#
✕ nick:sven
#asahi · 2025-09-07 22:11
svenwe *could* buy a second hand m1 and see what happens :>
#asahi · 2025-09-07 05:01
sventhe phy side, sure. but that doesn’t mean you get anything useful because the rest is still missing
#asahi · 2025-09-02 11:20
svenit really depends on what sets up the pcie bar space, there's also a mode where firmware does most of the work instead of the linux kernel
#asahi · 2025-09-02 11:18
svenif you run into the bugs i ran into
#asahi · 2025-09-02 11:18
svensometimes, yes
#asahi · 2025-09-02 11:16
svendo you have experience with kernel development? pcie hotplug at least two years ago had quite some issues when you plugged in long chains. x86 relies heavily on boot-time setup but on apple machines everything will be handled as hot plug
#asahi · 2025-09-02 11:12
svenmore or less. if you really want to help take a look at pcie hotplug
#asahi · 2025-09-02 10:59
svenany x86 machine will do
#asahi · 2025-09-02 10:57
svenuse a x86 machine with tbt and fix any issues you run into with pcie hotplug, especially with complex chains like dock -> dock -> tbt screen -> tbt enclosure
#asahi · 2025-08-31 16:53
svenpretty much
#asahi · 2025-08-26 18:48
svenlol
#asahi · 2025-08-26 18:44
svenbut that's hardly surprising after ~5 years
#asahi · 2025-08-26 18:44
svenmy battery is slowly starting to show its age
#asahi · 2025-08-26 18:43
svenyeah. and hopefully by then we have m3 working :D
#asahi · 2025-08-26 18:39
svenand it’s almost 5 years old now
#asahi · 2025-08-26 18:38
svenI feel the same about a base(!) m1 MacBook Air
#asahi · 2025-08-20 11:45
svenwhat exactly are you trying to achieve?
#asahi · 2025-08-20 11:41
sven*x86 linux machine
#asahi · 2025-08-20 11:40
svenbut i've seen usb4net with a x86 machine connected to a M1 running macos
#asahi · 2025-08-20 11:40
svenyeah, only with that atcphy wip branch I think I actually had it successfully negotiate tbt host to host mode. usb4net itself then failed because the tbt code still has issues :D
#asahi · 2025-08-20 11:32
sventbt cables are usually short and quite sturdy
#asahi · 2025-08-20 11:31
svenmight just not be a tbt compatible cable then
#asahi · 2025-08-20 11:31
svenif it's not tbt or usb4 compatible the macos side might fall back to device mode i guess. but no idea what device it'll expose
#asahi · 2025-08-20 11:29
svenso assuming you have a thunderbolt cable, a thunderbolt capable linux machine and are running normal macOS it should negotiate a thunderbolt connection and setup usb4net
#asahi · 2025-08-20 11:28
svenhrm, i'm pretty sure 1tr is 1 true recovery. but doesn't matter
#asahi · 2025-08-20 11:27
sveni assume recoveryos is just whatever runs in 1TR
#asahi · 2025-08-20 11:27
sventhese things should support USB4NET but maybe that's not enabled in recoveryOS
#asahi · 2025-07-27 09:42
svenand something changed between m3 and m4 there
#asahi · 2025-07-27 09:42
svenoh, and for m4 you also need to figure out what happened to PMGR because iirc most SoCs blocks are off and you need to enable them before you can use the HV
#asahi · 2025-07-27 09:41
svenand then probably create a shadow copy of all page tables and track changes and translate them to normal ARM pagetable
#asahi · 2025-07-27 09:41
sveni.e. stuff them both into EL1 and trap on genter/gexit and translate that to a switch to the other VM
#asahi · 2025-07-27 09:40
sveni'd virtualize both SPTM and XNU because drivers like DART live inside SPTM
#asahi · 2025-07-27 09:40
sven*ones
#asahi · 2025-07-27 09:39
sventhe very cheap used one usually "fell from a truck" and you don't have machine owner credentials for these which you need to even install custom boot objects
#asahi · 2025-07-27 09:39
svenalso, make sure the m4 you buy is actually new and not cheap from ebay or whatever
#asahi · 2025-05-27 17:19
svenbut I’m willing to help with that as long as someone signs up for the initial dt (plus maybe aic) work including sending that to LKML
#asahi · 2025-05-27 17:18
svenfair, those might need chicken bits added to m1n1
#asahi · 2025-05-27 16:55
sveni can help a bit with m3, I won’t have the time to help with the m4 hv stuff though
#asahi · 2025-05-27 16:54
svenm4 is much more complicated. the first (very big) step can be done on m2+ though: disable all custom apple instruction and virtualizate xnu with nested virt (or find some other way to run xnu under the hv when stuff like gxf and sprr is disabled)
#asahi · 2025-05-27 16:53
sven*sent
#asahi · 2025-05-27 16:52
svenor getting macOS to boot under the hv
#asahi · 2025-05-27 16:52
svenfor m3 the first step is either a minimal Linux device tree with basic stuff like aic and wdt ideally went directly upstream
#asahi · 2025-05-27 15:27
svenplus we need for wait for the iboot fix so that raw boot objects work again
#asahi · 2025-05-27 15:26
svenyeah, it’s doable, just *much* more annoying than bringing up m3
#asahi · 2025-05-18 07:59
sven*doubles
#asahi · 2025-05-18 07:58
svendouble the estimate again
#asahi · 2025-05-17 17:35
svencrap like that is the reason why getting usb3/tbt/etc working took three of us all this time so far
#asahi · 2025-05-17 14:11
sventhat entire thing is a mess
#asahi · 2025-05-17 14:10
svenhttps://github.com/AsahiLinux/linux/commit/b43bdb4fff6bd60a7b5a87e44a536fd6f5a810ab
#asahi · 2025-04-26 10:27
svenif you have proof of purchase that apple accepts they may be able to remove it

plus ›