#asahi-dev · 2025-09-16 13:03
nickchanmust be a leftover from when m1 is everything#asahi-dev · 2025-09-16 13:02
nickchansven: it's not a big deal but "usb: dwc3: Add Apple Silicon DWC3 glue layer driver" in the wip/atcphy branch mentions Apple M1 specifically which is probably not appropriate#asahi-dev · 2025-09-16 12:59
nickchana11 and t2 spmi series should be ready again now after the generic compatible incident#asahi-dev · 2025-09-16 12:48
nickchanthat i have no idea how to drive yet#asahi-dev · 2025-09-16 12:48
nickchannow there's also usb3 phys on some a9x and a10x ipads but as far as I can tell that's a completely different animal#asahi-dev · 2025-09-16 12:44
nickchanit also has complications of dealing with 32-bit dma devices but without a dart#asahi-dev · 2025-09-16 12:43
svenokay, great#asahi-dev · 2025-09-16 12:42
nickchansven: correct there's some similarities but not enough to use one driver for both#asahi-dev · 2025-09-16 08:53
svennickchan: the pre-M1 usb2phy only had similar registers but at different offsets, etc, didn't it?#asahi-dev · 2025-09-16 07:22
jannauI'd expect it only applies to new devices. Not sure if the old HW even would support it. There might be a similar thing to use more bandwidth for thunderbolt tunneled DP but that doesn't use multiplexing but tb's capability to tunnel 2 DP connections#asahi-dev · 2025-09-16 07:15
fl0_idmaettu102 who knows#asahi-dev · 2025-09-16 06:55
maettu102If they add mst support would that only affect newer machines or do you think they would ship firmware updates for the older ones as well?#asahi-dev · 2025-09-16 06:19
jannauit won't be noticable if using macos users as base but it might be compared to the current asahi user base#asahi-dev · 2025-09-16 05:39
sveneven if we supported m3/4 already #asahi-dev · 2025-09-16 05:38
svenI think we’re too niche for that#asahi-dev · 2025-09-16 05:29
nicolas17I honestly expect Liquid Glass and Apple Intelligence to bring more users to Asahi btw :D#asahi-dev · 2025-09-16 05:18
nicolas17maybe they deal with APNS keepalives and other stuff without waking the CPU now#asahi-dev · 2025-09-16 05:17
nicolas17I saw mentions of "AOP2 offload" and "ultra low power networking" on the iOS 26 push notification daemon, maybe it's only for devices with N1#asahi-dev · 2025-09-16 05:15
jannauN1 seems to use rtkit so I guess coproc and possibly CPU faacing parts of the network interface are on the SoC#asahi-dev · 2025-09-16 05:13
jannauI was not expecting that after that added an dcp to M4. I guess they got too many complaints that display outputs on hubs are not working (or are terribly late with implementing MST support)#asahi-dev · 2025-09-16 04:33
nicolas17probably means nothing for iPhone, but it hints M5 Macs could finally get displayport MST#asahi-dev · 2025-09-16 04:31
nicolas17apparently the A19 DCP firmware mentions MST#asahi-dev · 2025-09-15 19:04
jannaunot that I couldn't take a central scrutinizer board with but just a single cable is more convenient#asahi-dev · 2025-09-15 18:59
svenah :)#asahi-dev · 2025-09-15 18:59
jannautraveling early to XDC in vienna #asahi-dev · 2025-09-15 18:58
jannauunfortunately serial isn't reliable. the linux serial device seems to reset quickly on the host#asahi-dev · 2025-09-15 18:57
sven10 days?#asahi-dev · 2025-09-15 18:57
chaos_princess(i also was getting impatient and thinking of doing exactly that)#asahi-dev · 2025-09-15 18:57
jannauand I have use for it in 10 days#asahi-dev · 2025-09-15 18:57
chaos_princesslol#asahi-dev · 2025-09-15 18:57
sven\o/#asahi-dev · 2025-09-15 18:57
jannauhttps://paste.debian.net/1396756/#asahi-dev · 2025-09-15 18:56
jannauI grew impatient on https://github.com/AsahiLinux/linux/pull/399#asahi-dev · 2025-09-15 14:25
nickchanjannau: ah then im mistaken#asahi-dev · 2025-09-15 14:17
jannauwhy? the t8103 darts all have "dart,t8020" as single comaptible string in the ADT#asahi-dev · 2025-09-15 14:11
nickchanerm actually dart in general would be broken on t8103#asahi-dev · 2025-09-15 14:11
nickchancause that code matches against dart,t8020#asahi-dev · 2025-09-15 14:11
nickchanin that case the dart lock code must be broken right now on t8103#asahi-dev · 2025-09-15 14:09
jannaurust adt code hasn't the concept of string lists#asahi-dev · 2025-09-15 14:02
nickchanand i don't know about rust to fix that#asahi-dev · 2025-09-15 13:58
nickchanthis is a bug#asahi-dev · 2025-09-15 13:58
nickchanso like calling adt_is_compatible(adt, offset, "usb-complex,s5l8960x") on a node that has a compatible property of usb-complex,t8012\0usb-complex,s5l8960x\0 doesn't match anymore#asahi-dev · 2025-09-15 13:52
nickchanchadmed: the new adt_is_compatible() doesn't match against fallback compatibles i think#asahi-dev · 2025-09-15 09:37
mazthey could also be choosing to have a PIPT behaviour by halving the cache size to avoid aliasing. that's a pretty common trick.#asahi-dev · 2025-09-15 09:34
maz*macos's problem.#asahi-dev · 2025-09-15 09:34
maznickchan: possible. but that gives *more* guarantees to the guest, which means the hypervisor has to work harder to avoid aliasing. in any case, that's macos'#asahi-dev · 2025-09-15 09:25
nickchanmaz on a11 and m2 it's 0x8444c004 in HW... looks more like a workaround to me#asahi-dev · 2025-09-15 08:27
mazPretends that the I$ is PIPT, which allows SW to elide CMOs in a number of cases. they must have ways to handle that in the hypervisor somehow to avoid aliasing.#asahi-dev · 2025-09-15 08:14
nickchanbut why?#asahi-dev · 2025-09-15 08:13
nickchanon t6000 ctr_el0 is 0x84448004 but in a vm under macos xnu seems to trap it and become 0x8444c004