platform/kernel/linux-rpi.git
4 years agoperf/core: Fix race in the perf_mmap_close() function
Jiri Olsa [Wed, 16 Sep 2020 11:53:11 +0000 (13:53 +0200)]
perf/core: Fix race in the perf_mmap_close() function

There's a possible race in perf_mmap_close() when checking ring buffer's
mmap_count refcount value. The problem is that the mmap_count check is
not atomic because we call atomic_dec() and atomic_read() separately.

  perf_mmap_close:
  ...
   atomic_dec(&rb->mmap_count);
   ...
   if (atomic_read(&rb->mmap_count))
      goto out_put;

   <ring buffer detach>
   free_uid

out_put:
  ring_buffer_put(rb); /* could be last */

The race can happen when we have two (or more) events sharing same ring
buffer and they go through atomic_dec() and then they both see 0 as refcount
value later in atomic_read(). Then both will go on and execute code which
is meant to be run just once.

The code that detaches ring buffer is probably fine to be executed more
than once, but the problem is in calling free_uid(), which will later on
demonstrate in related crashes and refcount warnings, like:

  refcount_t: addition on 0; use-after-free.
  ...
  RIP: 0010:refcount_warn_saturate+0x6d/0xf
  ...
  Call Trace:
  prepare_creds+0x190/0x1e0
  copy_creds+0x35/0x172
  copy_process+0x471/0x1a80
  _do_fork+0x83/0x3a0
  __do_sys_wait4+0x83/0x90
  __do_sys_clone+0x85/0xa0
  do_syscall_64+0x5b/0x1e0
  entry_SYSCALL_64_after_hwframe+0x44/0xa9

Using atomic decrease and check instead of separated calls.

Tested-by: Michael Petlan <mpetlan@redhat.com>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Wade Mealing <wmealing@redhat.com>
Fixes: 9bb5d40cd93c ("perf: Fix mmap() accounting hole");
Link: https://lore.kernel.org/r/20200916115311.GE2301783@krava
4 years agoperf/x86: Fix n_metric for cancelled txn
Peter Zijlstra [Mon, 5 Oct 2020 08:10:24 +0000 (10:10 +0200)]
perf/x86: Fix n_metric for cancelled txn

When a group that has TopDown members is failed to be scheduled, any
later TopDown groups will not return valid values.

Here is an example.

A background perf that occupies all the GP counters and the fixed
counter 1.
 $perf stat -e "{cycles,cycles,cycles,cycles,cycles,cycles,cycles,
                 cycles,cycles}:D" -a

A user monitors a TopDown group. It works well, because the fixed
counter 3 and the PERF_METRICS are available.
 $perf stat -x, --topdown -- ./workload
   retiring,bad speculation,frontend bound,backend bound,
   18.0,16.1,40.4,25.5,

Then the user tries to monitor a group that has TopDown members.
Because of the cycles event, the group is failed to be scheduled.
 $perf stat -x, -e '{slots,topdown-retiring,topdown-be-bound,
                     topdown-fe-bound,topdown-bad-spec,cycles}'
                     -- ./workload
    <not counted>,,slots,0,0.00,,
    <not counted>,,topdown-retiring,0,0.00,,
    <not counted>,,topdown-be-bound,0,0.00,,
    <not counted>,,topdown-fe-bound,0,0.00,,
    <not counted>,,topdown-bad-spec,0,0.00,,
    <not counted>,,cycles,0,0.00,,

The user tries to monitor a TopDown group again. It doesn't work anymore.
 $perf stat -x, --topdown -- ./workload

    ,,,,,

In a txn, cancel_txn() is to truncate the event_list for a canceled
group and update the number of events added in this transaction.
However, the number of TopDown events added in this transaction is not
updated. The kernel will probably fail to add new Topdown events.

Fixes: 7b2c05a15d29 ("perf/x86/intel: Generic support for hardware TopDown metrics")
Reported-by: Andi Kleen <ak@linux.intel.com>
Reported-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Kan Liang <kan.liang@linux.intel.com>
Link: https://lkml.kernel.org/r/20201005082611.GH2628@hirez.programming.kicks-ass.net
4 years agoperf/x86: Fix n_pair for cancelled txn
Peter Zijlstra [Mon, 5 Oct 2020 08:09:06 +0000 (10:09 +0200)]
perf/x86: Fix n_pair for cancelled txn

Kan reported that n_metric gets corrupted for cancelled transactions;
a similar issue exists for n_pair for AMD's Large Increment thing.

The problem was confirmed and confirmed fixed by Kim using:

  sudo perf stat -e "{cycles,cycles,cycles,cycles}:D" -a sleep 10 &

  # should succeed:
  sudo perf stat -e "{fp_ret_sse_avx_ops.all}:D" -a workload

  # should fail:
  sudo perf stat -e "{fp_ret_sse_avx_ops.all,fp_ret_sse_avx_ops.all,cycles}:D" -a workload

  # previously failed, now succeeds with this patch:
  sudo perf stat -e "{fp_ret_sse_avx_ops.all}:D" -a workload

Fixes: 5738891229a2 ("perf/x86/amd: Add support for Large Increment per Cycle Events")
Reported-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Kim Phillips <kim.phillips@amd.com>
Link: https://lkml.kernel.org/r/20201005082516.GG2628@hirez.programming.kicks-ass.net
4 years agox86/events/amd/iommu: Fix sizeof mismatch
Colin Ian King [Thu, 1 Oct 2020 11:39:00 +0000 (12:39 +0100)]
x86/events/amd/iommu: Fix sizeof mismatch

An incorrect sizeof is being used, struct attribute ** is not correct,
it should be struct attribute *. Note that since ** is the same size as
* this is not causing any issues.  Improve this fix by using sizeof(*attrs)
as this allows us to not even reference the type of the pointer.

Addresses-Coverity: ("Sizeof not portable (SIZEOF_MISMATCH)")
Fixes: 51686546304f ("x86/events/amd/iommu: Fix sysfs perf attribute groups")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20201001113900.58889-1-colin.king@canonical.com
4 years agoperf/x86/intel: Check perf metrics feature for each CPU
Kan Liang [Thu, 1 Oct 2020 21:17:11 +0000 (14:17 -0700)]
perf/x86/intel: Check perf metrics feature for each CPU

It might be possible that different CPUs have different CPU metrics on a
platform. In this case, writing the GLOBAL_CTRL_EN_PERF_METRICS bit to
the GLOBAL_CTRL register of a CPU, which doesn't support the TopDown
perf metrics feature, causes MSR access error.

Current TopDown perf metrics feature is enumerated using the boot CPU's
PERF_CAPABILITIES MSR. The MSR only indicates the boot CPU supports this
feature.

Check the PERF_CAPABILITIES MSR for each CPU. If any CPU doesn't support
the perf metrics feature, disable the feature globally.

Fixes: 59a854e2f3b9 ("perf/x86/intel: Support TopDown metrics on Ice Lake")
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20201001211711.25708-1-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Fix Ice Lake event constraint table
Kan Liang [Mon, 28 Sep 2020 13:47:26 +0000 (06:47 -0700)]
perf/x86/intel: Fix Ice Lake event constraint table

An error occues when sampling non-PEBS INST_RETIRED.PREC_DIST(0x01c0)
event.

  perf record -e cpu/event=0xc0,umask=0x01/ -- sleep 1
  Error:
  The sys_perf_event_open() syscall returned with 22 (Invalid argument)
  for event (cpu/event=0xc0,umask=0x01/).
  /bin/dmesg | grep -i perf may provide additional information.

The idxmsk64 of the event is set to 0. The event never be successfully
scheduled.

The event should be limit to the fixed counter 0.

Fixes: 6017608936c1 ("perf/x86/intel: Add Icelake support")
Reported-by: Yi, Ammy <ammy.yi@intel.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20200928134726.13090-1-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Fix the scale of the IMC free-running events
Kan Liang [Mon, 28 Sep 2020 13:32:40 +0000 (06:32 -0700)]
perf/x86/intel/uncore: Fix the scale of the IMC free-running events

The "MiB" result of the IMC free-running bandwidth events,
uncore_imc_free_running/read/ and uncore_imc_free_running/write/ are 16
times too small.

The "MiB" value equals the raw IMC free-running bandwidth counter value
times a "scale" which is inaccurate.

The IMC free-running bandwidth events should be incremented per 64B
cache line, not DWs (4 bytes). The "scale" should be 6.103515625e-5.
Fix the "scale" for both Snow Ridge and Ice Lake.

Fixes: 2b3b76b5ec67 ("perf/x86/intel/uncore: Add Ice Lake server uncore support")
Fixes: ee49532b38dd ("perf/x86/intel/uncore: Add IMC uncore support for Snow Ridge")
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200928133240.12977-1-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Fix for iio mapping on Skylake Server
Alexander Antonov [Mon, 28 Sep 2020 10:21:33 +0000 (13:21 +0300)]
perf/x86/intel/uncore: Fix for iio mapping on Skylake Server

Introduced early attributes /sys/devices/uncore_iio_<pmu_idx>/die* are
initialized by skx_iio_set_mapping(), however, for example, for multiple
segment platforms skx_iio_get_topology() returns -EPERM before a list of
attributes in skx_iio_mapping_group will have been initialized.
As a result the list is being NULL. Thus the warning
"sysfs: (bin_)attrs not set by subsystem for group: uncore_iio_*/" appears
and uncore_iio pmus are not available in sysfs. Clear IIO attr_update
to properly handle the cases when topology information cannot be
retrieved.

Fixes: bb42b3d39781 ("perf/x86/intel/uncore: Expose an Uncore unit to IIO PMON mapping")
Reported-by: Kyle Meyer <kyle.meyer@hpe.com>
Suggested-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Alexander Antonov <alexander.antonov@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Alexei Budankov <alexey.budankov@linux.intel.com>
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Link: https://lkml.kernel.org/r/20200928102133.61041-1-alexander.antonov@linux.intel.com
4 years agoperf/x86/msr: Add Jasper Lake support
Kan Liang [Mon, 28 Sep 2020 12:30:42 +0000 (05:30 -0700)]
perf/x86/msr: Add Jasper Lake support

The Jasper Lake processor is also a Tremont microarchitecture. From the
perspective of perf MSR, there is nothing changed compared with
Elkhart Lake.
Share the code path with Elkhart Lake.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1601296242-32763-2-git-send-email-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Add Jasper Lake support
Kan Liang [Mon, 28 Sep 2020 12:30:41 +0000 (05:30 -0700)]
perf/x86/intel: Add Jasper Lake support

The Jasper Lake processor is also a Tremont microarchitecture. From the
perspective of Intel PMU, there is nothing changed compared with
Elkhart Lake.
Share the perf code with Elkhart Lake.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1601296242-32763-1-git-send-email-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Reduce the number of CBOX counters
Kan Liang [Fri, 25 Sep 2020 13:49:05 +0000 (06:49 -0700)]
perf/x86/intel/uncore: Reduce the number of CBOX counters

An oops is triggered by the fuzzy test.

[  327.853081] unchecked MSR access error: RDMSR from 0x70c at rIP:
0xffffffffc082c820 (uncore_msr_read_counter+0x10/0x50 [intel_uncore])
[  327.853083] Call Trace:
[  327.853085]  <IRQ>
[  327.853089]  uncore_pmu_event_start+0x85/0x170 [intel_uncore]
[  327.853093]  uncore_pmu_event_add+0x1a4/0x410 [intel_uncore]
[  327.853097]  ? event_sched_in.isra.118+0xca/0x240

There are 2 GP counters for each CBOX, but the current code claims 4
counters. Accessing the invalid registers triggers the oops.

Fixes: 6e394376ee89 ("perf/x86/intel/uncore: Add Intel Icelake uncore support")
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200925134905.8839-3-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Update Ice Lake uncore units
Kan Liang [Fri, 25 Sep 2020 13:49:04 +0000 (06:49 -0700)]
perf/x86/intel/uncore: Update Ice Lake uncore units

There are some updates for the Icelake model specific uncore performance
monitors. (The update can be found at 10th generation intel core
processors families specification update Revision 004, ICL068)

1) Counter 0 of ARB uncore unit is not available for software use
2) The global 'enable bit' (bit 29) and 'freeze bit' (bit 31) of
   MSR_UNC_PERF_GLOBAL_CTRL cannot be used to control counter behavior.
   Needs to use local enable in event select MSR.

Accessing the modified bit/registers will be ignored by HW. Users may
observe inaccurate results with the current code.

The changes of the MSR_UNC_PERF_GLOBAL_CTRL imply that groups cannot be
read atomically anymore. Although the error of the result for a group
becomes a bit bigger, it still far lower than not using a group. The
group support is still kept. Only Remove the *_box() related
implementation.

Since the counter 0 of ARB uncore unit is not available, update the MSR
address for the ARB uncore unit.

There is no change for IMC uncore unit, which only include free-running
counters.

Fixes: 6e394376ee89 ("perf/x86/intel/uncore: Add Intel Icelake uncore support")
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200925134905.8839-2-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Split the Ice Lake and Tiger Lake MSR uncore support
Kan Liang [Fri, 25 Sep 2020 13:49:03 +0000 (06:49 -0700)]
perf/x86/intel/uncore: Split the Ice Lake and Tiger Lake MSR uncore support

Previously, the MSR uncore for the Ice Lake and Tiger Lake are
identical. The code path is shared. However, with recent update, the
global MSR_UNC_PERF_GLOBAL_CTRL register and ARB uncore unit are changed
for the Ice Lake. Split the Ice Lake and Tiger Lake MSR uncore support.

The changes only impact the MSR ops() and the ARB uncore unit. Other
codes can still be shared between the Ice Lake and the Tiger Lake.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200925134905.8839-1-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Support PCIe3 unit on Snow Ridge
Kan Liang [Mon, 14 Sep 2020 14:34:20 +0000 (07:34 -0700)]
perf/x86/intel/uncore: Support PCIe3 unit on Snow Ridge

The Snow Ridge integrated PCIe3 uncore unit can be used to collect
performance data, e.g. utilization, between PCIe devices, plugged into
the PCIe port, and the components (in M2IOSF) responsible for
translating and managing requests to/from the device. The performance
data is very useful for analyzing the performance of PCIe devices.

The device with the PCIe3 uncore PMON units is owned by the portdrv_pci
driver. Create a PCI sub driver for the PCIe3 uncore PMON units.

Here are some difference between PCIe3 uncore unit and other uncore
pci units.
- There may be several Root Ports on a system. But the uncore counters
  only exist in the Root Port A. A user can configure the channel mask
  to collect the data from other Root Ports.
- The event format of the PCIe3 uncore unit is the same as IIO unit of
  SKX.
- The Control Register of PCIe3 uncore unit is 64 bits.
- The offset of each counters is 8, which is the same as M2M unit of
  SNR.
- New MSR addresses for unit control, counter and counter config.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1600094060-82746-7-git-send-email-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Generic support for the PCI sub driver
Kan Liang [Mon, 14 Sep 2020 14:34:19 +0000 (07:34 -0700)]
perf/x86/intel/uncore: Generic support for the PCI sub driver

Some uncore counters may be located in the configuration space of a PCI
device, which already has a bonded driver. Currently, the uncore driver
cannot register a PCI uncore PMU for these counters, because, to
register a PCI uncore PMU, the uncore driver must be bond to the device.
However, one device can only have one bonded driver.

