ChanServ changed the topic of #linux-msm to:
ungeskriptet0 has joined #linux-msm
ungeskriptet has quit [Ping timeout: 480 seconds]
bamse has quit [Quit: Changing server]
bamse has joined #linux-msm
Daanct12 has joined #linux-msm
jhovold has joined #linux-msm
pespin has joined #linux-msm
kholk has joined #linux-msm
jhovold has quit [Quit: WeeChat 3.8]
Daanct12 has quit [Quit: WeeChat 4.0.4]
Danct12 has quit [Read error: Connection reset by peer]
jessica_24 has joined #linux-msm
Danct12 has joined #linux-msm
flto has quit [Read error: Connection reset by peer]
flto has joined #linux-msm
ungeskriptet0 has quit []
ungeskriptet0 has joined #linux-msm
pespin has quit [Remote host closed the connection]
<lumag> sboyd, could you please comment https://lore.kernel.org/linux-arm-msm/20230901081812.19121-1-krzysztof.kozlowski@linaro.org/T/ ? Which clocks should work under /clocks/?
<sboyd> lumag: I'm not aware of this at all
<sboyd> what is the question?
<lumag> sboyd, "gpio-gate-clock" doesn't have CLK_OF_DECLARE, so it doesn't work if it is just declared under /clocks/
<sboyd> CLK_OF_DECLARE macro works by looking at nodes
<lumag> The question is if it should work or not
<sboyd> The 'clocks' container node is not great
<sboyd> I did it years ago though
<sboyd> CLK_OF_DECLARE is built upon OF_DECLARE macros
<sboyd> we rely on core OF code to create device_node structures
<sboyd> I think that doesn't descend anything besides simple bus?
<sboyd> we're not going to change of_clk_init()
<lumag> That's where the fun begins. device_node instances are created, but struct device (well, struct platform_device) is not, unless there is "simple-bus" compat under /clocks
<sboyd> we don't want to expand the use of CLK_OF_DECLARE
<sboyd> it's to workaround the case that clks are needed for the timer/clockevent/clocksource of the system
<sboyd> or for the irqchip sometimes although I'm not sure anyone uses that
<lumag> ack. Then krzk's patch is fine from the clocks point of view. Thank you.
<sboyd> yes we should probably remove all 'clocks' container nodes, at least on qcom
<sboyd> it has no purpose
<lumag> sboyd, ack, thanks for the explanations!
<sboyd> lumag: no problem!