01:42 <Nexrem> apritzel: Thanks for the response, in fact it is unnecessary but I was throwing things at the wall. Despite that the pin is still not getting pulled as high as the others I have. As per the gpio documentation mode 17's bit field is active low, pull up and nothing else. This works fine if I switch the pin. I thought it might be hardware damage, but the pin works just fine if I set it up as output or if I pull it down. Its just the pull up tha

05:50 <wens> jernej: MediaTek has something similar, but using it kind of makes reasoning about DRM state change harder, i.e. when atomic_update returns, the changes could still be in the queue.

08:55 <apritzel> jernej: sounds like saving pennies on the hardware side, to make software developers' life harder :-(

09:00 <apritzel> ... and to answer my own question from yesterday: you can download a recovery image from the Teclast website, the "Product ID Number" of the Teclast P85T tablet is "P3M2"

12:39 <apritzel> if one can trust the DT in this image, it looks like AW at least fixed the huge PCI problem from the H6 in the A523: there is now a 16 MB config space window, 16 MB I/O window and 112 MB MMIO32 window

12:39 <apritzel> though the single PCI line shares a combo PHY with the USB 3.0 controller :-(

12:44 <maz> apritzel: that's pretty common (my amlogic boards have the same issue). I moved USB3.0 over PCIe!

16:15 <jernej> wens: meson driver (amlogic) has this implemented, so I don't see a reason why we couldn't use. TBH, BSP driver always executed writes at vsync time, at least from DE2 onwards.

16:17 <jernej> to be more specific, register writes are meant to be done in start delay period, which is programmed in TCON and starts at vsync

16:18 <jernej> since DE3.3 scaler unit accepts values only in that period, I can't see any reason why we can't use RCQ for it.

16:19 <jernej> IIRC there is async flip capability flag, which wouldn't be exposed then

20:10 <jernej> did AW just limited access to documentation and sdks on open source portal?

20:10 <jernej> in the past I saw a couple of SDKs and a few documents, but now I don't see anything