08:58 <PabloPL> mszyprow: Hi, are You going to support in future other exynos pci devices in that patchset which You did send ?

09:00 <PabloPL> for example 7420 - i saw that there are some differences, for example in calibration or amount of clocks (checked 5433 and 7420)

09:05 <mszyprow> PabloPL: not now and not in the near future...

09:06 <mszyprow> PabloPL: so feel free to add it if you need it

09:08 <PabloPL> mszyprow: Also probably there is no update related to MCT patches (it looks like they're still not accepted)? i also need them for 7420?

09:08 <mszyprow> PabloPL: they won't be accepted by maintainers :(

09:09 <mszyprow> PabloPL: I have no idea if the bootloader & trustzone does its job right on 7420 or not

09:10 <PabloPL> mszyprow: Event that https://patchwork.kernel.org/project/linux-samsung-soc/cover/20181017134159.9656-1-m.szyprowski@samsung.com/ ? Without that "not-fw-configured" patch?

09:10 <mszyprow> PabloPL: but if you see insonsistent time on various cpus (easily to notice if you enable prefixing kernel logs with uptime) then yes, you need them

09:11 <PabloPL> mszyprow: Yes i see that problem, and it was fixed by using that patches plus patch for not-fw-configured for arm64 (which i saw that will not be accepted)

09:11 <PabloPL> did look at sources of newer exynos and it looks like this "issue" could be also there :(

09:16 <mszyprow> PabloPL: in the current kernels you need only the following 3 patches: https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1781506.html https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1781500.html https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1781501.html

09:16 <mszyprow> PabloPL: the rest have been merged

09:16 <mszyprow> PabloPL: the last one has to be adapted to your boards dts

09:17 <mszyprow> PabloPL: feel free to resend those patches when you will send 7420 support ;)

09:17 <mszyprow> PabloPL: maybe if you send them as a community person not Samsung employe, someone will change his mind

09:19 <mszyprow> PabloPL: you will have to write a few unpleasant words about the crapy bootloader/tz ;)

09:20 <PabloPL> i need to also find someone with s2mps15 datasheet...

09:20 <PabloPL> which probably will not be possible :(

09:23 <PabloPL> it looks like there is a bug with range of one regulator in driver (comparing with vendor sources) - fixed it with https://github.com/PabloPL/linux/commit/9b0df7a78a391d4b1d799ff50a8b3067820549bc

09:24 <PabloPL> strange that this regulator has different value if booted from recovery vs force reboot (using power + vol down), one if them is outside of range in mainline which is causing ufs to not work

10:05 <PabloPL> mszyprow: Any way to get rid of this "fix" https://github.com/PabloPL/linux/commit/cbe70049c70de8226fef1488e959239f7b6a354a ?

10:05 <PabloPL> mszyprow: ufs probes before pmic (which is used to power ufs) and when getting DEFER it tries to disable that clock (which makes whole boot hang)

10:06 <PabloPL> mszyprow: with this "fix" applied, kernel later probes pmic and later ufs again (which works fine), so it's possible to boot from ufs

10:09 <PabloPL> hope that it'll be possible to do it without datasheet (as always...)

10:10 <mszyprow> PabloPL: the problem with clock enable/disable is probably caused by disabling the top pll that provides the clock to sclk_ufs

10:11 <mszyprow> PabloPL: please check if moving the critical flag to the pll fixes the issue

10:11 <mszyprow> PabloPL: keeping top plls as critical seems to be needed also on older exynos socs

10:11 <PabloPL> mszyprow: Thanks! will check it (when have a time)

10:12 <mszyprow> PabloPL: see for example commit 0212a0483b0a and 67f96ff7c8f0