#
✕ nick:sven
#asahi-dev · 2025-09-02 20:08
svenI don’t see that here
#asahi-dev · 2025-09-02 20:07
sventimed out waiting for ACIOPHY_TOP_PHY_STAT_LN0_UNK0 that’s also not great
#asahi-dev · 2025-09-02 20:07
sven:/
#asahi-dev · 2025-09-02 19:59
sventhere’s a second or two delay until the device appears here I think
#asahi-dev · 2025-09-02 19:58
svenugh… that doesn’t happen on my m1 :/
#asahi-dev · 2025-09-02 19:06
sven*if
#asahi-dev · 2025-09-02 19:06
svenit usually works with wrong fuses as well, you just sometimes lose superspeed and. and tbt is very allergic is they are wrong
#asahi-dev · 2025-09-02 19:05
sven:D
#asahi-dev · 2025-09-02 18:59
svenhah, i can test gadget mode through my tbt dock. seems to work now
#asahi-dev · 2025-09-02 18:16
svenoh, lol, i think i may have fixed it. can't test it here though
#asahi-dev · 2025-09-02 17:53
svenneed to wait for a new a->c cable to arrive before i can look into fixing gadget mode again
#asahi-dev · 2025-09-02 17:52
svenhttps://git.kernel.org/pub/scm/linux/kernel/git/sven/linux.git/log/?h=wip/atcphy gadget mode possibly broken, but otherwise rather happy with this now
#asahi-dev · 2025-09-02 15:50
svenchaos_princess: thanks!
#asahi-dev · 2025-09-02 15:47
svenanything attached to atcphy should work
#asahi-dev · 2025-09-02 15:47
svenweird :D
#asahi-dev · 2025-09-02 15:45
sven3 would be better because i want to see the pipehandler setup as well, but 2 should be fine
#asahi-dev · 2025-09-02 15:45
svenusb 2 only is fine as well
#asahi-dev · 2025-09-02 15:45
svenno, both ports short work :/
#asahi-dev · 2025-09-02 15:39
sventhe only reason they have a different ADT compatible is that they hardcode the fuse offsets inside the kext iirc
#asahi-dev · 2025-09-02 15:39
svenatcphy is the same for t8103 and t6000 and all the other supported SoCs. i think apple hides any differences inside the tunables anyway
#asahi-dev · 2025-09-02 15:37
svenbefore that the approach was macbook air -> c-to-a adapter -> a-to-c cable -> mac mini
#asahi-dev · 2025-09-02 15:37
sveneven under macos
#asahi-dev · 2025-09-02 15:36
sveni have a somewhat weird setup here. c-to-c cable to another (x86) computer running linux and that seems to first negotiate host and then switch to device. I think my a-to-c cable that I used before broke because it doesn't negotiate anything anymore
#asahi-dev · 2025-09-02 15:36
svenanything is fine
#asahi-dev · 2025-09-02 15:33
svencan anyone run trace_atc.py while connecting the target to something that makes it switch to device mode? my setup seems to first negotiate host and then do the switch to device mode and i'd like to see something that directly goes to device mode
#asahi-dev · 2025-09-02 14:27
sveneither when converting to the glue driver or when cleaning that one up
#asahi-dev · 2025-09-02 14:27
sveni've also apparently managed to break gadget mode somehow again :/
#asahi-dev · 2025-09-02 14:26
svenso pipehandler as part of dwc3 glue makes everything worse, it'll stay inside atcphy
#asahi-dev · 2025-09-01 15:41
sveni don't think you can force a role switch though
#asahi-dev · 2025-09-01 15:40
svenyeah
#asahi-dev · 2025-08-31 20:09
svenhm, so maybe b4 to export everything and then mess with the CC headers manually and then run git send-email. sounds annoying though
#asahi-dev · 2025-08-31 20:06
svenno idea if b4 supports that though
#asahi-dev · 2025-08-31 20:06
svenother alternative would be to send each commit with different ccs and only the cover letter to everyone, I’ve seen that done for large series touching many subsystems
#asahi-dev · 2025-08-31 20:04
svenI assume that if the maintainer acks a commit it means we can take them
#asahi-dev · 2025-08-31 19:59
svenyeah, or that
#asahi-dev · 2025-08-31 18:24
svennickchan: i'd also wait until the t602x series is actually reviewed / picked up before sending stuff that depends on it. gets confusing rather quickly otherwise
#asahi-dev · 2025-08-31 18:00
svenmainly thinking about this because i'm trying to bring up a usb3 via usb4 tunnel right now and that needs additional synchronization for when the pipehandler switch happens
#asahi-dev · 2025-08-31 17:59
svenhm.. i think i'll try what happens if i move the pipehandler stuff to the dwc3 glue. most of the synchronization mess comes from that I think and it may fit better inside the dwc3 driver
#asahi-dev · 2025-08-31 13:23
svenasahi prevents burnout from $work and vice versa for me. also other hobbies that are unrelated to computers
#asahi-dev · 2025-08-27 18:10
svenshould've just adjusted the address manually instead of trying to fix that script :D
#asahi-dev · 2025-08-27 18:10
svenupdated
#asahi-dev · 2025-08-27 17:49
svengood catch
#asahi-dev · 2025-08-27 17:49
sventhat's the issue
#asahi-dev · 2025-08-27 17:49
svenyeah
#asahi-dev · 2025-08-27 17:48
svenoh, but i'm using it wrong i think
#asahi-dev · 2025-08-27 17:46
sventhe script uses libfdt via python bindings on the .dtb file fwiw
#asahi-dev · 2025-08-27 17:44
svenor did we mess up the fdt?
#asahi-dev · 2025-08-27 17:44
svenhrm.. did i mess something up in the python script then
#asahi-dev · 2025-08-27 17:39
svenpushed
#asahi-dev · 2025-08-27 17:38
sveni'll add a comment

plus ›