#asahi-dev · 2026-05-06 12:25
yukaThe question is: Based on which cpu feature I should decide whether to clear SYS_IMP_APL_L2C_ERR_STS? cyc_ovrd? or introduce a new one#asahi-dev · 2026-05-06 12:25
yukaI usually just comment out the respective lines, but I would like to fix this more properly.#asahi-dev · 2026-05-06 12:25
yukaSmall question: On M4 and A18 Pro, m1n1 can currently never recover since the exception handler itself throws another SError while clearing SYS_IMP_APL_L2C_ERR_STS.#asahi-dev · 2026-05-04 21:26
yukaStill, I would expect if things work somehow, it would enumerate the wifi device#asahi-dev · 2026-05-04 21:25
yukaSome things which might seem like a problem (like "no iommu-map translation for id") are also present on my M1 Pro, so I ignored them#asahi-dev · 2026-05-04 21:24
yukaI don't have an M3 boot log to compare to#asahi-dev · 2026-05-04 21:23
yukaSo for now I'm going with the last remaining option, 39X028000#asahi-dev · 2026-05-04 21:23
yuka39X024000 ready wait timeout#asahi-dev · 2026-05-04 21:22
yuka39X044000 SError#asahi-dev · 2026-05-04 21:22
yuka39X048000 Serror#asahi-dev · 2026-05-04 21:22
yuka39X03C000 SError#asahi-dev · 2026-05-04 21:21
yukaI tried all plausible regs for portX#asahi-dev · 2026-05-04 21:21
yukahttps://termbin.com/axk3#asahi-dev · 2026-05-04 21:18
yuka> bridge configuration invalid ([bus 00-00]), reconfiguring#asahi-dev · 2026-05-04 21:18
yukaand:#asahi-dev · 2026-05-04 21:18
yuka> link didn't come up#asahi-dev · 2026-05-04 21:18
yukahttps://termbin.com/qgkp#asahi-dev · 2026-05-04 17:06
yukaOk I can work around this, but it might need a new device tree field to represent this#asahi-dev · 2026-05-04 17:04
yukaBut gpio 0x86 is indeed the one we can not read#asahi-dev · 2026-05-04 16:51
yukaAh but that's not new. it just has a slightly different format now#asahi-dev · 2026-05-04 16:50
yukathe adt has "no-restore-gpio-pins = [0x0000003100000086, 0x0000003300000032]"#asahi-dev · 2026-05-04 16:48
yukathe regmap cache does that in pinctrl probe#asahi-dev · 2026-05-04 16:48
yukaattempting to read32(0x31a000218) SErrors#asahi-dev · 2026-05-04 16:47
yukaps_gpio is on#asahi-dev · 2026-05-04 16:47
yukaThe pinctrl_ap is supposedly at 0x31a000000#asahi-dev · 2026-05-04 16:46
yukahmmmm#asahi-dev · 2026-05-04 15:57
yukahttps://yuyu.yuka.dev/public/2026-05-04-17_56-j700-dump_pmgr2.log#asahi-dev · 2026-05-04 15:57
yuka(index)#asahi-dev · 2026-05-04 15:57
yukaand before that is the reg#asahi-dev · 2026-05-04 15:57
yukaOk I think I got it: the last 24 bits of unk2_1 are the offset#asahi-dev · 2026-05-04 15:27
yukayep#asahi-dev · 2026-05-04 15:22
yukaThe thing it errors on is DEBUG_GATED which has an offset (unk2_1) out of range of the primary pmgr and was also in pmgr_mini on M3#asahi-dev · 2026-05-04 15:17
yukathis is looking excellent#asahi-dev · 2026-05-04 15:16
yukathe things I get up to the SError match what I would expect as the primary pmgr nodes#asahi-dev · 2026-05-04 15:15
yukaAha#asahi-dev · 2026-05-04 14:08
yukaso I just disable boot security, go to the preboot or whatever partition and start editing the devicetree binary?#asahi-dev · 2026-05-04 14:07
yukauhm#asahi-dev · 2026-05-04 14:06
yukacan you point me somewhere for how to do this?#asahi-dev · 2026-05-04 14:06
yuka> enabling i_can_has_debugger#asahi-dev · 2026-05-04 13:55
yuka(obviously not since it starts SErroring halfway into the dump)#asahi-dev · 2026-05-04 13:53
yukanot saying it's correct or the entire solution#asahi-dev · 2026-05-04 13:53
yukaunk2_1 seemed interesting, and using it as an offset into pmgr.get_regs(0)[0] looks promising#asahi-dev · 2026-05-04 13:52
yukaid2 is just counting up as previously#asahi-dev · 2026-05-04 13:52
yukaI looked at the properties of the pmgr devices entries#asahi-dev · 2026-05-04 13:52
yukahttps://yuyu.yuka.dev/public/2026-05-04-15_51-j700-dump_pmgr.log#asahi-dev · 2026-05-04 13:50
yukaI was going to send the text file#asahi-dev · 2026-05-04 13:50
yukaWhoops#asahi-dev · 2026-05-04 11:41
yukaAh nice so when enabling the debugusb you can choose which pins to put it on, and currently tuxvdmtool simply puts it on the normal data pins but you could also put it on some unused pins to continue using the data pins normally?#asahi-dev · 2026-05-04 11:14
yukacan you explain that a bit more? is this something we can do through software(#asahi-dev · 2026-05-04 11:13
yuka> map debugusb to the non-connected usb2 pins