#asahi-dev · 2026-01-17 01:47
fl0_id___and I also put anohter printk which is also not shown#asahi-dev · 2026-01-17 01:38
fl0_id___there should at least be pr_notice("Kernel command line: %s\n", saved_command_line); before the problem section#asahi-dev · 2026-01-17 01:33
fl0_id___this is my cs serial output https://www.pastery.net/ganapx/#asahi-dev · 2026-01-17 01:30
yukaI see#asahi-dev · 2026-01-17 01:11
fl0_id___I should prob use the cs m1n1 connection though#asahi-dev · 2026-01-17 01:10
fl0_id___yuka, no this stacktrace from hv over m1n1 usb#asahi-dev · 2026-01-17 01:09
yukaSo possibly not available out of the box, but it should be possible to be inspired by this code to make my own debug_putc work after mmu init#asahi-dev · 2026-01-17 01:08
yukaThere is code related to it in the arm64 head.S#asahi-dev · 2026-01-17 01:07
yukaIs this stacktrace from earlycon output?#asahi-dev · 2026-01-17 01:07
fl0_id___yuka, it seems like this is not available for arm64, no?#asahi-dev · 2026-01-17 00:58
yukaIt seems CONFIG_DEBUG_LL would be useful for getting even earlier uart#asahi-dev · 2026-01-17 00:56
fl0_id___https://www.pastery.net/yenhwv/#asahi-dev · 2026-01-17 00:56
fl0_id___at least I have symbols now after disabling random base#asahi-dev · 2026-01-17 00:56
fl0_id___yes, after vectoring to next stage#asahi-dev · 2026-01-17 00:56
fl0_id___though I didn't trace it to head.S but build_all_zonelists#asahi-dev · 2026-01-17 00:55
yukaJust getting no output when starting the kernel?#asahi-dev · 2026-01-17 00:55
fl0_id___yuka, this seems very similar to what i am seeing#asahi-dev · 2026-01-17 00:49
yuka(Or maybe enabling the mmu is preventing debug_putc from working afterwards)#asahi-dev · 2026-01-17 00:47
yukaWith the help of debug_putc, I traced it into the __enable_mmu function in head.S of the kernel#asahi-dev · 2026-01-17 00:47
yukaAs a reminder, I was not seeing any output after "Vectoring to next stage" despite having the serial console enabled in the device tree and setting earlycon#asahi-dev · 2026-01-17 00:46
yukaI debugged my minimal device tree with a single cpu core on t8132 baremetal further. Copied the debug_putc into the kernel and added debug prints in the very first code of the kernel.#asahi-dev · 2026-01-16 22:48
svenso you get about 5 more seconds before the SoC resets :D#asahi-dev · 2026-01-16 22:48
svenso the problem is that un-gracefully shutting down the co-processor by plugging in something for the second time into that port also triggers a nasty watchdog#asahi-dev · 2026-01-16 22:48
svenlol#asahi-dev · 2026-01-16 22:47
noopwafelsven: who needs multiple cables :) sounds promising#asahi-dev · 2026-01-16 22:43
sventhe moment I shut that co-processor down via rtkit I can't access the IOMMU's MMIO range anymore#asahi-dev · 2026-01-16 22:43
svenat this point i'm also sure that the iommu as a child of acio makes sense: I think the IOMMU is actually part of that co-processor and we only get a view into its MMIO range#asahi-dev · 2026-01-16 22:30
sven*for#asahi-dev · 2026-01-16 22:30
svenand iirc that's all you need to cross-domain USB4#asahi-dev · 2026-01-16 22:30
sveni never traced a connection like that, but the TX and RX rings of the NHI should be up#asahi-dev · 2026-01-16 22:29
svenfair enough, no promises it'll work though#asahi-dev · 2026-01-16 22:29
chaos_princessthen that is not useless :P#asahi-dev · 2026-01-16 22:29
svenunless there's some vendor specific apple stuff missing it should work#asahi-dev · 2026-01-16 22:29
svenmaybe, haven't tested that#asahi-dev · 2026-01-16 22:29
chaos_princessout of curiosity, will host-to-host networking work?#asahi-dev · 2026-01-16 22:28
svenstill need to fix that assert_id=1333 crash on the second cable that's plugged in though#asahi-dev · 2026-01-16 22:27
leionevertheless I think that this is something to get very excited about ;)#asahi-dev · 2026-01-16 22:25
sven-tpp#asahi-dev · 2026-01-16 22:25
sven(before people get too excited: RFC, and the NHI only. so no pcie tunneling, etc. at best you get too your device in boltctl and it will still be useless)#asahi-dev · 2026-01-16 22:24
svenok. if all goes well i'll just send whatever i have on sunday or so and cc then people from that submission#asahi-dev · 2026-01-16 22:22
jannauwithout thunderbolt support though from reading the commit message closely#asahi-dev · 2026-01-16 22:21
jannausven: as far as I'm aware yes. you could ask in #aarch64-laptops. slightly related support for the retimer/mux is already upstream: https://lore.kernel.org/linux-arm-msm/20250206-x1e80100-ps8830-v6-2-60b1e49cfa8d@linaro.org/#asahi-dev · 2026-01-16 22:08
svenhttps://lore.kernel.org/linux-arm-msm/20250916-topic-qcom_usb4_bindings-v1-1-943ecb2c0fa7@oss.qualcomm.com/ is that the only thunderbolt / usb4 patch from qualcomm or did I miss any follow up?#asahi-dev · 2026-01-16 21:06
noopwafelnow time to try working out what is actually needed from this mess..#asahi-dev · 2026-01-16 21:01
noopwafelI'm just kind of surprised it's reproducible#asahi-dev · 2026-01-16 21:01
Shiznop sled? try AAAA-sled#asahi-dev · 2026-01-16 21:01
svenunderstandable :D#asahi-dev · 2026-01-16 21:01
Shiz(but it does work without)#asahi-dev · 2026-01-16 21:01
Shizat this point the padding is fully placebo and thus makes us feel better#asahi-dev · 2026-01-16 20:58
svenalso, lol, i guess you still need to pad boot arguments? :D