#asahi · 2026-11-08 22:39
kitlith*nods* -- I'll do what I can, but it's almost certainly not under apple care anymore, and certainly never under my name.#asahi · 2026-11-08 22:36
kitlithfl0_id: it's an M1 MacBook Pro. *shrugs*#asahi · 2026-11-08 22:23
kitlithupdate on my end: the fan does spin if I attempt to turn it on, but no there's audio chime or display output.#asahi · 2026-11-08 20:07
kitlithfl0_id: places. it was plugged into the wall overnight. I suppose it is also possible that the brick gave up instead of the laptop, though.#asahi · 2026-11-08 19:24
kitlith(it's plausible that the burnt smell is unrelated since I don't usually have cause to smell my laptop, but...)#asahi · 2026-11-08 19:22
kitlithany recommendations for m1 repair? I left my laptop out and running last night, and came back to it today with it turned off, the power light on the charging cable not turning on, and a burnt smell.#asahi · 2026-09-14 20:07
kitlithchaos_princess: I think I hadn't internalized the existence of -funsafe-math-optimizations (which -ffast-math includes), so I didn't catch your allusion to it. #asahi · 2026-09-14 19:56
kitlith(e-core didn't have slowdown on my machine, this is probably not news)#asahi · 2026-09-14 19:26
kitlithhm. rust-denormals-bench mentions a run on m2 max having no slowdown, but i don't know if that's p-core or e-core. (probably p-core)#asahi · 2026-09-14 19:20
kitlith(if that was tongue in cheek, my apologies)#asahi · 2026-09-14 19:17
kitliththe way -ffast-math sets it is not particularly safe <https://moyix.blogspot.com/2022/09/someones-been-messing-with-my-subnormals.html>, i think rust considers having those set to be UB since compiler optimizations assume that they're not set.#asahi · 2026-09-14 19:13
kitlithlinkmauve: almost, -ffast-math also tends to set DTZ/DAZ ("denormals to zero"/"denormals as zero")#asahi · 2026-05-24 03:36
kitlithI could use some help tracking down why let/const behaves so poorly in <https://gist.github.com/kitlith/7c773cdf03e0201c9c0ddf742972a8a3> compared to var on asahi. (maybe mesa in general? but i don't have AMD hardware to confirm against)#asahi · 2026-04-21 08:21
kitlithsigh#asahi · 2026-04-21 08:21
kitlithah, yes, ofc, LOS22 (android 15) doesn't actually include the changes to ART in AOSP android 15 to support 16K.#asahi · 2026-04-21 08:14
kitlithhmmm. "Loading odex files is only supported on devices with 4K page size"#asahi · 2026-04-21 07:53
kitlith<https://github.com/waydroid/waydroid/issues/692#issuecomment-2906983573> -- idk.#asahi · 2026-04-21 07:51
kitlithWhile waydroid does not officially support android 15 (currently on android 13, i think?) there is someone providing unofficial android 15 & 16 ROMs for waydroid, and my poking around at android 15 is based on their work.#asahi · 2026-04-21 07:50
kitlithchaos_princess, barusu: waydroid has provided "arm64_only" ROMs for almost a year at this point, so 32-bit support is no longer necessary. Starting with android 15 or so, there's a couple of build flags to state that you want to build ELFs to support 16k page size.#asahi · 2026-04-20 06:10
kitlithI've dug a little bit into both approaches (custom kernel for muvm a while ago, and more recently a custom android 15 rom), and both times I've gotten bogged down at "well I got the container running, and now I can't tell where the actual issue is in the boot logs are"#asahi · 2026-04-20 06:08
kitlithbarusu: waydroid doesn't have a ROM available that is capable of a 16k page size. muvm is missing a bunch of kernel config/modules in order to be able to run waydroid. Neither of these are insurmountable afaik, it just needs someone to dig in (and maybe be familiar with android's boot process).#asahi · 2026-04-02 02:21
kitlithon the bootloader hardening side of things after installing asahi, you could look into password protecting grub & u-boot, or disabling the u-boot cmdline entirely. (I don't know if I'd reccomend the last one, myself.)#asahi · 2026-03-30 19:36
