#asahi-dev · 2026-01-04 15:45
sveniirc we e.g. split the MMIO for the co-processors into half for linux (one half to the device, other half to mailbox)#asahi-dev · 2026-01-04 15:44
svenADT is known to lie or omit things#asahi-dev · 2026-01-04 15:44
svenyuka: by trying to guess how the hardware works and putting what makes sense in there :D#asahi-dev · 2026-01-04 15:41
svenmight make sense to just select the endpoint based on the smc or ans or whatever compatible#asahi-dev · 2026-01-04 10:53
sveni much prefer to finally work on tbt and pcie hotplug#asahi-dev · 2026-01-04 10:53
svenyup#asahi-dev · 2026-01-04 10:49
svenwhich works, but i was too lazy to do all that especially since I don't have any t2 machines to test :D#asahi-dev · 2026-01-04 10:49
sveni think his suggestion was to move the apple silicon backend to drivers/soc/apple and the x86 backend to some other place#asahi-dev · 2026-01-04 10:47
svenhaven't heard from the t2 people in a long while either, they wanted to add support for their chips at some point as well#asahi-dev · 2026-01-04 10:44
sventhe mfd maintainer didn't like those inside drives/mfd in the original submission fwiw#asahi-dev · 2026-01-04 10:34
svenI have to catch up on reviewing and merging things for our tree as well #asahi-dev · 2026-01-04 10:33
svenusually quicker than having to send another revision via email 🫠#asahi-dev · 2026-01-04 10:33
svenfeel free to mention the branch here once you’re done, I usually do the same and people then catch small issues to be fixed before submitting #asahi-dev · 2026-01-04 10:32
svennice!#asahi-dev · 2026-01-03 17:42
svenstill great to get started 😁#asahi-dev · 2026-01-03 10:45
svenI expect them to hide most tbt5 changes in the tunables #asahi-dev · 2026-01-03 09:45
svens/usb4/tbt5/#asahi-dev · 2026-01-03 09:44
svenso i think we can have two cursed streams running in parallel: 1) bring up m4 by guessing and luck and 2) virtualize sptm on m3/m4 :D#asahi-dev · 2026-01-03 09:44
sveniirc the linux tbt driver already has usb4 support so i suspect it won't be that hard with a m3 ultra#asahi-dev · 2026-01-03 09:17
sveni also still think we can probably bring up m4 except for dcp, gpu, and other complex co-processors almost blindly without the hv :D#asahi-dev · 2026-01-03 09:12
svenmight also be useful to reverse engineer iboot and figure out how exactly the lock out works#asahi-dev · 2026-01-03 09:12
svenit might be its own knob but I don't think we can set it individually#asahi-dev · 2026-01-03 09:10
svenno, but then we're running under SPTM#asahi-dev · 2026-01-03 09:09
svenstill leaves the issue of all those other apple isa extensions that are disabled#asahi-dev · 2026-01-03 09:09
svenhm.. so thinking about virtualizing sptm: i think we don't even need two separate EL1 VMs or even two separate sets of pagetables: it might be enough to "just" build a single shadow page table that has has max(gl2, el2) permissions set and fake genter/gexit transitions#asahi-dev · 2026-01-02 19:39
svenyeah, alright, let's mute that thing#asahi-dev · 2026-01-02 17:32
svennow that's already a good start for the matrix bridge again *sigh*#asahi-dev · 2026-01-02 12:26
svensure!#asahi-dev · 2026-01-02 12:18
svenintegralpilot: if you still need it i can try and grab a nvme trace this weekend#asahi-dev · 2026-01-02 12:16
svennickchan / integralpilot: the mutex isn't the issue, the actual reason was that -power needs cleanup before submitting it upstream. iirc there's quite a bunch of debug stuff left in there#asahi-dev · 2025-12-31 15:22
svenintegralpilot: ouch, that sucks! what i usually do is commit everything i do and push it to a private github repo if i'm not ready to share it just yet#asahi-dev · 2025-12-30 09:09
svenyes, about 50 minutes left#asahi-dev · 2025-12-30 09:09
sventhanks!#asahi-dev · 2025-12-29 13:26
sveni think he's more interested in the stuff required to get linux booting#asahi-dev · 2025-12-29 12:15
svenjannau: the stickers are all gone now, people really wanted them :D#asahi-dev · 2025-12-29 09:17
svenintegralpilot: gonna use neofetch and doom :)#asahi-dev · 2025-12-29 08:59
svenhahahahah, nice :D#asahi-dev · 2025-12-29 08:56
svendepending on time (i.e. how fast i'll talk :D) i can probably show 2-3#asahi-dev · 2025-12-29 08:56
svennice!!#asahi-dev · 2025-12-29 08:50
svengreat, i just wanna show how much black magic that thing is :)#asahi-dev · 2025-12-29 08:48
svenperfect, thanks! :)#asahi-dev · 2025-12-29 08:40
svenanyone happen to have a atcphy trace when plugging and unplugging some usb device laying around? want to show something from there on my slide but i'll create my own later otherwise#asahi-dev · 2025-12-28 22:15
svenPerfect 😁#asahi-dev · 2025-12-28 20:42
svenstill need to work on my (so far crappy...) slides#asahi-dev · 2025-12-28 20:42
svenlet's take a look tomorrow :D#asahi-dev · 2025-12-28 20:42
svenmaybe they do both in the single pcie device with separate BARs or something strange like that#asahi-dev · 2025-12-28 20:41
svenhuh, i don't see the subfunction in lspci for bt like we do on apple machines#asahi-dev · 2025-12-28 20:38
svenlet's take a closer look tomorrow then#asahi-dev · 2025-12-28 20:38
svenhm... strange