sched: replace cpumask_weight with cpumask_empty where appropriate

In some places, kernel/sched code calls cpumask_weight() to check if
any bit of a given cpumask is set. We can do it more efficiently with
cpumask_empty() because cpumask_empty() stops traversing the cpumask as
soon as it finds first set bit, while cpumask_weight() counts all bits
unconditionally.

Signed-off-by: Yury Norov <yury.norov@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220210224933.379149-23-yury.norov@gmail.com

sched/numa-balancing: Move some document to make it consistent with the code

After commit 8a99b6833c88 ("sched: Move SCHED_DEBUG sysctl to
debugfs"), some NUMA balancing sysctls enclosed with SCHED_DEBUG has
been moved to debugfs.  This patch move the document for these
sysctls from

  Documentation/admin-guide/sysctl/kernel.rst

to

  Documentation/scheduler/sched-debug.rst

to make the document consistent with the code.

Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Link: https://lkml.kernel.org/r/20220210052514.3038279-1-ying.huang@intel.com

sched/fair: Adjust the allowed NUMA imbalance when SD_NUMA spans multiple LLCs

Commit 7d2b5dd0bcc4 ("sched/numa: Allow a floating imbalance between NUMA
nodes") allowed an imbalance between NUMA nodes such that communicating
tasks would not be pulled apart by the load balancer. This works fine when
there is a 1:1 relationship between LLC and node but can be suboptimal
for multiple LLCs if independent tasks prematurely use CPUs sharing cache.

Zen* has multiple LLCs per node with local memory channels and due to
the allowed imbalance, it's far harder to tune some workloads to run
optimally than it is on hardware that has 1 LLC per node. This patch
allows an imbalance to exist up to the point where LLCs should be balanced
between nodes.

On a Zen3 machine running STREAM parallelised with OMP to have on instance
per LLC the results and without binding, the results are

                            5.17.0-rc0             5.17.0-rc0
                               vanilla       sched-numaimb-v6
MB/sec copy-16    162596.94 (   0.00%)   580559.74 ( 257.05%)
MB/sec scale-16   136901.28 (   0.00%)   374450.52 ( 173.52%)
MB/sec add-16     157300.70 (   0.00%)   564113.76 ( 258.62%)
MB/sec triad-16   151446.88 (   0.00%)   564304.24 ( 272.61%)

STREAM can use directives to force the spread if the OpenMP is new
enough but that doesn't help if an application uses threads and
it's not known in advance how many threads will be created.

Coremark is a CPU and cache intensive benchmark parallelised with
threads. When running with 1 thread per core, the vanilla kernel
allows threads to contend on cache. With the patch;

                               5.17.0-rc0             5.17.0-rc0
                                  vanilla       sched-numaimb-v5
Min       Score-16   368239.36 (   0.00%)   389816.06 (   5.86%)
Hmean     Score-16   388607.33 (   0.00%)   427877.08 *  10.11%*
Max       Score-16   408945.69 (   0.00%)   481022.17 (  17.62%)
Stddev    Score-16    15247.04 (   0.00%)    24966.82 ( -63.75%)
CoeffVar  Score-16        3.92 (   0.00%)        5.82 ( -48.48%)

It can also make a big difference for semi-realistic workloads
like specjbb which can execute arbitrary numbers of threads without
advance knowledge of how they should be placed. Even in cases where
the average performance is neutral, the results are more stable.

                               5.17.0-rc0             5.17.0-rc0
                                  vanilla       sched-numaimb-v6
Hmean     tput-1      71631.55 (   0.00%)    73065.57 (   2.00%)
Hmean     tput-8     582758.78 (   0.00%)   556777.23 (  -4.46%)
Hmean     tput-16   1020372.75 (   0.00%)  1009995.26 (  -1.02%)
Hmean     tput-24   1416430.67 (   0.00%)  1398700.11 (  -1.25%)
Hmean     tput-32   1687702.72 (   0.00%)  1671357.04 (  -0.97%)
Hmean     tput-40   1798094.90 (   0.00%)  2015616.46 *  12.10%*
Hmean     tput-48   1972731.77 (   0.00%)  2333233.72 (  18.27%)
Hmean     tput-56   2386872.38 (   0.00%)  2759483.38 (  15.61%)
Hmean     tput-64   2909475.33 (   0.00%)  2925074.69 (   0.54%)
Hmean     tput-72   2585071.36 (   0.00%)  2962443.97 (  14.60%)
Hmean     tput-80   2994387.24 (   0.00%)  3015980.59 (   0.72%)
Hmean     tput-88   3061408.57 (   0.00%)  3010296.16 (  -1.67%)
Hmean     tput-96   3052394.82 (   0.00%)  2784743.41 (  -8.77%)
Hmean     tput-104  2997814.76 (   0.00%)  2758184.50 (  -7.99%)
Hmean     tput-112  2955353.29 (   0.00%)  2859705.09 (  -3.24%)
Hmean     tput-120  2889770.71 (   0.00%)  2764478.46 (  -4.34%)
Hmean     tput-128  2871713.84 (   0.00%)  2750136.73 (  -4.23%)
Stddev    tput-1       5325.93 (   0.00%)     2002.53 (  62.40%)
Stddev    tput-8       6630.54 (   0.00%)    10905.00 ( -64.47%)
Stddev    tput-16     25608.58 (   0.00%)     6851.16 (  73.25%)
Stddev    tput-24     12117.69 (   0.00%)     4227.79 (  65.11%)
Stddev    tput-32     27577.16 (   0.00%)     8761.05 (  68.23%)
Stddev    tput-40     59505.86 (   0.00%)     2048.49 (  96.56%)
Stddev    tput-48    168330.30 (   0.00%)    93058.08 (  44.72%)
Stddev    tput-56    219540.39 (   0.00%)    30687.02 (  86.02%)
Stddev    tput-64    121750.35 (   0.00%)     9617.36 (  92.10%)
Stddev    tput-72    223387.05 (   0.00%)    34081.13 (  84.74%)
Stddev    tput-80    128198.46 (   0.00%)    22565.19 (  82.40%)
Stddev    tput-88    136665.36 (   0.00%)    27905.97 (  79.58%)
Stddev    tput-96    111925.81 (   0.00%)    99615.79 (  11.00%)
Stddev    tput-104   146455.96 (   0.00%)    28861.98 (  80.29%)
Stddev    tput-112    88740.49 (   0.00%)    58288.23 (  34.32%)
Stddev    tput-120   186384.86 (   0.00%)    45812.03 (  75.42%)
Stddev    tput-128    78761.09 (   0.00%)    57418.48 (  27.10%)

Similarly, for embarassingly parallel problems like NPB-ep, there are
improvements due to better spreading across LLC when the machine is not
fully utilised.

                              vanilla       sched-numaimb-v6
Min       ep.D       31.79 (   0.00%)       26.11 (  17.87%)
Amean     ep.D       31.86 (   0.00%)       26.17 *  17.86%*
Stddev    ep.D        0.07 (   0.00%)        0.05 (  24.41%)
CoeffVar  ep.D        0.22 (   0.00%)        0.20 (   7.97%)
Max       ep.D       31.93 (   0.00%)       26.21 (  17.91%)

Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20220208094334.16379-3-mgorman@techsingularity.net

sched/fair: Improve consistency of allowed NUMA balance calculations

There are inconsistencies when determining if a NUMA imbalance is allowed
that should be corrected.

o allow_numa_imbalance changes types and is not always examining
  the destination group so both the type should be corrected as
  well as the naming.
o find_idlest_group uses the sched_domain's weight instead of the
  group weight which is different to find_busiest_group
o find_busiest_group uses the source group instead of the destination
  which is different to task_numa_find_cpu
o Both find_idlest_group and find_busiest_group should account
  for the number of running tasks if a move was allowed to be
  consistent with task_numa_find_cpu

Fixes: 7d2b5dd0bcc4 ("sched/numa: Allow a floating imbalance between NUMA nodes")
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Link: https://lore.kernel.org/r/20220208094334.16379-2-mgorman@techsingularity.net

selftests/rseq: Change type of rseq_offset to ptrdiff_t

Just before the 2.35 release of glibc, the __rseq_offset userspace ABI
was changed from int to ptrdiff_t.

Adapt to this change in the kernel selftests.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://sourceware.org/pipermail/libc-alpha/2022-February/136024.html

sched: move autogroup sysctls into its own file

move autogroup sysctls to autogroup.c and use the new
register_sysctl_init() to register the sysctl interface.

Signed-off-by: Zhen Ni <nizhen@uniontech.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220128095025.8745-1-nizhen@uniontech.com

selftests/rseq: x86-32: use %gs segment selector for accessing rseq thread area

Rather than use rseq_get_abi() and pass its result through a register to
the inline assembler, directly access the per-thread rseq area through a
memory reference combining the %gs segment selector, the constant offset
of the field in struct rseq, and the rseq_offset value (in a register).

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-16-mathieu.desnoyers@efficios.com

selftests/rseq: x86-64: use %fs segment selector for accessing rseq thread area

Rather than use rseq_get_abi() and pass its result through a register to
the inline assembler, directly access the per-thread rseq area through a
memory reference combining the %fs segment selector, the constant offset
of the field in struct rseq, and the rseq_offset value (in a register).

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-15-mathieu.desnoyers@efficios.com

selftests/rseq: Fix: work-around asm goto compiler bugs

gcc and clang each have their own compiler bugs with respect to asm
goto. Implement a work-around for compiler versions known to have those
bugs.

gcc prior to 4.8.2 miscompiles asm goto.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58670

gcc prior to 8.1.0 miscompiles asm goto at O1.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103908

clang prior to version 13.0.1 miscompiles asm goto at O2.
https://github.com/llvm/llvm-project/issues/52735

Work around these issues by adding a volatile inline asm with
memory clobber in the fallthrough after the asm goto and at each
label target.  Emit this for all compilers in case other similar
issues are found in the future.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-14-mathieu.desnoyers@efficios.com

selftests/rseq: Remove arm/mips asm goto compiler work-around

The arm and mips work-around for asm goto size guess issues are not
properly documented, and lack reference to specific compiler versions,
upstream compiler bug tracker entry, and reproducer.

