09:09 <hexdump0815> looks like the samsung galaxy book go lte is out with a 7280c, but i guess it would be as ugly to get working as the other windows arm notebooks ...

09:18 <HdkR> At least the 7c already has a Chromebok

09:19 <HdkR> So a little closer than usual

11:23 <macc24> bamse: what is "ipa"?

11:44 <vkoul> macc24: IP Accelerator (IPA)

11:44 <vkoul> See Documentation/devicetree/bindings/net/qcom,ipa.yaml & drivers/net/ipa/

11:44 <macc24> vkoul: thanks

13:58 <robclark> hexdump0815: the samsung 7c laptop is very much not 7280.. other than dts file all the kernel support it needs is already upstream

14:09 <macc24> robclark: huh?

14:12 <robclark> not sure the part #, I want to say maybe sc7190 or something like that.. very much not sc7280.. compare the already upstream sc7280.dtsi to sc7180.dtsi ;-)

14:13 <macc24> hmmmmmmmmmmmmmmmm i kinda want to convince some arm laptop with efi to run on cadmium

15:53 <hexdump0815> robclark: oh - then i mixed up the soc names - i thought 7280c is 7c gen 2, but looks like its different then

15:54 <robclark> marketing and hw have different ideas of what a "generation" is :-P

15:56 <macc24> then what is 7280?

15:57 <robclark> it is 7280 :-P

15:57 <macc24> marketing didn't a bullshit name for it?

15:57 <macc24> * marketing didn't make a name for it?

15:57 <robclark> I'm sure they will eventually

15:57 <macc24> 7.5c

17:00 <bamse> the big question regarding 7c is how much work it would be to get upstream running with uefi and qc's tz...

17:03 <robclark> bamse: I think not much.. looks like we already have the right compatibles in the list in arm-smmu-qcom to deal with bootloader enabled display.. so I think all it should need is toplevel .dts and sorting out the smbios strings so DtbLoader does it's thing

17:03 <robclark> hmm, and maybe panel driver

17:03 <macc24> hmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

17:04 <robclark> hopefully they went with same edp bridge that 835/850 laptops used and sc7180 chromebooks use

17:04 <macc24> the WHAT

17:04 <robclark> dsi->edp bridge

17:04 <macc24> wtf

17:05 <robclark> ??

17:05 <macc24> why would they do that

17:05 <macc24> why not put edp thing into sco

17:05 <macc24> soc*

17:06 <robclark> they do in later ones, 835/850 were mobile chips (so no eDP).. and is still kinda "mobile-ish"

17:07 <robclark> it's kinda a transitional part

17:07 <macc24> oh ok

17:09 <robclark> it depends also a bit on form-factor.. some form-factors could actually use a DSI panel instead of eDP panel (but I don't think you'd see that in a traditional laptop form-factor since the cabling doesn't work out)

17:10 <bamse> robclark: right, once you get that high up in the stack it should be fine...but there's things below that's definitely affected

17:11 <bamse> robclark: like the clock drivers, cpuidle, remoteproc

17:11 <bamse> robclark: but you're certainly right in that past those details it should be a smooth sailin

17:12 <robclark> hmm, ok.. I was kinda assuming that since 845/850 worked with both flavors of fw, that it should all just work.. but maybe that just reflects fixes made for qc's tz..

17:14 <bamse> we don't do full on cpuidle on 845 yet...

17:14 <bamse> clock wise we did a bunch of work to describe the differences in dt

17:14 <bamse> and we have two sets of remoteproc drivers for 845

17:16 <robclark> oh.. ugg