_SUSPEND_TASK flag set to catch if a different task calls
thaw_processes than the one that called freeze_processes, leaving
a task with PF_SUSPEND_TASK permanently set on it.
Threads that spawn off a task with PF_SUSPEND_TASK set (which
swsusp does) will also have PF_SUSPEND_TASK set, preventing them
from freezing while they are helping with suspend, but they need
to be dead by the time suspend is triggered, otherwise they may
run when userspace is expected to be frozen. Add a WARN_ON in
thaw_processes if more than one thread has the PF_SUSPEND_TASK
flag set.
Reported-and-tested-by: Michael Leun <lkml20130126@newton.leun.net>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The wake-affine scheduler feature is currently always trying to pull
the wakee close to the waker. In theory this should be beneficial if
the waker's CPU caches hot data for the wakee, and it's also beneficial
in the extreme ping-pong high context switch rate case.
Testing shows it can benefit hackbench up to 15%.
However, the feature is somewhat blind, from which some workloads
such as pgbench suffer. It's also time-consuming algorithmically.
Testing shows it can damage pgbench up to 50% - far more than the
benefit it brings in the best case.
So wake-affine should be smarter and it should realize when to
stop its thankless effort at trying to find a suitable CPU to wake on.
This patch introduces 'wakee_flips', which will be increased each
time the task flips (switches) its wakee target.
So a high 'wakee_flips' value means the task has more than one
wakee, and the bigger the number, the higher the wakeup frequency.
Now when making the decision on whether to pull or not, pay attention to
the wakee with a high 'wakee_flips', pulling such a task may benefit
the wakee. Also imply that the waker will face cruel competition later,
it could be very cruel or very fast depends on the story behind
'wakee_flips', waker therefore suffers.
Furthermore, if waker also has a high 'wakee_flips', that implies that
multiple tasks rely on it, then waker's higher latency will damage all
of them, so pulling wakee seems to be a bad deal.
Thus, when 'waker->wakee_flips / wakee->wakee_flips' becomes
higher and higher, the cost of pulling seems to be worse and worse.
The patch therefore helps the wake-affine feature to stop its pulling
work when:
wakee->wakee_flips > factor &&
waker->wakee_flips > (factor * wakee->wakee_flips)
The 'factor' here is the number of CPUs in the current CPU's NUMA node,
so a bigger node will lead to more pulling since the trial becomes more
severe.
After applying the patch, pgbench shows up to 40% improvements and no regressions.
Tested with 12 cpu x86 server and tip 3.10.0-rc7.
The percentages in the final column highlight the areas with the biggest wins,
all other areas improved as well:
pgbench base smart
| db_size | clients | tps | | tps |
+---------+---------+-------+ +-------+
| 22 MB | 1 | 10598 | | 10796 |
| 22 MB | 2 | 21257 | | 21336 |
| 22 MB | 4 | 41386 | | 41622 |
| 22 MB | 8 | 51253 | | 57932 |
| 22 MB | 12 | 48570 | | 54000 |
| 22 MB | 16 | 46748 | | 55982 | +19.75%
| 22 MB | 24 | 44346 | | 55847 | +25.93%
| 22 MB | 32 | 43460 | | 54614 | +25.66%
| 7484 MB | 1 | 8951 | | 9193 |
| 7484 MB | 2 | 19233 | | 19240 |
| 7484 MB | 4 | 37239 | | 37302 |
| 7484 MB | 8 | 46087 | | 50018 |
| 7484 MB | 12 | 42054 | | 48763 |
| 7484 MB | 16 | 40765 | | 51633 | +26.66%
| 7484 MB | 24 | 37651 | | 52377 | +39.11%
| 7484 MB | 32 | 37056 | | 51108 | +37.92%
| 15 GB | 1 | 8845 | | 9104 |
| 15 GB | 2 | 19094 | | 19162 |
| 15 GB | 4 | 36979 | | 36983 |
| 15 GB | 8 | 46087 | | 49977 |
| 15 GB | 12 | 41901 | | 48591 |
| 15 GB | 16 | 40147 | | 50651 | +26.16%
| 15 GB | 24 | 37250 | | 52365 | +40.58%
| 15 GB | 32 | 36470 | | 50015 | +37.14%
Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/51D50057.9000809@linux.vnet.ibm.com
[ Improved the changelog. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Linux as a guest on KVM hypervisor, the only user of the pvclock
vsyscall interface, does not require notification on task migration
because:
1. cpu ID number maps 1:1 to per-CPU pvclock time info.