I can only find a loosely documented patch in my original LKML rseq post
refering to gcc < 7 on ARM, but it does not appear to be sufficient to
track the exact issue. Also, I am not sure MIPS really has the same
limitation.

Therefore, remove the work-around until we can properly document this.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/lkml/20171121141900.18471-17-mathieu.desnoyers@efficios.com/

selftests/rseq: Fix warnings about #if checks of undefined tokens

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-12-mathieu.desnoyers@efficios.com

selftests/rseq: Fix ppc32 offsets by using long rather than off_t

The semantic of off_t is for file offsets. We mean to use it as an
offset from a pointer. We really expect it to fit in a single register,
and not use a 64-bit type on 32-bit architectures.

Fix runtime issues on ppc32 where the offset is always 0 due to
inconsistency between the argument type (off_t -> 64-bit) and type
expected by the inline assembler (32-bit).

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-11-mathieu.desnoyers@efficios.com

selftests/rseq: Fix ppc32 missing instruction selection "u" and "x" for load/store

Building the rseq basic test  with
gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.12)
Target: powerpc-linux-gnu

leads to these errors:

/tmp/ccieEWxU.s: Assembler messages:
/tmp/ccieEWxU.s:118: Error: syntax error; found `,', expected `('
/tmp/ccieEWxU.s:118: Error: junk at end of line: `,8'
/tmp/ccieEWxU.s:121: Error: syntax error; found `,', expected `('
/tmp/ccieEWxU.s:121: Error: junk at end of line: `,8'
/tmp/ccieEWxU.s:626: Error: syntax error; found `,', expected `('
/tmp/ccieEWxU.s:626: Error: junk at end of line: `,8'
/tmp/ccieEWxU.s:629: Error: syntax error; found `,', expected `('
/tmp/ccieEWxU.s:629: Error: junk at end of line: `,8'
/tmp/ccieEWxU.s:735: Error: syntax error; found `,', expected `('
/tmp/ccieEWxU.s:735: Error: junk at end of line: `,8'
/tmp/ccieEWxU.s:738: Error: syntax error; found `,', expected `('
/tmp/ccieEWxU.s:738: Error: junk at end of line: `,8'
/tmp/ccieEWxU.s:741: Error: syntax error; found `,', expected `('
/tmp/ccieEWxU.s:741: Error: junk at end of line: `,8'
Makefile:581: recipe for target 'basic_percpu_ops_test.o' failed

Based on discussion with Linux powerpc maintainers and review of
the use of the "m" operand in powerpc kernel code, add the missing
%Un%Xn (where n is operand number) to the lwz, stw, ld, and std
instructions when used with "m" operands.

Using "WORD" to mean either a 32-bit or 64-bit type depending on
the architecture is misleading. The term "WORD" really means a
32-bit type in both 32-bit and 64-bit powerpc assembler. The intent
here is to wrap load/store to intptr_t into common macros for both
32-bit and 64-bit.

Rename the macros with a RSEQ_ prefix, and use the terms "INT"
for always 32-bit type, and "LONG" for architecture bitness-sized
type.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-10-mathieu.desnoyers@efficios.com

selftests/rseq: Fix ppc32: wrong rseq_cs 32-bit field pointer on big endian

ppc32 incorrectly uses padding as rseq_cs pointer field. Fix this by
using the rseq_cs.arch.ptr field.

Use this field across all architectures.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-9-mathieu.desnoyers@efficios.com

selftests/rseq: Uplift rseq selftests for compatibility with glibc-2.35

glibc-2.35 (upcoming release date 2022-02-01) exposes the rseq per-thread
data in the TCB, accessible at an offset from the thread pointer, rather
than through an actual Thread-Local Storage (TLS) variable, as the
Linux kernel selftests initially expected.

The __rseq_abi TLS and glibc-2.35's ABI for per-thread data cannot
actively coexist in a process, because the kernel supports only a single
rseq registration per thread.

Here is the scheme introduced to ensure selftests can work both with an
older glibc and with glibc-2.35+:

- librseq exposes its own "rseq_offset, rseq_size, rseq_flags" ABI.

- librseq queries for glibc rseq ABI (__rseq_offset, __rseq_size,
  __rseq_flags) using dlsym() in a librseq library constructor. If those
  are found, copy their values into rseq_offset, rseq_size, and
  rseq_flags.

- Else, if those glibc symbols are not found, handle rseq registration
  from librseq and use its own IE-model TLS to implement the rseq ABI
  per-thread storage.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-8-mathieu.desnoyers@efficios.com

selftests/rseq: Introduce thread pointer getters

This is done in preparation for the selftest uplift to become compatible
with glibc-2.35.

glibc-2.35 exposes the rseq per-thread data in the TCB, accessible
at an offset from the thread pointer.

The toolchains do not implement accessing the thread pointer on all
architectures. Provide thread pointer getters for ppc and x86 which
lack (or lacked until recently) toolchain support.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-7-mathieu.desnoyers@efficios.com

selftests/rseq: Introduce rseq_get_abi() helper

This is done in preparation for the selftest uplift to become compatible
with glibc-2.35.

glibc-2.35 exposes the rseq per-thread data in the TCB, accessible
at an offset from the thread pointer, rather than through an actual
Thread-Local Storage (TLS) variable, as the kernel selftests initially
expected.

Introduce a rseq_get_abi() helper, initially using the __rseq_abi
TLS variable, in preparation for changing this userspace ABI for one
which is compatible with glibc-2.35.

Note that the __rseq_abi TLS and glibc-2.35's ABI for per-thread data
cannot actively coexist in a process, because the kernel supports only
a single rseq registration per thread.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-6-mathieu.desnoyers@efficios.com

selftests/rseq: Remove volatile from __rseq_abi

This is done in preparation for the selftest uplift to become compatible
with glibc-2.35.

All accesses to the __rseq_abi fields are volatile, but remove the
volatile from the TLS variable declaration, otherwise we are stuck with
volatile for the upcoming rseq_get_abi() helper.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-5-mathieu.desnoyers@efficios.com

selftests/rseq: Remove useless assignment to cpu variable

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-4-mathieu.desnoyers@efficios.com

rseq: Remove broken uapi field layout on 32-bit little endian

The rseq rseq_cs.ptr.{ptr32,padding} uapi endianness handling is
entirely wrong on 32-bit little endian: a preprocessor logic mistake
wrongly uses the big endian field layout on 32-bit little endian
architectures.

Fortunately, those ptr32 accessors were never used within the kernel,
and only meant as a convenience for user-space.

Remove those and replace the whole rseq_cs union by a __u64 type, as
this is the only thing really needed to express the ABI. Document how
32-bit architectures are meant to interact with this field.

Fixes: ec9c82e03a74 ("rseq: uapi: Declare rseq_cs field as union, update includes")
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220127152720.25898-1-mathieu.desnoyers@efficios.com

selftests/rseq: introduce own copy of rseq uapi header

The Linux kernel rseq uapi header has a broken layout for the
rseq_cs.ptr field on 32-bit little endian architectures. The entire
rseq_cs.ptr field is planned for removal, leaving only the 64-bit
rseq_cs.ptr64 field available.

Both glibc and librseq use their own copy of the Linux kernel uapi
header, where they introduce proper union fields to access to the 32-bit
low order bits of the rseq_cs pointer on 32-bit architectures.

Introduce a copy of the Linux kernel uapi headers in the Linux kernel
selftests.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-2-mathieu.desnoyers@efficios.com

psi: Fix "no previous prototype" warnings when CONFIG_CGROUPS=n

When CONFIG_CGROUPS is disabled psi code generates the following warnings:

kernel/sched/psi.c:1112:21: warning: no previous prototype for 'psi_trigger_create' [-Wmissing-prototypes]
    1112 | struct psi_trigger *psi_trigger_create(struct psi_group *group,
         |                     ^~~~~~~~~~~~~~~~~~
kernel/sched/psi.c:1182:6: warning: no previous prototype for 'psi_trigger_destroy' [-Wmissing-prototypes]
    1182 | void psi_trigger_destroy(struct psi_trigger *t)
         |      ^~~~~~~~~~~~~~~~~~~
kernel/sched/psi.c:1249:10: warning: no previous prototype for 'psi_trigger_poll' [-Wmissing-prototypes]
    1249 | __poll_t psi_trigger_poll(void **trigger_ptr,
         |          ^~~~~~~~~~~~~~~~

Change declarations of these functions in the header to provide the
prototypes even when they are unused.

Fixes: 0e94682b73bf ("psi: introduce psi monitor")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220119223940.787748-2-surenb@google.com

psi: Fix "defined but not used" warnings when CONFIG_PROC_FS=n

When CONFIG_PROC_FS is disabled psi code generates the following warnings:

kernel/sched/psi.c:1364:30: warning: 'psi_cpu_proc_ops' defined but not used [-Wunused-const-variable=]
    1364 | static const struct proc_ops psi_cpu_proc_ops = {
         |                              ^~~~~~~~~~~~~~~~
kernel/sched/psi.c:1355:30: warning: 'psi_memory_proc_ops' defined but not used [-Wunused-const-variable=]
    1355 | static const struct proc_ops psi_memory_proc_ops = {
         |                              ^~~~~~~~~~~~~~~~~~~
kernel/sched/psi.c:1346:30: warning: 'psi_io_proc_ops' defined but not used [-Wunused-const-variable=]
    1346 | static const struct proc_ops psi_io_proc_ops = {
         |                              ^~~~~~~~~~~~~~~

Make definitions of these structures and related functions conditional on
CONFIG_PROC_FS config.

