#
✕ nick:amarioguy
#asahi-dev · 2022-07-14 23:02
amarioguyconsidering to enter the GXF ELs requires a custom instruction
#asahi-dev · 2022-07-14 23:01
amarioguyi mean the main issue i see with this approach as opposed to the EFI runtime service approach i've seen discussed is upstreaming
#asahi-dev · 2022-07-14 22:47
amarioguythink of GL1 and GL2 as lateral to EL1 and EL2 but otherwise acting as their own almost full featured exception levels
#asahi-dev · 2022-07-14 22:46
amarioguyGXF
#asahi-dev · 2022-07-14 22:46
amarioguyit's Apple's idea of it
#asahi-dev · 2022-07-14 22:46
amarioguynot TrustZone as per arm spec
#asahi-dev · 2022-07-14 22:40
amarioguyi'm sure this was already thought about, so if there's a reason we can't use the GXF stuff to get a "normal" PSCI implementation i'm all ears
#asahi-dev · 2022-07-14 22:39
amarioguyit'd be very similar to how other ARM systems use EL2 or EL3 to implement it
#asahi-dev · 2022-07-14 22:39
amarioguyi mean it's not like Linux is using GL2 for anything special at the moment, it should be somewhat simple to shove a PSCI handler in there and other than the custom instruction
#asahi-dev · 2022-07-14 22:38
amarioguywhy not use GL1 or GL2 for PSCI calls?
#asahi-dev · 2022-07-14 22:38
amarioguykinda wanted to put a little input into the psci stuff, i realize it may be a bit late now but still i figured i'd mention it
#asahi-dev · 2022-07-13 19:39
amarioguy(not like that'd be possible in any event)
#asahi-dev · 2022-07-13 19:38
amarioguyfeel free to check that out whenever (note it's mainly just slight changes to address the concerns in the PR so it's nothing earth shattering like raw binaries getting support for DWARF symbols lol)
#asahi-dev · 2022-07-13 19:36
amarioguymarcan: i realize you're out right now, just giving a heads up I addressed most of the concerns you mentioned on my pull request for m1n1
#asahi-dev · 2022-07-01 01:18
amarioguygotcha
#asahi-dev · 2022-07-01 01:12
amarioguybtw quick question does m2 use AICv1 or v2?
#asahi-dev · 2022-06-22 00:08
amarioguyalright the PR is up again, cleaner, less redundant code, 100% less git mistakes
#asahi-dev · 2022-06-21 23:58
amarioguynever mind, figured it out lol
#asahi-dev · 2022-06-21 23:45
amarioguynot too sure how i could reduce the redundant code into a separate method unless I can call another method within a method
#asahi-dev · 2022-06-21 23:28
amarioguyagain not the best git person out there so don't mean to make silly mistakes and all
#asahi-dev · 2022-06-21 23:27
amarioguytrue, should probably update the code just a bit more (had to close the PR but reopening it to do it more "properly")
#asahi-dev · 2022-06-21 23:18
amarioguyi actually was not familiar with that ngl (i usually just wait to commit until i've done initial polishing) but yes i'll note it for the future (if i can amend the PR, i'll do so asap)
#asahi-dev · 2022-06-21 23:14
amarioguy(this is my first pull request for the project please be understanding lol)
#asahi-dev · 2022-06-21 23:14
amarioguyi needed this for stuff i'm doing, and i wanted to contribute back to the project so i figured i'd put in a pull request for this so that I can give back to the community
#asahi-dev · 2022-06-21 23:13
amarioguysent a PR adding raw binary support to hv scripts (a lot of the "load_macho" logic ended up working perfectly for the raw binary logic so besides accounting for differences between macho and raw binaries it was a really simple change)
#asahi-dev · 2022-06-17 02:22
amarioguyused UUID and PARTUUID options along with declaring the root part flat out
#asahi-dev · 2022-06-17 02:20
amarioguychadmed: ah that's fair, though even with that built in it's not finding the filesystem, both in EL1 and EL2, it's probably some configuration i'm missing on my end gotta test it more
#asahi-dev · 2022-06-15 22:47
amarioguyah the apple nvme driver isn't even being compiled in LOL
#asahi-dev · 2022-06-15 22:43
amarioguy(my partitions are set up such that it uses the asahi linux desktop install option but with the backdoor proxy enabled)
#asahi-dev · 2022-06-15 22:42
amarioguybut i can't seem to point it towards the asahi linux desktop install root fs
#asahi-dev · 2022-06-15 22:42
amarioguyalright so my i2c patch *boots* when given an initramfs so i guess good news i don't encounter any immediate panics
#asahi-dev · 2022-06-14 17:23
amarioguyvdanielmo: asahi-re please
#asahi-dev · 2022-06-14 14:58
amarioguyjust confirmed definitely not the i2c patch causing the boot to hang lol
#asahi-dev · 2022-06-14 14:05
amarioguyriiight, though i'll probably still test unmodified linux to see if i get framebuffer output
#asahi-dev · 2022-06-14 14:04
amarioguyah gotcha
#asahi-dev · 2022-06-14 14:03
amarioguyare custom built linux kernels expecting EL1?
#asahi-dev · 2022-06-14 14:03
amarioguy_jannau_: i *think* i was chainloading to a newer m1n1 directly?
#asahi-dev · 2022-06-14 13:59
amarioguyah
#asahi-dev · 2022-06-14 13:59
amarioguythe guide says the DT and kernel are sufficient but i'm not 100% on this
#asahi-dev · 2022-06-14 13:58
amarioguyprobably a dumb question but do i need to send uboot when i do linux.py when tetherbooting?
#asahi-dev · 2022-06-14 13:58
amarioguygoing to check that now, just didn't have time yesterday but yes i was going to check that quickly
#asahi-dev · 2022-06-14 13:55
amarioguyit's odd i was expecting a panic at minimum especially (of note idk if this is linked but after kernel compile, i'm using the image at arch/arm64/boot/image.gz as the kernel image to boot since i assumed the compile process placed it there)
#asahi-dev · 2022-06-14 13:52
amarioguyboot args are earlycon debug rootwait root=UUID={uuidhere}
#asahi-dev · 2022-06-14 13:50
amarioguysince the use_irq flag *seems* to be being handled correctly?
#asahi-dev · 2022-06-14 13:50
amarioguythough i suspect the main issue is more so with the way i'm using the completion rather than the IRQ handling stuff
#asahi-dev · 2022-06-14 13:49
amarioguysven: i have a preliminary version of the i2c patch done (not ready for mainline at all lol, namely because i don't even get penguins on the framebuffer, nor do i get any TTY output over the m1n1 vUART port, also don't have real UART so can't investigate further atm)
#asahi-dev · 2022-06-13 00:28
amarioguynot too familiar with that ngl
#asahi-dev · 2022-06-13 00:27
amarioguyhow would i upload a git patch file here? apparently we use something called matrix?
#asahi-dev · 2022-06-11 18:28
amarioguyah gotcha
#asahi-dev · 2022-06-11 18:27
amarioguysven: quick question, we need to pass platform_get_irq a pointer to a pdev struct, but the pasemi i2c files only show an platform driver struct, do we just pass that or an element of that into the platform_get_irq func?

plus ›