#asahi-dev · 2023-06-22 18:36
svena mismatch usually results in obscure kernel panics #asahi-dev · 2023-06-22 18:36
svenmake sure that the kernel you feed to run_guest is the same version as the macOS installation where you installed m1n1#asahi-dev · 2023-06-20 15:49
svenI wonder if that’s a new bug or if I just moved things around enough to hit the issues people reported #asahi-dev · 2023-06-20 15:49
svenugh.. now with the new tipd code usb3 is unstable for me sometimes :/#asahi-dev · 2023-06-20 05:42
sventhere’s at least one that powers off all clusters#asahi-dev · 2023-06-20 05:41
sventhere are some „deeper“ sleep states that we don’t know much about yet #asahi-dev · 2023-06-18 10:47
svenand the reason that’s a bad idea is because when you’re in irq context („top half“) many (or even all) other interrupts are disabled and will be delayed #asahi-dev · 2023-06-18 10:45
sventhe issue happens when you wait for I/O over a slow bus like SPI where you usually have to sleep (or wait for an irq which effectively is sleeping anyway)#asahi-dev · 2023-06-18 10:44
svendoing sone I/O is totally fine in irq context. you usually *have* to do some I/O to e.g. read the irq and ack it in every handler #asahi-dev · 2023-06-18 10:43
svenstill orders of magnitude faster than reading from a SPI bus which is probably what you’re thinking of #asahi-dev · 2023-06-18 10:41
svenbut that readl is still plenty fast #asahi-dev · 2023-06-18 10:41
svenyou also can‘t really access devices behind a slow bus like SPI#asahi-dev · 2023-06-18 10:41
svenyou can’t sleep in irq context#asahi-dev · 2023-06-18 10:40
sventhat’s not true #asahi-dev · 2023-06-18 10:39
svenand isn’t too half the thing that runs in irq context while bottom half is the irq thread anyway? (not that that changes anything afaict)#asahi-dev · 2023-06-18 10:39
svenwhat do you mean with sleep issues? #asahi-dev · 2023-06-18 10:29
svenit’s just this dumb iPhone‘s „you pressed something slightly to the left of i and u is more likely than i in our ML model so I’m just gonna pretend you pressed u instead“ feature *sigh*#asahi-dev · 2023-06-18 10:27
svenand s/but/bit/#asahi-dev · 2023-06-18 10:27
svenyes#asahi-dev · 2023-06-18 10:24
svenand *bit #asahi-dev · 2023-06-18 10:23
sven*IRQ_FLAG#asahi-dev · 2023-06-18 10:23
svenjust guessing here because I don’t know the hardware but it looks like HW_FLAG is just „write but N to clear it“#asahi-dev · 2023-06-15 19:15
svenmarcan: anyway, I ripped out the thunderbolt mess and things still seem to work. I’ll clean up a bit but then I should have something ready on the weekend #asahi-dev · 2023-06-15 19:14
svenunclear, depends at which level their para virt works I guess #asahi-dev · 2023-06-15 19:11
svennice work from agraf as usual :-) now someone just needs to reverse engineer and implement that para virtualized gfx framework :D#asahi-dev · 2023-06-15 12:38
sventhe tipd side is mostly done #asahi-dev · 2023-06-15 12:38
svenuh, I’d have to extract the usb3 part from the thunderbolt mess#asahi-dev · 2023-06-09 13:32
svenyeah#asahi-dev · 2023-06-09 13:25
svenoh well.. i'll have to keep the fuses crap for t8103 then anyway#asahi-dev · 2023-06-09 13:24
svenyeah, probably#asahi-dev · 2023-06-09 13:20
svenI guess I just looked at some older firmware that still did fuses for m2#asahi-dev · 2023-06-09 13:19
svenhuh#asahi-dev · 2023-06-09 13:16
svenfrom what I can tell it should still use the 8112 type c kext which should read those :/#asahi-dev · 2023-06-09 13:15
svenweird#asahi-dev · 2023-06-09 13:13
svenor how did you figure out that they don't have them anymore?#asahi-dev · 2023-06-09 13:13
svenaren't the fuses at 0x000000009E2CC000?#asahi-dev · 2023-06-09 11:03
svenlooks like usb3-only -> usb3-dp transitions also work without resetting the entire dwc3 block :)#asahi-dev · 2023-06-09 10:30
svenmarcan: do you have a m2 pro/max ADT somewhere? or can you show me the atc-phy node?#asahi-dev · 2023-06-09 10:15
sven(i also managed to re-introduce my most favorite "the first thunderbolt connection just doesn't work" bug *sigh*)#asahi-dev · 2023-06-09 10:14
svenyeah#asahi-dev · 2023-06-09 10:14
sven*change#asahi-dev · 2023-06-09 10:14
svenhm... so either it's an actual HW changes or XNU saves/restores fuses then#asahi-dev · 2023-06-09 09:39
svenit looks like just applying fuses for t8103 from m1n1 is not enough for thunderbolt#asahi-dev · 2023-06-09 09:39
svenwhich chip got rid of the atc fuses again, m2 pro/max?#asahi-dev · 2023-06-07 11:36
svenIt’s sad that Broadcom doesn’t seem to care and instead seems to have written that other Android driver :/#asahi-dev · 2023-06-04 12:21
svenotherwise rebasing on torvald’s tree usually ends up being a huge mess #asahi-dev · 2023-06-04 12:21
svenjannau: I’ll cherry pick commits or just c/p DCP I to my local branch anyway#asahi-dev · 2023-06-03 19:14
svenI’ll just use that branch for now#asahi-dev · 2023-06-03 19:14
svensounds like my local setup :D#asahi-dev · 2023-06-03 15:17
svenjannau: what is your latest DCP(ext) branch? I’ll add DP mode switch to the tipd code next