06:32 <wens> make TF-A print messages in the debug build?

10:51 <apritzel> szemzoa: btw: you probably know that, but wait till -rc1 comes out on Sunday before sending a new version, then base on that

10:56 <szemzoa> apritzel: thanks, but I'm a littlebit lost in what should I send. The ccu-r driver is ok, the new one. There are still errors in ccu driver (ex. one RISCV reset -> clk gate). The new pinctrl driver depends on your patches, a bunch of dtbindigs is wrong, so on... :)

11:12 <apritzel> szemzoa: ah, I see, no worries, I just wanted to avoid a premature v2. I'd say sending something in like 2 weeks would be good, so we have some time to figure those things out

17:32 <BroderTuck> Am tempted to add a line to "Allwinner SoC Family" paga with 0x1821, sun8iw18p1, R328 based on a few finds.

17:41 <BroderTuck> (and 2xCortex-A7, based on https://www.cnx-software.com/2019/08/19/allwinner-r328-smart-speaker-system-on-module/ )

22:43 <apritzel> jernej: while working on the (CPU) PLL code, I realised we are pretty harsh to the DDR PLL in the DRAM code: the manual explicitly discourages clearing bit(31), instead say we should gate only

22:45 <apritzel> and I think in general we should leave the PLL alone as much as possible: it stays at the same frequency most of the time anyways, so pulling the plug and then re-locking it is pointless anyway

22:47 <apritzel> that applies to most SoCs, actually