On Saturday 19 August 2017 08:21 PM, Nishanth Menon wrote:
On 14:42-20170819, Matthijs van Duin wrote:
On 17 August 2017 at 18:34, Matthijs van Duin matthijsvanduin@gmail.com wrote:
See below for a forward-port of this patch. I haven't yet had a chance to test it.
I just had a chance to test it.
It doesn't work. The resulting kernel hangs early during boot without any error message.
So... what did I do wrong?
is'nt there like a dt way of doing things? http://git.ti.com/ti-linux-kernel/ti-linux-kernel/blobs/ti-linux-4.9.y/arch/...
I think. but that said, I am not sure if the configuration needs to be done in u-boot first.
Not necessary in u-boot.
Lokesh, do you see something wrong here: http://pastebin.ubuntu.com/25347386
Nothing. I just booted my omap5-uevm with this change and can see that input clock is changed to sys_clkin. http://pastebin.ubuntu.com/25360922/
Matthijs, Any chance you can enable earlyprintk or earlycon in your case and see where it is failing?
Thanks and regards, Lokesh
Context: In an IRC conversation, Nishanth Menon dug up that the ABE PLL issue is indeed known and the workaround is to use sys_clk as its input clock instead of using the 32 kHz clock:
http://gitorious.ti.com/lcpd-agross/omapdrm/commit/b55d24d8b7bcf7bc53261b82d...
Assuming this indeed fixes the problem, it seems like a good workaround for now. See below for a forward-port of this patch. I haven't yet had a chance to test it.
Perhaps in the future we might revisit this issue if we want to try entering an ultra low power mode in which the main osc is disabled and all clocks derive from the ABE PLL (apparently originally intended to allow music playback with minimal power consumption), but doing so would be quite challenging anyway due to the constraints imposed by this mode.
Matthijs