Use the new itrace 'q' option to add support for a mode of decoding that
ignores TNT, does not walk object code, but gets the ip from FUP and TIP
packets.

Example:

 $ perf record -e intel_pt//u grep -rI pudding drivers
 [ perf record: Woken up 52 times to write data ]
 [ perf record: Captured and wrote 57.870 MB perf.data ]
 $ time perf script --itrace=bi | wc -l
 58948289

 real    1m23.863s
 user    1m23.251s
 sys     0m7.452s
 $ time perf script --itrace=biq | wc -l
 3385694

 real    0m4.453s
 user    0m4.455s
 sys     0m0.328s

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-12-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

The 'q' option is for modes of decoding that are quicker because they
skip or omit decoding some aspects of trace data.

If supported, the 'q' option may be repeated to increase the effect.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-11-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Change the debug logging (when used with the --time option) to time
filter logged perf events, but allow that to be overridden by using
"d+a" instead of plain "d".

That can reduce the size of the log file.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-10-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

The "d" option may be followed by flags which affect what debug messages
will or will not be logged. Each flag must be preceded by either '+' or
'-'. The flags support by Intel PT are:

		-a	Suppress logging of perf events

Suppressing perf events is useful for decreasing the size of the log.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-9-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Allow the 'd' option to be followed by flags which will affect what debug
messages will or will not be reported. Each flag must be preceded by either
'+' or '-'. The flags are:
	a	all perf events

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-8-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

The itrace "e" option may be followed by flags which affect what errors
will or will not be reported.  Each flag must be preceded by either '+' or '-'.
The flags supported by Intel PT are:

		-o	Suppress overflow errors
		-l	Suppress trace data lost errors
For example, for errors but not overflow or data lost errors:

	--itrace=e-o-l

Suppressing those errors can be useful for testing and debugging because
they are not due to decoding.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-7-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Allow the 'e' option to be followed by flags which will affect what errors
will or will not be reported. Each flag must be preceded by either '+' or
'-'. The flags are:
	o	overflow
	l	trace data lost

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-6-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Add missing itrace options o, G and L.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-5-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

For example:

 Before:
   $ perf record -e '{intel_pt/branch=0/,branch-loads/aux-output/ppp}' -- ls -l
   Error:
   branch-loads: PMU Hardware doesn't support sampling/overflow-interrupts. Try 'perf stat'

 After:
   $ perf record -e '{intel_pt/branch=0/,branch-loads/aux-output/ppp}' -- ls -l
   Error:
   branch-loads: PMU Hardware doesn't support 'aux_output' feature

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lore.kernel.org/lkml/20200710151104.15137-4-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

CBR events can result in a duplicate branch event, because the state
type defaults to a branch. Fix by clearing the state type.

Example: trace 'sleep' and hope for a frequency change

 Before:

   $ perf record -e intel_pt//u sleep 0.1
   [ perf record: Woken up 1 times to write data ]
   [ perf record: Captured and wrote 0.034 MB perf.data ]
   $ perf script --itrace=bpe > before.txt

 After:

   $ perf script --itrace=bpe > after.txt
   $ diff -u before.txt after.txt
   --- before.txt  2020-07-07 14:42:18.191508098 +0300
   +++ after.txt   2020-07-07 14:42:36.587891753 +0300
   @@ -29673,7 +29673,6 @@
               sleep 93431 [007] 15411.619905:          1  branches:u:                 0 [unknown] ([unknown]) =>     7f0818abb2e0 clock_nanosleep@@GLIBC_2.17+0x0 (/usr/lib/x86_64-linux-gnu/libc-2.31.so)
               sleep 93431 [007] 15411.619905:          1  branches:u:      7f0818abb30c clock_nanosleep@@GLIBC_2.17+0x2c (/usr/lib/x86_64-linux-gnu/libc-2.31.so) =>                0 [unknown] ([unknown])
               sleep 93431 [007] 15411.720069:         cbr:  cbr: 15 freq: 1507 MHz ( 56%)         7f0818abb30c clock_nanosleep@@GLIBC_2.17+0x2c (/usr/lib/x86_64-linux-gnu/libc-2.31.so)
   -           sleep 93431 [007] 15411.720069:          1  branches:u:      7f0818abb30c clock_nanosleep@@GLIBC_2.17+0x2c (/usr/lib/x86_64-linux-gnu/libc-2.31.so) =>                0 [unknown] ([unknown])
               sleep 93431 [007] 15411.720076:          1  branches:u:                 0 [unknown] ([unknown]) =>     7f0818abb30e clock_nanosleep@@GLIBC_2.17+0x2e (/usr/lib/x86_64-linux-gnu/libc-2.31.so)
               sleep 93431 [007] 15411.720077:          1  branches:u:      7f0818abb323 clock_nanosleep@@GLIBC_2.17+0x43 (/usr/lib/x86_64-linux-gnu/libc-2.31.so) =>     7f0818ac0eb7 __nanosleep+0x17 (/usr/lib/x86_64-linux-gnu/libc-2.31.so)
               sleep 93431 [007] 15411.720077:          1  branches:u:      7f0818ac0ebf __nanosleep+0x1f (/usr/lib/x86_64-linux-gnu/libc-2.31.so) =>     55cb7e4c2827 rpl_nanosleep+0x97 (/usr/bin/sleep)

Fixes: 91de8684 ("perf intel-pt: Cater for CBR change in PSB+")
Fixes: abe5a1d3

 ("perf intel-pt: Decoder to output CBR changes immediately")
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: stable@vger.kernel.org
Link: http://lore.kernel.org/lkml/20200710151104.15137-3-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

