#asahi-dev · 2025-08-15 10:32
chadmedadt_first_child_offset has load bearing UB in that if you return an error when the currently-scoped node does not have any children it will crash, so somewhere on the C side relies on it miscalculating the next node in memory as a "child" of the current node even if it is a sibling or completely unrelated#asahi-dev · 2025-08-15 10:30
chadmedsmdh im trying to make the public adt api more sound and im finding soooo much load-bearing UB on the C side#asahi-dev · 2025-08-15 06:09
LarstiQjannau: if I search for that sentence I get hits on multiple sites, nice#asahi-dev · 2025-08-14 20:10
fl0_idjannau try a blockchain prediction market *ducksÜ#asahi-dev · 2025-08-14 19:42
jannauodds are that has already happened#asahi-dev · 2025-08-14 19:41
chaos_princesslol. they should be fired and replaced with chatgpt (/s)#asahi-dev · 2025-08-14 19:41
jannau> The findings reveal the internal identifier t8152, which matches the “M4 Ultra” scheme and also goes by the codename “Hidra.”#asahi-dev · 2025-08-14 19:40
jannauone of the "journalists" claimaing that#asahi-dev · 2025-08-14 19:37
chaos_princesswho would even take a counter-bet on that, ofc it won't be#asahi-dev · 2025-08-14 19:33
jannauwhere can I bet on mac rumors? "t8152" is not going to be m4 ultra#asahi-dev · 2025-08-14 10:11
jannaudrat!#asahi-dev · 2025-08-14 09:33
kettenisthat's a funny typo ;)#asahi-dev · 2025-08-14 08:43
jannaudrat 4 level page table support sent#asahi-dev · 2025-08-13 13:41
robherjannau: Yeah, checkpatch needs to be updated. It just greps the bindings tree for a match. So "the" would be a "documented" compatible. I have a patch to make it use schemas, but then it is dependent on build artifacts. Maybe we just drop it as we have other ways to check now.#asahi-dev · 2025-08-13 12:04
kettenisso the changes you propose should be fine#asahi-dev · 2025-08-13 12:04
kettenisconstraints set by the bus-range properties I mean#asahi-dev · 2025-08-13 12:02
kettenisjannau: regarding bus-range and iommu, OpenBSD will try to program the bus numbers in a way that meets the bus-range properties#asahi-dev · 2025-08-13 11:56
sveniirc we already had this discussion 2+ years ago when the first upstream attempt for SMC was made #asahi-dev · 2025-08-13 11:55
svenkettenis: even mfd will probe the device in Linux without the node. I’ll try to reply later this week#asahi-dev · 2025-08-13 11:54
mazI'd be happy to help reviving this effort if at all possible.#asahi-dev · 2025-08-13 11:53
mazkettenis: in my case, I'm dependent on the FW loader in u-boot, as all my stuff lives on external storage connected to the USB-A port.#asahi-dev · 2025-08-13 11:51
kettenisthere are bits that don't really make sense#asahi-dev · 2025-08-13 11:50
kettenismaz, jannau: right; I tried to upstream the FW loader at some point but feedback made me realize the linux driver needs some TLC#asahi-dev · 2025-08-13 11:49
kettenisOpenBSD uses this binding already and defenitely doesn't use the additional node to instantiate a driver ;)#asahi-dev · 2025-08-13 11:48
kettenisyeah, I think some push-back is in order#asahi-dev · 2025-08-13 09:00
svennickchan: let’s wait and see what happens. reboot also only has nvmem resources and there the separate subnode was fine #asahi-dev · 2025-08-13 08:58
jannaushould "pci[0-9a-f]{4}," added as NOT vendor property to the list at the begiining of vendor-prefixes.yaml (which explicitly says it should not be extended) or shall I add pci14e4 as broadcom vendor prefix?#asahi-dev · 2025-08-13 08:58
nickchanas long as rtc_offset existence is checked in nvmem-cell-names I suppose it is fine#asahi-dev · 2025-08-13 08:57
nickchanso upstream wants to have the smc rtc node folded into parent this might be possible without busting the "apple,smc" compatible#asahi-dev · 2025-08-13 08:56
jannaurobher: checkpatch complains about "WARNING: DT compatible string vendor "pci14e4" appears un-documented". using "pci${vendor_code},${device_code}" as compatible is not uncommon#asahi-dev · 2025-08-13 08:55
nickchanfor smc rtc support, there are 3 variants: 1. none at all (A11), 2. 4 byte offset (T2), 3. 6 byte offset (M1-M2)#asahi-dev · 2025-08-13 08:49
nickchan-does not like it when the IRC server still had my previous session registered when I timed out and needed to reconnect which makes the original nickname unavailable#asahi-dev · 2025-08-13 08:38
jannausigh, brcmfmac BCM4388 suport is not upstream, I guess I need to strip that from the t6020 submission#asahi-dev · 2025-08-13 05:14
nickchan-(particularly smc-hid but smc-rtc has some uses to me too)#asahi-dev · 2025-08-13 05:13
nickchan-awesome I can definitely make use of more upstream readdy smc drivers#asahi-dev · 2025-08-13 02:18
chadmedso hopefully that goes away once all that is gone#asahi-dev · 2025-08-13 02:17
chadmedm1n1 stage 2 feels a couple hundred ms slower to load with the rust changes but i feel like thats all due to having to relocate pointers every time C asks for an ADT offset #asahi-dev · 2025-08-12 19:03
sveni guess smc-hid is next and then i'll rebase usb3/atcphy and write a cover letter on the weekend and send that one out as well#asahi-dev · 2025-08-12 18:30
jannaumaz: I think the fw loader was never in upstream u-boot#asahi-dev · 2025-08-12 18:26
svenhttps://lore.kernel.org/asahi/20250812-wip-smc-rtc-v1-0-66a8e96dad60@kernel.org/ SMC RTC driver#asahi-dev · 2025-08-12 17:46
mazjannau: hasn't the ASM FW loader been ripped out of u-boot? I haven't updated mine in ages because of that...#asahi-dev · 2025-08-12 17:03
jannauu-boot doesn't care as it has no ethernet driver. could be an issue for the asm hcdon m2 pro mac mini though. I was more concerned about *bsd#asahi-dev · 2025-08-12 12:35
kettenisu-boot doesn't look at it at all#asahi-dev · 2025-08-12 12:34
kettenislooks like linux only uses bus-range for the host bridge#asahi-dev · 2025-08-12 10:29
mazjust checked on my box, and all ports are indeed enabled.#asahi-dev · 2025-08-12 10:27
jannauI think only t8103-j457 is affected in mainline. I'm currently preparing the t602x device trees fior upstream#asahi-dev · 2025-08-12 10:25
jannauI think that probably just doesn't have the port disabled#asahi-dev · 2025-08-12 10:24
mazhow comes this doesn't show? my M2 pro mini seems fine (but I'm running with a 2023-vintage DT, so I may be missing a lot of things...)#asahi-dev · 2025-08-12 10:23
mazhuh.#asahi-dev · 2025-08-12 10:22
jannauyes, iommu-map-ping assigns an disabled dart