Add an uncore PCI sub driver to support such kind of devices.

The sub driver doesn't own the device. In initialization, the sub
driver searches the device via pci_get_device(), and register the
corresponding PMU for the device. In the meantime, the sub driver
registers a PCI bus notifier, which is used to notify the sub driver
once the device is removed. The sub driver can unregister the PMU
accordingly.

The sub driver only searches the devices defined in its id table. The
id table varies on different platforms, which will be implemented in the
following platform-specific patch.

Suggested-by: Bjorn Helgaas <helgaas@kernel.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1600094060-82746-6-git-send-email-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Factor out uncore_pci_pmu_unregister()
Kan Liang [Mon, 14 Sep 2020 14:34:18 +0000 (07:34 -0700)]
perf/x86/intel/uncore: Factor out uncore_pci_pmu_unregister()

The PMU unregistration in the uncore PCI sub driver is similar as the
normal PMU unregistration for a PCI device. The codes to unregister a
PCI PMU can be shared.

Factor out uncore_pci_pmu_unregister(), which will be used later.

Use uncore_pci_get_dev_die_info() to replace the codes which retrieve
the socket and die informaion.

The pci_set_drvdata() is not included in uncore_pci_pmu_unregister() as
well, because the uncore PCI sub driver will not touch the private
driver data pointer of the device.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1600094060-82746-5-git-send-email-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Factor out uncore_pci_pmu_register()
Kan Liang [Mon, 14 Sep 2020 14:34:17 +0000 (07:34 -0700)]
perf/x86/intel/uncore: Factor out uncore_pci_pmu_register()

The PMU registration in the uncore PCI sub driver is similar as the
normal PMU registration for a PCI device. The codes to register a PCI
PMU can be shared.

Factor out uncore_pci_pmu_register(), which will be used later.

The pci_set_drvdata() is not included in uncore_pci_pmu_register(). The
uncore PCI sub driver doesn't own the PCI device. It will not touch the
private driver data pointer for the device.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1600094060-82746-4-git-send-email-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Factor out uncore_pci_find_dev_pmu()
Kan Liang [Mon, 14 Sep 2020 14:34:16 +0000 (07:34 -0700)]
perf/x86/intel/uncore: Factor out uncore_pci_find_dev_pmu()

When an uncore PCI sub driver gets a remove notification, the
corresponding PMU has to be retrieved and unregistered. The codes, which
find the corresponding PMU by comparing the pci_device_id table, can be
shared.

Factor out uncore_pci_find_dev_pmu(), which will be used later.

There is no functional change.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1600094060-82746-3-git-send-email-kan.liang@linux.intel.com
4 years agoperf/x86/intel/uncore: Factor out uncore_pci_get_dev_die_info()
Kan Liang [Mon, 14 Sep 2020 14:34:15 +0000 (07:34 -0700)]
perf/x86/intel/uncore: Factor out uncore_pci_get_dev_die_info()

The socket and die information is required to register/unregister a PMU
in the uncore PCI sub driver. The codes, which get the socket and die
information from a BUS number, can be shared.

Factor out uncore_pci_get_dev_die_info(), which will be used later.

There is no functional change.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1600094060-82746-2-git-send-email-kan.liang@linux.intel.com
4 years agoperf/amd/uncore: Inform the user how many counters each uncore PMU has
Kim Phillips [Mon, 21 Sep 2020 14:43:30 +0000 (09:43 -0500)]
perf/amd/uncore: Inform the user how many counters each uncore PMU has

Previously, the uncore driver would say "NB counters detected" on F17h
machines, which don't have NorthBridge (NB) counters.  They have Data
Fabric (DF) counters.  Just use the pmu.name to inform users which pmu
to use and its associated counter count.

F17h dmesg BEFORE:

amd_uncore: AMD NB counters detected
amd_uncore: AMD LLC counters detected

F17h dmesg AFTER:

amd_uncore: 4 amd_df counters detected
amd_uncore: 6 amd_l3 counters detected

Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200921144330.6331-5-kim.phillips@amd.com
4 years agoperf/amd/uncore: Allow F19h user coreid, threadmask, and sliceid specification
Kim Phillips [Mon, 21 Sep 2020 14:43:29 +0000 (09:43 -0500)]
perf/amd/uncore: Allow F19h user coreid, threadmask, and sliceid specification

On Family 19h, the driver checks for a populated 2-bit threadmask in
order to establish that the user wants to measure individual slices,
individual cores (only one can be measured at a time), and lets
the user also directly specify enallcores and/or enallslices if
desired.

Example F19h invocation to measure L3 accesses (event 4, umask 0xff)
by the first thread (id 0 -> mask 0x1) of the first core (id 0) on the
first slice (id 0):

perf stat -a -e instructions,amd_l3/umask=0xff,event=0x4,coreid=0,threadmask=1,sliceid=0,enallcores=0,enallslices=0/ <workload>

Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200921144330.6331-4-kim.phillips@amd.com
4 years agoperf/amd/uncore: Allow F17h user threadmask and slicemask specification
Kim Phillips [Mon, 21 Sep 2020 14:43:28 +0000 (09:43 -0500)]
perf/amd/uncore: Allow F17h user threadmask and slicemask specification

Continue to fully populate either one of threadmask or slicemask if the
user doesn't.

Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200921144330.6331-3-kim.phillips@amd.com
4 years agoperf/amd/uncore: Prepare to scale for more attributes that vary per family
Kim Phillips [Mon, 21 Sep 2020 14:43:27 +0000 (09:43 -0500)]
perf/amd/uncore: Prepare to scale for more attributes that vary per family

Replace AMD_FORMAT_ATTR with the more apropos DEFINE_UNCORE_FORMAT_ATTR
stolen from arch/x86/events/intel/uncore.h.  This way we can clearly
see the bit-variants of each of the attributes that want to have
the same name across families.

Also unroll AMD_ATTRIBUTE because we are going to separately add
new attributes that differ between DF and L3.

Also clean up the if-Family 17h-else logic in amd_uncore_init.

This is basically a rewrite of commit da6adaea2b7e
("perf/x86/amd/uncore: Update sysfs attributes for Family17h processors").

No functional changes.

Tested F17h+ /sys/bus/event_source/devices/amd_{l3,df}/format/*
content remains unchanged:

/sys/bus/event_source/devices/amd_l3/format/event:config:0-7
/sys/bus/event_source/devices/amd_l3/format/umask:config:8-15
/sys/bus/event_source/devices/amd_df/format/event:config:0-7,32-35,59-60
/sys/bus/event_source/devices/amd_df/format/umask:config:8-15

Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200921144330.6331-2-kim.phillips@amd.com
4 years agokprobes: Use module_name() macro
Jarkko Sakkinen [Tue, 22 Sep 2020 01:24:25 +0000 (21:24 -0400)]
kprobes: Use module_name() macro

It is advised to use module_name() macro instead of dereferencing mod->name
directly. This makes sense for consistencys sake and also it prevents a
hard dependency to CONFIG_MODULES.

Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Masami Hiramatsu <mhiramat@kernel.org>
Link: https://lkml.kernel.org/r/20200818050857.117998-1-jarkko.sakkinen@linux.intel.com
4 years agoarch/x86/amd/ibs: Fix re-arming IBS Fetch
Kim Phillips [Tue, 8 Sep 2020 21:47:36 +0000 (16:47 -0500)]
arch/x86/amd/ibs: Fix re-arming IBS Fetch

Stephane Eranian found a bug in that IBS' current Fetch counter was not
being reset when the driver would write the new value to clear it along
with the enable bit set, and found that adding an MSR write that would
first disable IBS Fetch would make IBS Fetch reset its current count.

Indeed, the PPR for AMD Family 17h Model 31h B0 55803 Rev 0.54 - Sep 12,
2019 states "The periodic fetch counter is set to IbsFetchCnt [...] when
IbsFetchEn is changed from 0 to 1."

Explicitly set IbsFetchEn to 0 and then to 1 when re-enabling IBS Fetch,
so the driver properly resets the internal counter to 0 and IBS
Fetch starts counting again.

A family 15h machine tested does not have this problem, and the extra
wrmsr is also not needed on Family 19h, so only do the extra wrmsr on
families 16h through 18h.

Reported-by: Stephane Eranian <stephane.eranian@google.com>
Signed-off-by: Kim Phillips <kim.phillips@amd.com>
[peterz: optimized]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://bugzilla.kernel.org/show_bug.cgi?id=206537
4 years agoperf/x86/rapl: Add AMD Fam19h RAPL support
Kim Phillips [Tue, 8 Sep 2020 21:47:40 +0000 (16:47 -0500)]
perf/x86/rapl: Add AMD Fam19h RAPL support

Family 19h RAPL support did not change from Family 17h; extend
the existing Fam17h support to work on Family 19h too.

Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200908214740.18097-8-kim.phillips@amd.com
4 years agoperf/x86/amd/ibs: Support 27-bit extended Op/cycle counter
Kim Phillips [Tue, 8 Sep 2020 21:47:39 +0000 (16:47 -0500)]
perf/x86/amd/ibs: Support 27-bit extended Op/cycle counter

IBS hardware with the OpCntExt feature gets a 7-bit wider internal
counter.  Both the maximum and current count bitfields in the
IBS_OP_CTL register are extended to support reading and writing it.

No changes are necessary to the driver for handling the extra
contiguous current count bits (IbsOpCurCnt), as the driver already
passes through 32 bits of that field.  However, the driver has to do
some extra bit manipulation when converting from a period to the
non-contiguous (although conveniently aligned) extra bits in the
IbsOpMaxCnt bitfield.

This decreases IBS Op interrupt overhead when the period is over
1,048,560 (0xffff0), which would previously activate the driver's
software counter.  That threshold is now 134,217,712 (0x7fffff0).

Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200908214740.18097-7-kim.phillips@amd.com
4 years agoperf/x86/amd/ibs: Fix raw sample data accumulation
Kim Phillips [Tue, 8 Sep 2020 21:47:38 +0000 (16:47 -0500)]
perf/x86/amd/ibs: Fix raw sample data accumulation

Neither IbsBrTarget nor OPDATA4 are populated in IBS Fetch mode.
Don't accumulate them into raw sample user data in that case.

Also, in Fetch mode, add saving the IBS Fetch Control Extended MSR.

Technically, there is an ABI change here with respect to the IBS raw
sample data format, but I don't see any perf driver version information
being included in perf.data file headers, but, existing users can detect
whether the size of the sample record has reduced by 8 bytes to
determine whether the IBS driver has this fix.

Fixes: 904cb3677f3a ("perf/x86/amd/ibs: Update IBS MSRs and feature definitions")
Reported-by: Stephane Eranian <stephane.eranian@google.com>
Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20200908214740.18097-6-kim.phillips@amd.com
4 years agoperf/x86/amd/ibs: Don't include randomized bits in get_ibs_op_count()
Kim Phillips [Tue, 8 Sep 2020 21:47:37 +0000 (16:47 -0500)]
perf/x86/amd/ibs: Don't include randomized bits in get_ibs_op_count()

get_ibs_op_count() adds hardware's current count (IbsOpCurCnt) bits
to its count regardless of hardware's valid status.

According to the PPR for AMD Family 17h Model 31h B0 55803 Rev 0.54,
if the counter rolls over, valid status is set, and the lower 7 bits
of IbsOpCurCnt are randomized by hardware.

Don't include those bits in the driver's event count.

Fixes: 8b1e13638d46 ("perf/x86-ibs: Fix usage of IBS op current count")
Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://bugzilla.kernel.org/show_bug.cgi?id=206537
4 years agoperf/x86/amd: Fix sampling Large Increment per Cycle events
Kim Phillips [Tue, 8 Sep 2020 21:47:35 +0000 (16:47 -0500)]
perf/x86/amd: Fix sampling Large Increment per Cycle events

Commit 5738891229a2 ("perf/x86/amd: Add support for Large Increment
per Cycle Events") mistakenly zeroes the upper 16 bits of the count
in set_period().  That's fine for counting with perf stat, but not
sampling with perf record when only Large Increment events are being
sampled.  To enable sampling, we sign extend the upper 16 bits of the
merged counter pair as described in the Family 17h PPRs:

"Software wanting to preload a value to a merged counter pair writes the
high-order 16-bit value to the low-order 16 bits of the odd counter and
then writes the low-order 48-bit value to the even counter. Reading the
even counter of the merged counter pair returns the full 64-bit value."

Fixes: 5738891229a2 ("perf/x86/amd: Add support for Large Increment per Cycle Events")
Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://bugzilla.kernel.org/show_bug.cgi?id=206537
4 years agoperf/amd/uncore: Set all slices and threads to restore perf stat -a behaviour
Kim Phillips [Tue, 8 Sep 2020 21:47:34 +0000 (16:47 -0500)]
perf/amd/uncore: Set all slices and threads to restore perf stat -a behaviour

Commit 2f217d58a8a0 ("perf/x86/amd/uncore: Set the thread mask for
F17h L3 PMCs") inadvertently changed the uncore driver's behaviour
wrt perf tool invocations with or without a CPU list, specified with
-C / --cpu=.

Change the behaviour of the driver to assume the former all-cpu (-a)
case, which is the more commonly desired default.  This fixes
'-a -A' invocations without explicit cpu lists (-C) to not count
L3 events only on behalf of the first thread of the first core
in the L3 domain.

BEFORE:

Activity performed by the first thread of the last core (CPU#43) in
CPU#40's L3 domain is not reported by CPU#40:

sudo perf stat -a -A -e l3_request_g1.caching_l3_cache_accesses taskset -c 43 perf bench mem memcpy -s 32mb -l 100 -f default
...
CPU36                 21,835      l3_request_g1.caching_l3_cache_accesses
CPU40                 87,066      l3_request_g1.caching_l3_cache_accesses
CPU44                 17,360      l3_request_g1.caching_l3_cache_accesses
...

AFTER:

The L3 domain activity is now reported by CPU#40:

sudo perf stat -a -A -e l3_request_g1.caching_l3_cache_accesses taskset -c 43 perf bench mem memcpy -s 32mb -l 100 -f default
...
CPU36                354,891      l3_request_g1.caching_l3_cache_accesses
CPU40              1,780,870      l3_request_g1.caching_l3_cache_accesses
CPU44                315,062      l3_request_g1.caching_l3_cache_accesses
...