While walking code towards a FUP ip, the packet state is
INTEL_PT_STATE_FUP or INTEL_PT_STATE_FUP_NO_TIP. That was mishandled
resulting in the state becoming INTEL_PT_STATE_IN_SYNC prematurely.  The
result was an occasional lost EXSTOP event.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: stable@vger.kernel.org
Link: http://lore.kernel.org/lkml/20200710151104.15137-2-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

To sync headers, for instance, in this case tools/perf was ahead of
upstream till Linus merged tip/perf/core to get the
PERF_RECORD_TEXT_POKE changes:

  Warning: Kernel ABI header at 'tools/include/uapi/linux/perf_event.h' differs from latest version at 'include/uapi/linux/perf_event.h'
  diff -u tools/include/uapi/linux/perf_event.h include/uapi/linux/perf_event.h

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Pull networking updates from David Miller:

 1) Support 6Ghz band in ath11k driver, from Rajkumar Manoharan.

 2) Support UDP segmentation in code TSO code, from Eric Dumazet.

 3) Allow flashing different flash images in cxgb4 driver, from Vishal
    Kulkarni.

 4) Add drop frames counter and flow status to tc flower offloading,
    from Po Liu.

 5) Support n-tuple filters in cxgb4, from Vishal Kulkarni.

 6) Various new indirect call avoidance, from Eric Dumazet and Brian
    Vazquez.

 7) Fix BPF verifier failures on 32-bit pointer arithmetic, from
    Yonghong Song.

 8) Support querying and setting hardware address of a port function via
    devlink, use this in mlx5, from Parav Pandit.

 9) Support hw ipsec offload on bonding slaves, from Jarod Wilson.

10) Switch qca8k driver over to phylink, from Jonathan McDowell.

11) In bpftool, show list of processes holding BPF FD references to
    maps, programs, links, and btf objects. Fro...

Pull drm updates from Dave Airlie:
 "New xilinx displayport driver, AMD support for two new GPUs (more
  header files), i915 initial support for RocketLake and some work on
  their DG1 (discrete chip).

  The core also grew some lockdep annotations to try and constrain what
  drivers do with dma-fences, and added some documentation on why the
  idea of indefinite fences doesn't work.

  The long list is below.

  I do have some fixes trees outstanding, but I'll follow up with those
  later.

  core:
   - add user def flag to cmd line modes
   - dma_fence_wait added might_sleep
   - dma-fence lockdep annotations
   - indefinite fences are bad documentation
   - gem CMA functions used in more drivers
   - struct mutex removal
   - more drm_ debug macro usage
   - set/drop master api fixes
   - fix for drm/mm hole size comparison
   - drm/mm remove invalid entry optimization
   - optimise drm/mm hole handling
   - VRR debugfs added
   - uncompressed AFBC modifier support
   - multiple display id blocks in EDID
   - multiple driver sg handling fixes
   - __drm_atomic_helper_crtc_reset in all drivers
   - managed vram helpers

  ttm:
   - ttm_mem_reg handling cleanup
   - remove bo offset field
   - drop CMA memtype flag
   - drop mappable flag

  xilinx:
   - New Xilinx ZynqMP DisplayPort Subsystem driver

  nouveau:
   - add CRC support
   - start using NVIDIA published class header files
   - convert all push buffer emission to new macros
   - Proper push buffer space management for EVO/NVD channels.
   - firmware loading fixes
   - 2MiB system memory pages support on Pascal and newer

  vkms:
   - larger cursor support

  i915:
   - Rocketlake platform enablement
   - Early DG1 enablement
   - Numerous GEM refactorings
   - DP MST fixes
   - FBC, PSR, Cursor, Color, Gamma fixes
   - TGL, RKL, EHL workaround updates
   - TGL 8K display support fixes
   - SDVO/HDMI/DVI fixes

  amdgpu:
   - Initial support for Sienna Cichlid GPU
   - Initial support for Navy Flounder GPU
   - SI UVD/VCE support
   - expose rotation property
   - Add support for unique id on Arcturus
   - Enable runtime PM on vega10 boards that support BACO
   - Skip BAR resizing if the bios already did id
   - Major swSMU code cleanup
   - Fixes for DCN bandwidth calculations

  amdkfd:
   - Track SDMA usage per process
   - SMI events interface

  radeon:
   - Default to on chip GART for AGP boards on all arches
   - Runtime PM reference count fixes

  msm:
   - headers regenerated causing churn
   - a650/a640 display and GPU enablement
   - dpu dither support for 6bpc panels
   - dpu cursor fix
   - dsi/mdp5 enablement for sdm630/sdm636/sdm66

  tegra:
   - video capture prep support
   - reflection support

  mediatek:
   - convert mtk_dsi to bridge API

  meson:
   - FBC support

  sun4i:
   - iommu support

  rockchip:
   - register locking fix
   - per-pixel alpha support PX30 VOP

  mgag200:
   - ported to simple and shmem helpers
   - device init cleanups
   - use managed pci functions
   - dropped hw cursor support

  ast:
   - use managed pci functions
   - use managed VRAM helpers
   - rework cursor support

  malidp:
   - dev_groups support

  hibmc:
   - refactor hibmc_drv_vdac:

  vc4:
   - create TXP CRTC

  imx:
   - error path fixes and cleanups

  etnaviv:
   - clock handling and error handling cleanups
   - use pin_user_pages"

