It was discovered that a constant stream of readers with occassional
writers pounding on a rwsem may cause many of the readers to enter the
slowpath unnecessarily thus increasing latency and lowering performance.
In the current code, a reader entering the slowpath critical section
will unconditionally set the WAITING_BIAS, if not set yet, and clear
its active count even if no one is in the wait queue and no writer
is present. This causes some incoming readers to observe the presence
of waiters in the wait queue and hence have to go into the slowpath
themselves.
With sufficient numbers of readers and a relatively short lock hold time,
the WAITING_BIAS may be repeatedly turned on and off and a substantial
portion of the readers will go into the slowpath sustaining a rather
long queue in the wait queue spinlock and repeated WAITING_BIAS on/off
cycle until the logjam is broken opportunistically.
To avoid this situation from happening, an additional check is added to
detect the special case that the reader in the critical section is the
only one in the wait queue and no writer is present. When that happens,
it can just exit the slowpath and return immediately as its active count
has already been set in the lock. Other incoming readers won't observe
the presence of waiters and so will not be forced into the slowpath.
The issue was found in a customer site where they had an application
that pounded on the pread64 syscalls heavily on an XFS filesystem. The
application was run in a recent 4-socket boxes with a lot of CPUs. They
saw significant spinlock contention in the rwsem_down_read_failed() call.
With this patch applied, the system CPU usage went down from 85% to 57%,
and the spinlock contention in the pread64 syscalls was gone.
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Davidlohr Bueso <dbueso@suse.de>
Acked-by: Will Deacon <will.deacon@arm.com>
Cc: Joe Mario <jmario@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1532459425-19204-1-git-send-email-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
It can happen that load_balance() finds a busiest group and then a
busiest rq but the calculated imbalance is in fact 0.
In such situation, detach_tasks() returns immediately and lets the
flag LBF_ALL_PINNED set. The busiest CPU is then wrongly assumed to
have pinned tasks and removed from the load balance mask. then, we
redo a load balance without the busiest CPU. This creates wrong load
balance situation and generates wrong task migration.
If the calculated imbalance is 0, it's useless to try to find a
busiest rq as no task will be migrated and we can return immediately.
This situation can happen with heterogeneous system or smp system when
RT tasks are decreasing the capacity of some CPUs.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: dietmar.eggemann@arm.com
Cc: jhugo@codeaurora.org
Link: http://lkml.kernel.org/r/1536306664-29827-1-git-send-email-vincent.guittot@linaro.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When a task which previously ran on a given CPU is remotely queued to
wake up on that same CPU, there is a period where the task's state is
TASK_WAKING and its vruntime is not normalized. This is not accounted
for in vruntime_normalized() which will cause an error in the task's
vruntime if it is switched from the fair class during this time.
For example if it is boosted to RT priority via rt_mutex_setprio(),
rq->min_vruntime will not be subtracted from the task's vruntime but
it will be added again when the task returns to the fair class. The
task's vruntime will have been erroneously doubled and the effective
priority of the task will be reduced.
Note this will also lead to inflation of all vruntimes since the doubled
vruntime value will become the rq's min_vruntime when other tasks leave
the rq. This leads to repeated doubling of the vruntime and priority
penalty.
Fix this by recognizing a WAKING task's vruntime as normalized only if
sched_remote_wakeup is true. This indicates a migration, in which case
the vruntime would have been normalized in migrate_task_rq_fair().
Based on a similar patch from John Dias <joaodias@google.com>.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Steve Muckle <smuckle@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Chris Redpath <Chris.Redpath@arm.com>
Cc: John Dias <joaodias@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Miguel de Dios <migueldedios@google.com>
Cc: Morten Rasmussen <Morten.Rasmussen@arm.com>
Cc: Patrick Bellasi <Patrick.Bellasi@arm.com>
Cc: Paul Turner <pjt@google.com>
Cc: Quentin Perret <quentin.perret@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Todd Kjos <tkjos@google.com>
Cc: kernel-team@android.com
Fixes: b5179ac70d ("sched/fair: Prepare to fix fairness problems on migration")
Link: http://lkml.kernel.org/r/20180831224217.169476-1-smuckle@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
update_blocked_averages() is called to periodiccally decay the stalled load
of idle CPUs and to sync all loads before running load balance.