Fixes: 0e94682b73bf ("psi: introduce psi monitor")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220119223940.787748-3-surenb@google.com

sched/uclamp: Fix iowait boost escaping uclamp restriction

iowait_boost signal is applied independently of util and doesn't take
into account uclamp settings of the rq. An io heavy task that is capped
by uclamp_max could still request higher frequency because
sugov_iowait_apply() doesn't clamp the boost via uclamp_rq_util_with()
like effective_cpu_util() does.

Make sure that iowait_boost honours uclamp requests by calling
uclamp_rq_util_with() when applying the boost.

Fixes: 982d9cdc22c9 ("sched/cpufreq, sched/uclamp: Add clamps for FAIR and RT tasks")
Signed-off-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20211216225320.2957053-3-qais.yousef@arm.com

sched/sugov: Ignore 'busy' filter when rq is capped by uclamp_max

sugov_update_single_{freq, perf}() contains a 'busy' filter that ensures
we don't bring the frqeuency down if there's no idle time (CPU is busy).

The problem is that with uclamp_max we will have scenarios where a busy
task is capped to run at a lower frequency and this filter prevents
applying the capping when this task starts running.

We handle this by skipping the filter when uclamp is enabled and the rq
is being capped by uclamp_max.

We introduce a new function uclamp_rq_is_capped() to help detecting when
this capping is taking effect. Some code shuffling was required to allow
using cpu_util_{cfs, rt}() in this new function.

On 2 Core SMT2 Intel laptop I see:

Without this patch:

uclampset -M 0 sysbench --test=cpu --threads = 4 run

produces a score of ~3200 consistently. Which is the highest possible.

Compiling the kernel also results in frequency running at max 3.1GHz all
the time - running uclampset -M 400 to cap it has no effect without this
patch.

With this patch:

uclampset -M 0 sysbench --test=cpu --threads = 4 run

produces a score of ~1100 with some outliers in ~1700. Uclamp max
aggregates the performance requirements, so having high values sometimes
is expected if some other task happens to require that frequency starts
running at the same time.

When compiling the kernel with uclampset -M 400 I can see the
frequencies mostly in the ~2GHz region. Helpful to conserve power and
prevent heating when not plugged in.

Fixes: 982d9cdc22c9 ("sched/cpufreq, sched/uclamp: Add clamps for FAIR and RT tasks")
Signed-off-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20211216225320.2957053-2-qais.yousef@arm.com

sched/core: Export pelt_thermal_tp

We can't use this tracepoint in modules without having the symbol
exported first, fix that.

Fixes: 765047932f15 ("sched/pelt: Add support to track thermal pressure")
Signed-off-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20211028115005.873539-1-qais.yousef@arm.com

MAINTAINERS: add Suren as psi co-maintainer

Suren wrote the poll() interface, which is a significant part of the
psi code and represents a large user of psi itself (Android). It's a
good idea to have him look at psi patches as well, and it's good to
have two people following things in case one of us is traveling.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220117120317.1581315-1-hannes@cmpxchg.org

sched/numa: initialize numa statistics when forking new task

The child processes will inherit numa_pages_migrated and
total_numa_faults from the parent. It means even if there is no numa
fault happen on the child, the statistics in /proc/$pid of the child
process might show huge amount. This is a bit weird. Let's initialize
them when do fork.

Signed-off-by: Honglei Wang <wanghonglei@didichuxing.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Mel Gorman <mgorman@suse.de>
Link: https://lore.kernel.org/r/20220113133920.49900-1-wanghonglei@didichuxing.com

sched/debug: Remove mpol_get/put and task_lock/unlock from sched_show_numa

The older format of /proc/pid/sched printed home node info which
required the mempolicy and task lock around mpol_get(). However
the format has changed since then and there is no need for
sched_show_numa() any more to have mempolicy argument,
asssociated mpol_get/put and task_lock/unlock. Remove them.

Fixes: 397f2378f1361 ("sched/numa: Fix numa balancing stats in /proc/pid/sched")
Signed-off-by: Bharata B Rao <bharata@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Link: https://lore.kernel.org/r/20220118050515.2973-1-bharata@amd.com

Linux 5.17-rc1

Merge tag 'perf-tools-for-v5.17-2022-01-22' of git://git./linux/kernel/git/acme/linux

Pull more perf tools updates from Arnaldo Carvalho de Melo:

 - Fix printing 'phys_addr' in 'perf script'.

 - Fix failure to add events with 'perf probe' in ppc64 due to not
   removing leading dot (ppc64 ABIv1).

 - Fix cpu_map__item() python binding building.

 - Support event alias in form foo-bar-baz, add pmu-events and
   parse-event tests for it.

 - No need to setup affinities when starting a workload or attaching to
   a pid.

 - Use path__join() to compose a path instead of ad-hoc snprintf()
   equivalent.

 - Override attr->sample_period for non-libpfm4 events.

 - Use libperf cpumap APIs instead of accessing the internal state
   directly.

 - Sync x86 arch prctl headers and files changed by the new
   set_mempolicy_home_node syscall with the kernel sources.

 - Remove duplicate include in cpumap.h.

 - Remove redundant err variable.

* tag 'perf-tools-for-v5.17-2022-01-22' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux:
  perf tools: Remove redundant err variable
  perf test: Add parse-events test for aliases with hyphens
  perf test: Add pmu-events test for aliases with hyphens
  perf parse-events: Support event alias in form foo-bar-baz
  perf evsel: Override attr->sample_period for non-libpfm4 events
  perf cpumap: Remove duplicate include in cpumap.h
  perf cpumap: Migrate to libperf cpumap api
  perf python: Fix cpu_map__item() building
  perf script: Fix printing 'phys_addr' failure issue
  tools headers UAPI: Sync files changed by new set_mempolicy_home_node syscall
  tools headers UAPI: Sync x86 arch prctl headers with the kernel sources
  perf machine: Use path__join() to compose a path instead of snprintf(dir, '/', filename)
  perf evlist: No need to setup affinities when disabling events for pid targets
  perf evlist: No need to setup affinities when enabling events for pid targets
  perf stat: No need to setup affinities when starting a workload
  perf affinity: Allow passing a NULL arg to affinity__cleanup()
  perf probe: Fix ppc64 'perf probe add events failed' case

Merge tag 'trace-v5.17-3' of git://git./linux/kernel/git/rostedt/linux-trace

Pull ftrace fix from Steven Rostedt:
 "Fix s390 breakage from sorting mcount tables.

  The latest merge of the tracing tree sorts the mcount table at build
  time. But s390 appears to do things differently (like always) and
  replaces the sorted table back to the original unsorted one. As the
  ftrace algorithm depends on it being sorted, bad things happen when it
  is not, and s390 experienced those bad things.

  Add a new config to tell the boot if the mcount table is sorted or
  not, and allow s390 to opt out of it"

* tag 'trace-v5.17-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  ftrace: Fix assuming build time sort works for s390

ftrace: Fix assuming build time sort works for s390

To speed up the boot process, as mcount_loc needs to be sorted for ftrace
to work properly, sorting it at build time is more efficient than boot up
and can save milliseconds of time. Unfortunately, this change broke s390
as it will modify the mcount_loc location after the sorting takes place
and will put back the unsorted locations. Since the sorting is skipped at
boot up if it is believed that it was sorted at run time, ftrace can crash
as its algorithms are dependent on the list being sorted.

Add a new config BUILDTIME_MCOUNT_SORT that is set when
BUILDTIME_TABLE_SORT but not if S390 is set. Use this config to determine
if sorting should take place at boot up.

Link: https://lore.kernel.org/all/yt9dee51ctfn.fsf@linux.ibm.com/
Fixes: 72b3942a173c ("scripts: ftrace - move the sort-processing in ftrace_init")
Reported-by: Sven Schnelle <svens@linux.ibm.com>
Tested-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>

Merge tag 'kbuild-fixes-v5.17' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - Bring include/uapi/linux/nfc.h into the UAPI compile-test coverage

 - Revert the workaround of CONFIG_CC_IMPLICIT_FALLTHROUGH

 - Fix build errors in certs/Makefile

* tag 'kbuild-fixes-v5.17' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  certs: Fix build error when CONFIG_MODULE_SIG_KEY is empty
  certs: Fix build error when CONFIG_MODULE_SIG_KEY is PKCS#11 URI
  Revert "Makefile: Do not quote value for CONFIG_CC_IMPLICIT_FALLTHROUGH"
  usr/include/Makefile: add linux/nfc.h to the compile-test coverage

Merge tag 'bitmap-5.17-rc1' of git://github.com/norov/linux

Pull bitmap updates from Yury Norov:

 - introduce for_each_set_bitrange()

 - use find_first_*_bit() instead of find_next_*_bit() where possible

 - unify for_each_bit() macros

* tag 'bitmap-5.17-rc1' of git://github.com/norov/linux:
  vsprintf: rework bitmap_list_string
  lib: bitmap: add performance test for bitmap_print_to_pagebuf
  bitmap: unify find_bit operations
  mm/percpu: micro-optimize pcpu_is_populated()
  Replace for_each_*_bit_from() with for_each_*_bit() where appropriate
  find: micro-optimize for_each_{set,clear}_bit()
  include/linux: move for_each_bit() macros from bitops.h to find.h
  cpumask: replace cpumask_next_* with cpumask_first_* where appropriate
  tools: sync tools/bitmap with mother linux
  all: replace find_next{,_zero}_bit with find_first{,_zero}_bit where appropriate
  cpumask: use find_first_and_bit()
  lib: add find_first_and_bit()
  arch: remove GENERIC_FIND_FIRST_BIT entirely
  include: move find.h from asm_generic to linux
  bitops: move find_bit_*_le functions from le.h to find.h
  bitops: protect find_first_{,zero}_bit properly

perf tools: Remove redundant err variable

Return value from perf_event__process_tracing_data() directly instead
of taking this in another redundant variable.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Minghao Chi <chi.minghao@zte.com.cn>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lore.kernel.org/lkml/20220112080109.666800-1-chi.minghao@zte.com.cn
Signed-off-by: CGEL ZTE <cgel.zte@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf test: Add parse-events test for aliases with hyphens

Add a test which allows us to test parsing an event alias with hyphens.