* tag 'drm-next-2020-08-06' of git://anongit.freedesktop.org/drm/drm: (1747 commits)
  drm/msm: use kthread_create_worker instead of kthread_run
  drm/msm/mdp5: Add MDP5 configuration for SDM636/660
  drm/msm/dsi: Add DSI configuration for SDM660
  drm/msm/mdp5: Add MDP5 configuration for SDM630
  drm/msm/dsi: Add phy configuration for SDM630/636/660
  drm/msm/a6xx: add A640/A650 hwcg
  drm/msm/a6xx: hwcg tables in gpulist
  drm/msm/dpu: add SM8250 to hw catalog
  drm/msm/dpu: add SM8150 to hw catalog
  drm/msm/dpu: intf timing path for displayport
  drm/msm/dpu: set missing flush bits for INTF_2 and INTF_3
  drm/msm/dpu: don't use INTF_INPUT_CTRL feature on sdm845
  drm/msm/dpu: move some sspp caps to dpu_caps
  drm/msm/dpu: update UBWC config for sm8150 and sm8250
  drm/msm/dpu: use right setup_blend_config for sm8150 and sm8250
  drm/msm/a6xx: set ubwc config for A640 and A650
  drm/msm/adreno: un-open-code some packets
  drm/msm: sync generated headers
  drm/msm/a6xx: add build_bw_table for A640/A650
  drm/msm/a6xx: fix crashstate capture for A650
  ...

Pull LED updates from Pavel Machek:
 "Okay, so... this one is interesting. RGB LEDs are very common, and we
  need to have some kind of support for them. Multicolor is for
  arbitrary set of LEDs in one package, RGB is for LEDs that can produce
  full range of colors. We do not have real multicolor LED that is not
  RGB in the pipeline, so that one is disabled for now.

  You can expect this saga to continue with next pull requests"

* tag 'leds-5.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/pavel/linux-leds: (37 commits)
  MAINTAINERS: Remove myself as LED subsystem maintainer
  leds: disallow /sys/class/leds/*:multi:* for now
  leds: add RGB color option, as that is different from multicolor.
  Make LEDS_LP55XX_COMMON depend on I2C to fix build errors:
  Documentation: ABI: leds-turris-omnia: document sysfs attribute
  leds: initial support for Turris Omnia LEDs
  dt-bindings: leds: add cznic,turris-omnia-leds binding
  leds: pattern trigger -- check pattern for validity
  leds: Replace HTTP links with HTTPS ones
  leds: trigger: add support for LED-private device triggers
  leds: lp5521: Add multicolor framework multicolor brightness support
  leds: lp5523: Update the lp5523 code to add multicolor brightness function
  leds: lp55xx: Add multicolor framework support to lp55xx
  leds: lp55xx: Convert LED class registration to devm_*
  dt-bindings: leds: Convert leds-lp55xx to yaml
  leds: multicolor: Introduce a multicolor class definition
  leds: Add multicolor ID to the color ID list
  dt: bindings: Add multicolor class dt bindings documention
  leds: lp5523: Fix various formatting issues in the code
  leds: lp55xx: Fix file permissions to use DEVICE_ATTR macros
  ...

A VF's mailbox mutex is not getting initialized by nicvf_probe() until after
it is first used. And such usage is resulting in...

[   28.270927] ------------[ cut here ]------------
[   28.270934] DEBUG_LOCKS_WARN_ON(lock->magic != lock)
[   28.270980] WARNING: CPU: 9 PID: 675 at kernel/locking/mutex.c:938 __mutex_lock+0xdac/0x12f0
[   28.270985] Modules linked in: ast(+) nicvf(+) i2c_algo_bit drm_vram_helper drm_ttm_helper ttm nicpf(+) drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm ixgbe(+) sg thunder_bgx mdio i2c_thunderx mdio_thunder thunder_xcv mdio_cavium dm_mirror dm_region_hash dm_log dm_mod
[   28.271064] CPU: 9 PID: 675 Comm: systemd-udevd Not tainted 4.18.0+ #1
[   28.271070] Hardware name: GIGABYTE R120-T34-00/MT30-GS2-00, BIOS F02 08/06/2019
[   28.271078] pstate: 60000005 (nZCv daif -PAN -UAO)
[   28.271086] pc : __mutex_lock+0xdac/0x12f0
[   28.271092] lr : __mutex_lock+0xdac/0x12f0
[   28.271097] sp : ffff800d42146fb0
[   28.271103] x29: ffff800d42146fb0 x28: 0000000000000000
[   28.271113] x27: ffff800d24361180 x26: dfff200000000000
[   28.271122] x25: 0000000000000000 x24: 0000000000000002
[   28.271132] x23: ffff20001597cc80 x22: ffff2000139e9848
[   28.271141] x21: 0000000000000000 x20: 1ffff001a8428e0c
[   28.271151] x19: ffff200015d5d000 x18: 1ffff001ae0f2184
[   28.271160] x17: 0000000000000000 x16: 0000000000000000
[   28.271170] x15: ffff800d70790c38 x14: ffff20001597c000
[   28.271179] x13: ffff20001597cc80 x12: ffff040002b2f779
[   28.271189] x11: 1fffe40002b2f778 x10: ffff040002b2f778
[   28.271199] x9 : 0000000000000000 x8 : 00000000f1f1f1f1
[   28.271208] x7 : 00000000f2f2f2f2 x6 : 0000000000000000
[   28.271217] x5 : 1ffff001ae0f2186 x4 : 1fffe400027eb03c
[   28.271227] x3 : dfff200000000000 x2 : ffff1001a8428dbe
[   28.271237] x1 : c87fdfac7ea11d00 x0 : 0000000000000000
[   28.271246] Call trace:
[   28.271254]  __mutex_lock+0xdac/0x12f0
[   28.271261]  mutex_lock_nested+0x3c/0x50
[   28.271297]  nicvf_send_msg_to_pf+0x40/0x3a0 [nicvf]
[   28.271316]  nicvf_register_misc_interrupt+0x20c/0x328 [nicvf]
[   28.271334]  nicvf_probe+0x508/0xda0 [nicvf]
[   28.271344]  local_pci_probe+0xc4/0x180
[   28.271352]  pci_device_probe+0x3ec/0x528
[   28.271363]  driver_probe_device+0x21c/0xb98
[   28.271371]  device_driver_attach+0xe8/0x120
[   28.271379]  __driver_attach+0xe0/0x2a0
[   28.271386]  bus_for_each_dev+0x118/0x190
[   28.271394]  driver_attach+0x48/0x60
[   28.271401]  bus_add_driver+0x328/0x558
[   28.271409]  driver_register+0x148/0x398
[   28.271416]  __pci_register_driver+0x14c/0x1b0
[   28.271437]  nicvf_init_module+0x54/0x10000 [nicvf]
[   28.271447]  do_one_initcall+0x18c/0xc18
[   28.271457]  do_init_module+0x18c/0x618
[   28.271464]  load_module+0x2bc0/0x4088
[   28.271472]  __se_sys_finit_module+0x110/0x188
[   28.271479]  __arm64_sys_finit_module+0x70/0xa0
[   28.271490]  el0_svc_handler+0x15c/0x380
[   28.271496]  el0_svc+0x8/0xc
[   28.271502] irq event stamp: 52649
[   28.271513] hardirqs last  enabled at (52649): [<ffff200011b4d790>] _raw_spin_unlock_irqrestore+0xc0/0xd8
[   28.271522] hardirqs last disabled at (52648): [<ffff200011b4d3c4>] _raw_spin_lock_irqsave+0x3c/0xf0
[   28.271530] softirqs last  enabled at (52330): [<ffff200010082af4>] __do_softirq+0xacc/0x117c
[   28.271540] softirqs last disabled at (52313): [<ffff20001019b354>] irq_exit+0x3cc/0x500
[   28.271545] ---[ end trace a9b90324c8a0d4ee ]---