When cfs rq is idle, it trigs a load balance during pick_next_task_fair()
in order to potentially pull tasks and to use this newly idle CPU. This
load balance happens whereas prev task from another class has not been put
and its utilization updated yet. This may lead to wrongly account running
time as idle time for RT or DL classes.
Test that no RT or DL task is running when updating their utilization in
update_blocked_averages().
We still update RT and DL utilization instead of simply skipping them to
make sure that all metrics are synced when used during load balance.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 371bf42732 ("sched/rt: Add rt_rq utilization tracking")
Fixes: 3727e0e163 ("sched/dl: Add dl_rq utilization tracking")
Link: http://lkml.kernel.org/r/1535728975-22799-1-git-send-email-vincent.guittot@linaro.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull perf/urgent fixes from Arnaldo Carvalho de Melo:
Kernel:
- Modify breakpoint fixes (Jiri Olsa)
perf annotate:
- Fix parsing aarch64 branch instructions after objdump update (Kim Phillips)
- Fix parsing indirect calls in 'perf annotate' (Martin Liška)
perf probe:
- Ignore SyS symbols irrespective of endianness on PowerPC (Sandipan Das)
perf trace:
- Fix include path for asm-generic/unistd.h on arm64 (Kim Phillips)
Core libraries:
- Fix potential null pointer dereference in perf_evsel__new_idx() (Hisao Tanabe)
- Use fixed size string for comms instead of scanf("%m"), that is
not present in the bionic libc and leads to a crash (Chris Phlipot)
- Fix bad memory access in trace info on 32-bit systems, we were reading
8 bytes from a 4-byte long variable when saving the command line in the
perf.data file. (Chris Phlipot)
Build system:
- Streamline bpf examples and headers installation, clarifying
some install messages. (Arnaldo Carvalho de Melo)
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull timekeeping fixes from Thomas Gleixner:
"Two fixes for timekeeping:
- Revert to the previous kthread based update, which is unfortunately
required due to lock ordering issues. The removal caused boot
failures on old Core2 machines. Add a proper comment why the thread
needs to stay to prevent accidental removal in the future.
- Fix a silly typo in a function declaration"
* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
clocksource: Revert "Remove kthread"
timekeeping: Fix declaration of read_persistent_wall_and_boot_offset()
Pull cpu hotplug fixes from Thomas Gleixner:
"Two fixes for the hotplug state machine code:
- Move the misplaces smb() in the hotplug thread function to the
proper place, otherwise a half update control struct could be
observed
- Prevent state corruption on error rollback, which causes the state
to advance by one and as a consequence skip it in the bringup
sequence"
* 'smp-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
cpu/hotplug: Prevent state corruption on error rollback
cpu/hotplug: Adjust misplaced smb() in cpuhp_thread_fun()
This goes through a lot of hooks just to call arch_teardown_dma_ops.
Replace it with a direct call instead.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
There is no good reason for this indirection given that the method
always exists.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Pull tracing fixes from Steven Rostedt:
"This fixes two annoying bugs:
- The first one is a side effect caused by using SRCU for rcuidle
tracepoints. It seems that the perf was depending on the rcuidle
tracepoints to make RCU watch when it wasn't.
The real fix will be to have perf use SRCU instead of depending on
RCU watching, but that can't be done until SRCU is safe to use in
NMI context (Paul's working on that).
- The second bug fix is for a bug that's been periodically making my
tests fail randomly for some time. I haven't had time to track it
down, but finally have. It has to do with stressing NMIs (via perf)
while enabling or disabling ftrace function handling with lockdep
enabled.
If an interrupt happens and just as it returns, it sets lockdep
back to "interrupts enabled" but before it returns an NMI is
triggered, and if this happens while printk_nmi_enter has a
breakpoint attached to it (because ftrace is converting it to or
from nop to call fentry), the breakpoint trap also calls into
lockdep, and since returning from the NMI to a interrupt handler,
interrupts were disabled when the NMI went off, lockdep keeps its
state as interrupts disabled when it returns back from the
interrupt handler where interrupts are enabled.
