Hi Diego, On 2024-02-13 at 06:59:05 +0100, Diego Roversi <die...@tiscali.it> wrote: > On Mon, 12 Feb 2024 14:58:22 +0100 > Frank Oltmanns <fr...@oltmanns.dev> wrote: > >> Could you please send the content of >> /sys/kernel/debug/clk/clk_summary >> for >> - 6.4 or earlier, >> - 6.5 (assuming you can log in via ssh/serial console) >> - 6.6 (same) > > I've not tested 6.6, but only 6.7.2 with and without CLK_SET_RATE_PARENT. > I've some other problem with 6.6 kernels atm, so I can't check them. > > Right now I can give you clk_summary from this kernel: > > 6.1.0-17-arm64 - standard debian kernel > 6.5.13 - compiled from mainline repo > 6.7.2 - compiled from mainline repo > 6.7.2-dirty - compiled from mainline repo without CLK_SET_RATE_PARENT
Thank you so much for the data. It shows that PLL-MIPI is set to the following rates: - 6.5.13: 9.702 GHz. - 6.7.2: 7.7184 GHz - 6.7.2-dirty: 764.4 MHz So the kernel tries to set a rate that PLL-MIPI does not support. Adding an upper (and while we're at it also lower) limit should be the real fix here. That the removal of the CLK_SET_RATE_PARENT flag fixes the issue for you is just a lucky coincidence. Fortunately, I already have a patch series [1] in the works, that also covers this part (thanks to Jernej's persistence and foresight). The relevant patches are PATCH 3 and 4, but they do not apply cleanly without patch 1 and 2 (but those have already been reviewed by Jernej). There are some open remarks about the patches 3 and 4, but those are more of an architectural nature and shouldn't affect the functionality. Therefore, could you please try to apply patches 1 through 4 to an otherwise clean 6.7 kernel to see if the TERES-I works then. (Actually, you could apply the whole series, because PATCH 5 is pinephone-specific and PATCH 6 touches the GPU and is unrelated to your specific issue.) I'd very much appreciate if you could report back here with your results including the content of /sys/kernel/debug/clk/clk_summary. [1]: https://lore.kernel.org/all/20240205-pinephone-pll-fixes-v2-0-96a46a2d8...@oltmanns.dev/ Cheers, Frank > > Also I noticed that in 6.5.13 serial port have some issue, may be a clock > problem too. I need to make some more tests, because it looks that some time > the > problem it's still there if I reboot in 6.1.0 kernel. But no problem if I > poweroff, before booting with the same kernel. Unfortunately, I have no explanation for that. > > > Regards, > Diego. -- You received this message because you are subscribed to the Google Groups "linux-sunxi" group. To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscr...@googlegroups.com. To view this discussion on the web, visit https://groups.google.com/d/msgid/linux-sunxi/87cysvld27.fsf%40oltmanns.dev.
