Messages in this thread | | | From | Sven Van Asbroeck <> | Date | Mon, 29 Jun 2020 09:40:38 -0400 | Subject | Re: [PATCH v4 2/2] ARM: imx6plus: enable internal routing of clk_enet_ref where possible |
| |
Hi Fabio,
On Mon, Jun 29, 2020 at 9:10 AM Fabio Estevam <festevam@gmail.com> wrote: > > I have tested this series on an imx6qp sabresd and unfortunately, it > breaks Ethernet as I can no longer get an IP address from the DHCP > server.
Thank you for testing this out on a different platform !
I had a look at how things are done in the Freescale fork of the kernel (5.4.24_2.1.0) and I noticed that this kernel has almost the same behaviour as this proposed patch: the GPR5 bit is _always_ set on a plus. The code does not check how the enet clock is generated.
https://source.codeaurora.org/external/imx/linux-imx/tree/arch/arm/mach-imx/mach-imx6q.c?h=rel_imx_5.4.24_2.1.0&id=babac008e5cf168abca1a85bda2e8071ca27a5c0#n269
Now, I'm assuming that the sabresd-plus can run on the Freescale kernel fork. The GPR5 bit will always be set there.
So why won't mainline work with this patch? What have I overlooked?
I'm sure you've checked that sabresd ethernet works ok on mainline even without this patch?
| |