This problem is resolved by moving the call to mutex_init() up earlier
in nicvf_probe().

Fixes: 609ea65c

 ("net: thunderx: add mutex to protect mailbox from concurrent calls for same VF")
Signed-off-by: Dean Nelson <dnelson@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Oliver Neukum says:

====================
misc bug fixes for the hso driver

1. Code reuse led to an unregistration of a net driver that has not been
registered
2. The kernel complains generically if kmalloc with GFP_KERNEL fails
3. A race that can lead to an URB that is in use being reused or
a use after free
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

This check an inherent race. It opens a race where
an error code has already been set or cleared yet
the URB has not been given back. We cannot do
such an optimization and must unlink unconditionally.

Signed-off-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If this fails, kmalloc() will print a report including
a stack trace. There is no need for a separate complaint.

Signed-off-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The driver tries to reuse code for disconnect in case
of a failed probe.
If resources need to be freed after an error in probe, the
netdev must not be freed because it has never been registered.
Fix it by telling the helper which path we are in.

Signed-off-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

It don't have enough time for reviewing patches and thus don't
want to be listed as regular LED maintainer. Nonetheless I may still
give a review from time to time.

Signed-off-by: Jacek Anaszewski <jacek.anaszewski@gmail.com>
Signed-off-by: Pavel Machek <pavel@ucw.cz>

Pull hmm updates from Jason Gunthorpe:
 "Ralph has been working on nouveau's use of hmm_range_fault() and
  migrate_vma() which resulted in this small series. It adds reporting
  of the page table order from hmm_range_fault() and some optimization
  of migrate_vma():

   - Report the size of the page table mapping out of hmm_range_fault().

     This makes it easier to establish a large/huge/etc mapping in the
     device's page table.

   - Allow devices to ignore the invalidations during migration in cases
     where the migration is not going to change pages.

     For instance migrating pages to a device does not require the
     device to invalidate pages already in the device.

   - Update nouveau and hmm_tests to use the above"

* tag 'for-linus-hmm' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  mm/hmm/test: use the new migration invalidation
  nouveau/svm: use the new migration invalidation
  mm/notifier: add migration invalidation type
  mm/migrate: add a flags parameter to migrate_vma
  nouveau: fix storing invalid ptes
  nouveau/hmm: support mapping large sysmem pages
  nouveau: fix mapping 2MB sysmem pages
  nouveau/hmm: fault one page at a time
  mm/hmm: add tests for hmm_pfn_to_map_order()
  mm/hmm: provide the page mapping order in hmm_range_fault()

Pull MMC updates from Ulf Hansson:
 "MMC core:

   - Add a new host cap bit and a corresponding DT property, to support
     power cycling of the card by FW at system suspend/resume.

