#asahi-dev · 2026-03-04 15:40
svenah, that diff may have been broken#asahi-dev · 2026-03-04 15:39
sventhe /dev/cu.kis thing should show up even without any m1n1 patch#asahi-dev · 2026-03-04 15:39
sven0x1824606 should be the correct actionid, do you use that in macvdmtool?#asahi-dev · 2026-03-04 15:38
svenchaos_princess: does /dev/cu.kis-100000-ch-0 should up?#asahi-dev · 2026-03-04 15:37
sveni haven't updated my stage1 since years#asahi-dev · 2026-03-04 15:37
svenshould work with chainload#asahi-dev · 2026-03-04 15:34
svenyeah#asahi-dev · 2026-03-04 15:33
svenchaos_princess: lol, cursed but that should work#asahi-dev · 2026-03-04 15:28
svencould also solder to some uart hw testpoint that surely exist otherwise if we really have to#asahi-dev · 2026-03-04 15:27
svenjannau: or maybe we're really really lucky and the special debug port is the usb2 one and we can at least trace via slow hw serial :D#asahi-dev · 2026-03-04 15:22
svenso maybe someone has to implement dwc2 support in m1n1 then :(#asahi-dev · 2026-03-04 15:21
svenyeah, didn't think about tracing at all#asahi-dev · 2026-03-04 15:21
sventrue :/#asahi-dev · 2026-03-04 15:18
svenguess m1n1 will maybe only work on that single usb3 port then. not great, but also not a big issue#asahi-dev · 2026-03-04 15:17
svenoh wait.. the second type-c is usb2 only? ugh#asahi-dev · 2026-03-04 15:16
sven*then#asahi-dev · 2026-03-04 15:15
svendoesn't a18 pro have usb3 support? should have dwc3 thren#asahi-dev · 2026-03-04 07:14
svenchadmed: hope it also comes with sharklets for even higher efficiency;)#asahi-dev · 2026-03-04 07:13
svenchadmed: New Engine Option :D#asahi-dev · 2026-03-04 06:29
svencan you reproduce this on Fedora?#asahi-dev · 2026-03-04 06:28
svener wait… usb starts to work later if you boot from NVMe?#asahi-dev · 2026-03-04 06:27
sven*on#asahi-dev · 2026-03-04 06:27
svenit shouldn’t happen in fedora #asahi-dev · 2026-03-04 06:26
svenplease file the bug with your distro #asahi-dev · 2026-03-04 06:26
svensounds like it, yeah #asahi-dev · 2026-02-27 13:15
svenwhen I write from the host without reading on the device tx_free keeps decreasing #asahi-dev · 2026-02-27 13:15
sveninteresting. there’s something in the middle then that reads from that dockchannel I guess #asahi-dev · 2026-02-27 09:41
svensure, but I’m talking about the UART path #asahi-dev · 2026-02-27 06:44
svenit looked like the kis usb interface was just MMIO read/write but maybe there’s actually something inbetween and there are two dockchannel fifos#asahi-dev · 2026-02-27 06:43
svenI can check later if sending decreases the available fifo space #asahi-dev · 2026-02-27 06:43
svenhuh, weird #asahi-dev · 2026-02-27 06:27
sven0x8000 on m1#asahi-dev · 2026-02-27 06:27
sventhe fifo buffer is quite large #asahi-dev · 2026-02-26 16:44
sventhat's fine for the HW uart because it drains automatically anyway#asahi-dev · 2026-02-26 16:43
svenpretty sure there are also unguarded calls to iodev_write#asahi-dev · 2026-02-26 16:43
sventhat's called unconditionally at least in the hv#asahi-dev · 2026-02-26 16:42
sveni'm talking about the _iodev_console_write_helper path#asahi-dev · 2026-02-26 16:40
svenif you don't it'll fill up the 0x8000 or so FIFO and then hang i think#asahi-dev · 2026-02-26 16:39
svenbecause you were reading from the other end#asahi-dev · 2026-02-26 16:39
sveni think the bigger issue is that this is going to block eventually#asahi-dev · 2026-02-26 16:38
sven*shrug*#asahi-dev · 2026-02-26 16:38
svenannoying messages?#asahi-dev · 2026-02-26 16:31
svenprobably still useful for people without hardware serial fallback#asahi-dev · 2026-02-26 16:30
svenah damn... i thought the slowness was only caused by my python hacks :(#asahi-dev · 2026-02-26 16:30
svenhrm, though i guess we should make the writes non-blocking#asahi-dev · 2026-02-26 16:29
svenhttps://github.com/AsahiLinux/m1n1/pull/546 proper PR then, github UI doesn't allow me to request a review from nickchan for some reason#asahi-dev · 2026-02-26 16:25
svenah, receiving didn't work. don't think i ever tested that#asahi-dev · 2026-02-25 09:17
svennickchan: I think I only tested with my own kis implementation but I can take a look if I pushed the right thing once in back from my $work trip#asahi-dev · 2026-02-24 19:21
svenbut iirc this worked once