[Feature] Skylake support in Perf

Has support landed upstream or should we retarget to 15.10?

It hasn't landed upstream yet and retarget to 15.10

Part of perf event are merged into maintainer's tree while another are waiting for SDM

This will be implemented in 4.3 kernel, pleas back port the following patch.
40997d6 perf report: Display cycles in branch sort mode
a18b027 perf top: Add branch annotation code to top
f8f4aae perf annotate: Finally display IPC and cycle accounting
30e863b perf annotate: Compute IPC and basic block cycles
57849998 perf report: Add processing for cycle histograms
d495763 perf report: Add infrastructure for a cycles histogram
98df858 perf report: Add flag for non ANY branch mode
0e332f0 perf tools: Add support for cycles, weight branch_info field
8c4fe70 perf/x86/intel: Use 0x11 as extra reg test value
47732d8 perf/x86: Make merge_attr() global to use from perf_event_intel
90405aa perf/x86/intel/lbr: Limit LBR accesses to TOS in callstack mode
e057336 perf/x86/intel/lbr: Use correct index to save/restore LBR_INFO with call stack
9a92e16 perf/x86/intel: Add Intel Skylake PMU support
425507f perf/x86/intel/lbr: Optimize v4 LBR unfreezing
0f29e57 perf/x86/intel: Move PMU ACK to after LBR read
d8020be perf/x86/intel: Handle new arch perfmon v4 status bits
50eab8f perf/x86/intel/lbr: Add support for LBRv5
71ef3c6 perf: Add cycles to branch_info
b83ff1c x86: Add new MSRs and MSR bits used for Intel Skylake PMU support
a7b58d2 perf/x86/intel/lbr: Allow time stamp for free running PEBSv3
2f7ebf2 perf/x86/intel: Add support for PEBSv3 profiling
a94cab2 perf/x86: Add a native_perf_sched_clock_from_tsc()

This bug was fixed in the package linux - 4.2.0-10.11

---------------
linux (4.2.0-10.11) wily; urgency=low

  [ Andy Whitcroft ]

  * Release Tracking Bug
    - LP: #1495208
  * No change rebuild to avoid issues with leading zero bytes on
    internal module signing key. (see launchpad bug 1494943)

  [ Ming Lei ]

  * [Config] SERIAL_8250_DW=y for arm64

  [ Upstream Kernel Changes ]

  * drm/i915: Move WaBarrierPerformanceFixDisable:skl to skl code from chv
    code
    - LP: #1484486
  * drm/i915/gen8: Add infrastructure to initialize WA batch buffers
    - LP: #1484486
  * drm/i915/gen8: Re-order init pipe_control in lrc mode
    - LP: #1484486
  * drm/i915/gen8: Add WaDisableCtxRestoreArbitration workaround
    - LP: #1484486
  * drm/i915/gen8: Add WaFlushCoherentL3CacheLinesAtContextSwitch
    workaround
    - LP: #1484486
  * drm/i915: Bail out early if WA batch is not available for given Gen
    - LP: #1484486
  * drm/i915/gen8: Add WaClearSlmSpaceAtContextSwitch workaround
    - LP: #1484486
  * drm/i915: Update WaFlushCoherentL3CacheLinesAtContextSwitch
    - LP: #1484486
  * drm/i915/gen9: Implement WaDisableKillLogic for gen 9
    - LP: #1484486
  * drm/i915: Enable WA batch buffers for Gen9
    - LP: #1484486
  * drm/i915/gen9: Add WaDisableCtxRestoreArbitration workaround
    - LP: #1484486
  * drm/i915: Update wa_ctx_emit() macro as per kernel coding guidelines
    - LP: #1484486
  * drm/i915/gen9: Add WaFlushCoherentL3CacheLinesAtContextSwitch
    workaround
    - LP: #1484486
  * drm/i915/gen9: Add
    WaSetDisablePixMaskCammingAndRhwoInCommonSliceChicken
    - LP: #1484486
  * drm/i915:skl: Add WaEnableGapsTsvCreditFix
    - LP: #1484486
  * drm/i915/skl: revert duplicated WaBarrierPerformanceFixDisable:skl
    - LP: #1484486
  * drm/i915/skl: Don't expose the top most plane on gen9 display
    - LP: #1484486
  * drm/i915/skl WaDisableSbeCacheDispatchPortSharing
    - LP: #1484486
  * drm/i915/skl: WaIgnoreDDIAStrap is forever, always init DDI A
    - LP: #1484486
  * drm/i915: reduce indent in i9xx_hpd_irq_handler
    - LP: #1484531
  * drm/i915: reduce duplicate conditions in i9xx_hpd_irq_handler
    - LP: #1484531
  * drm/i915: reduce indent in intel_hpd_irq_handler
    - LP: #1484531
  * drm/i915: group all hotplug related fields into a new struct in
    dev_priv
    - LP: #1484531
  * drm/i915: add for_each_hpd_pin to iterate over hotplug pins
    - LP: #1484531
  * drm/i915: simplify conditions for skipping the 2nd hpd loop iterations
    - LP: #1484531
  * drm/i915: put back the indent in intel_hpd_irq_handler
    - LP: #1484531
  * drm/i915: merge the two hpd loops in intel_hpd_irq_handler to one
    - LP: #1484531
  * drm/i915: simplify condition for digital port
    - LP: #1484531
  * drm/i915: abstract away platform specific parts from hpd handling
    - LP: #1484531
  * drm/i915: Handle HPD when it has actually occurred
    - LP: #1484531
  * drm/i915: Set power domain for DDI-E
    - LP: #1484531
  * drm/i915: Set alternate aux for DDI-E
    - LP: #1484531
  * drm/i915/skl: enable DDI-E hotplug
    - LP: #1484531
  * drm/i915/bxt: fix DDI PHY vswing scale value setting
    - LP: #1494163
  * drm/i915/skl: Buf...