   - Fix clock rate setting for SDIO in SDR12/SDR25 speed-mode

   - Fix switch to 1/4-bit mode at system suspend/resume for SD-combo
     cards

   - Convert the mmc-pwrseq DT bindings to the json-schema

   - Always allow the card detect uevent to be consumed by userspace

  MMC host controllers:

   - Convert a few DT bindings to the json-schema

   - mtk-sd:
      - Add support for command queue through cqhci
      - Add support for the MT6779 variant

   - renesas_sdhi_internal_dmac:
      - Fix dma unmapping in the error path

   - sdhci_am654:
      - Add support for the AM65x PG2.0 variant
      - Extend support for phys/clocks

   - sdhci-cadence:
      - Drop incorrect HW tuning for SD mode

   - sdhci-msm:
      - Add support for interconnect bandwidth scaling
      - Enable internal voltage control
      - Enable low power state for pinctrls

   - sdhci-of-at91:
      - Ludovic Desroches handovers maintenance to Eugen Hristev

   - sdhci-pci-gli:
      - Improve clock handling for GL975x

   - sdhci-pci-o2micro:
      - Add HW tuning for SDR104 mode
      - Fix support for O2 host controller Seabird1"

* tag 'mmc-v5.9' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc: (66 commits)
  mmc: mediatek: make function msdc_cqe_disable() static
  MAINTAINERS: mmc: sdhci-of-at91: handover maintenance to Eugen Hristev
  dt-bindings: mmc: mediatek: Add document for mt6779
  mmc: mediatek: command queue support
  mmc: mediatek: refine msdc timeout api
  mmc: mediatek: add MT6779 MMC driver support
  mmc: sdhci-pci-o2micro: Add HW tuning for SDR104 mode
  mmc: sdhci-pci-o2micro: Bug fix for O2 host controller Seabird1
  mmc: via-sdmmc: use generic power management
  memstick: jmb38x_ms: use generic power management
  mmc: sdhci-cadence: do not use hardware tuning for SD mode
  mmc: sdhci-pci-gli: Set SDR104's clock to 205MHz and enable SSC for GL975x
  mmc: cqhci: Fix a print format for the task descriptor
  mmc: sdhci-of-arasan: fix timings allocation code
  mmc: sdhci: Fix a potential uninitialized variable
  dt-bindings: mmc: renesas,sdhi: convert to YAML
  dt-bindings: mmc: convert arasan sdhci bindings to yaml
  mmc: sdhci: Fix potential null pointer access while accessing vqmmc
  mmc: core: Add MMC_CAP2_FULL_PWR_CYCLE_IN_SUSPEND
  dt-bindings: mmc: Add full-pwr-cycle-in-suspend property
  ...

Pull hwmon updates from Guenter Roeck:
 "Highlights:
   - New driver for Sparx5 SoC temperature sensot
   - New driver for Corsair Commander Pro
   - MAX20710 support added to max20730 driver

  Enhancements:
   - max6697: Allow max6581 to create tempX_offset attributes
   - gsc (Gateworks System Controller): add 16bit pre-scaled voltage mode
   - adm1275: Enable adm1278 ADM1278_TEMP1_EN
   - dell-smm: Add Latitude 5480 to fan control whitelist

  Fixes:
   - adc128d818: Fix advanced configuration register init
   - pmbus/core: Use s64 instead of long for calculations to fix
     overflow issues with 32-bit architectures

  Plus various cleanups in several drivers"

* tag 'hwmon-for-v5.9' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging: (32 commits)
  hwmon: (adc128d818) Fix advanced configuration register init
  hwmon: (axi-fan-control) remove duplicate macros
  hwmon: (i5k_amb, vt8231) Drop uses of pci_read_config_*() return value
  hwmon: (sparx5) Make symbol 's5_temp_match' static
  hwmon: (corsair-cpro) add reading pwm values
  hwmon: sparx5: Add Sparx5 SoC temperature driver
  dt-bindings: hwmon: Add Sparx5 temperature sensor
  hwmon: (tmp401) Replace HTTP links with HTTPS ones
  hwmon: (lm95234) Replace HTTP links with HTTPS ones
  hwmon: (lm90) Replace HTTP links with HTTPS ones
  hwmon: (k8temp) Replace HTTP links with HTTPS ones
  hwmon: (jc42) Replace HTTP links with HTTPS ones
  hwmon: (ina2xx) Replace HTTP links with HTTPS ones
  hwmon: (ina209) Replace HTTP links with HTTPS ones
  hwmon: Replace HTTP links with HTTPS ones
  docs: hwmon: Replace HTTP links with HTTPS ones
  hwmon: (adm1025) Replace HTTP links with HTTPS ones
  hwmon: add Corsair Commander Pro driver
  hwmon: (max6697) Allow max6581 to create tempX_offset
  hwmon: (tmmp513) Replace HTTP links with HTTPS links
  ...

Pull Devicetree updates from Rob Herring:

 - Improve device links cycle detection and breaking. Add more bindings
   for device link dependencies.

 - Refactor parsing 'no-map' in __reserved_mem_alloc_size()

 - Improve DT unittest 'ranges' and 'dma-ranges' test case to check
   differing cell sizes

 - Various http to https link conversions

 - Add a schema check to prevent 'syscon' from being used by itself
   without a more specific compatible

 - A bunch more DT binding conversions to schema

* tag 'devicetree-for-5.9' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux: (55 commits)
  of: reserved-memory: remove duplicated call to of_get_flat_dt_prop() for no-map node
  of: unittest: Use bigger address cells to catch parser regressions
  dt-bindings: memory-controllers: Convert mmdc to json-schema
  dt-bindings: mtd: Convert imx nand to json-schema
  dt-bindings: mtd: Convert gpmi nand to json-schema
  dt-bindings: iio: io-channel-mux: Fix compatible string in example code
  of: property: Add device link support for pinctrl-0 through pinctrl-8
  of: property: Add device link support for multiple DT bindings
  dt-bindings: phy: ti: phy-gmii-sel: convert bindings to json-schema
  dt-bindings: mux: mux.h: drop a duplicated word
  dt-bindings: misc: Convert olpc,xo1.75-ec to json-schema
  dt-bindings: aspeed-lpc: Replace HTTP links with HTTPS ones
  dt-bindings: drm/bridge: Replace HTTP links with HTTPS ones
  drm/tilcdc: Replace HTTP links with HTTPS ones
  dt-bindings: iommu: renesas,ipmmu-vmsa: Add r8a774e1 support
  dt-bindings: fpga: Replace HTTP links with HTTPS ones
  dt-bindings: virtio: Replace HTTP links with HTTPS ones
  dt-bindings: media: imx274: Add optional input clock and supplies
  dt-bindings: i2c-gpio: Use 'deprecated' keyword on deprecated properties
  dt-bindings: interrupt-controller: Fix typos in loongson,liointc.yaml
  ...

