#asahi-dev · 2026-01-27 10:44
svenimho we should focus on getting m3 support actually useable and installable before starting to put a lot of effort into the gpu#asahi-dev · 2026-01-27 10:37
svenyeah. it's also a good idea to remind people that the userspace driver alone is something like 50k lines of code#asahi-dev · 2026-01-27 10:28
svenare people asking about gpu yet without realizing just how many orders of magnitude that's gonna be more complex? :p#asahi-dev · 2026-01-27 10:18
svenyeah, and the base (?) m3 still has the tipd chip anyway#asahi-dev · 2026-01-27 10:01
svenjust need sign-offs and some cleanup#asahi-dev · 2026-01-27 10:01
svenhttps://github.com/mildsunrise/linux/compare/43cb782cdd...11313a6cff#asahi-dev · 2026-01-27 09:51
svenit's using regmap anyway so adding spmi support shouldn't be too hard#asahi-dev · 2026-01-27 09:50
sventhere's some commits by mildsunrise improving the upstream SPMI driver and some notes how the new pd controller works#asahi-dev · 2026-01-26 22:01
svennoopwafel: nice!#asahi-dev · 2026-01-26 18:01
svenand we probably need to adjust the pmgr bindings for the new compatibles as well #asahi-dev · 2026-01-26 18:01
svenat least the aic bindings also need to be reviewed before #asahi-dev · 2026-01-26 18:00
svenyeah, I’m not gonna pick up anything for this merge window so it makes sense to wait for -rc1 after we’ve collected all the device trees for the various models anyway #asahi-dev · 2026-01-26 10:44
svenyeah, llvmpipe is surprisingly useable. it’ll probably burn through your battery though #asahi-dev · 2026-01-26 09:44
svenhasn’t worked so far but I’ll continue after fosdem#asahi-dev · 2026-01-26 09:43
svenyup :D#asahi-dev · 2026-01-26 09:41
svenIt’s been like this all along, probably a subtle difference to what xnu does somewhere or maybe a missing delay that I can’t easily see in the MMIO trace #asahi-dev · 2026-01-26 09:40
svenwelcome to the wonderful world of type-c and usb :D#asahi-dev · 2026-01-26 09:40
svenhttps://github.com/AsahiLinux/linux/commits/sven/tbt-wip/ Is the branch that doesn’t work, needs a small m1n1 change for some node name #asahi-dev · 2026-01-26 09:39
svenI though it was just that dwc3 isn’t brought up because the usb bit is not set in tipd but that hasn’t fixed anything #asahi-dev · 2026-01-26 09:38
svenyeah, that is fixed but I didn’t notice that only the root complex comes up but nothing connected to the cable #asahi-dev · 2026-01-26 09:38
svenas usual. something’s wrong in atcphy that makes the link not come up#asahi-dev · 2026-01-26 09:37
svenwe’ll have to move some of that stuff to the kernel for the tbt pcie controller. that’s also gonna be fun #asahi-dev · 2026-01-26 09:28
sventhat makes more sense :)#asahi-dev · 2026-01-26 09:22
svenjust a bunch of stuff m1n1 copies from the ADT and which is then passed along to the WiFi chip#asahi-dev · 2026-01-26 09:22
svenmagic runes required to make the high frequency witchcraft work correctly ;)#asahi-dev · 2026-01-26 09:21
svenwhy would wifi tunables be fun? i thought m3 had roughly the same Broadcom chip as m2#asahi-dev · 2026-01-25 23:26
sven*not#asahi-dev · 2026-01-25 23:26
svenhuh, but no when installed in anything else?#asahi-dev · 2026-01-25 23:17
sventhere's a fallback to re-use some random data from apple's device tree iirc#asahi-dev · 2026-01-25 23:17
svenit means something changed on M3 with SEP that we have to figure out#asahi-dev · 2026-01-25 23:17
svenno, that's not normal#asahi-dev · 2026-01-25 23:03
svenyup#asahi-dev · 2026-01-25 23:02
svensounds good to me#asahi-dev · 2026-01-25 22:58
svenwe've had enough pain with that and we're still not through with everything#asahi-dev · 2026-01-25 22:58
sventhe only thing we should definitely avoid is adding un-reviewed device tree bindings and nodes if it's more than a simple compatible#asahi-dev · 2026-01-25 22:51
svenack#asahi-dev · 2026-01-25 22:47
sventhen once we have that base structure we can start extending them with all the other hw as it’s brought up #asahi-dev · 2026-01-25 22:46
svenideally we just send an initial series with all the m3 variants. can just be aic, serial and pmgr and similar very basic stuff #asahi-dev · 2026-01-25 21:52
svenbut sure, feel free to post a tree (i.e. not just the device tree, the entire branch based on the latest -rc1)#asahi-dev · 2026-01-25 21:51
sventhe merge window will also open next monday, submitting after that probably makes more sense#asahi-dev · 2026-01-25 21:51
svencan't really send that initial dt upstream before the AIC bindings are at least reviewed#asahi-dev · 2026-01-25 21:32
svennow how about usb and then thunderbolt? 😇#asahi-dev · 2026-01-25 15:44
sventhe protocol looks simple enough but i think i'm missing some "actually enable serial" packet#asahi-dev · 2026-01-25 15:44
svenbah, this usb trace isn't complete afaict, missing all the host->kis data. time to get a hardware usb sniffer again :D#asahi-dev · 2026-01-25 14:09
sveni assume it exists#asahi-dev · 2026-01-25 14:09
svenyeah, but we can figure that one out#asahi-dev · 2026-01-25 14:01
svenjust that one is hooked up to KIS on the other end and the other to MTP#asahi-dev · 2026-01-25 14:01
svenso this all really looks like two instantiations of the same HW block