(04-19-2016, 11:30 AM)tkaiser Wrote: Did you ever had a look whether it's 'just' an issue like it was with Ethernet on the 2GB variant?
We're currently dealing with Allwinner's BSP kernel and do know that we already ran in issues with obviously untested 32-bit code (the aforementioned Ethernet issue). Could be the same with the display: EDID/HDMI contains an integer overflow, you connect one of your 3 displays --> the driver crashes and no signal on the HDMI port until next reboot (and crash )
Did you had a look already whether a driver crashes in your situation?
[ 28.711244] libphy: 1c30000.eth: probed
[ 28.715547] sunxi-gmac 1c30000.eth eth0: eth0: PHY ID 001cc915 at 0 IRQ poll (1c30000.eth-0:00)
[ 28.771985] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
[ 29.420764] libphy: 1c30000.eth: probed
[ 29.425065] sunxi-gmac 1c30000.eth eth0: eth0: PHY ID 001cc915 at 0 IRQ poll (1c30000.eth-0:00)
[ 29.480764] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
[ 34.339313] libphy: 1c30000.eth-0:00 - Link is Up - 1000/Full
[ 34.345621] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[ 378.466140] hdmi_clk_enable_prepare()L161
As you can see it was sitting idle till I hooked up the hdmi. Literally thats the only thing I see come over the serial console, is there anything else I should try/check to beat on it?
Disconnect and reconnect nets me;
[ 668.940193] hdmi_clk_disable_prepare()L176
[ 673.692845] hdmi_clk_enable_prepare()L161