Fixes: 2f217d58a8a0 ("perf/x86/amd/uncore: Set the thread mask for F17h L3 PMCs")
Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20200908214740.18097-2-kim.phillips@amd.com
4 years agoperf/core: Pull pmu::sched_task() into perf_event_context_sched_out()
Kan Liang [Fri, 21 Aug 2020 19:57:53 +0000 (12:57 -0700)]
perf/core: Pull pmu::sched_task() into perf_event_context_sched_out()

The pmu::sched_task() is a context switch callback. It passes the
cpuctx->task_ctx as a parameter to the lower code. To find the
cpuctx->task_ctx, the current code iterates a cpuctx list.
The same context will iterated in perf_event_context_sched_out() soon.
Share the cpuctx->task_ctx can avoid the unnecessary iteration of the
cpuctx list.

The pmu::sched_task() is also required for the optimization case for
equivalent contexts.

The task_ctx_sched_out() will eventually disable and reenable the PMU
when schedule out events. Add perf_pmu_disable() and perf_pmu_enable()
around task_ctx_sched_out() don't break anything.

Drop the cpuctx->ctx.lock for the pmu::sched_task(). The lock is for
per-CPU context, which is not necessary for the per-task context
schedule.

No one uses sched_cb_entry, perf_sched_cb_usages, sched_cb_list, and
perf_pmu_sched_task() any more.

Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200821195754.20159-2-kan.liang@linux.intel.com
4 years agoperf/core: Pull pmu::sched_task() into perf_event_context_sched_in()
Kan Liang [Fri, 21 Aug 2020 19:57:52 +0000 (12:57 -0700)]
perf/core: Pull pmu::sched_task() into perf_event_context_sched_in()

The pmu::sched_task() is a context switch callback. It passes the
cpuctx->task_ctx as a parameter to the lower code. To find the
cpuctx->task_ctx, the current code iterates a cpuctx list.

The same context was just iterated in perf_event_context_sched_in(),
which is invoked right before the pmu::sched_task().

Reuse the cpuctx->task_ctx from perf_event_context_sched_in() can avoid
the unnecessary iteration of the cpuctx list.

Both pmu::sched_task and perf_event_context_sched_in() have to disable
PMU. Pull the pmu::sched_task into perf_event_context_sched_in() can
also save the overhead from the PMU disable and reenable.

The new and old tasks may have equivalent contexts. The current code
optimize this case by swapping the context, which avoids the scheduling.
For this case, pmu::sched_task() is still required, e.g., restore the
LBR content.

Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200821195754.20159-1-kan.liang@linux.intel.com
4 years agoperf/x86/intel/ds: Fix x86_pmu_stop warning for large PEBS
Kan Liang [Wed, 2 Sep 2020 21:06:49 +0000 (14:06 -0700)]
perf/x86/intel/ds: Fix x86_pmu_stop warning for large PEBS

A warning as below may be triggered when sampling with large PEBS.

[  410.411250] perf: interrupt took too long (72145 > 71975), lowering
kernel.perf_event_max_sample_rate to 2000
[  410.724923] ------------[ cut here ]------------
[  410.729822] WARNING: CPU: 0 PID: 16397 at arch/x86/events/core.c:1422
x86_pmu_stop+0x95/0xa0
[  410.933811]  x86_pmu_del+0x50/0x150
[  410.937304]  event_sched_out.isra.0+0xbc/0x210
[  410.941751]  group_sched_out.part.0+0x53/0xd0
[  410.946111]  ctx_sched_out+0x193/0x270
[  410.949862]  __perf_event_task_sched_out+0x32c/0x890
[  410.954827]  ? set_next_entity+0x98/0x2d0
[  410.958841]  __schedule+0x592/0x9c0
[  410.962332]  schedule+0x5f/0xd0
[  410.965477]  exit_to_usermode_loop+0x73/0x120
[  410.969837]  prepare_exit_to_usermode+0xcd/0xf0
[  410.974369]  ret_from_intr+0x2a/0x3a
[  410.977946] RIP: 0033:0x40123c
[  411.079661] ---[ end trace bc83adaea7bb664a ]---

In the non-overflow context, e.g., context switch, with large PEBS, perf
may stop an event twice. An example is below.

  //max_samples_per_tick is adjusted to 2
  //NMI is triggered
  intel_pmu_handle_irq()
     handle_pmi_common()
       drain_pebs()
         __intel_pmu_pebs_event()
           perf_event_overflow()
             __perf_event_account_interrupt()
               hwc->interrupts = 1
               return 0
  //A context switch happens right after the NMI.
  //In the same tick, the perf_throttled_seq is not changed.
  perf_event_task_sched_out()
     perf_pmu_sched_task()
       intel_pmu_drain_pebs_buffer()
         __intel_pmu_pebs_event()
           perf_event_overflow()
             __perf_event_account_interrupt()
               ++hwc->interrupts >= max_samples_per_tick
               return 1
           x86_pmu_stop();  # First stop
     perf_event_context_sched_out()
       task_ctx_sched_out()
         ctx_sched_out()
           event_sched_out()
             x86_pmu_del()
               x86_pmu_stop();  # Second stop and trigger the warning

Perf should only invoke the perf_event_overflow() in the overflow
context.

Current drain_pebs() is called from:
- handle_pmi_common() -- overflow context
- intel_pmu_pebs_sched_task() -- non-overflow context
- intel_pmu_pebs_disable() -- non-overflow context
- intel_pmu_auto_reload_read() -- possible overflow context
  With PERF_SAMPLE_READ + PERF_FORMAT_GROUP, the function may be
  invoked in the NMI handler. But, before calling the function, the
  PEBS buffer has already been drained. The __intel_pmu_pebs_event()
  will not be called in the possible overflow context.

To fix the issue, an indicator is required to distinguish between the
overflow context aka handle_pmi_common() and other cases.
The dummy regs pointer can be used as the indicator.

In the non-overflow context, perf should treat the last record the same
as other PEBS records, and doesn't invoke the generic overflow handler.

Fixes: 21509084f999 ("perf/x86/intel: Handle multiple records in the PEBS buffer")
Reported-by: Like Xu <like.xu@linux.intel.com>
Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Like Xu <like.xu@linux.intel.com>
Link: https://lkml.kernel.org/r/20200902210649.2743-1-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Support per-thread RDPMC TopDown metrics
Kan Liang [Thu, 23 Jul 2020 17:11:14 +0000 (10:11 -0700)]
perf/x86/intel: Support per-thread RDPMC TopDown metrics

Starts from Ice Lake, the TopDown metrics are directly available as
fixed counters and do not require generic counters. Also, the TopDown
metrics can be collected per thread. Extend the RDPMC usage to support
per-thread TopDown metrics.

The RDPMC index of the PERF_METRICS will be output if RDPMC users ask
for the RDPMC index of the metrics events.

To support per thread RDPMC TopDown, the metrics and slots counters have
to be saved/restored during the context switching.

The last_period and period_left are not used in the counting mode. Use
the fields for saved_metric and saved_slots.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-12-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Support TopDown metrics on Ice Lake
Kan Liang [Thu, 23 Jul 2020 17:11:13 +0000 (10:11 -0700)]
perf/x86/intel: Support TopDown metrics on Ice Lake

Ice Lake supports the hardware TopDown metrics feature, which can free
up the scarce GP counters.

Update the event constraints for the metrics events. The metric counters
do not exist, which are mapped to a dummy offset. The sharing between
multiple users of the same metric without multiplexing is not allowed.

Implement set_topdown_event_period for Ice Lake. The values in
PERF_METRICS MSR are derived from the fixed counter 3. Both registers
should start from zero.

Implement update_topdown_event for Ice Lake. The metric is reported by
multiplying the metric (fraction) with slots. To maintain accurate
measurements, both registers are cleared for each update. The fixed
counter 3 should always be cleared before the PERF_METRICS.

Implement td_attr for the new metrics events and the new slots fixed
counter. Make them visible to the perf user tools.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-11-kan.liang@linux.intel.com
4 years agoperf/x86: Add a macro for RDPMC offset of fixed counters
Kan Liang [Thu, 23 Jul 2020 17:11:12 +0000 (10:11 -0700)]
perf/x86: Add a macro for RDPMC offset of fixed counters

The RDPMC base offset of fixed counters is hard-code. Use a meaningful
name to replace the magic number to improve the readability of the code.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-10-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Generic support for hardware TopDown metrics
Kan Liang [Thu, 23 Jul 2020 17:11:11 +0000 (10:11 -0700)]
perf/x86/intel: Generic support for hardware TopDown metrics

Intro
=====

The TopDown Microarchitecture Analysis (TMA) Method is a structured
analysis methodology to identify critical performance bottlenecks in
out-of-order processors. Current perf has supported the method.

The method works well, but there is one problem. To collect the TopDown
events, several GP counters have to be used. If a user wants to collect
other events at the same time, the multiplexing probably be triggered,
which impacts the accuracy.

To free up the scarce GP counters, the hardware TopDown metrics feature
is introduced from Ice Lake. The hardware implements an additional
"metrics" register and a new Fixed Counter 3 that measures pipeline
"slots". The TopDown events can be calculated from them instead.

Events
======

The level 1 TopDown has four metrics. There is no event-code assigned to
the TopDown metrics. Four metric events are exported as separate perf
events, which map to the internal "metrics" counter register. Those
events do not exist in hardware, but can be allocated by the scheduler.

For the event mapping, a special 0x00 event code is used, which is
reserved for fake events. The metric events start from umask 0x10.

When setting up the metric events, they point to the Fixed Counter 3.
They have to be specially handled.
- Add the update_topdown_event() callback to read the additional metrics
  MSR and generate the metrics.
- Add the set_topdown_event_period() callback to initialize metrics MSR
  and the fixed counter 3.
- Add a variable n_metric_event to track the number of the accepted
  metrics events. The sharing between multiple users of the same metric
  without multiplexing is not allowed.
- Only enable/disable the fixed counter 3 when there are no other active
  TopDown events, which avoid the unnecessary writing of the fixed
  control register.
- Disable the PMU when reading the metrics event. The metrics MSR and
  the fixed counter 3 are read separately. The values may be modified by
  an NMI.

All four metric events don't support sampling. Since they will be
handled specially for event update, a flag PERF_X86_EVENT_TOPDOWN is
introduced to indicate this case.

The slots event can support both sampling and counting.
For counting, the flag is also applied.
For sampling, it will be handled normally as other normal events.

Groups
======

The slots event is required in a Topdown group.
To avoid reading the METRICS register multiple times, the metrics and
slots value can only be updated by slots event in a group.
All active slots and metrics events will be updated one time.
Therefore, the slots event must be before any metric events in a Topdown
group.

NMI
======

The METRICS related register may be overflow. The bit 48 of the STATUS
register will be set. If so, PERF_METRICS and Fixed counter 3 are
required to be reset. The patch also update all active slots and
metrics events in the NMI handler.

The update_topdown_event() has to read two registers separately. The
values may be modified by an NMI. PMU has to be disabled before calling
the function.

RDPMC
======

RDPMC is temporarily disabled. A later patch will enable it.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-9-kan.liang@linux.intel.com
4 years agoperf/core: Add a new PERF_EV_CAP_SIBLING event capability
Kan Liang [Thu, 23 Jul 2020 17:11:10 +0000 (10:11 -0700)]
perf/core: Add a new PERF_EV_CAP_SIBLING event capability

Current perf assumes that events in a group are independent. Close an
event doesn't impact the value of the other events in the same group.
If the closed event is a member, after the event closure, other events
are still running like a group. If the closed event is a leader, other
events are running as singleton events.

Add PERF_EV_CAP_SIBLING to allow events to indicate they require being
part of a group, and when the leader dies they cannot exist
independently.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-8-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Use switch in intel_pmu_disable/enable_event
Kan Liang [Thu, 23 Jul 2020 17:11:09 +0000 (10:11 -0700)]
perf/x86/intel: Use switch in intel_pmu_disable/enable_event

Currently, the if-else is used in the intel_pmu_disable/enable_event to
check the type of an event. It works well, but with more and more types
added later, e.g., perf metrics, compared to the switch statement, the
if-else may impair the readability of the code.

There is no harm to use the switch statement to replace the if-else
here. Also, some optimizing compilers may compile a switch statement
into a jump-table which is more efficient than if-else for a large
number of cases. The performance gain may not be observed for now,
because the number of cases is only 5, but the benefits may be observed
with more and more types added in the future.

Use switch to replace the if-else in the intel_pmu_disable/enable_event.

If the idx is invalid, print a warning.

For the case INTEL_PMC_IDX_FIXED_BTS in intel_pmu_disable_event, don't
need to check the event->attr.precise_ip. Use return for the case.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-7-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Fix the name of perf METRICS
Kan Liang [Thu, 23 Jul 2020 17:11:08 +0000 (10:11 -0700)]
perf/x86/intel: Fix the name of perf METRICS

Bit 15 of the PERF_CAPABILITIES MSR indicates that the perf METRICS
feature is supported. The perf METRICS is not a PEBS feature.

Rename pebs_metrics_available perf_metrics.

The bit is not used in the current code. It will be used in a later
patch.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-6-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Move BTS index to 47
Kan Liang [Thu, 23 Jul 2020 17:11:07 +0000 (10:11 -0700)]
perf/x86/intel: Move BTS index to 47

The bit 48 in the PERF_GLOBAL_STATUS is used to indicate the overflow
status of the PERF_METRICS counters.

Move the BTS index to the bit 47.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-5-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Introduce the fourth fixed counter
Kan Liang [Thu, 23 Jul 2020 17:11:06 +0000 (10:11 -0700)]
perf/x86/intel: Introduce the fourth fixed counter

The fourth fixed counter, TOPDOWN.SLOTS, is introduced in Ice Lake to
measure the level 1 TopDown events.

Add MSR address and macros for the new fixed counter, which will be used
in a later patch.

Add comments to explain the event encoding rules for the fixed counters.

Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-4-kan.liang@linux.intel.com
4 years agoperf/x86/intel: Name the global status bit in NMI handler
Kan Liang [Thu, 23 Jul 2020 17:11:05 +0000 (10:11 -0700)]
perf/x86/intel: Name the global status bit in NMI handler

Magic numbers are used in the current NMI handler for the global status
bit. Use a meaningful name to replace the magic numbers to improve the
readability of the code.

Remove a Tab for all GLOBAL_STATUS_* and INTEL_PMC_IDX_FIXED_BTS macros
to reduce the length of the line.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-3-kan.liang@linux.intel.com
4 years agoperf/x86: Use event_base_rdpmc for the RDPMC userspace support
Kan Liang [Thu, 23 Jul 2020 17:11:04 +0000 (10:11 -0700)]
perf/x86: Use event_base_rdpmc for the RDPMC userspace support

The RDPMC index is always re-calculated for the RDPMC userspace support,
which is unnecessary.

The RDPMC index value is stored in the variable event_base_rdpmc for
the kernel usage, which can be used for RDPMC userspace support as well.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200723171117.9918-2-kan.liang@linux.intel.com
4 years agoLinux 5.9-rc1
Linus Torvalds [Sun, 16 Aug 2020 20:04:57 +0000 (13:04 -0700)]
Linux 5.9-rc1

4 years agoMerge tag 'io_uring-5.9-2020-08-15' of git://git.kernel.dk/linux-block
Linus Torvalds [Sun, 16 Aug 2020 17:55:12 +0000 (10:55 -0700)]
Merge tag 'io_uring-5.9-2020-08-15' of git://git.kernel.dk/linux-block

Pull io_uring fixes from Jens Axboe:
 "A few differerent things in here.