Since these events typically do not exist on most host systems, add the
alias to the fake pmu.

Function perf_pmu__test_parse_init() has terms added to match known test
aliases.

Signed-off-by: John Garry <john.garry@huawei.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qi Liu <liuqi115@huawei.com>
Cc: Shaokun Zhang <zhangshaokun@hisilicon.com>
Cc: linuxarm@huawei.com
Link: https://lore.kernel.org/r/1642432215-234089-4-git-send-email-john.garry@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf test: Add pmu-events test for aliases with hyphens

Add a test for aliases with hyphens in the name to ensure that the
pmu-events tables are as expects. There should be no reason why these sort
of aliases would be treated differently, but no harm in checking.

Signed-off-by: John Garry <john.garry@huawei.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qi Liu <liuqi115@huawei.com>
Cc: Shaokun Zhang <zhangshaokun@hisilicon.com>
Cc: linuxarm@huawei.com
Link: https://lore.kernel.org/r/1642432215-234089-3-git-send-email-john.garry@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf parse-events: Support event alias in form foo-bar-baz

Event aliasing for events whose name in the form foo-bar-baz is not
supported, while foo-bar, foo_bar_baz, and other combinations are, i.e.
two hyphens are not supported.

The HiSilicon D06 platform has events in such form:

  $ ./perf list sdir-home-migrate

  List of pre-defined events (to be used in -e):

  uncore hha:
    sdir-home-migrate
   [Unit: hisi_sccl,hha]

  $ sudo ./perf stat -e sdir-home-migrate
  event syntax error: 'sdir-home-migrate'
                          \___ parser error
  Run 'perf list' for a list of valid events

   Usage: perf stat [<options>] [<command>]

   -e, --event <event>event selector. use 'perf list' to list available events

To support, add an extra PMU event symbol type for "baz", and add a new
rule in the bison file.

Signed-off-by: John Garry <john.garry@huawei.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qi Liu <liuqi115@huawei.com>
Cc: Shaokun Zhang <zhangshaokun@hisilicon.com>
Cc: linuxarm@huawei.com
Link: https://lore.kernel.org/r/1642432215-234089-2-git-send-email-john.garry@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf evsel: Override attr->sample_period for non-libpfm4 events

A previous patch preventing "attr->sample_period" values from being
overridden in pfm events changed a related behaviour in arm-spe.

Before said patch:

  perf record -c 10000 -e arm_spe_0// -- sleep 1

Would yield an SPE event with period=10000. After the patch, the period
in "-c 10000" was being ignored because the arm-spe code initializes
sample_period to a non-zero value.

This patch restores the previous behaviour for non-libpfm4 events.

Fixes: ae5dcc8abe31 (“perf record: Prevent override of attr->sample_period for libpfm4 events”)
Reported-by: Chase Conklin <chase.conklin@arm.com>
Signed-off-by: German Gomez <german.gomez@arm.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: KP Singh <kpsingh@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Martin KaFai Lau <kafai@fb.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Song Liu <songliubraving@fb.com>
Cc: Stephane Eranian <eranian@google.com>
Cc: Yonghong Song <yhs@fb.com>
Cc: bpf@vger.kernel.org
Cc: netdev@vger.kernel.org
Link: http://lore.kernel.org/lkml/20220118144054.2541-1-german.gomez@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf cpumap: Remove duplicate include in cpumap.h

Remove all but the first include of stdbool.h from cpumap.h.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Lv Ruyi <lv.ruyi@zte.com.cn>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@arm.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20220117083730.863200-1-lv.ruyi@zte.com.cn
Signed-off-by: CGEL ZTE <cgel.zte@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf cpumap: Migrate to libperf cpumap api

Switch from directly accessing the perf_cpu_map to using the appropriate
libperf API when possible. Using the API simplifies the job of
refactoring use of perf_cpu_map.

Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexey Bayduraev <alexey.v.bayduraev@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: André Almeida <andrealmeid@collabora.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Darren Hart <dvhart@infradead.org>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Dmitriy Vyukov <dvyukov@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: German Gomez <german.gomez@arm.com>
Cc: James Clark <james.clark@arm.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Miaoqian Lin <linmq006@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Riccardo Mancini <rickyman7@gmail.com>
Cc: Shunsuke Nakamura <nakamura.shun@fujitsu.com>
Cc: Song Liu <song@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Stephen Brennan <stephen.s.brennan@oracle.com>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Yury Norov <yury.norov@gmail.com>
Link: http://lore.kernel.org/lkml/20220122045811.3402706-3-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf python: Fix cpu_map__item() building

Value should be built as an integer.

Switch some uses of perf_cpu_map to use the library API.

Fixes: 6d18804b963b78dc ("perf cpumap: Give CPUs their own type")
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexey Bayduraev <alexey.v.bayduraev@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: André Almeida <andrealmeid@collabora.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Darren Hart <dvhart@infradead.org>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Dmitriy Vyukov <dvyukov@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: German Gomez <german.gomez@arm.com>
Cc: Ian Rogers <irogers@google.com>
Cc: James Clark <james.clark@arm.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Miaoqian Lin <linmq006@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Riccardo Mancini <rickyman7@gmail.com>
Cc: Shunsuke Nakamura <nakamura.shun@fujitsu.com>
Cc: Song Liu <song@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Stephen Brennan <stephen.s.brennan@oracle.com>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Yury Norov <yury.norov@gmail.com>
Link: http://lore.kernel.org/lkml/20220122045811.3402706-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf script: Fix printing 'phys_addr' failure issue

Perf script was failed to print the phys_addr for SPE profiling.
One 'dummy' event is added by SPE profiling but it doesn't have PHYS_ADDR
attribute set, perf script then exits with error.

Now referring to 'addr', use evsel__do_check_stype() to check the type.

Before:

  # perf record -e arm_spe_0/branch_filter=0,ts_enable=1,pa_enable=1,load_filter=1,jitter=0,\
store_filter=0,min_latency=0,event_filter=2/ -p 4064384 -- sleep 3
  # perf script -F pid,tid,addr,phys_addr
  Samples for 'dummy:u' event do not have PHYS_ADDR attribute set. Cannot print 'phys_addr' field.

After:

  # perf record -e arm_spe_0/branch_filter=0,ts_enable=1,pa_enable=1,load_filter=1,jitter=0,\
store_filter=0,min_latency=0,event_filter=2/ -p 4064384 -- sleep 3
  # perf script -F pid,tid,addr,phys_addr
  4064384/4064384 ffff802f921be0d0      2f921be0d0
  4064384/4064384 ffff802f921be0d0      2f921be0d0

Reviewed-by: German Gomez <german.gomez@arm.com>
Signed-off-by: Yao Jin <jinyao5@huawei.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Hanjun Guo <guohanjun@huawei.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lore.kernel.org/lkml/20220121065954.2121900-1-liwei391@huawei.com
Signed-off-by: Wei Li <liwei391@huawei.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

certs: Fix build error when CONFIG_MODULE_SIG_KEY is empty

Since b8c96a6b466c ("certs: simplify $(srctree)/ handling and remove
config_filename macro"), when CONFIG_MODULE_SIG_KEY is empty,
signing_key.x509 fails to build:

    CERT    certs/signing_key.x509
  Usage: extract-cert <source> <dest>
  make[1]: *** [certs/Makefile:78: certs/signing_key.x509] Error 2
  make: *** [Makefile:1831: certs] Error 2

Pass "" to the first argument of extract-cert to fix the build error.

Link: https://lore.kernel.org/linux-kbuild/20220120094606.2skuyb26yjlnu66q@lion.mk-sys.cz/T/#u
Fixes: b8c96a6b466c ("certs: simplify $(srctree)/ handling and remove config_filename macro")
Reported-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Tested-by: Michal Kubecek <mkubecek@suse.cz>

certs: Fix build error when CONFIG_MODULE_SIG_KEY is PKCS#11 URI

When CONFIG_MODULE_SIG_KEY is PKCS#11 URL (pkcs11:*), signing_key.x509
fails to build:

  certs/Makefile:77: *** target pattern contains no '%'.  Stop.

Due to the typo, $(X509_DEP) contains a colon.

Fix it.

Fixes: b8c96a6b466c ("certs: simplify $(srctree)/ handling and remove config_filename macro")
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Revert "Makefile: Do not quote value for CONFIG_CC_IMPLICIT_FALLTHROUGH"

This reverts commit cd8c917a56f20f48748dd43d9ae3caff51d5b987.

Commit 129ab0d2d9f3 ("kbuild: do not quote string values in
include/config/auto.conf") provided the final solution.

Now reverting the temporary workaround.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

usr/include/Makefile: add linux/nfc.h to the compile-test coverage

As linux/nfc.h userspace compilation was finally fixed by commits
79b69a83705e ("nfc: uapi: use kernel size_t to fix user-space builds")
and 7175f02c4e5f ("uapi: fix linux/nfc.h userspace compilation errors"),
there is no need to keep the compile-test exception for it in
usr/include/Makefile.

Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Merge branch 'akpm' (patches from Andrew)

Merge yet more updates from Andrew Morton:
 "This is the post-linux-next queue. Material which was based on or
  dependent upon material which was in -next.

  69 patches.