This causes lockdep_assert_irqs_enabled() to trigger a false
positive"
* tag 'trace-v4.19-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
printk/tracing: Do not trace printk_nmi_enter()
tracing: Add back in rcu_irq_enter/exit_irqson() for rcuidle tracepoints
I hit the following splat in my tests:
------------[ cut here ]------------
IRQs not enabled as expected
WARNING: CPU: 3 PID: 0 at kernel/time/tick-sched.c:982 tick_nohz_idle_enter+0x44/0x8c
Modules linked in: ip6t_REJECT nf_reject_ipv6 ip6table_filter ip6_tables ipv6
CPU: 3 PID: 0 Comm: swapper/3 Not tainted 4.19.0-rc2-test+ #2
Hardware name: MSI MS-7823/CSM-H87M-G43 (MS-7823), BIOS V1.6 02/22/2014
EIP: tick_nohz_idle_enter+0x44/0x8c
Code: ec 05 00 00 00 75 26 83 b8 c0 05 00 00 00 75 1d 80 3d d0 36 3e c1 00
75 14 68 94 63 12 c1 c6 05 d0 36 3e c1 01 e8 04 ee f8 ff <0f> 0b 58 fa bb a0
e5 66 c1 e8 25 0f 04 00 64 03 1d 28 31 52 c1 8b
EAX: 0000001c EBX: f26e7f8c ECX: 00000006 EDX: 00000007
ESI: f26dd1c0 EDI: 00000000 EBP: f26e7f40 ESP: f26e7f38
DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 EFLAGS: 00010296
CR0: 80050033 CR2: 0813c6b0 CR3: 2f342000 CR4: 001406f0
Call Trace:
do_idle+0x33/0x202
cpu_startup_entry+0x61/0x63
start_secondary+0x18e/0x1ed
startup_32_smp+0x164/0x168
irq event stamp: 18773830
hardirqs last enabled at (18773829): [<c040150c>] trace_hardirqs_on_thunk+0xc/0x10
hardirqs last disabled at (18773830): [<c040151c>] trace_hardirqs_off_thunk+0xc/0x10
softirqs last enabled at (18773824): [<c0ddaa6f>] __do_softirq+0x25f/0x2bf
softirqs last disabled at (18773767): [<c0416bbe>] call_on_stack+0x45/0x4b
---[ end trace b7c64aa79e17954a ]---
After a bit of debugging, I found what was happening. This would trigger
when performing "perf" with a high NMI interrupt rate, while enabling and
disabling function tracer. Ftrace uses breakpoints to convert the nops at
the start of functions to calls to the function trampolines. The breakpoint
traps disable interrupts and this makes calls into lockdep via the
trace_hardirqs_off_thunk in the entry.S code. What happens is the following:
do_idle {
[interrupts enabled]
<interrupt> [interrupts disabled]
TRACE_IRQS_OFF [lockdep says irqs off]
[...]
TRACE_IRQS_IRET
test if pt_regs say return to interrupts enabled [yes]
TRACE_IRQS_ON [lockdep says irqs are on]
<nmi>
nmi_enter() {
printk_nmi_enter() [traced by ftrace]
[ hit ftrace breakpoint ]
<breakpoint exception>
TRACE_IRQS_OFF [lockdep says irqs off]
[...]
TRACE_IRQS_IRET [return from breakpoint]
test if pt_regs say interrupts enabled [no]
[iret back to interrupt]
[iret back to code]
tick_nohz_idle_enter() {
lockdep_assert_irqs_enabled() [lockdep say no!]
Although interrupts are indeed enabled, lockdep thinks it is not, and since
we now do asserts via lockdep, it gives a false warning. The issue here is
that printk_nmi_enter() is called before lockdep_off(), which disables
lockdep (for this reason) in NMIs. By simply not allowing ftrace to see
printk_nmi_enter() (via notrace annotation) we keep lockdep from getting
confused.
Cc: stable@vger.kernel.org
Fixes: 42a0bb3f71 ("printk/nmi: generic solution for safe printk in NMI")
Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Acked-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Edward Cree says:
In check_mem_access(), for the PTR_TO_CTX case, after check_ctx_access()
has supplied a reg_type, the other members of the register state are set
appropriately. Previously reg.range was set to 0, but as it is in a
union with reg.map_ptr, which is larger, upper bytes of the latter were
left in place. This then caused the memcmp() in regsafe() to fail,
preventing some branches from being pruned (and occasionally causing the
same program to take a varying number of processed insns on repeated
verifier runs).