  Seems like syzbot got some more io_uring bits wired up, and we got a
  handful of reports and the associated fixes are in here.

  General fixes too, and a lot of them marked for stable.

  Lastly, a bit of fallout from the async buffered reads, where we now
  more easily trigger short reads. Some applications don't really like
  that, so the io_read() code now handles short reads internally, and
  got a cleanup along the way so that it's now easier to read (and
  documented). We're now passing tests that failed before"

* tag 'io_uring-5.9-2020-08-15' of git://git.kernel.dk/linux-block:
  io_uring: short circuit -EAGAIN for blocking read attempt
  io_uring: sanitize double poll handling
  io_uring: internally retry short reads
  io_uring: retain iov_iter state over io_read/io_write calls
  task_work: only grab task signal lock when needed
  io_uring: enable lookup of links holding inflight files
  io_uring: fail poll arm on queue proc failure
  io_uring: hold 'ctx' reference around task_work queue + execute
  fs: RWF_NOWAIT should imply IOCB_NOIO
  io_uring: defer file table grabbing request cleanup for locked requests
  io_uring: add missing REQ_F_COMP_LOCKED for nested requests
  io_uring: fix recursive completion locking on oveflow flush
  io_uring: use TWA_SIGNAL for task_work uncondtionally
  io_uring: account locked memory before potential error case
  io_uring: set ctx sq/cq entry count earlier
  io_uring: Fix NULL pointer dereference in loop_rw_iter()
  io_uring: add comments on how the async buffered read retry works
  io_uring: io_async_buf_func() need not test page bit

4 years agoparisc: fix PMD pages allocation by restoring pmd_alloc_one()
Mike Rapoport [Sun, 16 Aug 2020 14:24:03 +0000 (17:24 +0300)]
parisc: fix PMD pages allocation by restoring pmd_alloc_one()

Commit 1355c31eeb7e ("asm-generic: pgalloc: provide generic pmd_alloc_one()
and pmd_free_one()") converted parisc to use generic version of
pmd_alloc_one() but it missed the fact that parisc uses order-1 pages for
PMD.

Restore the original version of pmd_alloc_one() for parisc, just use
GFP_PGTABLE_KERNEL that implies __GFP_ZERO instead of GFP_KERNEL and
memset.

Fixes: 1355c31eeb7e ("asm-generic: pgalloc: provide generic pmd_alloc_one() and pmd_free_one()")
Reported-by: Meelis Roos <mroos@linux.ee>
Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
Tested-by: Meelis Roos <mroos@linux.ee>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Link: https://lkml.kernel.org/r/9f2b5ebd-e4a4-0fa1-6cd3-4b9f6892d1ad@linux.ee
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoMerge tag 'block-5.9-2020-08-14' of git://git.kernel.dk/linux-block
Linus Torvalds [Sun, 16 Aug 2020 03:36:42 +0000 (20:36 -0700)]
Merge tag 'block-5.9-2020-08-14' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:
 "A few fixes on the block side of things:

   - Discard granularity fix (Coly)

   - rnbd cleanups (Guoqing)

   - md error handling fix (Dan)

   - md sysfs fix (Junxiao)

   - Fix flush request accounting, which caused an IO slowdown for some
     configurations (Ming)

   - Properly propagate loop flag for partition scanning (Lennart)"

* tag 'block-5.9-2020-08-14' of git://git.kernel.dk/linux-block:
  block: fix double account of flush request's driver tag
  loop: unset GENHD_FL_NO_PART_SCAN on LOOP_CONFIGURE
  rnbd: no need to set bi_end_io in rnbd_bio_map_kern
  rnbd: remove rnbd_dev_submit_io
  md-cluster: Fix potential error pointer dereference in resize_bitmaps()
  block: check queue's limits.discard_granularity in __blkdev_issue_discard()
  md: get sysfs entry after redundancy attr group create

4 years agoMerge tag 'riscv-for-linus-5.9-mw1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 16 Aug 2020 01:54:42 +0000 (18:54 -0700)]
Merge tag 'riscv-for-linus-5.9-mw1' of git://git./linux/kernel/git/riscv/linux

Pull RISC-V fix from Palmer Dabbelt:
 "I collected a single fix during the merge window: we managed to break
  the early trap setup on !MMU, this fixes it"

* tag 'riscv-for-linus-5.9-mw1' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
  riscv: Setup exception vector for nommu platform

4 years agoMerge tag 'sh-for-5.9' of git://git.libc.org/linux-sh
Linus Torvalds [Sun, 16 Aug 2020 01:50:32 +0000 (18:50 -0700)]
Merge tag 'sh-for-5.9' of git://git.libc.org/linux-sh

Pull arch/sh updates from Rich Felker:
 "Cleanup, SECCOMP_FILTER support, message printing fixes, and other
  changes to arch/sh"

* tag 'sh-for-5.9' of git://git.libc.org/linux-sh: (34 commits)
  sh: landisk: Add missing initialization of sh_io_port_base
  sh: bring syscall_set_return_value in line with other architectures
  sh: Add SECCOMP_FILTER
  sh: Rearrange blocks in entry-common.S
  sh: switch to copy_thread_tls()
  sh: use the generic dma coherent remap allocator
  sh: don't allow non-coherent DMA for NOMMU
  dma-mapping: consolidate the NO_DMA definition in kernel/dma/Kconfig
  sh: unexport register_trapped_io and match_trapped_io_handler
  sh: don't include <asm/io_trapped.h> in <asm/io.h>
  sh: move the ioremap implementation out of line
  sh: move ioremap_fixed details out of <asm/io.h>
  sh: remove __KERNEL__ ifdefs from non-UAPI headers
  sh: sort the selects for SUPERH alphabetically
  sh: remove -Werror from Makefiles
  sh: Replace HTTP links with HTTPS ones
  arch/sh/configs: remove obsolete CONFIG_SOC_CAMERA*
  sh: stacktrace: Remove stacktrace_ops.stack()
  sh: machvec: Modernize printing of kernel messages
  sh: pci: Modernize printing of kernel messages
  ...

4 years agoio_uring: short circuit -EAGAIN for blocking read attempt
Jens Axboe [Sat, 15 Aug 2020 22:58:42 +0000 (15:58 -0700)]
io_uring: short circuit -EAGAIN for blocking read attempt

One case was missed in the short IO retry handling, and that's hitting
-EAGAIN on a blocking attempt read (eg from io-wq context). This is a
problem on sockets that are marked as non-blocking when created, they
don't carry any REQ_F_NOWAIT information to help us terminate them
instead of perpetually retrying.

Fixes: 227c0c9673d8 ("io_uring: internally retry short reads")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoio_uring: sanitize double poll handling
Jens Axboe [Sat, 15 Aug 2020 18:44:50 +0000 (11:44 -0700)]
io_uring: sanitize double poll handling

There's a bit of confusion on the matching pairs of poll vs double poll,
depending on if the request is a pure poll (IORING_OP_POLL_ADD) or
poll driven retry.

Add io_poll_get_double() that returns the double poll waitqueue, if any,
and io_poll_get_single() that returns the original poll waitqueue. With
that, remove the argument to io_poll_remove_double().

Finally ensure that wait->private is cleared once the double poll handler
has run, so that remove knows it's already been seen.

Cc: stable@vger.kernel.org # v5.8
Reported-by: syzbot+7f617d4a9369028b8a2c@syzkaller.appspotmail.com
Fixes: 18bceab101ad ("io_uring: allow POLL_ADD with double poll_wait() users")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
4 years agoMerge tag 'perf-tools-2020-08-14' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 15 Aug 2020 18:17:15 +0000 (11:17 -0700)]
Merge tag 'perf-tools-2020-08-14' of git://git./linux/kernel/git/acme/linux

Pull more perf tools updates from Arnaldo Carvalho de Melo:
 "Fixes:
   - Fixes for 'perf bench numa'.

   - Always memset source before memcpy in 'perf bench mem'.

   - Quote CC and CXX for their arguments to fix build in environments
     using those variables to pass more than just the compiler names.

   - Fix module symbol processing, addressing regression detected via
     "perf test".

   - Allow multiple probes in record+script_probe_vfs_getname.sh 'perf
     test' entry.

  Improvements:
   - Add script to autogenerate socket family name id->string table from
     copy of kernel header, used so far in 'perf trace'.

   - 'perf ftrace' improvements to provide similar options for this
     utility so that one can go from 'perf record', 'perf trace', etc to
     'perf ftrace' just by changing the name of the subcommand.

   - Prefer new "sched:sched_waking" trace event when it exists in 'perf
     sched' post processing.

   - Update POWER9 metrics to utilize other metrics.

   - Fall back to querying debuginfod if debuginfo not found locally.

  Miscellaneous:
   - Sync various kvm headers with kernel sources"

* tag 'perf-tools-2020-08-14' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux: (40 commits)
  perf ftrace: Make option description initials all capital letters
  perf build-ids: Fall back to debuginfod query if debuginfo not found
  perf bench numa: Remove dead code in parse_nodes_opt()
  perf stat: Update POWER9 metrics to utilize other metrics
  perf ftrace: Add change log
  perf: ftrace: Add set_tracing_options() to set all trace options
  perf ftrace: Add option --tid to filter by thread id
  perf ftrace: Add option -D/--delay to delay tracing
  perf: ftrace: Allow set graph depth by '--graph-opts'
  perf ftrace: Add support for trace option tracing_thresh
  perf ftrace: Add option 'verbose' to show more info for graph tracer
  perf ftrace: Add support for tracing option 'irq-info'
  perf ftrace: Add support for trace option funcgraph-irqs
  perf ftrace: Add support for trace option sleep-time
  perf ftrace: Add support for tracing option 'func_stack_trace'
  perf tools: Add general function to parse sublevel options
  perf ftrace: Add option '--inherit' to trace children processes
  perf ftrace: Show trace column header
  perf ftrace: Add option '-m/--buffer-size' to set per-cpu buffer size
  perf ftrace: Factor out function write_tracing_file_int()
  ...

4 years agoMerge tag 'x86-urgent-2020-08-15' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 15 Aug 2020 17:38:03 +0000 (10:38 -0700)]
Merge tag 'x86-urgent-2020-08-15' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar:
 "Misc fixes and small updates all around the place:

   - Fix mitigation state sysfs output

   - Fix an FPU xstate/sxave code assumption bug triggered by
     Architectural LBR support

   - Fix Lightning Mountain SoC TSC frequency enumeration bug

   - Fix kexec debug output

   - Fix kexec memory range assumption bug

   - Fix a boundary condition in the crash kernel code

   - Optimize porgatory.ro generation a bit

   - Enable ACRN guests to use X2APIC mode

   - Reduce a __text_poke() IRQs-off critical section for the benefit of
     PREEMPT_RT"

* tag 'x86-urgent-2020-08-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/alternatives: Acquire pte lock with interrupts enabled
  x86/bugs/multihit: Fix mitigation reporting when VMX is not in use
  x86/fpu/xstate: Fix an xstate size check warning with architectural LBRs
  x86/purgatory: Don't generate debug info for purgatory.ro
  x86/tsr: Fix tsc frequency enumeration bug on Lightning Mountain SoC
  kexec_file: Correctly output debugging information for the PT_LOAD ELF header
  kexec: Improve & fix crash_exclude_mem_range() to handle overlapping ranges
  x86/crash: Correct the address boundary of function parameters
  x86/acrn: Remove redundant chars from ACRN signature
  x86/acrn: Allow ACRN guest to use X2APIC mode

4 years agoMerge tag 'sched-urgent-2020-08-15' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 15 Aug 2020 17:36:40 +0000 (10:36 -0700)]
Merge tag 'sched-urgent-2020-08-15' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Ingo Molnar:
 "Two fixes: fix a new tracepoint's output value, and fix the formatting
  of show-state syslog printouts"

* tag 'sched-urgent-2020-08-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/debug: Fix the alignment of the show-state debug output
  sched: Fix use of count for nr_running tracepoint

4 years agoMerge tag 'perf-urgent-2020-08-15' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 15 Aug 2020 17:34:24 +0000 (10:34 -0700)]
Merge tag 'perf-urgent-2020-08-15' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar:
 "Misc fixes, an expansion of perf syscall access to CAP_PERFMON
  privileged tools, plus a RAPL HW-enablement for Intel SPR platforms"

* tag 'perf-urgent-2020-08-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/rapl: Add support for Intel SPR platform
  perf/x86/rapl: Support multiple RAPL unit quirks
  perf/x86/rapl: Fix missing psys sysfs attributes
  hw_breakpoint: Remove unused __register_perf_hw_breakpoint() declaration
  kprobes: Remove show_registers() function prototype
  perf/core: Take over CAP_SYS_PTRACE creds to CAP_PERFMON capability

4 years agoMerge tag 'locking-urgent-2020-08-15' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 15 Aug 2020 17:32:18 +0000 (10:32 -0700)]
Merge tag 'locking-urgent-2020-08-15' of git://git./linux/kernel/git/tip/tip

Pull locking fixlets from Ingo Molnar:
 "A documentation fix and a 'fallthrough' macro update"

* tag 'locking-urgent-2020-08-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  futex: Convert to use the preferred 'fallthrough' macro
  Documentation/locking/locktypes: Fix a typo

4 years agoMerge tag '9p-for-5.9-rc1' of git://github.com/martinetd/linux
Linus Torvalds [Sat, 15 Aug 2020 15:34:36 +0000 (08:34 -0700)]
Merge tag '9p-for-5.9-rc1' of git://github.com/martinetd/linux

Pull 9p updates from Dominique Martinet:

 - some code cleanup

 - a couple of static analysis fixes

 - setattr: try to pick a fid associated with the file rather than the
   dentry, which might sometimes matter

* tag '9p-for-5.9-rc1' of git://github.com/martinetd/linux:
  9p: Remove unneeded cast from memory allocation
  9p: remove unused code in 9p
  net/9p: Fix sparse endian warning in trans_fd.c
  9p: Fix memory leak in v9fs_mount
  9p: retrieve fid from file when file instance exist.