  Subsystems affected by this patch series: mm (migration and zsmalloc),
  sysctl, proc, and lib"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (69 commits)
  mm: hide the FRONTSWAP Kconfig symbol
  frontswap: remove support for multiple ops
  mm: mark swap_lock and swap_active_head static
  frontswap: simplify frontswap_register_ops
  frontswap: remove frontswap_test
  mm: simplify try_to_unuse
  frontswap: remove the frontswap exports
  frontswap: simplify frontswap_init
  frontswap: remove frontswap_curr_pages
  frontswap: remove frontswap_shrink
  frontswap: remove frontswap_tmem_exclusive_gets
  frontswap: remove frontswap_writethrough
  mm: remove cleancache
  lib/stackdepot: always do filter_irq_stacks() in stack_depot_save()
  lib/stackdepot: allow optional init and stack_table allocation by kvmalloc()
  proc: remove PDE_DATA() completely
  fs: proc: store PDE()->data into inode->i_private
  zsmalloc: replace get_cpu_var with local_lock
  zsmalloc: replace per zpage lock with pool->migrate_lock
  locking/rwlocks: introduce write_lock_nested
  ...

Merge tag '5.17-rc-part2-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs fixes from Steve French:

 - multichannel fixes, addressing additional reconnect and DFS scenarios

 - reenabling fscache support (indexing rewrite, metadata caching e.g.)

 - send additional version information during NTLMSSP negotiate to
   improve debugging

 - fix for a mount race

 - DFS fixes

 - fix for a memory leak for stable

* tag '5.17-rc-part2-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: update internal module number
  smb3: send NTLMSSP version information
  cifs: Support fscache indexing rewrite
  cifs: cifs_ses_mark_for_reconnect should also update reconnect bits
  cifs: update tcpStatus during negotiate and sess setup
  cifs: make status checks in version independent callers
  cifs: remove repeated state change in dfs tree connect
  cifs: fix the cifs_reconnect path for DFS
  cifs: remove unused variable ses_selected
  cifs: protect all accesses to chan_* with chan_lock
  cifs: fix the connection state transitions with multichannel
  cifs: check reconnects for channels of active tcons too
  smb3: add new defines from protocol specification
  cifs: serialize all mount attempts
  cifs: quirk for STATUS_OBJECT_NAME_INVALID returned for non-ASCII dfs refs
  cifs: alloc_path_with_tree_prefix: do not append sep. if the path is empty
  cifs: clean up an inconsistent indenting
  cifs: free ntlmsspblob allocated in negotiate

Merge tag 'xfs-5.17-merge-7' of git://git./fs/xfs/xfs-linux

Pull xfs fixes from Darrick Wong:
 "One of the patches removes some dead code from xfs_ioctl32.h and the
  other fixes broken workqueue flushing in the inode garbage collector.

   - Minor cleanup of ioctl32 cruft

   - Clean up open coded inodegc workqueue function calls"

* tag 'xfs-5.17-merge-7' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  xfs: flush inodegc workqueue tasks before cancel
  xfs: remove unused xfs_ioctl32.h declarations

Merge tag 'fscache-fixes-20220121' of git://git./linux/kernel/git/dhowells/linux-fs

Pull more fscache updates from David Howells:
 "A set of fixes and minor updates for the fscache rewrite:

   - Fix mishandling of volume collisions (the wait condition is
     inverted and so it was only waiting if the volume collision was
     already resolved).

   - Fix miscalculation of whether there's space available in
     cachefiles.

   - Make sure a default cache name is set on a cache if the user hasn't
     set one by the time they bind the cache.

   - Adjust the way the backing inode is presented in tracepoints, add a
     tracepoint for mkdir and trace directory lookup.

   - Add a tracepoint for failure to set the active file mark.

   - Add an explanation of the checks made on the backing filesystem.

   - Check that the backing filesystem supports tmpfile.

   - Document how the page-release cancellation of the read-skip
     optimisation works.

  And I've included a change for netfslib:

   - Make ops->init_rreq() optional"

* tag 'fscache-fixes-20220121' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
  netfs: Make ops->init_rreq() optional
  fscache: Add a comment explaining how page-release optimisation works
  cachefiles: Check that the backing filesystem supports tmpfiles
  cachefiles: Explain checks in a comment
  cachefiles: Trace active-mark failure
  cachefiles: Make some tracepoint adjustments
  cachefiles: set default tag name if it's unspecified
  cachefiles: Calculate the blockshift in terms of bytes, not pages
  fscache: Fix the volume collision wait condition

Merge tag 'folio-5.17a' of git://git.infradead.org/users/willy/pagecache

Pull more folio updates from Matthew Wilcox:
 "Three small folio patches.

  One bug fix, one patch pulled forward from the patches destined for
  5.18 and then a patch to make use of that functionality"

* tag 'folio-5.17a' of git://git.infradead.org/users/willy/pagecache:
  filemap: Use folio_put_refs() in filemap_free_folio()
  mm: Add folio_put_refs()
  pagevec: Initialise folio_batch->percpu_pvec_drained

Merge tag 'scsi-misc' of git://git./linux/kernel/git/jejb/scsi

Pull more SCSI updates from James Bottomley:
 "This series is all the stragglers that didn't quite make the first
  merge window pull. It's mostly minor updates and bug fixes of merge
  window code"

* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: nsp_cs: Check of ioremap return value
  scsi: ufs: ufs-mediatek: Fix error checking in ufs_mtk_init_va09_pwr_ctrl()
  scsi: ufs: Modify Tactive time setting conditions
  scsi: efct: Remove useless DMA-32 fallback configuration
  scsi: message: fusion: mptctl: Use dma_alloc_coherent()
  scsi: message: fusion: mptsas: Use dma_alloc_coherent()
  scsi: message: fusion: Use dma_alloc_coherent() in mptsas_exp_repmanufacture_info()
  scsi: message: fusion: mptbase: Use dma_alloc_coherent()
  scsi: message: fusion: Use dma_alloc_coherent() in mpt_alloc_fw_memory()
  scsi: message: fusion: Remove usage of the deprecated "pci-dma-compat.h" API
  scsi: megaraid: Avoid mismatched storage type sizes
  scsi: hisi_sas: Remove unused variable and check in hisi_sas_send_ata_reset_each_phy()
  scsi: aic79xx: Remove redundant error variable
  scsi: pm80xx: Port reset timeout error handling correction
  scsi: mpi3mr: Fix formatting problems in some kernel-doc comments
  scsi: mpi3mr: Fix some spelling mistakes
  scsi: mpt3sas: Update persistent trigger pages from sysfs interface
  scsi: core: Fix scsi_mode_select() interface
  scsi: aacraid: Fix spelling of "its"
  scsi: qedf: Fix potential dereference of NULL pointer

Merge tag 'ata-5.17-rc1-part2' of git://git./linux/kernel/git/dlemoal/libata

Pull ATA fix from Damien Le Moal:
 "A single patch to fix a compilation error in the pata_octeon_cf driver
  (mips architecture), from me"

* tag 'ata-5.17-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/libata:
  ata: pata_octeon_cf: fix call to trace_ata_bmdma_stop()

Merge tag 'thermal-5.17-rc1-2' of git://git./linux/kernel/git/rafael/linux-pm

Pull more thermal control updates from Rafael Wysocki:
 "Add device IDs for Raptor Lake to the int340x thermal control driver
  (Srinivas Pandruvada)"

* tag 'thermal-5.17-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  thermal: int340x: Add Raptor Lake PCI device id
  thermal: int340x: Support Raptor Lake

Merge tag 'acpi-5.17-rc1-3' of git://git./linux/kernel/git/rafael/linux-pm

Pull extra ACPI updates from Rafael Wysocki:
 "These fix and clean up the ACPI CPPC driver on top of the recent
  changes in it merged previously and add some new device IDs to the
  ACPI DPTF driver.

  Specifics:

   - Fix a recently introduced endianness-related issue in the ACPI CPPC
     library and clean it up on top of that (Rafael Wysocki)

   - Add new device IDs for the Raptor Lake SoC to the ACPI DPTF driver
     (Srinivas Pandruvada)"

* tag 'acpi-5.17-rc1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI: DPTF: Support Raptor Lake
  ACPI: CPPC: Drop redundant local variable from cpc_read()
  ACPI: CPPC: Fix up I/O port access in cpc_read()

Merge tag 'devicetree-fixes-for-5.17-1' of git://git./linux/kernel/git/robh/linux

Pull devicetree fixes and cleanups from Rob Herring:

 - Fix a regression when probing a child device reusing the parent
   device's DT node pointer

 - Refactor of_parse_phandle*() variants to static inlines

 - Drop Enric Balletbo i Serra as a maintainer

 - Fix DT schemas with arrays incorrectly encoded as a matrix

 - Drop unneeded pinctrl properties from schemas

 - Add SPI peripheral schema to SPI based displays

 - Clean-up several schema examples

 - Clean-up trivial-devices.yaml comments

 - Add missing, in use vendor prefixes: Wingtech, Thundercomm, Huawei,
   F(x)tec, 8devices

* tag 'devicetree-fixes-for-5.17-1' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  dt-bindings: google,cros-ec: drop Enric Balletbo i Serra from maintainers
  dt-bindings: display: bridge: drop Enric Balletbo i Serra from maintainers
  of: Check 'of_node_reused' flag on of_match_device()
  of: property: define of_property_read_u{8,16,32,64}_array() unconditionally
  of: base: make small of_parse_phandle() variants static inline
  dt-bindings: mfd: cirrus,madera: Fix 'interrupts' in example
  dt-bindings: Fix array schemas encoded as matrices
  dt-bindings: Drop unnecessary pinctrl properties
  dt-bindings: rtc: st,stm32-rtc: Make each example a separate entry
  dt-bindings: mmc: arm,pl18x: Make each example a separate entry
  dt-bindings: display: Add SPI peripheral schema to SPI based displays
  scripts/dtc: dtx_diff: remove broken example from help text
  dt-bindings: trivial-devices: fix double spaces in comments
  dt-bindings: trivial-devices: fix swapped comments
  dt-bindings: vendor-prefixes: add Wingtech
  dt-bindings: vendor-prefixes: add Thundercomm
  dt-bindings: vendor-prefixes: add Huawei
  dt-bindings: vendor-prefixes: add F(x)tec
  dt-bindings: vendor-prefixes: add 8devices
  dt-bindings: power: reset: gpio-restart: Correct default priority

Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull more kvm updates from Paolo Bonzini:
 "Generic:

   - selftest compilation fix for non-x86

   - KVM: avoid warning on s390 in mark_page_dirty

 x86:

   - fix page write-protection bug and improve comments

   - use binary search to lookup the PMU event filter, add test

   - enable_pmu module parameter support for Intel CPUs

   - switch blocked_vcpu_on_cpu_lock to raw spinlock

   - cleanups of blocked vCPU logic

   - partially allow KVM_SET_CPUID{,2} after KVM_RUN (5.16 regression)

   - various small fixes"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (46 commits)
  docs: kvm: fix WARNINGs from api.rst
  selftests: kvm/x86: Fix the warning in lib/x86_64/processor.c
  selftests: kvm/x86: Fix the warning in pmu_event_filter_test.c
  kvm: selftests: Do not indent with spaces
  kvm: selftests: sync uapi/linux/kvm.h with Linux header
  selftests: kvm: add amx_test to .gitignore
  KVM: SVM: Nullify vcpu_(un)blocking() hooks if AVIC is disabled
  KVM: SVM: Move svm_hardware_setup() and its helpers below svm_x86_ops
  KVM: SVM: Drop AVIC's intermediate avic_set_running() helper
  KVM: VMX: Don't do full kick when handling posted interrupt wakeup
  KVM: VMX: Fold fallback path into triggering posted IRQ helper
  KVM: VMX: Pass desired vector instead of bool for triggering posted IRQ
  KVM: VMX: Don't do full kick when triggering posted interrupt "fails"
  KVM: SVM: Skip AVIC and IRTE updates when loading blocking vCPU
  KVM: SVM: Use kvm_vcpu_is_blocking() in AVIC load to handle preemption
  KVM: SVM: Remove unnecessary APICv/AVIC update in vCPU unblocking path
  KVM: SVM: Don't bother checking for "running" AVIC when kicking for IPIs
  KVM: SVM: Signal AVIC doorbell iff vCPU is in guest mode
  KVM: x86: Remove defunct pre_block/post_block kvm_x86_ops hooks
  KVM: x86: Unexport LAPIC's switch_to_{hv,sw}_timer() helpers
  ...

Merge tag 'for-5.17/parisc-2' of git://git./linux/kernel/git/deller/parisc-linux

Pull more parisc architecture updates from Helge Deller:
 "Fixes and enhancements:

   - a memory leak fix in an error path in pdc_stable (Miaoqian Lin)

   - two compiler warning fixes in the TOC code

   - added autodetection for currently used console type (serial or
     graphics) which inserts console=<type> if it's missing"

* tag 'for-5.17/parisc-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: pdc_stable: Fix memory leak in pdcs_register_pathentries
  parisc: Fix missing prototype for 'toc_intr' warning in toc.c
  parisc: Autodetect default output device and set console= kernel parameter
  parisc: Use safer strscpy() in setup_cmdline()
  parisc: Add visible flag to toc_stack variable

Merge tag 'riscv-for-linus-5.17-mw1' of git://git./linux/kernel/git/riscv/linux

Pull more RISC-V updates from Palmer Dabbelt:

 - Support for sv48 paging

 - Hart ID mappings are now sparse, which enables more CPUs to come up
   on systems with sparse hart IDs

 - A handful of cleanups and fixes

* tag 'riscv-for-linus-5.17-mw1' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux: (27 commits)
  RISC-V: nommu_virt: Drop unused SLAB_MERGE_DEFAULT
  RISC-V: Remove redundant err variable
  riscv: dts: sifive unmatched: Add gpio poweroff
  riscv: canaan: remove useless select of non-existing config SYSCON
  RISC-V: Do not use cpumask data structure for hartid bitmap
  RISC-V: Move spinwait booting method to its own config
  RISC-V: Move the entire hart selection via lottery to SMP
  RISC-V: Use __cpu_up_stack/task_pointer only for spinwait method
  RISC-V: Do not print the SBI version during HSM extension boot print
  RISC-V: Avoid using per cpu array for ordered booting
  riscv: default to CONFIG_RISCV_SBI_V01=n
  riscv: fix boolconv.cocci warnings
  riscv: Explicit comment about user virtual address space size
  riscv: Use pgtable_l4_enabled to output mmu_type in cpuinfo
  riscv: Implement sv48 support
  asm-generic: Prepare for riscv use of pud_alloc_one and pud_free
  riscv: Allow to dynamically define VA_BITS
  riscv: Introduce functions to switch pt_ops
  riscv: Split early kasan mapping to prepare sv48 introduction
  riscv: Move KASAN mapping next to the kernel mapping
  ...

Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fixes/cleanups from Catalin Marinas:
 "Some fixes that turned up during the merge window:

   - Add brackets to the io_stop_wc macro

   - Avoid -Warray-bounds warning with the LSE atomics inline asm

   - Apply __ro_after_init to memory_limit"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: mm: apply __ro_after_init to memory_limit
  arm64: atomics: lse: Dereference matching size
  asm-generic: Add missing brackets for io_stop_wc macro

Merge tag 'docs-5.17-2' of git://git.lwn.net/linux

Pull documentation fixes from Jonathan Corbet:
 "Three small documentation fixes"

* tag 'docs-5.17-2' of git://git.lwn.net/linux:
  Documentation: fix firewire.rst ABI file path error
  docs: ftrace: fix ambiguous sentence
  docs: staging/tee.rst: fix two typos found while reading

mm: hide the FRONTSWAP Kconfig symbol

Select FRONTSWAP from ZSWAP instead of prompting for it.

Link: https://lkml.kernel.org/r/20211224062246.1258487-14-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: remove support for multiple ops

There is only a single instance of frontswap ops in the kernel, so
simplify the frontswap code by removing support for multiple operations.

Link: https://lkml.kernel.org/r/20211224062246.1258487-13-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: mark swap_lock and swap_active_head static

swap_lock and swap_active_head are only used in swapfile.c, so mark them
static.

Link: https://lkml.kernel.org/r/20211224062246.1258487-12-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: simplify frontswap_register_ops

Given that frontswap_register_ops must be called from built-in code,
there is no need to handle the case of swapfiles coming online before or
during it, so delete the code that deals with that case.

Link: https://lkml.kernel.org/r/20211224062246.1258487-11-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: remove frontswap_test

frontswap_test is unused now, remove it.

Link: https://lkml.kernel.org/r/20211224062246.1258487-10-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: simplify try_to_unuse

Remove the unused frontswap and pages_to_unuse arguments, and mark the
function static now that the caller in frontswap is gone.

[akpm@linux-foundation.org: fix shmem_unuse() stub, per Matthew]

Link: https://lkml.kernel.org/r/20211224062246.1258487-9-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Cc: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: remove the frontswap exports

None of the frontswap API is called from modular code.

Link: https://lkml.kernel.org/r/20211224062246.1258487-8-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: simplify frontswap_init

Just use IS_ENABLED() and remove the __frontswap_init indirection.

Also remove the unused export.

Link: https://lkml.kernel.org/r/20211224062246.1258487-7-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: remove frontswap_curr_pages

frontswap_curr_pages is never called, so remove it.

Link: https://lkml.kernel.org/r/20211224062246.1258487-6-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: remove frontswap_shrink

frontswap_shrink is never called, so remove it.

Link: https://lkml.kernel.org/r/20211224062246.1258487-5-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: remove frontswap_tmem_exclusive_gets

frontswap_tmem_exclusive_gets is never called, so remove it.

Link: https://lkml.kernel.org/r/20211224062246.1258487-4-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frontswap: remove frontswap_writethrough

frontswap_writethrough is never called, so remove it.

Link: https://lkml.kernel.org/r/20211224062246.1258487-3-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: remove cleancache

Patch series "remove Xen tmem leftovers".

Since the removal of the Xen tmem driver in 2019, the cleancache hooks
are entirely unused, as are large parts of frontswap.  This series
against linux-next (with the folio changes included) removes
cleancaches, and cuts down frontswap to the bits actually used by zswap.

This patch (of 13):

The cleancache subsystem is unused since the removal of Xen tmem driver
in commit 814bbf49dcd0 ("xen: remove tmem driver").

[akpm@linux-foundation.org: remove now-unreachable code]

Link: https://lkml.kernel.org/r/20211224062246.1258487-1-hch@lst.de
Link: https://lkml.kernel.org/r/20211224062246.1258487-2-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Konrad Rzeszutek Wilk <Konrad.wilk@oracle.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/stackdepot: always do filter_irq_stacks() in stack_depot_save()

The non-interrupt portion of interrupt stack traces before interrupt
entry is usually arbitrary.  Therefore, saving stack traces of
interrupts (that include entries before interrupt entry) to stack depot
leads to unbounded stackdepot growth.

As such, use of filter_irq_stacks() is a requirement to ensure
stackdepot can efficiently deduplicate interrupt stacks.

Looking through all current users of stack_depot_save(), none (except
KASAN) pass the stack trace through filter_irq_stacks() before passing
it on to stack_depot_save().

Rather than adding filter_irq_stacks() to all current users of
stack_depot_save(), it became clear that stack_depot_save() should
simply do filter_irq_stacks().