2. per-CPU pvclock time info is updated if the
underlying CPU changes.
3. that version is increased whenever underlying CPU
changes.
Which is sufficient to guarantee nanoseconds counter
is calculated properly.
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Pull networking updates from David Miller:
"This is a re-do of the net-next pull request for the current merge
window. The only difference from the one I made the other day is that
this has Eliezer's interface renames and the timeout handling changes
made based upon your feedback, as well as a few bug fixes that have
trickeled in.
Highlights:
1) Low latency device polling, eliminating the cost of interrupt
handling and context switches. Allows direct polling of a network
device from socket operations, such as recvmsg() and poll().
Currently ixgbe, mlx4, and bnx2x support this feature.
Full high level description, performance numbers, and design in
commit 0a4db187a9 ("Merge branch 'll_poll'")
From Eliezer Tamir.
2) With the routing cache removed, ip_check_mc_rcu() gets exercised
more than ever before in the case where we have lots of multicast
addresses. Use a hash table instead of a simple linked list, from
Eric Dumazet.
3) Add driver for Atheros CQA98xx 802.11ac wireless devices, from
Bartosz Markowski, Janusz Dziedzic, Kalle Valo, Marek Kwaczynski,
Marek Puzyniak, Michal Kazior, and Sujith Manoharan.
4) Support reporting the TUN device persist flag to userspace, from
Pavel Emelyanov.
5) Allow controlling network device VF link state using netlink, from
Rony Efraim.
6) Support GRE tunneling in openvswitch, from Pravin B Shelar.
7) Adjust SOCK_MIN_RCVBUF and SOCK_MIN_SNDBUF for modern times, from
Daniel Borkmann and Eric Dumazet.
8) Allow controlling of TCP quickack behavior on a per-route basis,
from Cong Wang.
9) Several bug fixes and improvements to vxlan from Stephen
Hemminger, Pravin B Shelar, and Mike Rapoport. In particular,
support receiving on multiple UDP ports.
10) Major cleanups, particular in the area of debugging and cookie
lifetime handline, to the SCTP protocol code. From Daniel
Borkmann.
11) Allow packets to cross network namespaces when traversing tunnel
devices. From Nicolas Dichtel.
12) Allow monitoring netlink traffic via AF_PACKET sockets, in a
manner akin to how we monitor real network traffic via ptype_all.
From Daniel Borkmann.
13) Several bug fixes and improvements for the new alx device driver,
from Johannes Berg.
14) Fix scalability issues in the netem packet scheduler's time queue,
by using an rbtree. From Eric Dumazet.
15) Several bug fixes in TCP loss recovery handling, from Yuchung
Cheng.
16) Add support for GSO segmentation of MPLS packets, from Simon
Horman.
17) Make network notifiers have a real data type for the opaque
pointer that's passed into them. Use this to properly handle
network device flag changes in arp_netdev_event(). From Jiri
Pirko and Timo Teräs.
18) Convert several drivers over to module_pci_driver(), from Peter
Huewe.
19) tcp_fixup_rcvbuf() can loop 500 times over loopback, just use a
O(1) calculation instead. From Eric Dumazet.
20) Support setting of explicit tunnel peer addresses in ipv6, just
like ipv4. From Nicolas Dichtel.
21) Protect x86 BPF JIT against spraying attacks, from Eric Dumazet.
22) Prevent a single high rate flow from overruning an individual cpu
during RX packet processing via selective flow shedding. From
Willem de Bruijn.
