#asahi-dev · 2025-05-30 10:35
sveni wonder if this is just a bug inside the validator#asahi-dev · 2025-05-30 10:35
svenokay, allowing two iommus inside snps,dwc3.yaml seems to fix it#asahi-dev · 2025-05-30 10:26
sveni guess because of the snps,dwc3 compatible :/#asahi-dev · 2025-05-30 10:25
svenI wonder why it even tries to validate against snps,dwc3.yaml. I only include snps,dwc3-common.yaml##asahi-dev · 2025-05-30 10:12
kettenismaybe the generic binding just needs to allow for two iommus then#asahi-dev · 2025-05-30 10:09
svenso that's what I get even with the select stanza#asahi-dev · 2025-05-30 10:09
sven from schema $id: http://devicetree.org/schemas/usb/snps,dwc3.yaml##asahi-dev · 2025-05-30 10:09
sven from schema $id: http://devicetree.org/schemas/usb/snps,dwc3.yaml##asahi-dev · 2025-05-30 10:09
svenhome/speter/linux/arch/arm64/boot/dts/apple/t8103-j457.dtb: usb@502280000: Unevaluated properties are not allowed ('iommus' was unexpected)#asahi-dev · 2025-05-30 10:09
svenhome/speter/linux/arch/arm64/boot/dts/apple/t8103-j457.dtb: usb@502280000: iommus: [[104, 0], [105, 1]] is too long#asahi-dev · 2025-05-30 10:09
svenmeh#asahi-dev · 2025-05-30 10:09
sven from schema $id: http://devicetree.org/schemas/usb/snps,dwc3.yaml##asahi-dev · 2025-05-30 10:09
sven from schema $id: http://devicetree.org/schemas/usb/snps,dwc3.yaml##asahi-dev · 2025-05-30 10:04
kettenissven: maybe you need a "select" stanza like in rockchip,dwc3.yaml#asahi-dev · 2025-05-30 09:49
kettenisthat'll break u-boot and OpenBSD, so please try to add it back#asahi-dev · 2025-05-30 09:48
svenmaybe. i removed that snps compatible for now because I didn't want to deal with that :D#asahi-dev · 2025-05-30 09:48
kettenisIsn't it possible to override MaxItems for that in apple,dwc3.yaml?#asahi-dev · 2025-05-30 09:41
sveni can give it another try and see if i can fix it. iirc it was because the apple ones needs two iommus and that clashed with snps,dwc3-common.yaml#asahi-dev · 2025-05-30 09:40
svencouldn't get dtbs_check to pass with that one included :/#asahi-dev · 2025-05-30 09:39
kettenissven: I noticed that you dropped the snps,dwc3 compatible for tor the dwc3 controllers on that wip/atcphy branch#asahi-dev · 2025-05-30 09:09
mazoh, and there is Documentation/admin-guide/dynamic-debug-howto.rst (can't quite believe we have actual documentation!)#asahi-dev · 2025-05-30 09:07
maz /sys/kernel/debug/dynamic_debug/control has all the traces, and you can pick the ones you want to see.#asahi-dev · 2025-05-30 09:03
sveni think that's what i meant, dynamic debug sounds familiar #asahi-dev · 2025-05-30 09:02
mazIIRC, with CONFIG_DYNAMIC_DEBUG, you can enable debug on a call-site basis.#asahi-dev · 2025-05-30 09:02
svenor maybe it was the command line#asahi-dev · 2025-05-30 09:01
sventhere's some way to enable them just for rtkit with a write to some debugfs(?) node iirc#asahi-dev · 2025-05-30 09:00
svenwe can just turn dev_info into dev_dbg for syslog messages#asahi-dev · 2025-05-30 07:30
chadmed_we should probably make rtkit log prints far far far less verbose by default and hide them behind a module param#asahi-dev · 2025-05-30 06:36
svenI also don’t think it’s even related to PAN#asahi-dev · 2025-05-30 06:33
svenwe have multiple ways to work around it but that’s not the point #asahi-dev · 2025-05-30 04:51
nickchanand it turned out hacking data abort handler is less misreable so i didn't really use it#asahi-dev · 2025-05-30 04:47
nickchanbootleg cables with pi pico#asahi-dev · 2025-05-30 04:47
nickchanjust not a very pleasent experience using bootleg cables from pi pico with openocd#asahi-dev · 2025-05-30 04:46
nickchanI *can* use a swd debugger on a11 which has pan if that would help#asahi-dev · 2025-05-30 04:44
nickchana way to workaround is probably to patch the pan codes on startup instead of checking at runtime which is already done for the exception vectors but I suppose there's a deeper issue here#asahi-dev · 2025-05-29 22:21
jannaumy theory of how increasing dummy_stack avoids the problem is that something else gets overwritten#asahi-dev · 2025-05-29 22:19
jannauI haven't seen supports_pan being inlined. one variant of "fixed by too much debug" I've seen was saving LR at function entry#asahi-dev · 2025-05-29 21:58
svenwhat I don't understand is why increasing DUMMY_STACK makes the issue disappear as well#asahi-dev · 2025-05-29 21:56
svenand then it never stored/loaded lr from the stack#asahi-dev · 2025-05-29 21:56
sven*inlined#asahi-dev · 2025-05-29 21:56
sveni'll have to try again but I think what happened is that supports_pan was somehow inline when I added "too much debugging" to mmu_setup_secondary#asahi-dev · 2025-05-29 21:54
jannauaccording to my debugging SP was consistent within mmu_setup_secondary but I don't see how "too much" debugging inside mmu_setup_secondary would prevent something corrupting the secondary stack of the affected core#asahi-dev · 2025-05-29 21:42
sventrue#asahi-dev · 2025-05-29 21:41
jannaucould be memory contents#asahi-dev · 2025-05-29 21:40
svenand adding just slightly too much debugging made the exception disappear as well#asahi-dev · 2025-05-29 21:40
svenI didn't even find any instruction that would write 0x2000000 somewhere and would make sense in that context when I checked a few weeks ago#asahi-dev · 2025-05-29 21:38
jannauI still do not understand against what we could be racing though#asahi-dev · 2025-05-29 21:31
svenif this really is some timing issue i'm not sure even a SWD debugger for host m1n1 would help debugging this