ChanServ changed the topic of #linux-msm to:
svarbanov has quit [Ping timeout: 480 seconds]
svarbanov has joined #linux-msm
marvin24 has joined #linux-msm
marvin24_ has quit [Ping timeout: 480 seconds]
Daanct12 has joined #linux-msm
Daanct12 has quit [Remote host closed the connection]
Daanct12 has joined #linux-msm
jhovold has joined #linux-msm
Daanct12 has quit [Quit: Leaving]
pevik has joined #linux-msm
abhinav__ has quit [Quit: Ping timeout (120 seconds)]
abhinav__ has joined #linux-msm
jessica_24 has quit [Quit: Ping timeout (120 seconds)]
bhsharma has quit [Read error: Connection reset by peer]
jessica_24 has joined #linux-msm
bhsharma has joined #linux-msm
gpiccoli has quit [Quit: Bears...Beets...Battlestar Galactica]
gpiccoli has joined #linux-msm
<Marijn[m]> lumag: You want a spreadsheet with contributors mapped to SoC they own? Could be useful in general for testing :)
<lumag_> Marijn[m], well, I was thinking about something closer to old Arnd's arm-soc or to Debian's arch qualification tables. Which SoCs have been updated to use DT clocks (and who is caring about the conversion). But you might evolve this idea as you see fit
<Mis012[m]> wait, so converting msm-clk to using DT clocks is in the cards?
<Marijn[m]> lumag: That sounds like a better idea, I don't think everyone wants to share exactly what hardware they have access to, and list if it's (in)convenient for them to perform tests. Let's have a table about the current status-quo instead
<Marijn[m]> Mis012: We've already converted some drivers and new clock drivers are hopefully rejected if they rely on global names instead of being converted to DT
pespin has joined #linux-msm
<Mis012[m]> Marijn: wouldn't that need rewriting the msm-clk core?
<Mis012[m]> to do it like e.g omap
<Mis012[m]> `compatible = "qcom,fabia-pll"`
<Marijn[m]> Mis012: I don't have any `msm-clk` in my tree, what specifically are you referring to?
<Mis012[m]> clk-msm?
<Mis012[m]> I always forget which driver is which way around
<Marijn[m]> If it isn't obvious, we're referring to `gcc` and related drivers (dsi phy, dispcc, etc) still using hardcoded clock names instead of reading the dependency from the tree
<Mis012[m]> oh, I thought you were talking about different hardcoded things that should in the device tree
<Mis012[m]> my bad
<Mis012[m]> *should be
<Marijn[m]> Ah okay
<Mis012[m]> this for example looks like prime device tree material
<Marijn[m]> Mis012: You mean specifically pulling gplls out of the gcc driver, or every individual clock branch?
<Mis012[m]> every individual clock branch
<Mis012[m]> you could argue that it's more organized like this, but technically it belongs in the dt based on general principles of what goes in the dt :P
<Mis012[m]> I would expect this kind of hardcoding on ACPI systems because ACPI sucks, not sure if it's doable with the ACPI devicetree frankenstain spec but noone uses that anyway because Windows doesn't support it (afaik)
<lumag_> Mis012[m], it's parent_hws vs parent_names vs parent_data that we are fighting at this moment.
<lumag_> The general clock structure belongs to the driver (register, masks, ops, etc). It would be a nightmare to have everything in the DT
<Mis012[m]> well... it would be a nightmare, but it's ostensibly data
<Mis012[m]> fwiw I'm more interested in splitting gpio/pwm/i2c drivers off amdgpu and giving them dts bindings
<Mis012[m]> to sidestep dealing with atombios on devicetree systems when there's no ROM
<Mis012[m]> ACPI just has a binary blob for the atombios, but that's ACPI why would you expect it to not be ugly
<Mis012[m]> I'm sure that will be a fun conversation...
<Mis012[m]> especially since the x86 systems that actually shipped a DT were always intel ones
Danct12 has quit [Quit: Quitting]
<sboyd> lumag_: will you resend https://lore.kernel.org/r/20220426093608.2957210-1-dmitry.baryshkov@linaro.org? bamse merged the alternative version it seems
<sboyd> ok so I can slam that on top right now
<sboyd> oh I see why I didn't see it, no linux-clk Cc
<lumag_> ugh, missed that
<lumag_> excuse me
<sboyd> alright then maybe I'll just wait for bamse to pick it up
<lumag_> sboyd, bamse: whichever way works for you
jhovold has quit [Ping timeout: 480 seconds]
pespin_ has joined #linux-msm
pespin has quit [Ping timeout: 480 seconds]
lumag_ has quit [Ping timeout: 480 seconds]
pespin_ has quit []
pevik has quit [Ping timeout: 480 seconds]
lumag_ has joined #linux-msm
Danct12 has joined #linux-msm