23) Don't use spinlocks in TCP md5 signing fast paths, from Eric
Dumazet.
24) Don't just drop GSO packets which are above the TBF scheduler's
burst limit, chop them up so they are in-bounds instead. Also
from Eric Dumazet.
25) VLAN offloads are missed when configured on top of a bridge, fix
from Vlad Yasevich.
26) Support IPV6 in ping sockets. From Lorenzo Colitti.
27) Receive flow steering targets should be updated at poll() time
too, from David Majnemer.
28) Fix several corner case regressions in PMTU/redirect handling due
to the routing cache removal, from Timo Teräs.
29) We have to be mindful of ipv4 mapped ipv6 sockets in
upd_v6_push_pending_frames(). From Hannes Frederic Sowa.
30) Fix L2TP sequence number handling bugs, from James Chapman."
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1214 commits)
drivers/net: caif: fix wrong rtnl_is_locked() usage
drivers/net: enic: release rtnl_lock on error-path
vhost-net: fix use-after-free in vhost_net_flush
net: mv643xx_eth: do not use port number as platform device id
net: sctp: confirm route during forward progress
virtio_net: fix race in RX VQ processing
virtio: support unlocked queue poll
net/cadence/macb: fix bug/typo in extracting gem_irq_read_clear bit
Documentation: Fix references to defunct linux-net@vger.kernel.org
net/fs: change busy poll time accounting
net: rename low latency sockets functions to busy poll
bridge: fix some kernel warning in multicast timer
sfc: Fix memory leak when discarding scattered packets
sit: fix tunnel update via netlink
dt:net:stmmac: Add dt specific phy reset callback support.
dt:net:stmmac: Add support to dwmac version 3.610 and 3.710
dt:net:stmmac: Allocate platform data only if its NULL.
net:stmmac: fix memleak in the open method
ipv6: rt6_check_neigh should successfully verify neigh if no NUD information are available
net: ipv6: fix wrong ping_v6_sendmsg return value
...
Move __set_special_pids() from exit.c to sys.c close to its single caller
and make it static.
And rename it to set_special_pids(), another helper with this name has
gone away.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is intended for use in loops which read data protected by RCU and may
have a large number of iterations. Such an example is dumping the list of
connections known to IPVS: ip_vs_conn_array() and ip_vs_conn_seq_next().
The benefits are for CONFIG_PREEMPT_RCU=y where we save CPU cycles
by moving rcu_read_lock and rcu_read_unlock out of large loops
but still allowing the current task to be preempted after every
loop iteration for the CONFIG_PREEMPT_RCU=n case.
The call to cond_resched() is not needed when CONFIG_PREEMPT_RCU=y.
Thanks to Paul E. McKenney for explaining this and for the
final version that checks the context with CONFIG_DEBUG_ATOMIC_SLEEP=y
for all possible configurations.
The function can be empty in the CONFIG_PREEMPT_RCU case,
rcu_read_lock and rcu_read_unlock are not needed in this case
because the task can be preempted on indication from scheduler.
Thanks to Peter Zijlstra for catching this and for his help
in trying a solution that changes __might_sleep.
Initial cond_resched_rcu_lock() function suggested by Eric Dumazet.
Tested-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Pull audit changes from Eric Paris:
"Al used to send pull requests every couple of years but he told me to
just start pushing them to you directly.
