06:51 <jhovold> steev: yeah, we have one more known audio issue that you may be hitting and which causes the soundcard to fail to probe

06:51 <jhovold> i reported it to srinik in March and have hit it 4-5 times since, so not very frequently

06:53 <jhovold> i was hoping we'd have that fixed before anyone relly noticed but i'll go and add it to the wiki page

06:53 <jhovold> next time it happens, can you check your logs, there should be probe deferall message like:

06:53 <jhovold> platform sound: deferred probe pending

06:55 <jhovold> may be caused by a bug in the underlying remoteproc code as well, but haven't really had time look at this one yet

09:52 <HdkR> I see a whole bunch of patches for SM8650 \o/

10:08 <cenunix[m]> whats the status of hardware accel ? on steev's 6.6.0-rc5 commit:f407eddf8f89860256c17f85550631263581079b . I'm unsure how to check if it's working properly or not though

10:20 <Jasper[m]> <cenunix[m]> "whats the status of hardware..." <- Which kind of hardware acceleration? 3d? 2d? Video?

12:39 <Jasper[m]> Btw, out of curiousity. Is the qcvss8280.mbn fw file something that's device specific? I know the people trying to get Linux to boot on Volterra are messing around with fetching the file from Windows, but that feels a bit annoying because I kinda want to get rid of Windows on it

14:14 <HdkR> And now sc8380xp patches, woo \o/

14:19 <ajhalaney[m]> yeah I was about to say that's a good sign :)

14:19 <ajhalaney[m]> Also, from some qualcomm emails (I like seeing more qualcomm contributions on list)

14:29 <robclark> nice sc8380xp patches nearly the same day as announced

14:35 <HdkR> Not quite day zero support but getting better. Looking forward to playing with that 4.6TFLOP GPU :)

14:38 <robclark> I guess 0.5 day?

14:38 <robclark> tbf it is long before hw shipped

14:39 <robclark> yeah, gpu should be fun.. AFAICT a690 is still faster than any of the a7xx.. but an a7xx version of laptop gpu should be quite fun

14:40 <krzk> SM8650 had ~0.1-day-support :)

14:41 <HdkR> Also curious if those USB4 ports can handle a Radeon dGPU :D

14:41 <HdkR> s/dGPU/eGPU

14:42 <clover[m]> So new snapdragon socks were announced?

14:43 <robclark> indeed

14:43 <HdkR> Fresh socks, been wearing holes in the old ones

14:44 <robclark> I guess someone needs to step up and figure out open src way to do new fangled AI upscaling on these things

14:46 <HdkR> There's some Snapdragon solution that hasn't been talked about since April. Assuming it is GPU based rather than AI based from their previous marketing

14:47 <HdkR> Probably difficult to leverage the Hexagon for AI upscaling techniques for games?

14:48 <robclark> idk, AI upscaling seems to be the new hotness.. because AI.. and I could see wanting to do it on hexagon to offload the GPU .. although from a sw arch standpoint that is a bit terrifying

14:49 <HdkR> Indeed

14:54 <travmurav[m]> given this is primarily a windows-targeted SoC, one funny interpretation is that they convinced MS to ditch hyper-v

14:55 <travmurav[m]> but even if no, this sounds like no EL2 again, but "why are you complaining, we gave you our great new hypervisor"

14:55 <robclark> so, crosvm apparently has gunyah support.. so maybe that is a path for linux vm support on these things?

14:55 <travmurav[m]> and most people would not care

14:57 <robclark> I mean, they'd also be using it on android.. and I suspect if CrOS ever supports protected content on qc things then it would also need to switch to gunyah and "dom0" booting in EL1

15:03 <Dylanger> <travmurav[m]> "given this is primarily a..." <- Didn't they do virt-in-virt?

15:03 <Dylanger> I care, very much, tis why I didn't buy a X13s, give me damn EL2.

15:03 <travmurav[m]> microsoft has vendor-locked signature verified way to escalate to el2

15:04 <travmurav[m]> (had?)

15:04 <Jasper[m]> <travmurav[m]> "but even if no, this sounds like..." <- This^ what I was told

15:05 <Dylanger> Gunyah provides very good Qualcomm specific attack surface, the very best

15:05 <Dylanger> 👌

15:05 <Dylanger> Lol sorry I'm emotional about this topic

15:05 <Jasper[m]> I'm fine with qemu working

15:06 <Dylanger> Oh sc8380 is the super brand new SoC announced today?!

15:07 <Dylanger> Time to look for a PBL/BootROM exploit to go along with my new Qualcomm Laptop 🤦‍♂️

15:08 <travmurav[m]> I guess the only hope is that microsoft didn't cave in and the secure-launch bs is still there, then there is still a chance it can be abused to elevate to el2, I think it's possible

15:08 <HdkR> Dylanger: There were two new SoCs, need to specify which one :P

15:43 <clover[m]> two socks for too foots

15:46 <robclark> Dylanger: at least the hypervisor is open src.. which seems like an improvement over every other vendor's hyp

15:57 <robclark> maybe you could compare executable size?

15:57 <Dylanger> <robclark> "Dylanger: at least the hyperviso..." <- But it's not tho, I can't build my own hyp and run it on EL2, fork of Ganyah or otherwise, also iirc Qualcomm's Ganyah haven't been updated for like years

15:59 <robclark> the main thing I'm aware of that hyp does is stuff w/ mmu/smmu L2 translation to implement L1 protected content.. not sure if that is in ganyah github or not, but not really rocket science

19:15 <steev> https://github.com/quic/gunyah-hypervisor it was updated last month?

19:49 <konradybcio> yes

21:13 <Dylanger> Woah

21:13 <Dylanger> So that worked

21:13 <Dylanger> I have Fedora working on my Cherry

21:22 <Dylanger> I needed to extract the rootfs from the raw xz file from Fedora's website, copy it to my f2fs partition, patch the old UUID over the old one, remove the old entries for fstab as I'm not using btrfs and copy my kernel modules

21:22 <Dylanger> And it boots perfectly

21:22 <Dylanger> WiFi Works, BT Works