4 years agoMerge tag '5.9-rc-smb3-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sat, 15 Aug 2020 15:31:39 +0000 (08:31 -0700)]
Merge tag '5.9-rc-smb3-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs fixes from Steve French:
 "Three small cifs/smb3 fixes, one for stable fixing mkdir path with
  the 'idsfromsid' mount option"

* tag '5.9-rc-smb3-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6:
  SMB3: Fix mkdir when idsfromsid configured on mount
  cifs: Convert to use the fallthrough macro
  cifs: Fix an error pointer dereference in cifs_mount()

4 years agoMerge tag 'nfs-for-5.9-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Sat, 15 Aug 2020 15:26:55 +0000 (08:26 -0700)]
Merge tag 'nfs-for-5.9-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Stable fixes:
   - pNFS: Don't return layout segments that are being used for I/O
   - pNFS: Don't move layout segments off the active list when being used for I/O

  Features:
   - NFS: Add support for user xattrs through the NFSv4.2 protocol
   - NFS: Allow applications to speed up readdir+statx() using AT_STATX_DONT_SYNC
   - NFSv4.0 allow nconnect for v4.0

  Bugfixes and cleanups:
   - nfs: ensure correct writeback errors are returned on close()
   - nfs: nfs_file_write() should check for writeback errors
   - nfs: Fix getxattr kernel panic and memory overflow
   - NFS: Fix the pNFS/flexfiles mirrored read failover code
   - SUNRPC: dont update timeout value on connection reset
   - freezer: Add unsafe versions of freezable_schedule_timeout_interruptible for NFS
   - sunrpc: destroy rpc_inode_cachep after unregister_filesystem"

* tag 'nfs-for-5.9-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (32 commits)
  NFS: Fix flexfiles read failover
  fs: nfs: delete repeated words in comments
  rpc_pipefs: convert comma to semicolon
  nfs: Fix getxattr kernel panic and memory overflow
  NFS: Don't return layout segments that are in use
  NFS: Don't move layouts to plh_return_segs list while in use
  NFS: Add layout segment info to pnfs read/write/commit tracepoints
  NFS: Add tracepoints for layouterror and layoutstats.
  NFS: Report the stateid + status in trace_nfs4_layoutreturn_on_close()
  SUNRPC dont update timeout value on connection reset
  nfs: nfs_file_write() should check for writeback errors
  nfs: ensure correct writeback errors are returned on close()
  NFSv4.2: xattr cache: get rid of cache discard work queue
  NFS: remove redundant initialization of variable result
  NFSv4.0 allow nconnect for v4.0
  freezer: Add unsafe versions of freezable_schedule_timeout_interruptible for NFS
  sunrpc: destroy rpc_inode_cachep after unregister_filesystem
  NFSv4.2: add client side xattr caching.
  NFSv4.2: hook in the user extended attribute handlers
  NFSv4.2: add the extended attribute proc functions.
  ...

4 years agoMerge tag 'edac_updates_for_5.9_pt2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 15 Aug 2020 15:25:41 +0000 (08:25 -0700)]
Merge tag 'edac_updates_for_5.9_pt2' of git://git./linux/kernel/git/ras/ras

Pull edac fix from Tony Luck:
 "Fix for the ie31200 driver that missed the first pull"

* tag 'edac_updates_for_5.9_pt2' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras:
  EDAC/ie31200: Fallback if host bridge device is already initialized

4 years agoMerge tag 'devicetree-fixes-for-5.9' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 15 Aug 2020 15:19:58 +0000 (08:19 -0700)]
Merge tag 'devicetree-fixes-for-5.9' of git://git./linux/kernel/git/robh/linux

Pull devicetree fixes from Rob Herring:
 "Another round of 'allOf' removals and whitespace clean-ups of schemas"

* tag 'devicetree-fixes-for-5.9' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  dt-bindings: Remove more cases of 'allOf' containing a '$ref'
  dt-bindings: Whitespace clean-ups in schema files

4 years agoMerge tag 'acpi-5.9-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sat, 15 Aug 2020 15:18:22 +0000 (08:18 -0700)]
Merge tag 'acpi-5.9-rc1-2' of git://git./linux/kernel/git/rafael/linux-pm

Pull more ACPI updates from Rafael Wysocki:
 "Add new hardware support to the ACPI driver for AMD SoCs, the x86 clk
  driver and the Designware i2c driver (changes from Akshu Agrawal and
  Pu Wen)"

* tag 'acpi-5.9-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  clk: x86: Support RV architecture
  ACPI: APD: Add a fmw property is_raven
  clk: x86: Change name from ST to FCH
  ACPI: APD: Change name from ST to FCH
  i2c: designware: Add device HID for Hygon I2C controller

4 years agoMerge tag 'pm-5.9-rc1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sat, 15 Aug 2020 15:17:01 +0000 (08:17 -0700)]
Merge tag 'pm-5.9-rc1-3' of git://git./linux/kernel/git/rafael/linux-pm

Pull one more power management update from Rafael Wysocki:
 "Modify the intel_pstate driver to allow it to work in the passive mode
  with hardware-managed P-states (HWP) enabled"

* tag 'pm-5.9-rc1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  cpufreq: intel_pstate: Implement passive mode with HWP enabled

4 years agoMerge tag 'mfd-next-5.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd
Linus Torvalds [Sat, 15 Aug 2020 15:09:38 +0000 (08:09 -0700)]
Merge tag 'mfd-next-5.9-1' of git://git./linux/kernel/git/lee/mfd

Pull MFD updates from Lee Jones:
 "Core Frameworks
   - Make better attempt at matching device with the correct OF node
   - Allow batch removal of hierarchical sub-devices

  New Drivers
   - Add STM32 Clocksource driver
   - Add support for Khadas System Control Microcontroller

  Driver Removal
   - Remove unused driver for TI's SMSC ECE1099

  New Device Support
   - Add support for Intel Emmitsburg PCH to Intel LPSS PCI
   - Add support for Intel Tiger Lake PCH-H to Intel LPSS PCI
   - Add support for Dialog DA revision to Dialog DA9063

  New Functionality
   - Add support for AXP803 to be probed by I2C

  Fix-ups
   - Numerous W=1 warning fixes
   - Device Tree changes (stm32-lptimer, gateworks-gsc, khadas,mcu, stmfx, cros-ec, j721e-system-controller)
   - Enabled Regmap 'fast I/O' in stm32-lptimer
   - Change BUG_ON to WARN_ON in arizona-core
   - Remove superfluous code/initialisation (madera, max14577)
   - Trivial formatting/spelling issues (madera-core, madera-i2c, da9055, max77693-private)
   - Switch to of_platform_populate() in sprd-sc27xx-spi
   - Expand out set/get brightness/pwm macros in lm3533-ctrlbank
   - Disable IRQs on suspend in motorola-cpcap
   - Clean-up error handling in intel_soc_pmic_mrfld
   - Ensure correct removal order of sub-devices in madera
   - Many s/HTTP/HTTPS/ link changes
   - Ensure name used with Regmap is unique in syscon

  Bug Fixes
   - Properly 'put' clock on unbind and error in arizona-core
   - Fix revision handling in da9063
   - Fix 'assignment of read-only location' error in kempld-core
   - Avoid using the Regmap API when atomic in rn5t618
   - Redefine volatile register description in rn5t618
   - Use locking to protect event handler in dln2"

* tag 'mfd-next-5.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd: (76 commits)
  mfd: syscon: Use a unique name with regmap_config
  mfd: Replace HTTP links with HTTPS ones
  mfd: dln2: Run event handler loop under spinlock
  mfd: madera: Improve handling of regulator unbinding
  mfd: mfd-core: Add mechanism for removal of a subset of children
  mfd: intel_soc_pmic_mrfld: Simplify the return expression of intel_scu_ipc_dev_iowrite8()
  mfd: max14577: Remove redundant initialization of variable current_bits
  mfd: rn5t618: Fix caching of battery related registers
  mfd: max77693-private: Drop a duplicated word
  mfd: da9055: pdata.h: Drop a duplicated word
  mfd: rn5t618: Make restart handler atomic safe
  mfd: kempld-core: Fix 'assignment of read-only location' error
  mfd: axp20x: Allow the AXP803 to be probed by I2C
  mfd: da9063: Add support for latest DA silicon revision
  mfd: da9063: Fix revision handling to correctly select reg tables
  dt-bindings: mfd: st,stmfx: Remove I2C unit name
  dt-bindings: mfd: ti,j721e-system-controller.yaml: Add J721e system controller
  mfd: motorola-cpcap: Disable interrupt for suspend
  mfd: smsc-ece1099: Remove driver
  mfd: core: Add OF_MFD_CELL_REG() helper
  ...

4 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Sat, 15 Aug 2020 15:02:03 +0000 (08:02 -0700)]
Merge branch 'akpm' (patches from Andrew)

Merge more updates from Andrew Morton:
 "Subsystems affected by this patch series: mm/hotfixes, lz4, exec,
  mailmap, mm/thp, autofs, sysctl, mm/kmemleak, mm/misc and lib"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (35 commits)
  virtio: pci: constify ioreadX() iomem argument (as in generic implementation)
  ntb: intel: constify ioreadX() iomem argument (as in generic implementation)
  rtl818x: constify ioreadX() iomem argument (as in generic implementation)
  iomap: constify ioreadX() iomem argument (as in generic implementation)
  sh: use generic strncpy()
  sh: clkfwk: remove r8/r16/r32
  include/asm-generic/vmlinux.lds.h: align ro_after_init
  mm: annotate a data race in page_zonenum()
  mm/swap.c: annotate data races for lru_rotate_pvecs
  mm/rmap: annotate a data race at tlb_flush_batched
  mm/mempool: fix a data race in mempool_free()
  mm/list_lru: fix a data race in list_lru_count_one
  mm/memcontrol: fix a data race in scan count
  mm/page_counter: fix various data races at memsw
  mm/swapfile: fix and annotate various data races
  mm/filemap.c: fix a data race in filemap_fault()
  mm/swap_state: mark various intentional data races
  mm/page_io: mark various intentional data races
  mm/frontswap: mark various intentional data races
  mm/kmemleak: silence KCSAN splats in checksum
  ...

4 years agovirtio: pci: constify ioreadX() iomem argument (as in generic implementation)
Krzysztof Kozlowski [Sat, 15 Aug 2020 00:32:20 +0000 (17:32 -0700)]
virtio: pci: constify ioreadX() iomem argument (as in generic implementation)

The ioreadX() helpers have inconsistent interface.  On some architectures
void *__iomem address argument is a pointer to const, on some not.

Implementations of ioreadX() do not modify the memory under the address so
they can be converted to a "const" version for const-safety and
consistency among architectures.

Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: Allen Hubbe <allenbh@gmail.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Dave Jiang <dave.jiang@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Helge Deller <deller@gmx.de>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Jon Mason <jdmason@kudzu.us>
Cc: Kalle Valo <kvalo@codeaurora.org>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Rich Felker <dalias@libc.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Link: http://lkml.kernel.org/r/20200709072837.5869-5-krzk@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agontb: intel: constify ioreadX() iomem argument (as in generic implementation)
Krzysztof Kozlowski [Sat, 15 Aug 2020 00:32:15 +0000 (17:32 -0700)]
ntb: intel: constify ioreadX() iomem argument (as in generic implementation)

The ioreadX() helpers have inconsistent interface.  On some architectures
void *__iomem address argument is a pointer to const, on some not.

Implementations of ioreadX() do not modify the memory under the address so
they can be converted to a "const" version for const-safety and
consistency among architectures.

Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Cc: Allen Hubbe <allenbh@gmail.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Helge Deller <deller@gmx.de>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Jon Mason <jdmason@kudzu.us>
Cc: Kalle Valo <kvalo@codeaurora.org>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Rich Felker <dalias@libc.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Link: http://lkml.kernel.org/r/20200709072837.5869-4-krzk@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agortl818x: constify ioreadX() iomem argument (as in generic implementation)
Krzysztof Kozlowski [Sat, 15 Aug 2020 00:32:11 +0000 (17:32 -0700)]
rtl818x: constify ioreadX() iomem argument (as in generic implementation)

The ioreadX() helpers have inconsistent interface.  On some architectures
void *__iomem address argument is a pointer to const, on some not.

Implementations of ioreadX() do not modify the memory under the address so
they can be converted to a "const" version for const-safety and
consistency among architectures.

Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Kalle Valo <kvalo@codeaurora.org>
Cc: Allen Hubbe <allenbh@gmail.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Dave Jiang <dave.jiang@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Helge Deller <deller@gmx.de>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Jon Mason <jdmason@kudzu.us>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Rich Felker <dalias@libc.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Link: http://lkml.kernel.org/r/20200709072837.5869-3-krzk@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoiomap: constify ioreadX() iomem argument (as in generic implementation)
Krzysztof Kozlowski [Sat, 15 Aug 2020 00:32:07 +0000 (17:32 -0700)]
iomap: constify ioreadX() iomem argument (as in generic implementation)

Patch series "iomap: Constify ioreadX() iomem argument", v3.

The ioread8/16/32() and others have inconsistent interface among the
architectures: some taking address as const, some not.

It seems there is nothing really stopping all of them to take pointer to
const.

This patch (of 4):

The ioreadX() and ioreadX_rep() helpers have inconsistent interface.  On
some architectures void *__iomem address argument is a pointer to const,
on some not.

Implementations of ioreadX() do not modify the memory under the address so
they can be converted to a "const" version for const-safety and
consistency among architectures.

[krzk@kernel.org: sh: clk: fix assignment from incompatible pointer type for ioreadX()]
Link: http://lkml.kernel.org/r/20200723082017.24053-1-krzk@kernel.org
[akpm@linux-foundation.org: fix drivers/mailbox/bcm-pdc-mailbox.c]
Link: http://lkml.kernel.org/r/202007132209.Rxmv4QyS%25lkp@intel.com
Suggested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Matt Turner <mattst88@gmail.com>
Cc: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>
Cc: Helge Deller <deller@gmx.de>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Rich Felker <dalias@libc.org>
Cc: Kalle Valo <kvalo@codeaurora.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Dave Jiang <dave.jiang@intel.com>
Cc: Jon Mason <jdmason@kudzu.us>
Cc: Allen Hubbe <allenbh@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Link: http://lkml.kernel.org/r/20200709072837.5869-1-krzk@kernel.org
Link: http://lkml.kernel.org/r/20200709072837.5869-2-krzk@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agosh: use generic strncpy()
Kuninori Morimoto [Sat, 15 Aug 2020 00:32:04 +0000 (17:32 -0700)]
sh: use generic strncpy()

Current SH will get below warning at strncpy()

In file included from ${LINUX}/arch/sh/include/asm/string.h:3,
                 from ${LINUX}/include/linux/string.h:20,
                 from ${LINUX}/include/linux/bitmap.h:9,
                 from ${LINUX}/include/linux/nodemask.h:95,
                 from ${LINUX}/include/linux/mmzone.h:17,
                 from ${LINUX}/include/linux/gfp.h:6,
                 from ${LINUX}/innclude/linux/slab.h:15,
                 from ${LINUX}/linux/drivers/mmc/host/vub300.c:38:
${LINUX}/drivers/mmc/host/vub300.c: In function 'new_system_port_status':
${LINUX}/arch/sh/include/asm/string_32.h:51:42: warning: array subscript\
  80 is above array bounds of 'char[26]' [-Warray-bounds]
   : "0" (__dest), "1" (__src), "r" (__src+__n)
                                     ~~~~~^~~~

In general, strncpy() should behave like below.

char dest[10];
char *src = "12345";

strncpy(dest, src, 10);
// dest = {'1', '2', '3', '4', '5',
           '\0','\0','\0','\0','\0'}

But, current SH strnpy() has 2 issues.
1st is it will access to out-of-memory (= src + 10).
2nd is it needs big fixup for it, and maintenance __asm__
code is difficult.

