[Intel-gfx] Skylake Linux 4.2 FIFO underruns

[Daniel Drake]

Hi,

Using Linux 4.2 on a new Skylake laptop, I frequently see the display
go momentarily black, with a coinciding message in the logs:

[drm:intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun

Testing linus master, the problem is gone. Does anyone know what
commit might have fixed this? I would like to backport to 4.2.

I already tried cherry-picking the patches that mention skl that apply
trivially, and while that fixed some GPU hangs I was seeing, the FIFO
underrun is still present with 4.2 plus:

 drm/i915/skl: Retrieve the Rpe value from Pcode
 drm/i915/skl: WaIgnoreDDIAStrap is forever, always init DDI A
 drm/i915/skl WaDisableSbeCacheDispatchPortSharing
 drm/i915:skl: Add WaEnableGapsTsvCreditFix
 drm/i915/skl: Restrict the ring frequency table programming to SKL
 drm/i915/skl: Ring frequency table programming changes
 drm/i915/skl: Don't warn if reading back DPLL0 is disabled
 drm/i915: Move WaBarrierPerformanceFixDisable:skl to skl code from chv code
 drm/i915/skl: Don't try to store the wrong central frequency
 drm/i915/skl: Remove unnecessary () used with abs_diff()
 drm/i915/skl: Remove unnecessary () used with div_u64()
 drm/i915/skl: Factor out computing the DPLL paramaters from the dividers
 drm/i915/skl: Use a more idomatic early return
 drm/i915/skl: Propagate the error if we fail to find a suitable DPLL divider
 drm/i915/skl: Display the WRPLL frequency we couldn't accomodate when failing
 drm/i915/skl: Make sure to break when not finding suitable PLL dividers

Thanks
Daniel