Link: https://lkml.kernel.org/r/20211130095727.2378739-1-elver@google.com
Signed-off-by: Marco Elver <elver@google.com>
Reviewed-by: Alexander Potapenko <glider@google.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Vijayanand Jitta <vjitta@codeaurora.org>
Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org>
Cc: Imran Khan <imran.f.khan@oracle.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/stackdepot: allow optional init and stack_table allocation by kvmalloc()

Currently, enabling CONFIG_STACKDEPOT means its stack_table will be
allocated from memblock, even if stack depot ends up not actually used.
The default size of stack_table is 4MB on 32-bit, 8MB on 64-bit.

This is fine for use-cases such as KASAN which is also a config option
and has overhead on its own.  But it's an issue for functionality that
has to be actually enabled on boot (page_owner) or depends on hardware
(GPU drivers) and thus the memory might be wasted.  This was raised as
an issue [1] when attempting to add stackdepot support for SLUB's debug
object tracking functionality.  It's common to build kernels with
CONFIG_SLUB_DEBUG and enable slub_debug on boot only when needed, or
create only specific kmem caches with debugging for testing purposes.

It would thus be more efficient if stackdepot's table was allocated only
when actually going to be used.  This patch thus makes the allocation
(and whole stack_depot_init() call) optional:

 - Add a CONFIG_STACKDEPOT_ALWAYS_INIT flag to keep using the current
   well-defined point of allocation as part of mem_init(). Make
   CONFIG_KASAN select this flag.

 - Other users have to call stack_depot_init() as part of their own init
   when it's determined that stack depot will actually be used. This may
   depend on both config and runtime conditions. Convert current users
   which are page_owner and several in the DRM subsystem. Same will be
   done for SLUB later.

 - Because the init might now be called after the boot-time memblock
   allocation has given all memory to the buddy allocator, change
   stack_depot_init() to allocate stack_table with kvmalloc() when
   memblock is no longer available. Also handle allocation failure by
   disabling stackdepot (could have theoretically happened even with
   memblock allocation previously), and don't unnecessarily align the
   memblock allocation to its own size anymore.

[1] https://lore.kernel.org/all/CAMuHMdW=eoVzM1Re5FVoEN87nKfiLmM2+Ah7eNu2KXEhCvbZyA@mail.gmail.com/

Link: https://lkml.kernel.org/r/20211013073005.11351-1-vbabka@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Dmitry Vyukov <dvyukov@google.com>
Reviewed-by: Marco Elver <elver@google.com> # stackdepot
Cc: Marco Elver <elver@google.com>
Cc: Vijayanand Jitta <vjitta@codeaurora.org>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Oliver Glitta <glittao@gmail.com>
Cc: Imran Khan <imran.f.khan@oracle.com>
From: Colin Ian King <colin.king@canonical.com>
Subject: lib/stackdepot: fix spelling mistake and grammar in pr_err message

There is a spelling mistake of the work allocation so fix this and
re-phrase the message to make it easier to read.

Link: https://lkml.kernel.org/r/20211015104159.11282-1-colin.king@canonical.com
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
From: Vlastimil Babka <vbabka@suse.cz>
Subject: lib/stackdepot: allow optional init and stack_table allocation by kvmalloc() - fixup

On FLATMEM, we call page_ext_init_flatmem_late() just before
kmem_cache_init() which means stack_depot_init() (called by page owner
init) will not recognize properly it should use kvmalloc() and not
memblock_alloc().  memblock_alloc() will also not issue a warning and
return a block memory that can be invalid and cause kernel page fault when
saving stacks, as reported by the kernel test robot [1].

Fix this by moving page_ext_init_flatmem_late() below kmem_cache_init() so
that slab_is_available() is true during stack_depot_init().  SPARSEMEM
doesn't have this issue, as it doesn't do page_ext_init_flatmem_late(),
but a different page_ext_init() even later in the boot process.

Thanks to Mike Rapoport for pointing out the FLATMEM init ordering issue.

While at it, also actually resolve a checkpatch warning in stack_depot_init()
from DRM CI, which was supposed to be in the original patch already.

[1] https://lore.kernel.org/all/20211014085450.GC18719@xsang-OptiPlex-9020/

Link: https://lkml.kernel.org/r/6abd9213-19a9-6d58-cedc-2414386d2d81@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reported-by: kernel test robot <oliver.sang@intel.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
From: Vlastimil Babka <vbabka@suse.cz>
Subject: lib/stackdepot: allow optional init and stack_table allocation by kvmalloc() - fixup3

Due to cd06ab2fd48f ("drm/locking: add backtrace for locking contended
locks without backoff") landing recently to -next adding a new stack depot
user in drivers/gpu/drm/drm_modeset_lock.c we need to add an appropriate
call to stack_depot_init() there as well.

Link: https://lkml.kernel.org/r/2a692365-cfa1-64f2-34e0-8aa5674dce5e@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Naresh Kamboju <naresh.kamboju@linaro.org>
Cc: Marco Elver <elver@google.com>
Cc: Vijayanand Jitta <vjitta@codeaurora.org>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Oliver Glitta <glittao@gmail.com>
Cc: Imran Khan <imran.f.khan@oracle.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
From: Vlastimil Babka <vbabka@suse.cz>
Subject: lib/stackdepot: allow optional init and stack_table allocation by kvmalloc() - fixup4

Due to 4e66934eaadc ("lib: add reference counting tracking
infrastructure") landing recently to net-next adding a new stack depot
user in lib/ref_tracker.c we need to add an appropriate call to
stack_depot_init() there as well.

Link: https://lkml.kernel.org/r/45c1b738-1a2f-5b5f-2f6d-86fab206d01c@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Cc: Jiri Slab <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

proc: remove PDE_DATA() completely

Remove PDE_DATA() completely and replace it with pde_data().

[akpm@linux-foundation.org: fix naming clash in drivers/nubus/proc.c]
[akpm@linux-foundation.org: now fix it properly]

Link: https://lkml.kernel.org/r/20211124081956.87711-2-songmuchun@bytedance.com
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Alexey Gladkov <gladkov.alexey@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs: proc: store PDE()->data into inode->i_private

PDE_DATA(inode) is introduced to get user private data and hide the
layout of struct proc_dir_entry.  The inode->i_private is used to do the
same thing as well.  Save a copy of user private data to inode->
i_private when proc inode is allocated.  This means the user also can
get their private data by inode->i_private.

Introduce pde_data() to wrap inode->i_private so that we can remove
PDE_DATA() from fs/proc/generic.c and make PTE_DATE() as a wrapper of
pde_data().  It will be easier if we decide to remove PDE_DATE() in the
future.

Link: https://lkml.kernel.org/r/20211124081956.87711-1-songmuchun@bytedance.com
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Alexey Gladkov <gladkov.alexey@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

zsmalloc: replace get_cpu_var with local_lock

The usage of get_cpu_var() in zs_map_object() is problematic because it
disables preemption and makes it impossible to acquire any sleeping lock
on PREEMPT_RT such as a spinlock_t.

Replace the get_cpu_var() usage with a local_lock_t which is embedded
struct mapping_area.  It ensures that the access the struct is
synchronized against all users on the same CPU.

[minchan: remove the bit_spin_lock part and change the title]

Link: https://lkml.kernel.org/r/20211115185909.3949505-10-minchan@kernel.org
Signed-off-by: Mike Galbraith <umgwanakikbuti@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Minchan Kim <minchan@kernel.org>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

zsmalloc: replace per zpage lock with pool->migrate_lock

The zsmalloc has used a bit for spin_lock in zpage handle to keep zpage
object alive during several operations.  However, it causes the problem
for PREEMPT_RT as well as introducing too complicated.

This patch replaces the bit spin_lock with pool->migrate_lock rwlock.
It could make the code simple as well as zsmalloc work under PREEMPT_RT.

The drawback is the pool->migrate_lock is bigger granuarity than per
zpage lock so the contention would be higher than old when both
IO-related operations(i.e., zsmalloc, zsfree, zs_[map|unmap]) and
compaction(page/zpage migration) are going in parallel(*, the
migrate_lock is rwlock and IO related functions are all read side lock
so there is no contention).  However, the write-side is fast
enough(dominant overhead is just page copy) so it wouldn't affect much.
If the lock granurity becomes more problem later, we could introduce
table locks based on handle as a hash value.

Link: https://lkml.kernel.org/r/20211115185909.3949505-9-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

locking/rwlocks: introduce write_lock_nested

In preparation for converting bit_spin_lock to rwlock in zsmalloc so
that multiple writers of zspages can run at the same time but those
zspages are supposed to be different zspage instance.  Thus, it's not
deadlock.  This patch adds write_lock_nested to support the case for
LOCKDEP.

[minchan@kernel.org: fix write_lock_nested for RT]
Link: https://lkml.kernel.org/r/YZfrMTAXV56HFWJY@google.com
[bigeasy@linutronix.de: fixup write_lock_nested() implementation]
Link: https://lkml.kernel.org/r/20211123170134.y6xb7pmpgdn4m3bn@linutronix.de
Link: https://lkml.kernel.org/r/20211115185909.3949505-8-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

zsmalloc: remove zspage isolation for migration

zspage isolation for migration introduced additional exceptions to be
dealt with since the zspage was isolated from class list.  The reason
why I isolated zspage from class list was to prevent race between
obj_malloc and page migration via allocating zpage from the zspage
further.  However, it couldn't prevent object freeing from zspage so it
needed corner case handling.

This patch removes the whole mess.  Now, we are fine since class->lock
and zspage->lock can prevent the race.

Link: https://lkml.kernel.org/r/20211115185909.3949505-7-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

zsmalloc: move huge compressed obj from page to zspage

The flag aims for zspage, not per page.  Let's move it to zspage.

Link: https://lkml.kernel.org/r/20211115185909.3949505-6-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

zsmalloc: introduce obj_allocated

The usage pattern for obj_to_head is to check whether the zpage is
allocated or not.  Thus, introduce obj_allocated.