Pull GPIO updates from Linus Walleij:
 "This is the bulk of GPIO changes for the v5.9 kernel cycle.

  There is nothing too exciting in it, but a new macro that fixes a
  build failure on a minor ARM32 platform that appeared yesterday is
  part of it so we better merge it.

  Core changes:

   - Introduce the for_each_requested_gpio() macro to help in dependent
     code all over the place. Also patch a few locations to use it while
     we are at it.

   - Split out the sysfs code into its own file.

   - Split out the character device code into its own file, then make a
     set of refactorings and improvements to this code. We are setting
     the stage to revamp the userspace API a bit in the next cycle.

   - Fix a whole slew of kerneldoc that was wrong or missing.

  New drivers:

   - The PCA953x driver now supports the PCAL9535.

  Driver improvements:

   - A host of incremental modernizations and improvements to the
     PCA953x driver.

   - Incremental improvements to the Xilinx Zynq driver.

   - Some improvements to the GPIO aggregator driver.

   - I ran all over the place switching all threaded and other drivers
     requesting their own IRQ while using the core GPIO IRQ helpers to
     pass the GPIO irq chip as a template instead of calling the
     explicit set-up functions. Next merge window we may retire the old
     code altogether"

* tag 'gpio-v5.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio: (97 commits)
  gpio: wcove: Request IRQ after all initialisation done
  gpio: crystalcove: Free IRQ on error path
  gpio: pca953x: Request IRQ after all initialisation done
  gpio: don't use same lockdep class for all devm_gpiochip_add_data users
  gpio: max732x: Use irqchip template
  gpio: stmpe: Move chip registration
  gpio: rcar: Use irqchip template
  gpio: regmap: fix type clash
  gpio: Correct kernel-doc inconsistency
  gpio: pci-idio-16: Use irqchip template
  gpio: pcie-idio-24: Use irqchip template
  gpio: 104-idio-16: Use irqchip template
  gpio: 104-idi-48: Use irqchip template
  gpio: 104-dio-48e: Use irqchip template
  gpio: ws16c48: Use irqchip template
  gpio: omap: improve coding style for pin config flags
  gpio: dln2: Use irqchip template
  gpio: sch: Add a blank line between declaration and code
  gpio: sch: changed every 'unsigned' to 'unsigned int'
  gpio: ich: changed every 'unsigned' to 'unsigned int'
  ...

This is hopefully the final piece of the crazy puzzle with random.h
dependencies.

And by "hopefully" I obviously mean "Linus is a hopeless optimist".

Reported-and-tested-by: Daniel Díaz <daniel.diaz@linaro.org>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

On architectures defining _HAVE_ARCH_IPV6_CSUM, we get
csum_ipv6_magic() defined by means of arch checksum.h headers. On
other architectures, we actually need to include net/ip6_checksum.h
to be able to use it.

Without this include, building with defconfig breaks at least for
s390.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Fixes: 4cb47a86

 ("tunnels: PMTU discovery support for directly bridged IP packets")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The msg_zerocopy test pins the sender and receiver threads to separate
cores to reduce variance between runs.

But it hardcodes the cores and skips core 0, so it fails on machines
with the selected cores offline, or simply fewer cores.

The test mainly gives code coverage in automated runs. The throughput
of zerocopy ('-z') and non-zerocopy runs is logged for manual
inspection.

Continue even when sched_setaffinity fails. Just log to warn anyone
interpreting the data.

Fixes: 07b65c5b

 ("test: add msg_zerocopy test")
Reported-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Acked-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Nicolas reported the following oops:

