From: Simon Goldschmidt Date: Mon, 29 Apr 2019 21:50:21 +0200 Subject: [U-Boot] [PATCH] dts: arm: socfpga: fix socfpga_de10_nano console In-Reply-To: <8bec4cd1-435b-3d86-73c9-4c5e2e22b226@denx.de> References: <20190429183227.8210-1-simon.k.r.goldschmidt@gmail.com> <0bdbacc0-a3b3-909d-dd25-f3782599942e@gmail.com> <8bec4cd1-435b-3d86-73c9-4c5e2e22b226@denx.de> Message-ID: List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit To: u-boot@lists.denx.de Am 29.04.2019 um 21:23 schrieb Marek Vasut: > On 4/29/19 8:53 PM, Simon Goldschmidt wrote: >> Am 29.04.2019 um 20:33 schrieb Marek Vasut: >>> On 4/29/19 8:32 PM, Simon Goldschmidt wrote: >>>> Booting this board failed as the initial console isn't found since >>>> commit c402e8170245 ("dts: arm: socfpga: merge gen5 devicetrees from >>>> linux") >>>> >>>> The uart0 devicetree entry was missing "clock-frequency = <100000000>:" >>>> since that commit >>>> >>>> Fixes: c402e8170245 ("dts: arm: socfpga: merge gen5 devicetrees from >>>> linux") >>>> Reported-by: rafael mello >>>> Signed-off-by: Simon Goldschmidt >>>> --- >>>> >>>>   arch/arm/dts/socfpga_cyclone5_de10_nano.dts | 1 + >>>>   1 file changed, 1 insertion(+) >>>> >>>> diff --git a/arch/arm/dts/socfpga_cyclone5_de10_nano.dts >>>> b/arch/arm/dts/socfpga_cyclone5_de10_nano.dts >>>> index b620dd8dda..4be4083941 100644 >>>> --- a/arch/arm/dts/socfpga_cyclone5_de10_nano.dts >>>> +++ b/arch/arm/dts/socfpga_cyclone5_de10_nano.dts >>>> @@ -77,6 +77,7 @@ >>>>   }; >>>>     &uart0 { >>>> +    clock-frequency = <100000000>; >>>>       u-boot,dm-pre-reloc; >>>>   }; >>>> >>> Applied, thanks. >> >> Wow, that was fast! >> >>> >>> While at it, can we do something about Gen5 clock driver ? >> >> Which Gen5 clock driver, haha? >> >> I might find the time. Would you think the A10 driver would be a good >> point to start? > > For a read-only clock driver, based off DT, yes. Well, that would be a start, using the handoff values from Quartus. Would it make sense to convert the qts header output for gen5 to a dtsi format like A10? Not only the clock settings, but all information living in the qts directory would be better off in a devicetree. And in addition, different FPGA configurations require different pinmux/iocsr configurations, which is a thing we haven't covered, yet. I don't want to update SPL just because an FPGA image decides to use loan-IO instead of a HPS GPIO... However, the bindings from A10 handoff dtsi don't seem in a state to just copy them. (How does that handof dtsi get included anyway?) Regards, Simon