Link: https://lkml.kernel.org/r/20211115185909.3949505-5-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

zsmalloc: decouple class actions from zspage works

This patch moves class stat update out of obj_malloc since it's not
related to zspage operation.  This is a preparation to introduce new
lock scheme in next patch.

Link: https://lkml.kernel.org/r/20211115185909.3949505-4-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

zsmalloc: rename zs_stat_type to class_stat_type

The stat aims for class stat, not zspage so rename it.

Link: https://lkml.kernel.org/r/20211115185909.3949505-3-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

zsmalloc: introduce some helper functions

Patch series "zsmalloc: remove bit_spin_lock", v2.

zsmalloc uses bit_spin_lock to minimize space overhead since it's zpage
granularity lock.  However, it causes zsmalloc non-working under
PREEMPT_RT as well as adding too much complication.

This patchset tries to replace the bit_spin_lock with per-pool rwlock.
It also removes unnecessary zspage isolation logic from class, which was
the other part too much complication added into zsmalloc.

Last patch changes the get_cpu_var to local_lock to make it work in
PREEMPT_RT.

This patch (of 9):

get_zspage_mapping returns fullness as well as class_idx.  However, the
fullness is usually not used since it could be stale in some contexts.
It causes misleading as well as unnecessary instructions so this patch
introduces zspage_class.

obj_to_location also produces page and index but we don't need always
the index, either so this patch introduces obj_to_page.

Link: https://lkml.kernel.org/r/20211115185909.3949505-1-minchan@kernel.org
Link: https://lkml.kernel.org/r/20211115185909.3949505-2-minchan@kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysctl: returns -EINVAL when a negative value is passed to proc_doulongvec_minmax

When we pass a negative value to the proc_doulongvec_minmax() function,
the function returns 0, but the corresponding interface value does not
change.

we can easily reproduce this problem with the following commands:

    cd /proc/sys/fs/epoll
    echo -1 > max_user_watches; echo $?; cat max_user_watches

This function requires a non-negative number to be passed in, so when a
negative number is passed in, -EINVAL is returned.

Link: https://lkml.kernel.org/r/20211220092627.3744624-1-libaokun1@huawei.com
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reported-by: Hulk Robot <hulkci@huawei.com>
Acked-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/sysctl.c: remove unused variable ten_thousand

The const variable ten_thousand is not used, it is redundant and can be
removed.

Cleans up clang warning:

  kernel/sysctl.c:99:18: warning: unused variable 'ten_thousand' [-Wunused-const-variable]
  static const int ten_thousand = 10000;

Link: https://lkml.kernel.org/r/20211221184501.574670-1-colin.i.king@gmail.com
Fixes: c26da54dc8ca ("printk: move printk sysctl to printk/sysctl.c")
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Acked-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kprobe: move sysctl_kprobes_optimization to kprobes.c

kernel/sysctl.c is a kitchen sink where everyone leaves their dirty
dishes, this makes it very difficult to maintain.

To help with this maintenance let's start by moving sysctls to places
where they actually belong.  The proc sysctl maintainers do not want to
know what sysctl knobs you wish to add for your own piece of code, we
just care about the core logic.

Move sysctl_kprobes_optimization from kernel/sysctl.c to
kernel/kprobes.c.  Use register_sysctl() to register the sysctl
interface.

[mcgrof@kernel.org: fix compile issue when CONFIG_OPTPROBES is disabled]

Link: https://lkml.kernel.org/r/20211129211943.640266-7-mcgrof@kernel.org
Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Eric Biggers <ebiggers@google.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: "Naveen N. Rao" <naveen.n.rao@linux.ibm.com>
Cc: Stephen Kitt <steve@sk2.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs/coredump: move coredump sysctls into its own file

This moves the fs/coredump.c respective sysctls to its own file.

Link: https://lkml.kernel.org/r/20211129211943.640266-6-mcgrof@kernel.org
Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Eric Biggers <ebiggers@google.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: "Naveen N. Rao" <naveen.n.rao@linux.ibm.com>
Cc: Stephen Kitt <steve@sk2.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

printk: fix build warning when CONFIG_PRINTK=n

build warning when CONFIG_PRINTK=n

kernel/printk/printk.c:175:5: warning: no previous prototype for
 'devkmsg_sysctl_set_loglvl' [-Wmissing-prototypes]

devkmsg_sysctl_set_loglvl() is only used in sysctl.c when
CONFIG_PRINTK=y, but it participates in the build when CONFIG_PRINTK=n.
So add compile dependency CONFIG_PRINTK=y && CONFIG_SYSCTL=y to fix the
build warning.

Link: https://lkml.kernel.org/r/20211129211943.640266-5-mcgrof@kernel.org
Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Eric Biggers <ebiggers@google.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: "Naveen N. Rao" <naveen.n.rao@linux.ibm.com>
Cc: Stephen Kitt <steve@sk2.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/sysctl.c: rename sysctl_init() to sysctl_init_bases()

Rename sysctl_init() to sysctl_init_bases() so to reflect exactly what
this is doing.

Link: https://lkml.kernel.org/r/20211129211943.640266-4-mcgrof@kernel.org
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Eric Biggers <ebiggers@google.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: "Naveen N. Rao" <naveen.n.rao@linux.ibm.com>
Cc: Stephen Kitt <steve@sk2.org>
Cc: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs: move namespace sysctls and declare fs base directory

This moves the namespace sysctls to its own file as part of the
kernel/sysctl.c spring cleaning

Since we have now removed all sysctls for "fs", we now have to declare
it on the filesystem code, we do that using the new helper, which
reduces boiler plate code.

We rename init_fs_shared_sysctls() to init_fs_sysctls() to reflect that
now fs/sysctls.c is taking on the burden of being the first to register
the base directory as well.

Lastly, since init code will load in the order in which we link it we
have to move the sysctl code to be linked in early, so that its early
init routine runs prior to other fs code.  This way, other filesystem
code can register their own sysctls using the helpers after this:

  * register_sysctl_init()
  * register_sysctl()

Link: https://lkml.kernel.org/r/20211129211943.640266-3-mcgrof@kernel.org
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Eric Biggers <ebiggers@google.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: "Naveen N. Rao" <naveen.n.rao@linux.ibm.com>
Cc: Stephen Kitt <steve@sk2.org>
Cc: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysctl: add and use base directory declarer and registration helper

Patch series "sysctl: add and use base directory declarer and
registration helper".

In this patch series we start addressing base directories, and so we
start with the "fs" sysctls.  The end goal is we end up completely
moving all "fs" sysctl knobs out from kernel/sysctl.

This patch (of 6):

Add a set of helpers which can be used to declare and register base
directory sysctls on their own.  We do this so we can later move each of
the base sysctl directories like "fs", "kernel", etc, to their own
respective files instead of shoving the declarations and registrations
all on kernel/sysctl.c.  The lazy approach has caught up and with this,
we just end up extending the list of base directories / sysctls on one
file and this makes maintenance difficult due to merge conflicts from
many developers.

The declarations are used first by kernel/sysctl.c for registration its
own base which over time we'll try to clean up.  It will be used in the
next patch to demonstrate how to cleanly deal with base sysctl
directories.

[mcgrof@kernel.org: null-terminate the ctl_table arrays]
Link: https://lkml.kernel.org/r/YafJY3rXDYnjK/gs@bombadil.infradead.org
Link: https://lkml.kernel.org/r/20211129211943.640266-1-mcgrof@kernel.org
Link: https://lkml.kernel.org/r/20211129211943.640266-2-mcgrof@kernel.org
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Kees Cook <keescook@chromium.org>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Xiaoming Ni <nixiaoming@huawei.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Stephen Kitt <steve@sk2.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Eric Biggers <ebiggers@google.com>
Cc: "Naveen N. Rao" <naveen.n.rao@linux.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs: move pipe sysctls to is own file

kernel/sysctl.c is a kitchen sink where everyone leaves their dirty
dishes, this makes it very difficult to maintain.

To help with this maintenance let's start by moving sysctls to places
where they actually belong.  The proc sysctl maintainers do not want to
know what sysctl knobs you wish to add for your own piece of code, we
just care about the core logic.

So move the pipe sysctls to its own file.

Link: https://lkml.kernel.org/r/20211129205548.605569-10-mcgrof@kernel.org
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Stephen Kitt <steve@sk2.org>
Cc: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs: move fs/exec.c sysctls into its own file

kernel/sysctl.c is a kitchen sink where everyone leaves their dirty
dishes, this makes it very difficult to maintain.

To help with this maintenance let's start by moving sysctls to places
where they actually belong.  The proc sysctl maintainers do not want to
know what sysctl knobs you wish to add for your own piece of code, we
just care about the core logic.

So move the fs/exec.c respective sysctls to its own file.

Since checkpatch complains about style issues with the old code, this
move also fixes a few of those minor style issues:

  * Use pr_warn() instead of prink(WARNING
  * New empty lines are wanted at the beginning of routines

Link: https://lkml.kernel.org/r/20211129205548.605569-9-mcgrof@kernel.org
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Stephen Kitt <steve@sk2.org>
Cc: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs: move namei sysctls to its own file

kernel/sysctl.c is a kitchen sink where everyone leaves their dirty
dishes, this makes it very difficult to maintain.

To help with this maintenance let's start by moving sysctls to places
where they actually belong.  The proc sysctl maintainers do not want to
know what sysctl knobs you wish to add for your own piece of code, we
just care about the core logic.

So move namei's own sysctl knobs to its own file.

Other than the move we also avoid initializing two static variables to 0
as this is not needed:

  * sysctl_protected_symlinks
  * sysctl_protected_hardlinks

Link: https://lkml.kernel.org/r/20211129205548.605569-8-mcgrof@kernel.org
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lukas Middendorf <kernel@tuxforce.de>
Cc: Stephen Kitt <steve@sk2.org>
Cc: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>