#asahi-dev · 2024-10-28 20:59
jannaukettenis: yes. that not merged for v2024.10. I spotted it while searching in the u-boot mailing list for iommu#asahi-dev · 2024-10-28 20:57
kettenisjannau: that is different breakage from that dev_get_priv() thing though isn't it?#asahi-dev · 2024-10-28 20:51
nicolas17the iMac M4 "system block diagram" was leaked btw (FCC screwed up), but I guess that's nowhere near as interesting as schematics#asahi-dev · 2024-10-28 20:47
jannauit looks like it will break things. if it doesn't we're probably only saved by phys memory starting high#asahi-dev · 2024-10-28 20:45
kettenisah, I was afraid it would break things#asahi-dev · 2024-10-28 20:44
jannaubreaking patch is https://lore.kernel.org/u-boot/20240826115940.3233167-6-sughosh.ganu@linaro.org/#asahi-dev · 2024-10-28 20:44
jannaudo you know what's the status of https://lore.kernel.org/u-boot/20240826115940.3233167-1-sughosh.ganu@linaro.org/ is? That seems to break apple_dart because it tries to lump the per dart lmb use of iova space into a single global LMB#asahi-dev · 2024-10-28 20:40
kettenisI can look into that, but it might take me a few days#asahi-dev · 2024-10-28 20:37
kettenisso maybe implementing .connect() and a new .disconnect() and drop the remove#asahi-dev · 2024-10-28 20:29
jannauyes, doing that. -EPROBE_DEFER in remove does't seem to have the intended effect. remove is not called again#asahi-dev · 2024-10-28 20:22
ketteniscould keep track of the mappings and hope the drivers properly remove them#asahi-dev · 2024-10-28 20:15
jannaucan it know that? iommu_ops has .connect() but no .disconnect()#asahi-dev · 2024-10-28 20:12
kettenismaybe the dart driver needs to return -EPROBE_DEFER if there are still active users#asahi-dev · 2024-10-28 20:11
kettenisand the iommu references go sideways in the tree#asahi-dev · 2024-10-28 20:10
kettenisI think the u-boot device model is weaker than the linux version#asahi-dev · 2024-10-28 20:10
jannauindeed. I thought the device model would prevent that#asahi-dev · 2024-10-28 20:04
kettenismaybe apple_dart_remove() gets called before xhci_{dwc3|pci}_remove()?#asahi-dev · 2024-10-28 20:00
jannaukettenis: can you think of a reason why dev_get_priv() in apple_dart_unmap() for the usb xhci returns NULL on the studio in v2024.10. reproduced in v2024.10-rc2, v2024.07 was ok#asahi-dev · 2024-10-28 19:54
jannauyes, apparently they are doing daily announcements until they run out of devices#asahi-dev · 2024-10-28 19:52
kettenisso we get an m4 imac but nothing else yet?#asahi-dev · 2024-10-28 18:04
chaos_princessthose are useless, since you can't enter 1TR on those#asahi-dev · 2024-10-28 18:03
aakThere those mac mini stadiums that you can rent one, but again cant test hardware#asahi-dev · 2024-10-28 18:03
chaos_princessit is not a vm in a classic sense, more that it uses vm-adjacent technology to allow you to puppeteer macos from another machine#asahi-dev · 2024-10-28 18:02
nicolas17well there are no VMs that emulate Apple's custom hardware#asahi-dev · 2024-10-28 18:02
aakthe problem is with touch id and hardware related stuff its trickier, although maybe you can have emulators?#asahi-dev · 2024-10-28 18:02
aakI thought there were vm's where you could test this stuff#asahi-dev · 2024-10-28 18:02
chaos_princessbecause the standard method involves two machine debugging#asahi-dev · 2024-10-28 18:02
chaos_princessif you want to do development on kernel components you most likely want another machine, yes (any computer works, does not have to be a mac)#asahi-dev · 2024-10-28 18:01
nicolas17it can be done from Linux, it's just more of a hassle#asahi-dev · 2024-10-28 18:01
aakin case i need to re flash#asahi-dev · 2024-10-28 18:01
aakYea I don't have another mac so probably a good idea to get one#asahi-dev · 2024-10-28 18:00
aakI think I'll try to get a spare machine from ebay, even if it's fixable screwing up my one and only machine would be costly#asahi-dev · 2024-10-28 18:00
chaos_princesspmuch that. The reflash is kind of annoying to do if you do not have another mac (intel is fine, just needs to run macos).#asahi-dev · 2024-10-28 17:59
nicolas17aak: if you go out of your way to screw up the system bootloader or something, you'll have to connect it to another computer to reflash it, that's the worst that can happen#asahi-dev · 2024-10-28 17:57
aakah right#asahi-dev · 2024-10-28 17:57
chaos_princessslang for "you are on your own if you need to fix it"#asahi-dev · 2024-10-28 17:57
aak> get to keep the pieces?#asahi-dev · 2024-10-28 17:56
chaos_princessin fact, you are required to keep a macos install, while uninstalling it would work, this configuration is unsupported for various reasons, and if you do it and something breaks - you get to keep the pieces#asahi-dev · 2024-10-28 17:55
aakI can boot back and forth between asahi and macos? I thought that only worked on intel macs#asahi-dev · 2024-10-28 17:52
chaos_princessSo, just install a second copy of the os, it will be fine#asahi-dev · 2024-10-28 17:52
chaos_princessAnd even then it is unlikely to happen#asahi-dev · 2024-10-28 17:52
chaos_princessbut unless you are doing things like messing with partition tables, you are unlikely to destroy a different os other than one you are actively working on#asahi-dev · 2024-10-28 17:51
chaos_princesswith few specific exceptions, you can't damage those machines from software, and while an os recovery is kind of a pain, it works.#asahi-dev · 2024-10-28 17:50
aakI don't want to brick my main one#asahi-dev · 2024-10-28 17:50
aakWhat do people use for testing features, other than having a spare machine#asahi-dev · 2024-10-26 12:52
chaos_princesslinux runs directly on hardware, with no hypervisor present#asahi-dev · 2024-10-26 12:52
aakmakes sense, ty!#asahi-dev · 2024-10-26 12:51
chaos_princessthe hypervisor part is only for exploring how macos does things and for some debugging, correct.#asahi-dev · 2024-10-26 12:51
aakah so the hypervisor is just temporary, got it#asahi-dev · 2024-10-26 12:51
svenand all performance critical parts of the hypervisor are implemented in C anyway