[ 1521.392541] BUG: kernel NULL pointer dereference, address: 00000000000000c0
[ 1521.394189] #PF: supervisor read access in kernel mode
[ 1521.395376] #PF: error_code(0x0000) - not-present page
[ 1521.396607] PGD 0 P4D 0
[ 1521.397156] Oops: 0000 [#1] SMP PTI
[ 1521.398020] CPU: 0 PID: 22986 Comm: kworker/0:2 Not tainted 5.8.0-rc4+ #109
[ 1521.399618] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1ubuntu1 04/01/2014
[ 1521.401728] Workqueue: events mptcp_worker
[ 1521.402651] RIP: 0010:mptcp_subflow_create_socket+0xf1/0x1c0
[ 1521.403954] Code: 24 08 89 44 24 04 48 8b 7a 18 e8 2a 48 d4 ff 8b 44 24 04 85 c0 75 7a 48 8b 8b 78 02 00 00 48 8b 54 24 08 48 8d bb 80 00 00 00 <48> 8b 89 c0 00 00 00 48 89 8a c0 00 00 00 48 8b 8b 78 02 00 00 8b
[ 1521.408201] RSP: 0000:ffffabc4002d3c60 EFLAGS: 00010246
[ 1521.409433] RAX: 0000000000000000 RBX: ffffa0b9ad8c9a00 RCX: 0000000000000000
[ 1521.411096] RDX: ffffa0b9ae78a300 RSI: 00000000fffffe01 RDI: ffffa0b9ad8c9a80
[ 1521.412734] RBP: ffffa0b9adff2e80 R08: ffffa0b9af02d640 R09: ffffa0b9ad923a00
[ 1521.414333] R10: ffffabc4007139f8 R11: fefefefefefefeff R12: ffffabc4002d3cb0
[ 1521.415918] R13: ffffa0b9ad91fa58 R14: ffffa0b9ad8c9f9c R15: 0000000000000000
[ 1521.417592] FS:  0000000000000000(0000) GS:ffffa0b9af000000(0000) knlGS:0000000000000000
[ 1521.419490] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1521.420839] CR2: 00000000000000c0 CR3: 000000002951e006 CR4: 0000000000160ef0
[ 1521.422511] Call Trace:
[ 1521.423103]  __mptcp_subflow_connect+0x94/0x1f0
[ 1521.425376]  mptcp_pm_create_subflow_or_signal_addr+0x200/0x2a0
[ 1521.426736]  mptcp_worker+0x31b/0x390
[ 1521.431324]  process_one_work+0x1fc/0x3f0
[ 1521.432268]  worker_thread+0x2d/0x3b0
[ 1521.434197]  kthread+0x117/0x130
[ 1521.435783]  ret_from_fork+0x22/0x30

on some unconventional configuration.

The MPTCP protocol is trying to create a subflow for an
unaccepted server socket. That is allowed by the RFC, even
if subflow creation will likely fail.
Unaccepted sockets have still a NULL sk_socket field,
avoid the issue by failing earlier.

Reported-and-tested-by: Nicolas Rybowski <nicolas.rybowski@tessares.net>
Fixes: 7d14b0d2

 ("mptcp: set correct vfs info for subflows")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Po-Hsu Lin says:

====================
selftests: rtnetlink: Fix for false-negative return values

This patchset will address the false-negative return value issue
caused by the following:
  1. The return value "ret" in this script will be reset to 0 from
     the beginning of each sub-test in rtnetlink.sh, therefore this
     rtnetlink test will always pass if the last sub-test has passed.
  2. The test result from two sub-tests in kci_test_encap() were not
     being processed, thus they will not affect the final test result
     of this test.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

kci_test_encap() is actually composed by two different sub-tests,
kci_test_encap_vxlan() and kci_test_encap_fou()

Therefore we should check the test result of these two in
kci_test_encap() to let the script be aware of the pass / fail status.
Otherwise it will generate false-negative result like below:
    $ sudo ./test.sh
    PASS: policy routing
    PASS: route get
    PASS: preferred_lft addresses have expired
    PASS: promote_secondaries complete
    PASS: tc htb hierarchy
    PASS: gre tunnel endpoint
    PASS: gretap
    PASS: ip6gretap
    PASS: erspan
    PASS: ip6erspan
    PASS: bridge setup
    PASS: ipv6 addrlabel
    PASS: set ifalias 5b193daf-0a08-46d7-af2c-e7aadd422ded for test-dummy0
    PASS: vrf
    PASS: vxlan
    FAIL: can't add fou port 7777, skipping test
    PASS: macsec
    PASS: bridge fdb get
    PASS: neigh get
    $ echo $?
    0

Signed-off-by: Po-Hsu Lin <po-hsu.lin@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The return value "ret" will be reset to 0 from the beginning of each
sub-test in rtnetlink.sh, therefore this test will always pass if the
last sub-test has passed:
    $ sudo ./rtnetlink.sh
    PASS: policy routing
    PASS: route get
    PASS: preferred_lft addresses have expired
    PASS: promote_secondaries complete
    PASS: tc htb hierarchy
    PASS: gre tunnel endpoint
    PASS: gretap
    PASS: ip6gretap
    PASS: erspan
    PASS: ip6erspan
    PASS: bridge setup
    PASS: ipv6 addrlabel
    PASS: set ifalias a39ee707-e36b-41d3-802f-63179ed4d580 for test-dummy0
    PASS: vrf
    PASS: vxlan
    FAIL: can't add fou port 7777, skipping test
    PASS: macsec
    PASS: ipsec
    3,7c3,7
    < sa[0]    spi=0x00000009 proto=0x32 salt=0x64636261 crypt=1
    < sa[0]    key=0x31323334 35363738 39303132 33343536
    < sa[1] rx ipaddr=0x00000000 00000000 00000000 c0a87b03
    < sa[1]    spi=0x00000009 proto=0x32 salt=0x64636261 crypt=1
    < sa[1]    key=0x31323334 35363738 39303132 33343536
    ---
    > sa[0]    spi=0x00000009 proto=0x32 salt=0x61626364 crypt=1
    > sa[0]    key=0x34333231 38373635 32313039 36353433
    > sa[1] rx ipaddr=0x00000000 00000000 00000000 037ba8c0
    > sa[1]    spi=0x00000009 proto=0x32 salt=0x61626364 crypt=1
    > sa[1]    key=0x34333231 38373635 32313039 36353433
    FAIL: ipsec_offload incorrect driver data
    FAIL: ipsec_offload
    PASS: bridge fdb get
    PASS: neigh get
    $ echo $?
    0

Make "ret" become a local variable for all sub-tests.
Also, check the sub-test results in kci_test_rtnl() and return the
final result for this test.

