TODO

-check kernel blocking still is done properly
-merge sun7i nand improvements
-fix emac phy setup when cable gets inserted after boot (sun7i improvements)
 -drop link wait ?
 -the sun7i code dump has many interesting improvements to the emac code, also
  see emac-sun7i-diff-wip (which is incomplete)
-get wifi on auxtek-t004 to work

-look into suspend/resume
 -fails due to mmc card power off too soon causing rootfs errors

-look into using
 https://github.com/ssvb/xf86-video-sunxifb/commits/master

-look into ir support

-install to HD
-install to NAND

-lower fbmem, disable double buffer by default. Freeing 10 MB (useful on
 low mem devices mostly)

-try 16bpp framebuffer -> better performance ?

-rebase rootfs-resize patch, mail to ctyler

-rtc-sun4i, make BACKUP_PWM a runtime check based on ic ver

-sound/soc/sun4i improvements:
 -remove dead ENFORCE_RATES code from sound/soc/sun4i
 -remove dead sun4i_sndhdmi_startup / sun4i_sndhdmi_shutdown from
  sound/soc/sun4i/hdmiaudio/sun4i-sndhdmi.c
 -remove bogus input code for hdmi / spdif ?
 -add full-duplex support to audio-codec + i2s code (take from a13/a10s/a20
  source dumps)
 -according to a10_source_1.5 sun4i hdmi has no rx/tx ctrl just like sun7i

-take photos of devices + create wiki pages so people can easily identify
 which device they have

-mali clock div is too low in many fex files when not using PLL6 for PLL4