lanefu changed the topic of #armbian-rockchip to: Armbian - Linux for ARM development boards | Rockchip SoC | www.armbian.com | This channel is relayed to the equivalent Discord channel | this channel is logged
<DC-IRC>
<gambl0r> Hello, I'm looking to build kernel 6.3.2 for opi4lts using buildroot with armbian patches for uwe5622. It successfully adds the source for the module using the first patch (wireless-driver-for-uwe5622-allwinner-v6.3.patch) but the second patch fails (wireless-driver-for-uwe5622-allwinner-bugfix-v6.3.patch) and it looks like the second patch by xunlong is older than the first? Any assistance greatly appreciated.
<DC-IRC>
<amazingfate> @rpardini what is the good way to update the rockchip legacy kernel? I've just generated 50 patches from the current `rockchip-5.10` branch and applied it to the new rkr4 tag. Can we create a new branch to do this update?
<DC-IRC>
<rpardini> Hey, you mean rk released new tree and we gotta rebase on top of it?
<DC-IRC>
<amazingfate> Yes that should be the easy way. It seems hard to update directly from current rkr3.6 to rkr4.
<DC-IRC>
<rpardini> Yes I'd say create a new branch, say `rk-5.10-rkr4`, then cherry-pick our stuff on top.
<DC-IRC>
<rpardini> If all goes well we can switch the default branch on that repo, then switch on the .conf on armbian/build ?
<DC-IRC>
<amazingfate> A lot of conflicts.
<DC-IRC>
<amazingfate> Yeah that's the way I like, no need to deal with much of rockchip's code.
<DC-IRC>
<rpardini> I was not expecting rk to release anything new this year lol
<DC-IRC>
<rpardini> could we not pull those in and zero-out the armbian patch dir?
<DC-IRC>
<amazingfate> Yes we can. The new kernel failed to build because of some patch failing.
<DC-IRC>
<rpardini> I did super fast, but maybe that `stmmac: hack LED_FIX` is problematic?
<DC-IRC>
<rpardini> or is one of the armbian/build patches failing?
<DC-IRC>
<amazingfate> That LED_FIX is introduced from a vendor to our kernel tree. A drm patch `0002-dma-buf-Add-an-API-for-exporting-sync-files-v14.patch` is the casuse of failure.
<DC-IRC>
<amazingfate> I'm going to clean the patch dir.
<DC-IRC>
<rpardini> Yeah, I'd say pick from those patches into the new rkr4 tree, drop whatever doesn't make sense and clean out that patch dir
<DC-IRC>
<rpardini> only problem here with this strategy (rebase into new branch, switch) is that we need Igor admin-powers to switch default branch on the repo
<DC-IRC>
<EfeCTN> Are there any changelog for rkr4? Looks like mostly related to vop stuff
<DC-IRC>
<rpardini> looks like a double-sandwich Android rebase to 5.10.160, some vop stuff indeed
<DC-IRC>
<NicoD> Why do they still make new SoC's with old cores? All other RK35xx have A55 but then they make a RK3562 with A53. Will be cost-effective for low powered devices.
<DC-IRC>
<NicoD> Ain't as bad as the Tinker Board S2 R2.
<DC-IRC>
<Tenkawa> Inexpensive and easy to reuse established tech they already have and less
<DC-IRC>
<Tenkawa> Inexpensive and easy to reuse established tech they already have and less paperwork to do it.
<DC-IRC>
<Tenkawa> although yeah the 3562 appears "strangely" just to be newly built
<DC-IRC>
<Tenkawa> and extremely low performing
<DC-IRC>
<NicoD> I wouldn't want it. The S905X4 then is way more attractive to me.
<DC-IRC>
<Tenkawa> I'm looking at a comparison of it and a Unisoc Tiger
<DC-IRC>
<Tenkawa> and it looks underwhelming
<DC-IRC>
<NicoD> Yeah. Even at 2Ghz the A53 just isn't much.
<DC-IRC>
<Tenkawa> I mean look at this line between the two units
<DC-IRC>
<Tenkawa> Cache memory 2 MB 256 KB
<DC-IRC>
<Tenkawa> the lower being the 3562
<DC-IRC>
<Tenkawa> thats just not good
<DC-IRC>
<c0rnelius> Someone asked me about that rk3528 a few weeks ago. I was just as honest I could be at the time. I couldn't find it in any of the RK sources and it appears its just being used with Android devices.
<DC-IRC>
<Tonymac32> Yeah, the one before it (rk3328) was only a TV box processor, that's why it has shakier support than say the RK3288 and RK3399
<DC-IRC>
<Tonymac32> My understanding was the 3528 is an rk3328 with the cores swapped out
<DC-IRC>
<Tonymac32> True or not, no idea
<DC-IRC>
<Tonymac32> The article looks like the accidentally put the rk3328 specs up (mali450 really?), I can't believe they only changed the VOPs and called it a new SoC
<DC-IRC>
<Tonymac32> Ick
<DC-IRC>
<amazingfate> Finally get it running. Mali firmware has to get updated to g17p0.