05:27 <junari> I'm trying to add H728 support to u-boot, right now I only checked the memory card

05:27 <junari> https://gist.github.com/iuncuim/bed7c913dc48715ccaa27f1767c70754

05:28 <junari> oh, not memory card, but register map

05:28 <junari> The next thing to check is clk I think

05:30 <junari> If anyone has an opportunity to do something, join in and let me know about your work

10:48 <apritzel> junari, that's great, any contributions are welcome!

10:48 <apritzel> I have some branch, but it's not cleaned up, and without DRAM it's less useful

10:50 <apritzel> just note that it's an NCAT2 chip,so there shou

10:50 <apritzel> ...shouldn't be any need for a new memory map

10:53 <apritzel> look at the A133 patches for a guide what you'd need: https://github.com/apritzel/u-boot/commits/a133/

10:54 <apritzel> junari: if you have any code for DRAM that would be best of course

10:55 <apritzel> i think MasterR3C0RD and jernej had some code already?

14:08 <junari> apritzel: Thanks for the info, I'll take my time so I don't do double work

16:37 <apritzel_> brodertuck: sorry would love to but dont have the files on me. Maybe warpme has some binaries to share?

17:08 <warpme> brodertruck: how i can help you : -) ?

17:10 <warpme> apritzel : sorry not retuning to you with q about gmac-200 on mine x96pro+ (and t527): on x96pro+ my working gmac-200 is backport of bsp. Interesting is that it works with generic stmmac phy (so no maxio phy driver is used)

17:12 <apritzel> brodertuck: nevermind, found something: https://paste.c-net.org/VernonRemarry

17:12 <apritzel> contains all the files plus a Readme

17:14 <apritzel> warpme: yeah, saw that in your repo. I still have the feeling that GMAC-200 is not needed, its probably just some platform setup missing

17:15 <apritzel> that's good news about the PHY, which means it's nothing special

17:16 <apritzel> will have a closer look when I am back with my beloved boards...

17:18 <warpme> on t527 however i have interesting situation (with my bsp port of gmac-200 + yt8531 phy): in dhcp ip req, board eth port sends bcst to dhcp server, server replays with lease but gmac-200 receives nothing. It looks to me like yt8531 not correctly communicating with gmac-200. But tx works. issue seems to be only with rx. I recall yt8531 is really specific with correct reset procedure. Maybe this is issue?

17:32 <Soupborsh> Hello, I can compile working boot0 from sources. Branching to 0xffff0020 hungs up it.

17:32 <Soupborsh> I wrote a little assembly program and branching to that address there hungs up too, but if I branch to 0xffff2858 it works(lr in this program has this value).

17:32 <Soupborsh> sp is 0x00005e08.

17:33 <Soupborsh> Executing 0xffff2858 from FEL changes nothing, likely executes FEL mode. Executing 0xffff0020 seems to enter FEL after some time, it throws -7 usb error before that.

17:33 <Soupborsh> I found some related bsp src https://github.com/mirkerson/c600/tree/master/brandy/u-boot-2014.07.

17:33 <Soupborsh> I think PocketBook's repo is based on this one.

17:33 <Soupborsh> Also init_DRAM function is defined in uboot_b288/sunxi_spl/dram/sun8iw10p1/dram/libdram.o. It does not have source files unfortunately.

17:33 <Soupborsh> this c600 bsp repo has source files for it but it lacks some functions such as dram_enable_all_master.

17:35 <Soupborsh> Branching to 0xffff2858 in boot0 result in hang up.

17:36 <Soupborsh> With setting sp to 0x00005e08

17:39 <warpme> apritzel : i looked (for sec) over cpu dvfs on a523 - and i see bsp has cpu clock in ccu - but mainline seems not. is this because wip or maybe idea is to control cpu clocks exclusively via psci?

18:57 <Soupborsh> How to make U-Boot to enter FEL mode?

18:58 <Soupborsh> Sorry, I need it to enter to always command prompt

19:14 <Soupborsh> It it BSP U-Boot sources. Or maybe how to make it enter command prompt instead of reseting cpu when it fails(it fails to initialize nand)? I execute it from FEL.

20:38 <hlauer> <apritzel> "re: USB> the OTG port is a..." <- Thanks, in the DT EHCI and OHCI are listed separately, so the USB-A are counted as 4 roots

20:41 <hlauer> Have to check the power consumption of V1.0 and V1.1. OTOH the proprietary image lifts the 900mA limitation already in their U-Boot - so I assume, BPI people know about the power limitation

20:41 <BroderTuck> If the kernel_b288 bsp can be trusted, the B288 has SRAM A1 (16K) and SRAM C (36K) back-to-back for a total of 52K

22:37 <apritzel> warpme: the CPU clocks are in a separate CCU. I have some sketch code, but something was missing, and it isn't essential for now

22:38 <apritzel> what's more important than pushing out more code is actually review, that's completely missing by now :-(

22:39 <apritzel> clocks and pinctrl are essential for *every* other device, so they have to come first

22:40 <apritzel> then the rest of the drivers can be done independently and in parallel, by multiple people

22:55 <apritzel> hlauer: so the question is what we do with the patch? it seems like it's wrong now, since PH23 *is* the VUSB supply, and doesn't hurt on v1.0

22:57 <hlauer> Yes, it's wrong. Is there any formal way to retract it besides the Email I send already?

23:00 <hlauer> Oh, and best of course would be to detect the situation (it's available in the axp registers) and spit out an error instead of enabling PH23 forcing the board into reset...

23:47 <apritzel> are the AXP registers different between v1.0 and v1.1? Or are there any other known differences between the two we could detect?

23:50 <apritzel> NC pins are sometimes detectable, by using the internal pull-up/pull-down registers, if the other case (GPIO connected to the regulator) plays along

23:50 <apritzel> s/registers/resistors/