To solve these issues, this patch simply uses generic strncpy()
instead of architecture specific one.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alan Modra <amodra@gmail.com>
Cc: Bin Meng <bin.meng@windriver.com>
Cc: Chen Zhou <chenzhou10@huawei.com>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Cc: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Romain Naour <romain.naour@gmail.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Link: https://marc.info/?l=linux-renesas-soc&m=157664657013309
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agosh: clkfwk: remove r8/r16/r32
Kuninori Morimoto [Sat, 15 Aug 2020 00:32:00 +0000 (17:32 -0700)]
sh: clkfwk: remove r8/r16/r32

SH will get below warning

${LINUX}/drivers/sh/clk/cpg.c: In function 'r8':
${LINUX}/drivers/sh/clk/cpg.c:41:17: warning: passing argument 1 of 'ioread8'
 discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
  return ioread8(addr);
                 ^~~~
In file included from ${LINUX}/arch/sh/include/asm/io.h:21,
                 from ${LINUX}/include/linux/io.h:13,
                 from ${LINUX}/drivers/sh/clk/cpg.c:14:
${LINUX}/include/asm-generic/iomap.h:29:29: note: expected 'void *' but
argument is of type 'const void *'
 extern unsigned int ioread8(void __iomem *);
                             ^~~~~~~~~~~~~~

We don't need "const" for r8/r16/r32.  And we don't need r8/r16/r32
themselvs.  This patch cleanup these.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alan Modra <amodra@gmail.com>
Cc: Bin Meng <bin.meng@windriver.com>
Cc: Chen Zhou <chenzhou10@huawei.com>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Cc: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Romain Naour <romain.naour@gmail.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
X-MARC-Message: https://marc.info/?l=linux-renesas-soc&m=157852973916903
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoinclude/asm-generic/vmlinux.lds.h: align ro_after_init
Romain Naour [Sat, 15 Aug 2020 00:31:57 +0000 (17:31 -0700)]
include/asm-generic/vmlinux.lds.h: align ro_after_init

Since the patch [1], building the kernel using a toolchain built with
binutils 2.33.1 prevents booting a sh4 system under Qemu.  Apply the patch
provided by Alan Modra [2] that fix alignment of rodata.

[1] https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=ebd2263ba9a9124d93bbc0ece63d7e0fae89b40e
[2] https://www.sourceware.org/ml/binutils/2019-12/msg00112.html

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alan Modra <amodra@gmail.com>
Cc: Bin Meng <bin.meng@windriver.com>
Cc: Chen Zhou <chenzhou10@huawei.com>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Cc: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Cc: Rich Felker <dalias@libc.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: <stable@vger.kernel.org>
Link: https://marc.info/?l=linux-sh&m=158429470221261
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: annotate a data race in page_zonenum()
Qian Cai [Sat, 15 Aug 2020 00:31:53 +0000 (17:31 -0700)]
mm: annotate a data race in page_zonenum()

 BUG: KCSAN: data-race in page_cpupid_xchg_last / put_page

 write (marked) to 0xfffffc0d48ec1a00 of 8 bytes by task 91442 on cpu 3:
  page_cpupid_xchg_last+0x51/0x80
  page_cpupid_xchg_last at mm/mmzone.c:109 (discriminator 11)
  wp_page_reuse+0x3e/0xc0
  wp_page_reuse at mm/memory.c:2453
  do_wp_page+0x472/0x7b0
  do_wp_page at mm/memory.c:2798
  __handle_mm_fault+0xcb0/0xd00
  handle_pte_fault at mm/memory.c:4049
  (inlined by) __handle_mm_fault at mm/memory.c:4163
  handle_mm_fault+0xfc/0x2f0
  handle_mm_fault at mm/memory.c:4200
  do_page_fault+0x263/0x6f9
  do_user_addr_fault at arch/x86/mm/fault.c:1465
  (inlined by) do_page_fault at arch/x86/mm/fault.c:1539
  page_fault+0x34/0x40

 read to 0xfffffc0d48ec1a00 of 8 bytes by task 94817 on cpu 69:
  put_page+0x15a/0x1f0
  page_zonenum at include/linux/mm.h:923
  (inlined by) is_zone_device_page at include/linux/mm.h:929
  (inlined by) page_is_devmap_managed at include/linux/mm.h:948
  (inlined by) put_page at include/linux/mm.h:1023
  wp_page_copy+0x571/0x930
  wp_page_copy at mm/memory.c:2615
  do_wp_page+0x107/0x7b0
  __handle_mm_fault+0xcb0/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 69 PID: 94817 Comm: systemd-udevd Tainted: G        W  O L 5.5.0-next-20200204+ #6
 Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019

A page never changes its zone number. The zone number happens to be
stored in the same word as other bits which are modified, but the zone
number bits will never be modified by any other write, so it can accept
a reload of the zone bits after an intervening write and it don't need
to use READ_ONCE(). Thus, annotate this data race using
ASSERT_EXCLUSIVE_BITS() to also assert that there are no concurrent
writes to it.

Suggested-by: Marco Elver <elver@google.com>
Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jan Kara <jack@suse.cz>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Ira Weiny <ira.weiny@intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Link: http://lkml.kernel.org/r/1581619089-14472-1-git-send-email-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/swap.c: annotate data races for lru_rotate_pvecs
Qian Cai [Sat, 15 Aug 2020 00:31:50 +0000 (17:31 -0700)]
mm/swap.c: annotate data races for lru_rotate_pvecs

Read to lru_add_pvec->nr could be interrupted and then write to the same
variable.  The write has local interrupt disabled, but the plain reads
result in data races.  However, it is unlikely the compilers could do much
damage here given that lru_add_pvec->nr is a "unsigned char" and there is
an existing compiler barrier.  Thus, annotate the reads using the
data_race() macro.  The data races were reported by KCSAN,

 BUG: KCSAN: data-race in lru_add_drain_cpu / rotate_reclaimable_page

 write to 0xffff9291ebcb8a40 of 1 bytes by interrupt on cpu 23:
  rotate_reclaimable_page+0x2df/0x490
  pagevec_add at include/linux/pagevec.h:81
  (inlined by) rotate_reclaimable_page at mm/swap.c:259
  end_page_writeback+0x1b5/0x2b0
  end_swap_bio_write+0x1d0/0x280
  bio_endio+0x297/0x560
  dec_pending+0x218/0x430 [dm_mod]
  clone_endio+0xe4/0x2c0 [dm_mod]
  bio_endio+0x297/0x560
  blk_update_request+0x201/0x920
  scsi_end_request+0x6b/0x4a0
  scsi_io_completion+0xb7/0x7e0
  scsi_finish_command+0x1ed/0x2a0
  scsi_softirq_done+0x1c9/0x1d0
  blk_done_softirq+0x181/0x1d0
  __do_softirq+0xd9/0x57c
  irq_exit+0xa2/0xc0
  do_IRQ+0x8b/0x190
  ret_from_intr+0x0/0x42
  delay_tsc+0x46/0x80
  __const_udelay+0x3c/0x40
  __udelay+0x10/0x20
  kcsan_setup_watchpoint+0x202/0x3a0
  __tsan_read1+0xc2/0x100
  lru_add_drain_cpu+0xb8/0x3f0
  lru_add_drain+0x25/0x40
  shrink_active_list+0xe1/0xc80
  shrink_lruvec+0x766/0xb70
  shrink_node+0x2d6/0xca0
  do_try_to_free_pages+0x1f7/0x9a0
  try_to_free_pages+0x252/0x5b0
  __alloc_pages_slowpath+0x458/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x16e/0x6f0
  __handle_mm_fault+0xcd5/0xd40
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 read to 0xffff9291ebcb8a40 of 1 bytes by task 37761 on cpu 23:
  lru_add_drain_cpu+0xb8/0x3f0
  lru_add_drain_cpu at mm/swap.c:602
  lru_add_drain+0x25/0x40
  shrink_active_list+0xe1/0xc80
  shrink_lruvec+0x766/0xb70
  shrink_node+0x2d6/0xca0
  do_try_to_free_pages+0x1f7/0x9a0
  try_to_free_pages+0x252/0x5b0
  __alloc_pages_slowpath+0x458/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x16e/0x6f0
  __handle_mm_fault+0xcd5/0xd40
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 2 locks held by oom02/37761:
  #0: ffff9281e5928808 (&mm->mmap_sem#2){++++}, at: do_page_fault
  #1: ffffffffb3ade380 (fs_reclaim){+.+.}, at: fs_reclaim_acquire.part
 irq event stamp: 1949217
 trace_hardirqs_on_thunk+0x1a/0x1c
 __do_softirq+0x2e7/0x57c
 __do_softirq+0x34c/0x57c
 irq_exit+0xa2/0xc0

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 23 PID: 37761 Comm: oom02 Not tainted 5.6.0-rc3-next-20200226+ #6
 Hardware name: HP ProLiant BL660c Gen9, BIOS I38 10/17/2018

Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Marco Elver <elver@google.com>
Link: http://lkml.kernel.org/r/20200228044018.1263-1-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/rmap: annotate a data race at tlb_flush_batched
Qian Cai [Sat, 15 Aug 2020 00:31:47 +0000 (17:31 -0700)]
mm/rmap: annotate a data race at tlb_flush_batched

mm->tlb_flush_batched could be accessed concurrently as noticed by
KCSAN,

 BUG: KCSAN: data-race in flush_tlb_batched_pending / try_to_unmap_one

 write to 0xffff93f754880bd0 of 1 bytes by task 822 on cpu 6:
  try_to_unmap_one+0x59a/0x1ab0
  set_tlb_ubc_flush_pending at mm/rmap.c:635
  (inlined by) try_to_unmap_one at mm/rmap.c:1538
  rmap_walk_anon+0x296/0x650
  rmap_walk+0xdf/0x100
  try_to_unmap+0x18a/0x2f0
  shrink_page_list+0xef6/0x2870
  shrink_inactive_list+0x316/0x880
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  balance_pgdat+0x652/0xd90
  kswapd+0x396/0x8d0
  kthread+0x1e0/0x200
  ret_from_fork+0x27/0x50

 read to 0xffff93f754880bd0 of 1 bytes by task 6364 on cpu 4:
  flush_tlb_batched_pending+0x29/0x90
  flush_tlb_batched_pending at mm/rmap.c:682
  change_p4d_range+0x5dd/0x1030
  change_pte_range at mm/mprotect.c:44
  (inlined by) change_pmd_range at mm/mprotect.c:212
  (inlined by) change_pud_range at mm/mprotect.c:240
  (inlined by) change_p4d_range at mm/mprotect.c:260
  change_protection+0x222/0x310
  change_prot_numa+0x3e/0x60
  task_numa_work+0x219/0x350
  task_work_run+0xed/0x140
  prepare_exit_to_usermode+0x2cc/0x2e0
  ret_from_intr+0x32/0x42

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 4 PID: 6364 Comm: mtest01 Tainted: G        W    L 5.5.0-next-20200210+ #5
 Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019

flush_tlb_batched_pending() is under PTL but the write is not, but
mm->tlb_flush_batched is only a bool type, so the value is unlikely to be
shattered.  Thus, mark it as an intentional data race by using the data
race macro.

Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marco Elver <elver@google.com>
Link: http://lkml.kernel.org/r/1581450783-8262-1-git-send-email-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/mempool: fix a data race in mempool_free()
Qian Cai [Sat, 15 Aug 2020 00:31:44 +0000 (17:31 -0700)]
mm/mempool: fix a data race in mempool_free()

mempool_t pool.curr_nr could be accessed concurrently as noticed by
KCSAN,

 BUG: KCSAN: data-race in mempool_free / remove_element

 write to 0xffffffffa937638c of 4 bytes by task 6359 on cpu 113:
  remove_element+0x4a/0x1c0
  remove_element at mm/mempool.c:132
  mempool_alloc+0x102/0x210
  (inlined by) mempool_alloc at mm/mempool.c:399
  bio_alloc_bioset+0x106/0x2c0
  get_swap_bio+0x49/0x230
  __swap_writepage+0x680/0xc30
  swap_writepage+0x9c/0xf0
  pageout+0x33e/0xae0
  shrink_page_list+0x1f57/0x2870
  shrink_inactive_list+0x316/0x880
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290
  <snip>

 read to 0xffffffffa937638c of 4 bytes by interrupt on cpu 64:
  mempool_free+0x3e/0x150
  mempool_free at mm/mempool.c:492
  bio_free+0x192/0x280
  bio_put+0x91/0xd0
  end_swap_bio_write+0x1d8/0x280
  bio_endio+0x2c2/0x5b0
  dec_pending+0x22b/0x440 [dm_mod]
  clone_endio+0xe4/0x2c0 [dm_mod]
  bio_endio+0x2c2/0x5b0
  blk_update_request+0x217/0x940
  scsi_end_request+0x6b/0x4d0
  scsi_io_completion+0xb7/0x7e0
  scsi_finish_command+0x223/0x310
  scsi_softirq_done+0x1d5/0x210
  blk_mq_complete_request+0x224/0x250
  scsi_mq_done+0xc2/0x250
  pqi_raid_io_complete+0x5a/0x70 [smartpqi]
  pqi_irq_handler+0x150/0x1410 [smartpqi]
  __handle_irq_event_percpu+0x90/0x540
  handle_irq_event_percpu+0x49/0xd0
  handle_irq_event+0x85/0xca
  handle_edge_irq+0x13f/0x3e0
  do_IRQ+0x86/0x190
  <snip>

Since the write is under pool->lock but the read is done as lockless.
Even though the commit 5b990546e334 ("mempool: fix and document
synchronization and memory barrier usage") introduced the smp_wmb() and
smp_rmb() pair to improve the situation, it is adequate to protect it
from data races which could lead to a logic bug, so fix it by adding
READ_ONCE() for the read.

Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marco Elver <elver@google.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Link: http://lkml.kernel.org/r/1581446384-2131-1-git-send-email-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/list_lru: fix a data race in list_lru_count_one
Qian Cai [Sat, 15 Aug 2020 00:31:41 +0000 (17:31 -0700)]
mm/list_lru: fix a data race in list_lru_count_one

struct list_lru_one l.nr_items could be accessed concurrently as noticed
by KCSAN,

 BUG: KCSAN: data-race in list_lru_count_one / list_lru_isolate_move

 write to 0xffffa102789c4510 of 8 bytes by task 823 on cpu 39:
  list_lru_isolate_move+0xf9/0x130
  list_lru_isolate_move at mm/list_lru.c:180
  inode_lru_isolate+0x12b/0x2a0
  __list_lru_walk_one+0x122/0x3d0
  list_lru_walk_one+0x75/0xa0
  prune_icache_sb+0x8b/0xc0
  super_cache_scan+0x1b8/0x250
  do_shrink_slab+0x256/0x6d0
  shrink_slab+0x41b/0x4a0
  shrink_node+0x35c/0xd80
  balance_pgdat+0x652/0xd90
  kswapd+0x396/0x8d0
  kthread+0x1e0/0x200
  ret_from_fork+0x27/0x50

 read to 0xffffa102789c4510 of 8 bytes by task 6345 on cpu 56:
  list_lru_count_one+0x116/0x2f0
  list_lru_count_one at mm/list_lru.c:193
  super_cache_count+0xe8/0x170
  do_shrink_slab+0x95/0x6d0
  shrink_slab+0x41b/0x4a0
  shrink_node+0x35c/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x170/0x700
  __handle_mm_fault+0xc9f/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 56 PID: 6345 Comm: oom01 Tainted: G        W    L 5.5.0-next-20200205+ #4
 Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019

A shattered l.nr_items could affect the shrinker behaviour due to a data
race. Fix it by adding READ_ONCE() for the read. Since the writes are
aligned and up to word-size, assume those are safe from data races to
avoid readability issues of writing WRITE_ONCE(var, var + val).

Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marco Elver <elver@google.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Link: http://lkml.kernel.org/r/1581114679-5488-1-git-send-email-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/memcontrol: fix a data race in scan count
Qian Cai [Sat, 15 Aug 2020 00:31:37 +0000 (17:31 -0700)]
mm/memcontrol: fix a data race in scan count

struct mem_cgroup_per_node mz.lru_zone_size[zone_idx][lru] could be
accessed concurrently as noticed by KCSAN,

 BUG: KCSAN: data-race in lruvec_lru_size / mem_cgroup_update_lru_size

 write to 0xffff9c804ca285f8 of 8 bytes by task 50951 on cpu 12:
  mem_cgroup_update_lru_size+0x11c/0x1d0
  mem_cgroup_update_lru_size at mm/memcontrol.c:1266
  isolate_lru_pages+0x6a9/0xf30
  shrink_active_list+0x123/0xcc0
  shrink_lruvec+0x8fd/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x170/0x700
  __handle_mm_fault+0xc9f/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 read to 0xffff9c804ca285f8 of 8 bytes by task 50964 on cpu 95:
  lruvec_lru_size+0xbb/0x270
  mem_cgroup_get_zone_lru_size at include/linux/memcontrol.h:536
  (inlined by) lruvec_lru_size at mm/vmscan.c:326
  shrink_lruvec+0x1d0/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_current+0xa6/0x120
  alloc_slab_page+0x3b1/0x540
  allocate_slab+0x70/0x660
  new_slab+0x46/0x70
  ___slab_alloc+0x4ad/0x7d0
  __slab_alloc+0x43/0x70
  kmem_cache_alloc+0x2c3/0x420
  getname_flags+0x4c/0x230
  getname+0x22/0x30
  do_sys_openat2+0x205/0x3b0
  do_sys_open+0x9a/0xf0
  __x64_sys_openat+0x62/0x80
  do_syscall_64+0x91/0xb47
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 95 PID: 50964 Comm: cc1 Tainted: G        W  O L    5.5.0-next-20200204+ #6
 Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019

The write is under lru_lock, but the read is done as lockless.  The scan
count is used to determine how aggressively the anon and file LRU lists
should be scanned.  Load tearing could generate an inefficient heuristic,
so fix it by adding READ_ONCE() for the read.

Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Link: http://lkml.kernel.org/r/20200206034945.2481-1-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/page_counter: fix various data races at memsw
Qian Cai [Sat, 15 Aug 2020 00:31:34 +0000 (17:31 -0700)]
mm/page_counter: fix various data races at memsw

Commit 3e32cb2e0a12 ("mm: memcontrol: lockless page counters") could had
memcg->memsw->watermark and memcg->memsw->failcnt been accessed
concurrently as reported by KCSAN,

 BUG: KCSAN: data-race in page_counter_try_charge / page_counter_try_charge

 read to 0xffff8fb18c4cd190 of 8 bytes by task 1081 on cpu 59:
  page_counter_try_charge+0x4d/0x150 mm/page_counter.c:138
  try_charge+0x131/0xd50 mm/memcontrol.c:2405
  __memcg_kmem_charge_memcg+0x58/0x140
  __memcg_kmem_charge+0xcc/0x280
  __alloc_pages_nodemask+0x1e1/0x450
  alloc_pages_current+0xa6/0x120
  pte_alloc_one+0x17/0xd0
  __pte_alloc+0x3a/0x1f0
  copy_p4d_range+0xc36/0x1990
  copy_page_range+0x21d/0x360
  dup_mmap+0x5f5/0x7a0
  dup_mm+0xa2/0x240
  copy_process+0x1b3f/0x3460
  _do_fork+0xaa/0xa20
  __x64_sys_clone+0x13b/0x170
  do_syscall_64+0x91/0xb47
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

 write to 0xffff8fb18c4cd190 of 8 bytes by task 1153 on cpu 120:
  page_counter_try_charge+0x5b/0x150 mm/page_counter.c:139
  try_charge+0x131/0xd50 mm/memcontrol.c:2405
  mem_cgroup_try_charge+0x159/0x460
  mem_cgroup_try_charge_delay+0x3d/0xa0
  wp_page_copy+0x14d/0x930
  do_wp_page+0x107/0x7b0
  __handle_mm_fault+0xce6/0xd40
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 BUG: KCSAN: data-race in page_counter_try_charge / page_counter_try_charge

 write to 0xffff88809bbf2158 of 8 bytes by task 11782 on cpu 0:
  page_counter_try_charge+0x100/0x170 mm/page_counter.c:129
  try_charge+0x185/0xbf0 mm/memcontrol.c:2405
  __memcg_kmem_charge_memcg+0x4a/0xe0 mm/memcontrol.c:2837
  __memcg_kmem_charge+0xcf/0x1b0 mm/memcontrol.c:2877
  __alloc_pages_nodemask+0x26c/0x310 mm/page_alloc.c:4780

 read to 0xffff88809bbf2158 of 8 bytes by task 11814 on cpu 1:
  page_counter_try_charge+0xef/0x170 mm/page_counter.c:129
  try_charge+0x185/0xbf0 mm/memcontrol.c:2405
  __memcg_kmem_charge_memcg+0x4a/0xe0 mm/memcontrol.c:2837
  __memcg_kmem_charge+0xcf/0x1b0 mm/memcontrol.c:2877
  __alloc_pages_nodemask+0x26c/0x310 mm/page_alloc.c:4780

Since watermark could be compared or set to garbage due to a data race
which would change the code logic, fix it by adding a pair of READ_ONCE()
and WRITE_ONCE() in those places.

The "failcnt" counter is tolerant of some degree of inaccuracy and is only
used to report stats, a data race will not be harmful, thus mark it as an
intentional data race using the data_race() macro.

Fixes: 3e32cb2e0a12 ("mm: memcontrol: lockless page counters")
Reported-by: syzbot+f36cfe60b1006a94f9dc@syzkaller.appspotmail.com
Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Marco Elver <elver@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Link: http://lkml.kernel.org/r/1581519682-23594-1-git-send-email-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/swapfile: fix and annotate various data races
Qian Cai [Sat, 15 Aug 2020 00:31:31 +0000 (17:31 -0700)]
mm/swapfile: fix and annotate various data races

swap_info_struct si.highest_bit, si.swap_map[offset] and si.flags could
be accessed concurrently separately as noticed by KCSAN,

=== si.highest_bit ===

 write to 0xffff8d5abccdc4d4 of 4 bytes by task 5353 on cpu 24:
  swap_range_alloc+0x81/0x130
  swap_range_alloc at mm/swapfile.c:681
  scan_swap_map_slots+0x371/0xb90
  get_swap_pages+0x39d/0x5c0
  get_swap_page+0xf2/0x524
  add_to_swap+0xe4/0x1c0
  shrink_page_list+0x1795/0x2870
  shrink_inactive_list+0x316/0x880
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290

 read to 0xffff8d5abccdc4d4 of 4 bytes by task 6672 on cpu 70:
  scan_swap_map_slots+0x4a6/0xb90
  scan_swap_map_slots at mm/swapfile.c:892
  get_swap_pages+0x39d/0x5c0
  get_swap_page+0xf2/0x524
  add_to_swap+0xe4/0x1c0
  shrink_page_list+0x1795/0x2870
  shrink_inactive_list+0x316/0x880
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 70 PID: 6672 Comm: oom01 Tainted: G        W    L 5.5.0-next-20200205+ #3
 Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019

=== si.swap_map[offset] ===

 write to 0xffffbc370c29a64c of 1 bytes by task 6856 on cpu 86:
  __swap_entry_free_locked+0x8c/0x100
  __swap_entry_free_locked at mm/swapfile.c:1209 (discriminator 4)
  __swap_entry_free.constprop.20+0x69/0xb0
  free_swap_and_cache+0x53/0xa0
  unmap_page_range+0x7f8/0x1d70
  unmap_single_vma+0xcd/0x170
  unmap_vmas+0x18b/0x220
  exit_mmap+0xee/0x220
  mmput+0x10e/0x270
  do_exit+0x59b/0xf40
  do_group_exit+0x8b/0x180

 read to 0xffffbc370c29a64c of 1 bytes by task 6855 on cpu 20:
  _swap_info_get+0x81/0xa0
  _swap_info_get at mm/swapfile.c:1140
  free_swap_and_cache+0x40/0xa0
  unmap_page_range+0x7f8/0x1d70
  unmap_single_vma+0xcd/0x170
  unmap_vmas+0x18b/0x220
  exit_mmap+0xee/0x220
  mmput+0x10e/0x270
  do_exit+0x59b/0xf40
  do_group_exit+0x8b/0x180

=== si.flags ===

 write to 0xffff956c8fc6c400 of 8 bytes by task 6087 on cpu 23:
  scan_swap_map_slots+0x6fe/0xb50
  scan_swap_map_slots at mm/swapfile.c:887
  get_swap_pages+0x39d/0x5c0
  get_swap_page+0x377/0x524
  add_to_swap+0xe4/0x1c0
  shrink_page_list+0x1795/0x2870
  shrink_inactive_list+0x316/0x880
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290

 read to 0xffff956c8fc6c400 of 8 bytes by task 6207 on cpu 63:
  _swap_info_get+0x41/0xa0
  __swap_info_get at mm/swapfile.c:1114
  put_swap_page+0x84/0x490
  __remove_mapping+0x384/0x5f0
  shrink_page_list+0xff1/0x2870
  shrink_inactive_list+0x316/0x880
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290

The writes are under si->lock but the reads are not. For si.highest_bit
and si.swap_map[offset], data race could trigger logic bugs, so fix them
by having WRITE_ONCE() for the writes and READ_ONCE() for the reads
except those isolated reads where they compare against zero which a data
race would cause no harm. Thus, annotate them as intentional data races
using the data_race() macro.

For si.flags, the readers are only interested in a single bit where a
data race there would cause no issue there.

[cai@lca.pw: add a missing annotation for si->flags in memory.c]
Link: http://lkml.kernel.org/r/1581612647-5958-1-git-send-email-cai@lca.pw
Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marco Elver <elver@google.com>
Cc: Hugh Dickins <hughd@google.com>
Link: http://lkml.kernel.org/r/1581095163-12198-1-git-send-email-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/filemap.c: fix a data race in filemap_fault()
Kirill A. Shutemov [Sat, 15 Aug 2020 00:31:27 +0000 (17:31 -0700)]
mm/filemap.c: fix a data race in filemap_fault()

struct file_ra_state ra.mmap_miss could be accessed concurrently during
page faults as noticed by KCSAN,

 BUG: KCSAN: data-race in filemap_fault / filemap_map_pages

 write to 0xffff9b1700a2c1b4 of 4 bytes by task 3292 on cpu 30:
  filemap_fault+0x920/0xfc0
  do_sync_mmap_readahead at mm/filemap.c:2384
  (inlined by) filemap_fault at mm/filemap.c:2486
  __xfs_filemap_fault+0x112/0x3e0 [xfs]
  xfs_filemap_fault+0x74/0x90 [xfs]
  __do_fault+0x9e/0x220
  do_fault+0x4a0/0x920
  __handle_mm_fault+0xc69/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 read to 0xffff9b1700a2c1b4 of 4 bytes by task 3313 on cpu 32:
  filemap_map_pages+0xc2e/0xd80
  filemap_map_pages at mm/filemap.c:2625
  do_fault+0x3da/0x920
  __handle_mm_fault+0xc69/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 32 PID: 3313 Comm: systemd-udevd Tainted: G        W    L 5.5.0-next-20200210+ #1
 Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019

ra.mmap_miss is used to contribute the readahead decisions, a data race
could be undesirable.  Both the read and write is only under non-exclusive
mmap_sem, two concurrent writers could even underflow the counter.  Fix
the underflow by writing to a local variable before committing a final
store to ra.mmap_miss given a small inaccuracy of the counter should be
acceptable.

Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: Qian Cai <cai@lca.pw>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Marco Elver <elver@google.com>
Link: http://lkml.kernel.org/r/20200211030134.1847-1-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/swap_state: mark various intentional data races
Qian Cai [Sat, 15 Aug 2020 00:31:24 +0000 (17:31 -0700)]
mm/swap_state: mark various intentional data races

swap_cache_info.* could be accessed concurrently as noticed by
KCSAN,

 BUG: KCSAN: data-race in lookup_swap_cache / lookup_swap_cache

 write to 0xffffffff85517318 of 8 bytes by task 94138 on cpu 101:
  lookup_swap_cache+0x12e/0x460
  lookup_swap_cache at mm/swap_state.c:322
  do_swap_page+0x112/0xeb0
  __handle_mm_fault+0xc7a/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 read to 0xffffffff85517318 of 8 bytes by task 91655 on cpu 100:
  lookup_swap_cache+0x117/0x460
  lookup_swap_cache at mm/swap_state.c:322
  shmem_swapin_page+0xc7/0x9e0
  shmem_getpage_gfp+0x2ca/0x16c0
  shmem_fault+0xef/0x3c0
  __do_fault+0x9e/0x220
  do_fault+0x4a0/0x920
  __handle_mm_fault+0xc69/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 100 PID: 91655 Comm: systemd-journal Tainted: G        W  O L 5.5.0-next-20200204+ #6
 Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019

 write to 0xffffffff8d717308 of 8 bytes by task 11365 on cpu 87:
   __delete_from_swap_cache+0x681/0x8b0
   __delete_from_swap_cache at mm/swap_state.c:178

 read to 0xffffffff8d717308 of 8 bytes by task 11275 on cpu 53:
   __delete_from_swap_cache+0x66e/0x8b0
   __delete_from_swap_cache at mm/swap_state.c:178

Both the read and write are done as lockless. Since swap_cache_info.*
are only used to print out counter information, even if any of them
missed a few incremental due to data races, it will be harmless, so just
mark it as an intentional data race using the data_race() macro.