Our touching outside of core audit code is pretty straight forward. A
couple of interface changes which hit net/. A simple argument bug
calling audit functions in namei.c and the removal of some assembly
branch prediction code on ppc"
* git://git.infradead.org/users/eparis/audit: (31 commits)
audit: fix message spacing printing auid
Revert "audit: move kaudit thread start from auditd registration to kaudit init"
audit: vfs: fix audit_inode call in O_CREAT case of do_last
audit: Make testing for a valid loginuid explicit.
audit: fix event coverage of AUDIT_ANOM_LINK
audit: use spin_lock in audit_receive_msg to process tty logging
audit: do not needlessly take a lock in tty_audit_exit
audit: do not needlessly take a spinlock in copy_signal
audit: add an option to control logging of passwords with pam_tty_audit
audit: use spin_lock_irqsave/restore in audit tty code
helper for some session id stuff
audit: use a consistent audit helper to log lsm information
audit: push loginuid and sessionid processing down
audit: stop pushing loginid, uid, sessionid as arguments
audit: remove the old depricated kernel interface
audit: make validity checking generic
audit: allow checking the type of audit message in the user filter
audit: fix build break when AUDIT_DEBUG == 2
audit: remove duplicate export of audit_enabled
Audit: do not print error when LSMs disabled
...
Pull block driver updates from Jens Axboe:
"It might look big in volume, but when categorized, not a lot of
drivers are touched. The pull request contains:
- mtip32xx fixes from Micron.
- A slew of drbd updates, this time in a nicer series.
- bcache, a flash/ssd caching framework from Kent.
- Fixes for cciss"
* 'for-3.10/drivers' of git://git.kernel.dk/linux-block: (66 commits)
bcache: Use bd_link_disk_holder()
bcache: Allocator cleanup/fixes
cciss: bug fix to prevent cciss from loading in kdump crash kernel
cciss: add cciss_allow_hpsa module parameter
drivers/block/mg_disk.c: add CONFIG_PM_SLEEP to suspend/resume functions
mtip32xx: Workaround for unaligned writes
bcache: Make sure blocksize isn't smaller than device blocksize
bcache: Fix merge_bvec_fn usage for when it modifies the bvm
bcache: Correctly check against BIO_MAX_PAGES
bcache: Hack around stuff that clones up to bi_max_vecs
bcache: Set ra_pages based on backing device's ra_pages
bcache: Take data offset from the bdev superblock.
mtip32xx: mtip32xx: Disable TRIM support
mtip32xx: fix a smatch warning
bcache: Disable broken btree fuzz tester
bcache: Fix a format string overflow
bcache: Fix a minor memory leak on device teardown
bcache: Documentation updates
bcache: Use WARN_ONCE() instead of __WARN()
bcache: Add missing #include <linux/prefetch.h>
...
Pull 'full dynticks' support from Ingo Molnar:
"This tree from Frederic Weisbecker adds a new, (exciting! :-) core
kernel feature to the timer and scheduler subsystems: 'full dynticks',
or CONFIG_NO_HZ_FULL=y.
This feature extends the nohz variable-size timer tick feature from
idle to busy CPUs (running at most one task) as well, potentially
reducing the number of timer interrupts significantly.
This feature got motivated by real-time folks and the -rt tree, but
the general utility and motivation of full-dynticks runs wider than
that:
- HPC workloads get faster: CPUs running a single task should be able
to utilize a maximum amount of CPU power. A periodic timer tick at
HZ=1000 can cause a constant overhead of up to 1.0%. This feature
removes that overhead - and speeds up the system by 0.5%-1.0% on
typical distro configs even on modern systems.
- Real-time workload latency reduction: CPUs running critical tasks
should experience as little jitter as possible. The last remaining
source of kernel-related jitter was the periodic timer tick.
- A single task executing on a CPU is a pretty common situation,
especially with an increasing number of cores/CPUs, so this feature
helps desktop and mobile workloads as well.
The cost of the feature is mainly related to increased timer
reprogramming overhead when a CPU switches its tick period, and thus
slightly longer to-idle and from-idle latency.
Configuration-wise a third mode of operation is added to the existing
two NOHZ kconfig modes:
- CONFIG_HZ_PERIODIC: [formerly !CONFIG_NO_HZ], now explicitly named
as a config option. This is the traditional Linux periodic tick
design: there's a HZ tick going on all the time, regardless of
whether a CPU is idle or not.
- CONFIG_NO_HZ_IDLE: [formerly CONFIG_NO_HZ=y], this turns off the
periodic tick when a CPU enters idle mode.