Signed-off-by: Po-Hsu Lin <po-hsu.lin@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Although we can detect the chip revision 100% at runtime, it is useful
to specify it in the device tree compatible string too, because
otherwise there would be no way to assess the correctness of device tree
bindings statically, without booting a board (only some switch versions
have internal RGMII delays and/or an SGMII port).

But for testing the P/Q/R/S support, what I have is a reworked board
with the SJA1105T replaced by a pin-compatible SJA1105Q, and I don't
want to keep a separate device tree blob just for this one-off board.
Since just the chip has been replaced, its RGMII delay setup is
inherently the same (meaning: delays added by the PHY on the slave
ports, and by PCB traces on the fixed-link CPU port).

For this board, I'd rather have the driver shout at me, but go ahead and
use what it found even if it doesn't match what it's been told is there.

[    2.970826] sja1105 spi0.1: Device tree specifies chip SJA1105T but found SJA1105Q, please fix it!
[    2.980010] sja1105 spi0.1: Probed switch chip: SJA1105Q
[    3.005082] sja1105 spi0.1: Enabled switch tagging

Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Xin Long says:

====================
net: fix a mcast issue for tipc udp media

Patch 1 is to add a function to get the dev by source address,
which will be used by Patch 2.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

Without ub->ifindex set for ipv6 address in tipc_udp_enable(),
ipv6_sock_mc_join() may make the wrong dev join the multicast
address in enable_mcast(). This causes that tipc links would
never be created.

So fix it by getting the right netdev and setting ub->ifindex,
as it does for ipv4 address.

Reported-by: Shuang Li <shuali@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This is to add an ip_dev_find like function for ipv6, used to find
the dev by saddr.

It will be used by TIPC protocol. So also export it.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull USB/Thunderbolt updates from Greg KH:
 "Here is the large set of USB and Thunderbolt patches for 5.9-rc1.

  Nothing really magic/major in here, just lots of little changes and
  updates:

   - clean up language usages in USB core and some drivers

   - Thunderbolt driver updates and additions

   - USB Gadget driver updates

   - dwc3 driver updates (like always...)

   - build with "W=1" warning fixups

   - mtu3 driver updates

   - usb-serial driver updates and device ids

   - typec additions and updates for new hardware

   - xhci debug code updates for future platforms

   - cdns3 driver updates

   - lots of other minor driver updates and fixes and cleanups

  All of these have been in linux-next for a while with no reported
  issues"

* tag 'usb-5.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (330 commits)
  usb: common: usb-conn-gpio: Register charger
  usb: mtu3: simplify mtu3_req_complete()
  usb: mtu3: clear dual mode of u3port when disable device
  usb: mtu3: use MTU3_EP_WEDGE flag
  usb: mtu3: remove useless member @busy in mtu3_ep struct
  usb: mtu3: remove repeated error log
  usb: mtu3: add ->udc_set_speed()
  usb: mtu3: introduce a funtion to check maximum speed
  usb: mtu3: clear interrupts status when disable interrupts
  usb: mtu3: reinitialize CSR registers
  usb: mtu3: fix macro for maximum number of packets
  usb: mtu3: remove unnecessary pointer checks
  usb: xhci: Fix ASMedia ASM1142 DMA addressing
  usb: xhci: define IDs for various ASMedia host controllers
  usb: musb: convert to devm_platform_ioremap_resource_byname
  usb: gadget: tegra-xudc: convert to devm_platform_ioremap_resource_byname
  usb: gadget: r8a66597: convert to devm_platform_ioremap_resource_byname
  usb: dwc3: convert to devm_platform_ioremap_resource_byname
  usb: cdns3: convert to devm_platform_ioremap_resource_byname
  usb: phy: am335x: convert to devm_platform_ioremap_resource_byname
  ...

ovs_flow_tbl_destroy always is called from RCU callback
or error path. It is no need to check if rcu_read_lock
or lockdep_ovsl_is_held was held.

ovs_dp_cmd_fill_info always is called with ovs_mutex,
So use the rcu_dereference_ovsl instead of rcu_dereference
in ovs_flow_tbl_masks_cache_size.

Fixes: 9bf24f59

 ("net: openvswitch: make masks cache size configurable")
Cc: Eelco Chaudron <echaudro@redhat.com>
Reported-by:  <syzbot+c0eb9e7cdde04e4eb4be@syzkaller.appspotmail.com>
Reported-by:  <syzbot+f612c02823acb02ff9bc@syzkaller.appspotmail.com>
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This reverts commit 71130f29.

In commit 71130f29 ("vxlan: fix tos value before xmit") we want to
make sure the tos value are filtered by RT_TOS() based on RFC1349.

       0     1     2     3     4     5     6     7
    +-----+-----+-----+-----+-----+-----+-----+-----+
    |   PRECEDENCE    |          TOS          | MBZ |
    +-----+-----+-----+-----+-----+-----+-----+-----+

But RFC1349 has been obsoleted by RFC2474. The new DSCP field defined like

       0     1     2     3     4     5     6     7
    +-----+-----+-----+-----+-----+-----+-----+-----+
    |          DS FIELD, DSCP           | ECN FIELD |
    +-----+-----+-----+-----+-----+-----+-----+-----+

So with

IPTOS_TOS_MASK          0x1E
RT_TOS(tos)		((tos)&IPTOS_TOS_MASK)

the first 3 bits DSCP info will get lost.

To take all the DSCP info in xmit, we should revert the patch and just push
all tos bits to ip_tunnel_ecn_encap(), which will handling ECN field later.

Fixes: 71130f29

 ("vxlan: fix tos value before xmit")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Acked-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>