07:04 <krzk> aalm: using dummy regulators does not mean that there is no power. In fact, DWC3 works fine "without" this power (because providing regulators is not necessary if they are are enabled anyway).

07:06 <krzk> However it would be very good to add (not necessarily rename...) proper regulators for DWC3 block as this allows removal of always-on and for example save energy in certain cases.

07:06 <krzk> So feel free to send a patch :)

07:09 <forkbomb> mszyprow: should UART on exynos4412 still use DMA on 4.17 with your two patches applied?

07:10 <mszyprow> forkbomb: sadly no, a patch for pl330 is needed to fix reported capabilities, I will send it soon

07:10 <forkbomb> ok. but without the two patches it should work, right?

07:11 <mszyprow> forkbomb: yes, it will work fine

07:11 <forkbomb> i think my u-boot port is not doing something - even without your patches it's broken. I'm just not sure what it's missing

07:13 <mszyprow> forkbomb: franky I have no idea what might be missing in your uboot

07:13 <mszyprow> forkbomb: in case of my Trats2 and Odroid U3 everything works fine

07:14 <forkbomb> ok, thanks for the confirmation

07:26 <forkbomb> oh, does the TZPC default to everything being secure-mode access only? that might explain it

07:26 <mszyprow> frankly I have no idea

07:27 <forkbomb> i'll have to check

07:27 <mszyprow> btw, have you configured muic/extcon to uart path?

07:28 <forkbomb> mszyprow: yes, it's like that on boot

07:28 <mszyprow> because at least for marget i9300 it was by default configured to open path (micro-usb not connected at all)

07:28 <mszyprow> market

07:28 <forkbomb> mszyprow: yeah. i replaced S-BOOT with u-boot SPL+actual bootloader

07:29 <forkbomb> that is, SPL + u-boot proper

07:29 <mszyprow> okay

07:29 <mszyprow> what if you disable dma?

07:29 <mszyprow> does it work right?

07:29 <forkbomb> mszyprow: yeah

07:29 <mszyprow> hmm

07:29 <mszyprow> have you checked clocks?

07:30 <mszyprow> maybe pl330 is clocked from osc or so?

07:31 <forkbomb> clocks seem fine. it's not completely broken, either

07:31 <forkbomb> small amounts of tx/rx work

07:31 <forkbomb> but if i do ls / or something, the pl330 faults

07:31 <mszyprow> small transfers use pio

07:31 <forkbomb> ah, well that would explain it

07:31 <mszyprow> dma is used only for large blocks

07:32 <mszyprow> one more thing - do you use vanilla v4.17 or some -next based on v4.17?

07:32 <mszyprow> because there was a stupid patch in -next which broke pl330

07:32 <forkbomb> yeah, i saw the discussion about that. it's vanilla 4.17, pretty much

07:33 <forkbomb> ~30 patches on top for stuff that i need to mainline eventually

07:34 <mszyprow> well, go ahead :)

07:34 <mszyprow> I'm waiting especially for the modem drivers :)

07:35 <forkbomb> heh

07:35 <forkbomb> so much to do, so little time...

07:42 <forkbomb> mszyprow: yeah, so it turns out the UARTs (and I guess most things) are in secure mode on boot, so PDMA can't access them

07:45 <snawrocki> LiquidAcid: the patch you pointed out was abandoned, after some arguments with the soc/samsung maintainer... ;)

07:46 <snawrocki> now for Odroid U3 we use "hardkernel,odroid-xu4-audio" compatible, https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/arm/boot/dts/exynos4412-odroid-common.dtsi?id=68605101460ea4c62a966b1ad3e8db90d8fbaa31

07:47 <snawrocki> but indeed documentation and Kconfig should be updated, so it is clear Odroid U3 is also supported

07:48 <snawrocki> LiquidAcid: I will prepare a patch, thanks for pointing out