- CONFIG_NO_HZ_FULL: this new mode, in addition to turning off the
tick when a CPU is idle, also slows the tick down to 1 Hz (one
timer interrupt per second) when only a single task is running on a
CPU.
The .config behavior is compatible: existing !CONFIG_NO_HZ and
CONFIG_NO_HZ=y settings get translated to the new values, without the
user having to configure anything. CONFIG_NO_HZ_FULL is turned off by
default.
This feature is based on a lot of infrastructure work that has been
steadily going upstream in the last 2-3 cycles: related RCU support
and non-periodic cputime support in particular is upstream already.
This tree adds the final pieces and activates the feature. The pull
request is marked RFC because:
- it's marked 64-bit only at the moment - the 32-bit support patch is
small but did not get ready in time.
- it has a number of fresh commits that came in after the merge
window. The overwhelming majority of commits are from before the
merge window, but still some aspects of the tree are fresh and so I
marked it RFC.
- it's a pretty wide-reaching feature with lots of effects - and
while the components have been in testing for some time, the full
combination is still not very widely used. That it's default-off
should reduce its regression abilities and obviously there are no
known regressions with CONFIG_NO_HZ_FULL=y enabled either.
- the feature is not completely idempotent: there is no 100%
equivalent replacement for a periodic scheduler/timer tick. In
particular there's ongoing work to map out and reduce its effects
on scheduler load-balancing and statistics. This should not impact
correctness though, there are no known regressions related to this
feature at this point.
- it's a pretty ambitious feature that with time will likely be
enabled by most Linux distros, and we'd like you to make input on
its design/implementation, if you dislike some aspect we missed.
Without flaming us to crisp! :-)
Future plans:
- there's ongoing work to reduce 1Hz to 0Hz, to essentially shut off
the periodic tick altogether when there's a single busy task on a
CPU. We'd first like 1 Hz to be exposed more widely before we go
for the 0 Hz target though.
- once we reach 0 Hz we can remove the periodic tick assumption from
nr_running>=2 as well, by essentially interrupting busy tasks only
as frequently as the sched_latency constraints require us to do -
once every 4-40 msecs, depending on nr_running.
I am personally leaning towards biting the bullet and doing this in
v3.10, like the -rt tree this effort has been going on for too long -
but the final word is up to you as usual.
More technical details can be found in Documentation/timers/NO_HZ.txt"
* 'timers-nohz-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (39 commits)
sched: Keep at least 1 tick per second for active dynticks tasks
rcu: Fix full dynticks' dependency on wide RCU nocb mode
nohz: Protect smp_processor_id() in tick_nohz_task_switch()
nohz_full: Add documentation.
cputime_nsecs: use math64.h for nsec resolution conversion helpers
nohz: Select VIRT_CPU_ACCOUNTING_GEN from full dynticks config
nohz: Reduce overhead under high-freq idling patterns
nohz: Remove full dynticks' superfluous dependency on RCU tree
nohz: Fix unavailable tick_stop tracepoint in dynticks idle
nohz: Add basic tracing
nohz: Select wide RCU nocb for full dynticks
nohz: Disable the tick when irq resume in full dynticks CPU
nohz: Re-evaluate the tick for the new task after a context switch
nohz: Prepare to stop the tick on irq exit
nohz: Implement full dynticks kick
nohz: Re-evaluate the tick from the scheduler IPI
sched: New helper to prevent from stopping the tick in full dynticks
sched: Kick full dynticks CPU that have more than one task enqueued.
perf: New helper to prevent full dynticks CPUs from stopping tick
perf: Kick full dynticks CPU if events rotation is needed
...
Johannes Weiner
Oleg Nesterov
Linus Torvalds
Ingo Molnar
Radu Caragea
Colin Cross
Michael Wang
Marcelo Tosatti
Yacine Belkadi
Michel Lespinasse
Kamalesh Babulal
Alex Shi
Simon Horman
Kent Overstreet