[OpenWrt-Devel] [PATCH 2/4] ipq40xx: fix sleep clock

[Павел]

Hey,
I believe this patch can be merged as is. According to Sricharan R. [1]:
"It is [sleep clk] derived from a 48M wifi refclk

48M wifi ref clk -> [/2 divider] -> [/750 divider] -> sleep_clk (32000)"


[1] https://patchwork.kernel.org/comment/22721613/

пт, 17 мая 2019 г., 11:16 Павел <be.dissent at gmail.com>:

>
>
> чт, 16 мая 2019 г., 15:22 Sven Eckelmann <sven at narfation.org>:
>
>> On Tuesday, 14 May 2019 15:42:18 CEST Pavel Kubelun wrote:
>> > +--- a/arch/arm/boot/dts/qcom-ipq4019.dtsi
>> > ++++ b/arch/arm/boot/dts/qcom-ipq4019.dtsi
>> > +@@ -141,9 +141,9 @@
>> > +       };
>> > +
>> > +       clocks {
>> > +-              sleep_clk: sleep_clk {
>> > ++              sleep_clk: gcc_sleep_clk_src {
>> > +                       compatible = "fixed-clock";
>> > +-                      clock-frequency = <32768>;
>> > ++                      clock-frequency = <32000>;
>> > +                       #clock-cells = <0>;
>> > +               };
>>
>> On Thursday, 16 May 2019 13:18:14 CEST Павел wrote:
>> [...]
>> > > And maybe some of these guys also know how to find the ipq40xx clock
>> > > controller reference or hardware reference. Because I was only able to
>> > > verify
>> > > for IPQ8072 that it had a 32.768 KHz sleep clock. But the
>> > >
>> >
>> > If you are completely sure about that, then I guess that they have
>> > (un)intentionally messed with the clock in QSDK, because they state that
>> > ipq807x has the same 32000 khz crystal.
>> >
>> https://source.codeaurora.org/quic/qsdk/oss/kernel/linux-msm/tree/arch/arm64/boot/dts/qcom/qcom-ipq807x-soc.dtsi?h=eggplant#n2055
>>
>> Confidence is the wrong word. I can only state that this is written in
>> 80-YA727-13 Rev. D (IPQ8072.AP.HK07). Same for other devices like
>> IPQ8078 AP.HK02, IPQ8074 AP.HK01, ...
>>
>> But I found in the same document that they call it the "32 KHz sleep
>> clock in"
>> in one section and and in another table "32.768 KHz sleep clock input to
>> the
>> IPQ8072" (next to the name "...32K..."). So it is now to the reader to
>> find
>> out what they meant here in which reference document. So maybe they also
>> meant
>> 32.768 KHz when in the IPQ4019 Watchdog document when they wrote 32 Khz
>> sleep
>> clock... who knows.
>>
>
> Okay, here what I've found in oem firmware (Zyxel nbg6617)
>
> There are 2 clocks related to timer in oem firmware:
> 1. gcc_sleep_clk_src - 32.000
>  ^---USB3 sleep clock
>  ^---USB2 sleep clock
> 2. wifi_rtc_clk_src - 32.768
>  ^---wcnssN sleep clock
>  ^---wcnssN sleep clock
>
> In upstream we don't have wifi_rtc clk and all mentioned childs are
> connected to sleep clk.
>
> So sleep clock at 32000 rate looks like some kind of a workaround and not
> to break wifi they introduced a separate clock for that, the one that
> really represents the hardware.
>
> I'll drop the sleep clock rate change in the next version of patchset
> since we don't need that kind of workaround for now.
>
>
>> My gut feeling (sorry, not an HW guy) tell me that they are just using a
>> 32.768 KHz clock (from a standard 32.768 KHz oscillator) in all these
>> products
>> and just shortened it to 32K at some point in the document. And now
>> Gopinath
>> Sekar wrote 32000 instead of 32768. But I absolutely don't know what
>> actually
>> is there in HW.
>>
>> Kind regards,
>>         Sven
>>
>> [1]
>> https://source.codeaurora.org/quic/qsdk/oss/kernel/linux-msm/commit/?id=d92ec59973484acc86dd24b67f10f8911b4b4b7d
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.infradead.org/pipermail/openwrt-devel/attachments/20190708/35c7c04c/attachment.htm>
-------------- next part --------------
_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel