#
#asahi-dev · 2025-04-03 16:43
jannau_https://paste.debian.net/1367284/
#asahi-dev · 2025-04-03 16:43
jannau_yes, thank's to maz pcie changes
#asahi-dev · 2025-04-03 16:43
chaos_princesss/macos/m-chips/
#asahi-dev · 2025-04-03 16:43
chaos_princessi thought macos is vhe only
#asahi-dev · 2025-04-03 16:42
chaos_princesswhat
#asahi-dev · 2025-04-03 16:42
nickchanneed to be under VHE even macOS NV only does nVHE i think
#asahi-dev · 2025-04-03 16:41
nickchanNV GL1 (so it appears to be GL2) should not need bugs but that would be immense complexity I would imagine
#asahi-dev · 2025-04-03 16:41
chaos_princessi think i know where this is going :P
#asahi-dev · 2025-04-03 16:41
chaos_princessoh
#asahi-dev · 2025-04-03 16:41
j`eyPCIe GPU?
#asahi-dev · 2025-04-03 16:41
jannau_renders at least glxgears
#asahi-dev · 2025-04-03 16:41
nickchanalternative is NV GL1
#asahi-dev · 2025-04-03 16:40
jannau_OpenGL ES profile renderer: AMD RV710 (DRM 2.50.0 / 6.14.0-asahi+, LLVM 19.1.7)
#asahi-dev · 2025-04-03 16:40
jannau_I guess I need newer GPUs
#asahi-dev · 2025-04-03 16:39
nickchani already noticed some weirdness on m1 where the stock kernel would refuse to work under the custom boot object unless ctrr is disabled as well
#asahi-dev · 2025-04-03 16:38
chaos_princessnah, they will patch them if we use them in m1n1
#asahi-dev · 2025-04-03 16:37
nickchanbut I hope that bugs are not needed
#asahi-dev · 2025-04-03 16:37
nickchanchaos_princess: right every firmware version that would want to be supported for the hypervisor would need a bug, but then again having a fully hostile xnu is different from typical sptm bypass scenario where the attack will not even have xnu code exec so there may not be much incentive to patch them
#asahi-dev · 2025-04-03 16:34
chaos_princessbut then they patch them and we are still in the same boat for m5
#asahi-dev · 2025-04-03 16:33
nickchansven: in that case might actually need to abuse bugs to gain GL2 control... (though probably not *too* hard with ctrr and such disabled)
#asahi-dev · 2025-04-03 15:33
svenif that's true it's going to make hv bringup very very annoying
#asahi-dev · 2025-04-03 15:33
svenwhen configuring a macho object they'll start this SPTM thing in GL2 and then drop to EL2 and only then give control to the macho
#asahi-dev · 2025-04-03 15:33
sven(i.e. we're running under EL2 with GL2,etc. completely disabled)
#asahi-dev · 2025-04-03 15:32
svenwhat i think happens is that they disabled their custom ISA extensions when using a raw boot object
#asahi-dev · 2025-04-03 15:32
svenmischa85: so it looks like that XNU on M4 expects to be running in EL2 under what Apple calls SPTM in GL2, see e.g. https://github.com/apple-oss-distributions/xnu/blob/main/osfmk/arm64/sptm/start_sptm.s#L34
#asahi-dev · 2025-04-01 13:29
alyssamaz: very fair
#asahi-dev · 2025-04-01 08:19
mazbut on the other hand, I'd never run -stable.
#asahi-dev · 2025-04-01 08:17
mazalyssa: it's definitely at least 1, since bare upstream is all I'm running on my m1 mini. everything else requires tons of patches though.
#asahi-dev · 2025-03-31 20:27
mischa85another suggestion found in irc logs are pointing towards the page tables not being set up properly. common issue is the very first instruction failing. it mentions a translation fault which points to page tables setup, but i don't think that's the case here as mmu is deactivated
#asahi-dev · 2025-03-31 18:27
mischa85and msr(CYC_OVRD_EL1,1) gives a sync exception, so the register is indeed unwritable
#asahi-dev · 2025-03-31 18:04
mischa85i can read the registers tho: mrs(CYC_OVRD_EL1) = 0x0, mrs(ACC_OVRD_EL1) = 0x1800109000040212, mrs(ACC_EBLK_OVRD_EL1) = 0x8800
#asahi-dev · 2025-03-31 18:03
mischa85sven: not having any access yet to m1/m2, but while making a clean branch this one turned out to be suspicious. git blaming the m1n1 tree found that original commit mentioning FIQ
#asahi-dev · 2025-03-31 18:00
svenwhat makes you think that’s the cause?
#asahi-dev · 2025-03-31 17:58
mischa85so the question probably boils down to: how to access the OVRD register?
#asahi-dev · 2025-03-31 17:41
mischa85which doesn't seem to be possible on M4 Donan: https://github.com/yu-re-ka/m1n1/commit/c0a59fa10b311cfb29029ee3c84b917172a54872
#asahi-dev · 2025-03-31 17:32
mischa85i have a strong hunch the absence of this call, which crashes t6041, has something to do with my hv troubles: https://github.com/AsahiLinux/m1n1/commit/5a4e30a98d2ac9a998df49e4c5aa513a222b6509
#asahi-dev · 2025-03-31 16:09
alyssaI guess theoretically there's enough for a m1 mini server already, but I am aware of 0 people who actually do that with an unpatched upstream kernel :p
#asahi-dev · 2025-03-31 16:08
alyssamaz: personally i don't see the point of stable backports until the upstream is in a state where people are actually running torvalds' kernels
#asahi-dev · 2025-03-31 11:26
jannauthe change is referenced via Link: in the commit message
#asahi-dev · 2025-03-31 11:24
jannaufeel free to add my tested-by: I'll reply in a moment
#asahi-dev · 2025-03-31 11:17
mazI'll add the reference to this commit as a justification.
#asahi-dev · 2025-03-31 11:17
mazlet's see if we have better luck this time. if they are annoyed by that, they can drop the cc stable themselves.
#asahi-dev · 2025-03-31 11:14
jannausimilar situation as the m2 pro mac mini. When I tried to submit one of the PCI maintainers wasn't convinced it's needed at all
#asahi-dev · 2025-03-31 11:13
mazjannau: ah, fair enough. I'll reorder the series before reposting it then.
#asahi-dev · 2025-03-31 11:12
jannaumaz: upstream "needs" it for the imac with 2 usb-c ports (t8103-j457) since commit a0189fdfb73d ("arm64: dts: apple: t8103: Disable unused PCIe ports")
#asahi-dev · 2025-03-31 09:48
mazjannau: do we really need the backport to stable of "PCI: apple: Set only available ports up"? prior versions of Asahi don't need to take the backport, and upstream doesn't really care yet. but maybe you have other plans for it that would require backporting?
#asahi-dev · 2025-03-30 20:16
mischa85i will test this on m1 somewhere this week
#asahi-dev · 2025-03-30 20:16
mischa85here is a clean commit of the minimal changes i made to m1n1 for t6041: https://github.com/mischa85/m1n1/commit/a2f3b4a2473a19db577b5c38c7e99bbf3a213823
#asahi-dev · 2025-03-30 17:55
mischa85yes good one, i'll give it a go. first i'll redo the necessary changes on upstream m1n1
#asahi-dev · 2025-03-30 17:53
sveni'd like to know if the hv works on m1/m2 with your patches. if it does it's a HW issue with M4. if it doesn't one of your commits broke it.

plus ›