10:24 <indy> hi all, is there undergoing work on a527 and a733 upstreaming (u-boot/kernel)?

10:25 <indy> primarily a733

10:34 <dlan> indy: for u-boot, I've just pushed a branch, only work for SD and eMMC, pls check this https://github.com/dlan17/u-boot/blob/allwinner/A733/README.A733.md

10:35 <dlan> for kernel, apritzel has sent out patches for pinctrl, pmu/regulator.

10:41 <apritzel> indy: for A527, the basic support is mainline in both projects, with ongoing efforts to fill the gaps. Find more details here: https://linux-sunxi.org/A523

11:23 <indy> dlan, do you have working jtag? :)

11:55 <dlan> indy: no, I haven't tried jtag

12:08 <paulk> let me know if you need it, I'm preparing a patch series with sunxi support for openocd

12:09 <paulk> (although it's on hold for now, due to -EBUSY)

12:09 <paulk> I don't have a733 yet but it shouldn't be too hard to figure out

12:15 <indy> paulk, -ENOTENOUGHTIME is very frequent obstacle :)

12:16 <indy> paulk, do you have something somewhere?

12:17 <paulk> I can push my current stuff on github

12:20 <paulk> indy: https://github.com/paulkocialkowski/openocd-config-sunxi

12:23 <paulk> it's probably using ADIv6 with SoC-600

12:23 <indy> paulk, great

12:47 <indy> paulk, are jtag pins enable by detaul?

12:48 <indy> enabled

12:56 <paulk> indy: typically not, what I do is load the jtag-loop via fel to pinmux the sdcard pins to jtag

12:57 <paulk> the jtag pins may also be available on dedicated pins

12:57 <paulk> but AFAIK they all need explicit pinmux config

18:17 <electricworry> I'm trying to get KGDB working with an allwinner h618 with earlycon, but unable to do so. Should this be straightforward? If I run with `earlycon kgdboc=ttyS0,115200 kgdbwait` it works beautifully, but if I run `earlycon kgdboc_earlycon kgdboc=ttyS0,115200 kgdbwait` it won't work despite earlycon working fine. The uart is snps,dw-apb-uart which seems to be earlycon compatible.

18:48 <electricworry> Oh! I got it working. I had to add `=uart,mmio32,0x5000000` to the earlycon. That got it working. Now able to debug from before any other drivers load.

18:55 <apritzel> electricworry: that's odd, but it worked with just earlycon for the actual earlycon output? Does kgdboc require a fully qualified earlycon line then?

18:56 <paulk> maybe the board doesn't have serial0, IIRC that's how it works without an explicit base address

19:24 <electricworry> That's correct, the earlycon (without any params) was happily writing to the serial port and then even said "KGDB: Waiting for connection from remote gdb..." but connecting with GDB (via agent-proxy) just didn't do anything.

19:25 <electricworry> The board does have serial0 alias in the DT which points to uart0. So I'm flummoxed. But very happy it's working now.

19:33 <electricworry> If I can bother with another question; now KGDB is working, is it easy to get GDB to interrupt when running? agent-proxy doesn't seem to be successfully sending SysRq-g over serial even although CONFIG_MAGIC_SYSRQ_SERIAL=y. I can successfuly invoke SysRq-g with a keyboard connected to USB.