While at it, fix a checkpatch.pl warning,

WARNING: Single statement macros should not use a do {} while (0) loop

Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marco Elver <elver@google.com>
Link: http://lkml.kernel.org/r/20200207003715.1578-1-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/page_io: mark various intentional data races
Qian Cai [Sat, 15 Aug 2020 00:31:20 +0000 (17:31 -0700)]
mm/page_io: mark various intentional data races

struct swap_info_struct si.flags could be accessed concurrently as noticed
by KCSAN,

 BUG: KCSAN: data-race in scan_swap_map_slots / swap_readpage

 write to 0xffff9c77b80ac400 of 8 bytes by task 91325 on cpu 16:
  scan_swap_map_slots+0x6fe/0xb50
  scan_swap_map_slots at mm/swapfile.c:887
  get_swap_pages+0x39d/0x5c0
  get_swap_page+0x377/0x524
  add_to_swap+0xe4/0x1c0
  shrink_page_list+0x1740/0x2820
  shrink_inactive_list+0x316/0x8b0
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x170/0x700
  __handle_mm_fault+0xc9f/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 read to 0xffff9c77b80ac400 of 8 bytes by task 5422 on cpu 7:
  swap_readpage+0x204/0x6a0
  swap_readpage at mm/page_io.c:380
  read_swap_cache_async+0xa2/0xb0
  swapin_readahead+0x6a0/0x890
  do_swap_page+0x465/0xeb0
  __handle_mm_fault+0xc7a/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 Reported by Kernel Concurrency Sanitizer on:
 CPU: 7 PID: 5422 Comm: gmain Tainted: G        W  O L 5.5.0-next-20200204+ #6
 Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019

Other reads,

 read to 0xffff91ea33eac400 of 8 bytes by task 11276 on cpu 120:
  __swap_writepage+0x140/0xc20
  __swap_writepage at mm/page_io.c:289

 read to 0xffff91ea33eac400 of 8 bytes by task 11264 on cpu 16:
  swap_set_page_dirty+0x44/0x1f4
  swap_set_page_dirty at mm/page_io.c:442

The write is under &si->lock, but the reads are done as lockless.  Since
the reads only check for a specific bit in the flag, it is harmless even
if load tearing happens.  Thus, just mark them as intentional data races
using the data_race() macro.

[cai@lca.pw: add a missing annotation]
Link: http://lkml.kernel.org/r/1581612585-5812-1-git-send-email-cai@lca.pw
Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marco Elver <elver@google.com>
Link: http://lkml.kernel.org/r/20200207003601.1526-1-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/frontswap: mark various intentional data races
Qian Cai [Sat, 15 Aug 2020 00:31:17 +0000 (17:31 -0700)]
mm/frontswap: mark various intentional data races

There are a few information counters that are intentionally not protected
against increment races, so just annotate them using the data_race()
macro.

 BUG: KCSAN: data-race in __frontswap_store / __frontswap_store

 write to 0xffffffff8b7174d8 of 8 bytes by task 6396 on cpu 103:
  __frontswap_store+0x2d0/0x344
  inc_frontswap_failed_stores at mm/frontswap.c:70
  (inlined by) __frontswap_store at mm/frontswap.c:280
  swap_writepage+0x83/0xf0
  pageout+0x33e/0xae0
  shrink_page_list+0x1f57/0x2870
  shrink_inactive_list+0x316/0x880
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x170/0x700
  __handle_mm_fault+0xc9f/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 read to 0xffffffff8b7174d8 of 8 bytes by task 6405 on cpu 47:
  __frontswap_store+0x2b9/0x344
  inc_frontswap_failed_stores at mm/frontswap.c:70
  (inlined by) __frontswap_store at mm/frontswap.c:280
  swap_writepage+0x83/0xf0
  pageout+0x33e/0xae0
  shrink_page_list+0x1f57/0x2870
  shrink_inactive_list+0x316/0x880
  shrink_lruvec+0x8dc/0x1380
  shrink_node+0x317/0xd80
  do_try_to_free_pages+0x1f7/0xa10
  try_to_free_pages+0x26c/0x5e0
  __alloc_pages_slowpath+0x458/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x170/0x700
  __handle_mm_fault+0xc9f/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marco Elver <elver@google.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Link: http://lkml.kernel.org/r/1581114499-5042-1-git-send-email-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/kmemleak: silence KCSAN splats in checksum
Qian Cai [Sat, 15 Aug 2020 00:31:14 +0000 (17:31 -0700)]
mm/kmemleak: silence KCSAN splats in checksum

Even if KCSAN is disabled for kmemleak, update_checksum() could still call
crc32() (which is outside of kmemleak.c) to dereference object->pointer.
Thus, the value of object->pointer could be accessed concurrently as
noticed by KCSAN,

 BUG: KCSAN: data-race in crc32_le_base / do_raw_spin_lock

 write to 0xffffb0ea683a7d50 of 4 bytes by task 23575 on cpu 12:
  do_raw_spin_lock+0x114/0x200
  debug_spin_lock_after at kernel/locking/spinlock_debug.c:91
  (inlined by) do_raw_spin_lock at kernel/locking/spinlock_debug.c:115
  _raw_spin_lock+0x40/0x50
  __handle_mm_fault+0xa9e/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 read to 0xffffb0ea683a7d50 of 4 bytes by task 839 on cpu 60:
  crc32_le_base+0x67/0x350
  crc32_le_base+0x67/0x350:
  crc32_body at lib/crc32.c:106
  (inlined by) crc32_le_generic at lib/crc32.c:179
  (inlined by) crc32_le at lib/crc32.c:197
  kmemleak_scan+0x528/0xd90
  update_checksum at mm/kmemleak.c:1172
  (inlined by) kmemleak_scan at mm/kmemleak.c:1497
  kmemleak_scan_thread+0xcc/0xfa
  kthread+0x1e0/0x200
  ret_from_fork+0x27/0x50

If a shattered value was returned due to a data race, it will be corrected
in the next scan.  Thus, let KCSAN ignore all reads in the region to
silence KCSAN in case the write side is non-atomic.

Suggested-by: Marco Elver <elver@google.com>
Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Marco Elver <elver@google.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Link: http://lkml.kernel.org/r/20200317182754.2180-1-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoall arch: remove system call sys_sysctl
Xiaoming Ni [Sat, 15 Aug 2020 00:31:07 +0000 (17:31 -0700)]
all arch: remove system call sys_sysctl

Since commit 61a47c1ad3a4dc ("sysctl: Remove the sysctl system call"),
sys_sysctl is actually unavailable: any input can only return an error.

We have been warning about people using the sysctl system call for years
and believe there are no more users.  Even if there are users of this
interface if they have not complained or fixed their code by now they
probably are not going to, so there is no point in warning them any
longer.

So completely remove sys_sysctl on all architectures.

[nixiaoming@huawei.com: s390: fix build error for sys_call_table_emu]
Link: http://lkml.kernel.org/r/20200618141426.16884-1-nixiaoming@huawei.com
Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Will Deacon <will@kernel.org> [arm/arm64]
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Bin Meng <bin.meng@windriver.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: chenzefeng <chenzefeng2@huawei.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Christian Brauner <christian@brauner.io>
Cc: Chris Zankel <chris@zankel.net>
Cc: David Howells <dhowells@redhat.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Diego Elio Pettenò <flameeyes@flameeyes.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Helge Deller <deller@gmx.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: James Bottomley <James.Bottomley@HansenPartnership.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kars de Jong <jongk@linux-m68k.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Marco Elver <elver@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Martin K. Petersen <martin.petersen@oracle.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Miklos Szeredi <mszeredi@redhat.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Nick Piggin <npiggin@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Olof Johansson <olof@lixom.net>
Cc: Paul Burton <paulburton@kernel.org>
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@linux.ibm.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Rich Felker <dalias@libc.org>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Sami Tolvanen <samitolvanen@google.com>
Cc: Sargun Dhillon <sargun@sargun.me>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Cc: Sven Schnelle <svens@stackframe.org>
Cc: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Zhou Yanjie <zhouyanjie@wanyeetech.com>
Link: http://lkml.kernel.org/r/20200616030734.87257-1-nixiaoming@huawei.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agofs: autofs: delete repeated words in comments
Randy Dunlap [Sat, 15 Aug 2020 00:30:46 +0000 (17:30 -0700)]
fs: autofs: delete repeated words in comments

Drop duplicated words {the, at} in comments.

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Ian Kent <raven@themaw.net>
Link: http://lkml.kernel.org/r/20200811021817.24982-1-rdunlap@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: introduce offset_in_thp
Matthew Wilcox (Oracle) [Sat, 15 Aug 2020 00:30:43 +0000 (17:30 -0700)]
mm: introduce offset_in_thp

Mirroring offset_in_page(), this gives you the offset within this
particular page, no matter what size page it is.  It optimises down to
offset_in_page() if CONFIG_TRANSPARENT_HUGEPAGE is not set.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Link: http://lkml.kernel.org/r/20200629151959.15779-8-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: add thp_head
Matthew Wilcox (Oracle) [Sat, 15 Aug 2020 00:30:40 +0000 (17:30 -0700)]
mm: add thp_head

This is like compound_head() but compiles away when
CONFIG_TRANSPARENT_HUGEPAGE is not enabled.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Link: http://lkml.kernel.org/r/20200629151959.15779-7-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: replace hpage_nr_pages with thp_nr_pages
Matthew Wilcox (Oracle) [Sat, 15 Aug 2020 00:30:37 +0000 (17:30 -0700)]
mm: replace hpage_nr_pages with thp_nr_pages

The thp prefix is more frequently used than hpage and we should be
consistent between the various functions.

[akpm@linux-foundation.org: fix mm/migrate.c]

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Link: http://lkml.kernel.org/r/20200629151959.15779-6-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: add thp_size
Matthew Wilcox (Oracle) [Sat, 15 Aug 2020 00:30:33 +0000 (17:30 -0700)]
mm: add thp_size

This function returns the number of bytes in a THP.  It is like
page_size(), but compiles to just PAGE_SIZE if CONFIG_TRANSPARENT_HUGEPAGE
is disabled.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Link: http://lkml.kernel.org/r/20200629151959.15779-5-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: add thp_order
Matthew Wilcox (Oracle) [Sat, 15 Aug 2020 00:30:30 +0000 (17:30 -0700)]
mm: add thp_order

This function returns the order of a transparent huge page.  It compiles
to 0 if CONFIG_TRANSPARENT_HUGEPAGE is disabled.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Link: http://lkml.kernel.org/r/20200629151959.15779-4-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: move page-flags include to top of file
Matthew Wilcox (Oracle) [Sat, 15 Aug 2020 00:30:26 +0000 (17:30 -0700)]
mm: move page-flags include to top of file

Give up on the notion that we can remove page-flags.h from mm.h.  There
are currently 14 inline functions which use a PageFoo function.  Also, two
of the files directly included by mm.h include page-flags.h themselves,
and there are probably more indirect inclusions.  So just include it at
the top like any other header file.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Link: http://lkml.kernel.org/r/20200629151959.15779-3-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: store compound_nr as well as compound_order
Matthew Wilcox (Oracle) [Sat, 15 Aug 2020 00:30:23 +0000 (17:30 -0700)]
mm: store compound_nr as well as compound_order

Patch series "THP prep patches".

These are some generic cleanups and improvements, which I would like
merged into mmotm soon.  The first one should be a performance improvement
for all users of compound pages, and the others are aimed at getting code
to compile away when CONFIG_TRANSPARENT_HUGEPAGE is disabled (ie small
systems).  Also better documented / less confusing than the current prefix
mixture of compound, hpage and thp.

This patch (of 7):

This removes a few instructions from functions which need to know how many
pages are in a compound page.  The storage used is either page->mapping on
64-bit or page->index on 32-bit.  Both of these are fine to overlay on
tail pages.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Link: http://lkml.kernel.org/r/20200629151959.15779-1-willy@infradead.org
Link: http://lkml.kernel.org/r/20200629151959.15779-2-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomailmap: add entry for Greg Kurz
Greg Kurz [Sat, 15 Aug 2020 00:30:20 +0000 (17:30 -0700)]
mailmap: add entry for Greg Kurz

I had stopped using gkurz@linux.vnet.ibm.com a while back already but this
email address was shutdown last June when I quit IBM.  It's about time to
map it to groug@kaod.org.

Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Link: http://lkml.kernel.org/r/159724692879.76040.4938578139173154028.stgit@bahia.lan
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoselftests/exec: add file type errno tests
Kees Cook [Sat, 15 Aug 2020 00:30:17 +0000 (17:30 -0700)]
selftests/exec: add file type errno tests

Make sure execve() returns the expected errno values for non-regular
files.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marc Zyngier <maz@kernel.org>
Link: http://lkml.kernel.org/r/20200813231723.2725102-3-keescook@chromium.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoexec: restore EACCES of S_ISDIR execve()
Kees Cook [Sat, 15 Aug 2020 00:30:14 +0000 (17:30 -0700)]
exec: restore EACCES of S_ISDIR execve()

Patch series "Fix S_ISDIR execve() errno".

Fix an errno change for execve() of directories, noticed by Marc Zyngier.
Along with the fix, include a regression test to avoid seeing this return
in the future.

This patch (of 2):

The return code for attempting to execute a directory has always been
EACCES.  Adjust the S_ISDIR exec test to reflect the old errno instead of
the general EISDIR for other kinds of "open" attempts on directories.

Fixes: 633fb6ac3980 ("exec: move S_ISREG() check earlier")
Reported-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: Greg Kroah-Hartman <gregkh@android.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@google.com>
Link: http://lkml.kernel.org/r/20200813231723.2725102-2-keescook@chromium.org
Link: https://lore.kernel.org/lkml/20200813151305.6191993b@why
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agolz4: fix kernel decompression speed
Nick Terrell [Sat, 15 Aug 2020 00:30:10 +0000 (17:30 -0700)]
lz4: fix kernel decompression speed

This patch replaces all memcpy() calls with LZ4_memcpy() which calls
__builtin_memcpy() so the compiler can inline it.

LZ4 relies heavily on memcpy() with a constant size being inlined.  In x86
and i386 pre-boot environments memcpy() cannot be inlined because memcpy()
doesn't get defined as __builtin_memcpy().

An equivalent patch has been applied upstream so that the next import
won't lose this change [1].

I've measured the kernel decompression speed using QEMU before and after
this patch for the x86_64 and i386 architectures.  The speed-up is about
10x as shown below.

Code Arch Kernel Size Time Speed
v5.8 x86_64 11504832 B 148 ms  79 MB/s
patch x86_64 11503872 B  13 ms 885 MB/s
v5.8 i386  9621216 B  91 ms 106 MB/s
patch i386  9620224 B  10 ms 962 MB/s

I also measured the time to decompress the initramfs on x86_64, i386, and
arm.  All three show the same decompression speed before and after, as
expected.

[1] https://github.com/lz4/lz4/pull/890

Signed-off-by: Nick Terrell <terrelln@fb.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Yann Collet <yann.collet.73@gmail.com>
Cc: Gao Xiang <gaoxiang25@huawei.com>
Cc: Sven Schmidt <4sschmid@informatik.uni-hamburg.de>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Arvind Sankar <nivedita@alum.mit.edu>
Link: http://lkml.kernel.org/r/20200803194022.2966806-1-nickrterrell@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>