Fix the instability by clearing bpf_reg_state in __mark_reg_[un]known()
Fixes: f1174f77b5 ("bpf/verifier: rework value tracking")
Debugged-by: Edward Cree <ecree@solarflare.com>
Acked-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Merge misc fixes from Andrew Morton:
"17 fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
nilfs2: convert to SPDX license tags
drivers/dax/device.c: convert variable to vm_fault_t type
lib/Kconfig.debug: fix three typos in help text
checkpatch: add __ro_after_init to known $Attribute
mm: fix BUG_ON() in vmf_insert_pfn_pud() from VM_MIXEDMAP removal
uapi/linux/keyctl.h: don't use C++ reserved keyword as a struct member name
memory_hotplug: fix kernel_panic on offline page processing
checkpatch: add optional static const to blank line declarations test
ipc/shm: properly return EIDRM in shm_lock()
mm/hugetlb: filter out hugetlb pages if HUGEPAGE migration is not supported.
mm/util.c: improve kvfree() kerneldoc
tools/vm/page-types.c: fix "defined but not used" warning
tools/vm/slabinfo.c: fix sign-compare warning
kmemleak: always register debugfs file
mm: respect arch_dup_mmap() return value
mm, oom: fix missing tlb_finish_mmu() in __oom_reap_task_mm().
mm: memcontrol: print proper OOM header when no eligible victim left
Pull networking fixes from David Miller:
1) Must perform TXQ teardown before unregistering interfaces in
mac80211, from Toke Høiland-Jørgensen.
2) Don't allow creating mac80211_hwsim with less than one channel, from
Johannes Berg.
3) Division by zero in cfg80211, fix from Johannes Berg.
4) Fix endian issue in tipc, from Haiqing Bai.
5) BPF sockmap use-after-free fixes from Daniel Borkmann.
6) Spectre-v1 in mac80211_hwsim, from Jinbum Park.
7) Missing rhashtable_walk_exit() in tipc, from Cong Wang.
8) Revert kvzalloc() conversion of AF_PACKET, it breaks mmap() when
kvzalloc() tries to use kmalloc() pages. From Eric Dumazet.
9) Fix deadlock in hv_netvsc, from Dexuan Cui.
10) Do not restart timewait timer on RST, from Florian Westphal.
11) Fix double lwstate refcount grab in ipv6, from Alexey Kodanev.
12) Unsolicit report count handling is off-by-one, fix from Hangbin Liu.
13) Sleep-in-atomic in cadence driver, from Jia-Ju Bai.
14) Respect ttl-inherit in ip6 tunnel driver, from Hangbin Liu.
15) Use-after-free in act_ife, fix from Cong Wang.
16) Missing hold to meta module in act_ife, from Vlad Buslov.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (91 commits)
net: phy: sfp: Handle unimplemented hwmon limits and alarms
net: sched: action_ife: take reference to meta module
act_ife: fix a potential use-after-free
net/mlx5: Fix SQ offset in QPs with small RQ
tipc: correct spelling errors for tipc_topsrv_queue_evt() comments
tipc: correct spelling errors for struct tipc_bc_base's comment
bnxt_en: Do not adjust max_cp_rings by the ones used by RDMA.
bnxt_en: Clean up unused functions.
bnxt_en: Fix firmware signaled resource change logic in open.
sctp: not traverse asoc trans list if non-ipv6 trans exists for ipv6_flowlabel
sctp: fix invalid reference to the index variable of the iterator
net/ibm/emac: wrong emac_calc_base call was used by typo
net: sched: null actions array pointer before releasing action
vhost: fix VHOST_GET_BACKEND_FEATURES ioctl request definition
r8169: add support for NCube 8168 network card
ip6_tunnel: respect ttl inherit for ip6tnl
mac80211: shorten the IBSS debug messages
mac80211: don't Tx a deauth frame if the AP forbade Tx
mac80211: Fix station bandwidth setting after channel switch
mac80211: fix a race between restart and CSA flows
...
Introduce CONFIG_STACKLEAK_RUNTIME_DISABLE option, which provides
'stack_erasing' sysctl. It can be used in runtime to control kernel
stack erasing for kernels built with CONFIG_GCC_PLUGIN_STACKLEAK.
Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Alexander Popov <alex.popov@linux.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Introduce CONFIG_STACKLEAK_METRICS providing STACKLEAK information about
tasks via the /proc file system. In particular, /proc/<pid>/stack_depth
shows the maximum kernel stack consumption for the current and previous
syscalls. Although this information is not precise, it can be useful for
estimating the STACKLEAK performance impact for your workloads.
Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Alexander Popov <alex.popov@linux.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
The STACKLEAK feature erases the kernel stack before returning from
syscalls. That reduces the information which kernel stack leak bugs can
reveal and blocks some uninitialized stack variable attacks.
This commit introduces the STACKLEAK gcc plugin. It is needed for
tracking the lowest border of the kernel stack, which is important
for the code erasing the used part of the kernel stack at the end
of syscalls (comes in a separate commit).
The STACKLEAK feature is ported from grsecurity/PaX. More information at:
https://grsecurity.net/https://pax.grsecurity.net/
This code is modified from Brad Spengler/PaX Team's code in the last
public patch of grsecurity/PaX based on our understanding of the code.
Changes or omissions from the original code are ours and don't reflect
the original grsecurity/PaX code.
Signed-off-by: Alexander Popov <alex.popov@linux.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
The STACKLEAK feature (initially developed by PaX Team) has the following
benefits:
1. Reduces the information that can be revealed through kernel stack leak
bugs. The idea of erasing the thread stack at the end of syscalls is
similar to CONFIG_PAGE_POISONING and memzero_explicit() in kernel
crypto, which all comply with FDP_RIP.2 (Full Residual Information
Protection) of the Common Criteria standard.
2. Blocks some uninitialized stack variable attacks (e.g. CVE-2017-17712,
CVE-2010-2963). That kind of bugs should be killed by improving C
compilers in future, which might take a long time.
This commit introduces the code filling the used part of the kernel
stack with a poison value before returning to userspace. Full
STACKLEAK feature also contains the gcc plugin which comes in a
separate commit.
The STACKLEAK feature is ported from grsecurity/PaX. More information at:
https://grsecurity.net/https://pax.grsecurity.net/
This code is modified from Brad Spengler/PaX Team's code in the last
public patch of grsecurity/PaX based on our understanding of the code.
Changes or omissions from the original code are ours and don't reflect
the original grsecurity/PaX code.
Performance impact:
Hardware: Intel Core i7-4770, 16 GB RAM
Test #1: building the Linux kernel on a single core
0.91% slowdown
Test #2: hackbench -s 4096 -l 2000 -g 15 -f 25 -P
4.2% slowdown
So the STACKLEAK description in Kconfig includes: "The tradeoff is the
performance impact: on a single CPU system kernel compilation sees a 1%
slowdown, other systems and workloads may vary and you are advised to
test this feature on your expected workload before deploying it".
Signed-off-by: Alexander Popov <alex.popov@linux.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Pull dma-mapping fixes from Christoph Hellwig:
"A few fixes for the fallout of being a little more pedantic about dma
masks"
* tag 'dma-mapping-4.19-2' of git://git.infradead.org/users/hch/dma-mapping:
of/platform: initialise AMBA default DMA masks
sparc: set a default 32-bit dma mask for OF devices
kernel/dma/direct: take DMA offset into account in dma_direct_supported
Currently we check sk_user_data is non NULL to determine if the sk
exists in a map. However, this is not sufficient to ensure the psock
or the ULP ops are not in use by another user, such as kcm or TLS. To
avoid this when adding a sock to a map also verify it is of the
correct ULP type. Additionally, when releasing a psock verify that
it is the TCP_ULP_BPF type before releasing the ULP. The error case
where we abort an update due to ULP collision can cause this error
path.
For example,
__sock_map_ctx_update_elem()
[...]
err = tcp_set_ulp_id(sock, TCP_ULP_BPF) <- collides with TLS
if (err) <- so err out here
goto out_free
[...]
out_free:
smap_release_sock() <- calling tcp_cleanup_ulp releases the
TLS ULP incorrectly.
Fixes: 2f857d0460 ("bpf: sockmap, remove STRPARSER map_flags and add multi-map support")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Pull CPU hotplug fix from Thomas Gleixner:
"Remove the stale skip_onerr member from the hotplug states"
* 'smp-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
cpu/hotplug: Remove skip_onerr field from cpuhp_step structure
Pull core fixes from Thomas Gleixner:
"A small set of updates for core code:
- Prevent tracing in functions which are called from trace patching
via stop_machine() to prevent executing half patched function trace
entries.
- Remove old GCC workarounds
- Remove pointless includes of notifier.h"
* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
objtool: Remove workaround for unreachable warnings from old GCC
notifier: Remove notifier header file wherever not used
watchdog: Mark watchdog touch functions as notrace
When a device has a DMA offset the dma capable result will change due
to the difference between the physical and DMA address. Take that into
account.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
When notifiers were there, `skip_onerr` was used to avoid calling
particular step startup/teardown callbacks in the CPU up/down rollback
path, which made the hotplug asymmetric.
As notifiers are gone now after the full state machine conversion, the
`skip_onerr` field is no longer required.
Remove it from the structure and its usage.
Signed-off-by: Mukesh Ojha <mojha@codeaurora.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1535439294-31426-1-git-send-email-mojha@codeaurora.org
Allocating a list_head structure that is almost never used, and, when
used, is used only during early boot (rcu_init() and earlier), is a bit
wasteful. This commit therefore eliminates that list_head in favor of
the one in the work_struct structure. This is safe because the work_struct
structure cannot be used until after rcu_init() returns.
Reported-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Tested-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Event tracing is moving to SRCU in order to take advantage of the fact
that SRCU may be safely used from idle and even offline CPUs. However,
event tracing can invoke call_srcu() very early in the boot process,
even before workqueue_init_early() is invoked (let alone rcu_init()).
Therefore, call_srcu()'s attempts to queue work fail miserably.
This commit therefore detects this situation, and refrains from attempting
to queue work before rcu_init() time, but does everything else that it
would have done, and in addition, adds the srcu_struct to a global list.
The rcu_init() function now invokes a new srcu_init() function, which
is empty if CONFIG_SRCU=n. Otherwise, srcu_init() queues work for
each srcu_struct on the list. This all happens early enough in boot
that there is but a single CPU with interrupts disabled, which allows
synchronization to be dispensed with.
Of course, the queued work won't actually be invoked until after
workqueue_init() is invoked, which happens shortly after the scheduler
is up and running. This means that although call_srcu() may be invoked
any time after per-CPU variables have been set up, there is still a very
narrow window when synchronize_srcu() won't work, and this window
extends from the time that the scheduler starts until the time that
workqueue_init() returns. This can be fixed in a manner similar to
the fix for synchronize_rcu_expedited() and friends, but until someone
actually needs to use synchronize_srcu() during this window, this fix
is added churn for no benefit.
Finally, note that Tree SRCU's new srcu_init() function invokes
queue_work() rather than the queue_delayed_work() function that is
invoked post-boot. The reason is that queue_delayed_work() will (as you
would expect) post a timer, and timers have not yet been initialized.
So use of queue_work() avoids the complaints about use of uninitialized
spinlocks that would otherwise result. Besides, some delay is already
provide by the aforementioned fact that the queued work won't actually
be invoked until after the scheduler is up and running.
Requested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
1e64b15a4b ("rcu: Fix grace-period hangs due to race with CPU offline")
added spinlock_t ofl_lock to the rcu_state structure, then takes it with
preemption disabled during CPU offline, which gives the -rt patchset's
sleeping spinlock heartburn.
This commit therefore converts ->ofl_lock to raw_spinlock_t.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
The rcu_data structure's ->dynticks_fqs is incremented but never
accesses. Its ->cond_resched_completed field isn't used at all.
This commit therefore removes both fields.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit move ->dynticks from the rcu_dynticks structure to the
rcu_data structure, replacing the field of the same name. It also updates
the code to access ->dynticks from the rcu_data structure and to use the
rcu_data structure rather than following to now-gone ->dynticks field
to the now-gone rcu_dynticks structure. While in the area, this commit
also fixes up comments.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->dynticks_nesting and ->dynticks_nmi_nesting from
the rcu_dynticks structure and updates the code to access them from the
rcu_data structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->rcu_need_heavy_qs and ->rcu_urgent_qs from the
rcu_dynticks structure and updates the code to access them from the
rcu_data structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->all_lazy, ->nonlazy_posted and ->nonlazy_posted_snap
from the rcu_dynticks structure and updates the code to access them from
the rcu_data structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->last_accelerate and ->last_advance_all from the
rcu_dynticks structure and updates the code to access them from the
rcu_data structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->tick_nohz_enabled_snap from the rcu_dynticks
structure and updates the code to access it from the rcu_data
structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Now that there is only ever one rcu_data structure per CPU, there is no
need for a separate rcu_dynticks structure. This commit therefore adds
the rcu_dynticks fields into the rcu_data structure in preparation for
removing the rcu_dynticks structure entirely. Note that the ->dynticks
field will be handled specially because there is a field by that name
in both structures.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