kitlithTomus[m]: https://github.com/nix-community/nixos-apple-silicon/blob/main/apple-silicon-support/packages/linux-asahi/default.nix#asahi · 2026-01-21 05:00
kitlithsome amount of linear algebra may be useful for optimizations done in the shader compiler, but the specific optimization you're mentioning, involving diagonalization of a matrix, feels more like the domain of an optimization a shader author might apply. It is a different algorithm, if that makes sense.#asahi · 2026-01-21 04:57
kitlithbigmac: the rough division i'd expect is that the kernel space driver talks to hardware, the user space driver does compiliation (general purpose optimization), and the shader author does algorithm design (special purpose optimization).#asahi · 2024-10-10 02:09
Kitlithhm. and testdisk is also only finding the APFS partitions, and not the EFI partitions that I *know* are still present (because macos can see them)#asahi · 2024-10-10 01:18
KitlithI'm finally following up on something I was working on a couple weeks ago, where I took a raw disk image of my m1 macbook pro from recovery (via /dev/rdisk0), and transferred it to another computer for inspection. However, while I see a protective MBR present, I don't see a GPT? So I feel like I've missed something important.#asahi · 2024-09-21 06:04
Kitlithphoenix hyperspace has an interesting implementation, though, lol#asahi · 2024-09-21 06:03
Kitlith^ none of these things were aimed at any kind of technical user and therefore didn't include those things#asahi · 2024-09-21 05:48
Kitlithchadmed: cathode ray dude's been doing a series on youtube all about stuff like that, linux comes up a lot.#asahi · 2024-09-21 04:56
KitlithIn any case, thank you. I'll report back once I have a chance to poke around the disk image. (i.e. maybe sometime tomorrow)#asahi · 2024-09-21 04:49
Kitlithrecoveryos doesn't have man :) (but i figured it out)#asahi · 2024-09-21 04:45
Kitlithnicolas17: yeah, something in my network seems to be bottlenecking to around 25-26MiB/s, neither rdisk or bs are likely to be making much of a difference. as long as i can keep it from going to sleep, though, it should be fine to run overnight.#asahi · 2024-09-21 04:33
Kitlithchaos_princess: unforunately, i don't have a spare 1TB external drive lying around, otherwise i'd totally just use it.#asahi · 2024-09-21 04:08
Kitlith*nods nods*#asahi · 2024-09-21 04:06
Kitlithso "treat it like normal" and "do something to pipe it over the network", got it.#asahi · 2024-09-21 03:54
Kitlithnicolas17: alright, i'm in recovery. how should i go about making the disk backup?#asahi · 2024-09-21 03:47
Kitlithswitching computers, brb#asahi · 2024-09-21 03:18
KitlithI'll wipe the work partition (by letting diskutil wipe it by creating a noformat partition there) and then opt for a disk backup from recovery.#asahi · 2024-09-21 03:16
KitlithYeah, I'm not worried about the macOS partition (i figured that was already handled securely enough) but the "work asahi" partition i had setup.#asahi · 2024-09-21 03:11
Kitlith`diskutil list` lists 99.5GB and 491.6GB of free space, respectively#asahi · 2024-09-21 03:09
Kitlithnicolas17: It is free space, so far as I can tell. The EFI partitions are still present, so it shouldn't have been able to grow the macOS partition into that space (it would overlap with the EFI partitions)#asahi · 2024-09-21 00:57
KitlithI *very* much do not want to attempt to recover any material from the former workplace, but it seems useful to know what actually remains.#asahi · 2024-09-21 00:54
KitlithContext here, to avoid XY: My former workplace remotely wiped the laptop and released it from their provisioning recently after letting me go. I had two asahi installs -- one was used for rustconf, and contains no work material. The EFI partitions from both installs remain, but the data partitions are... missing.#asahi · 2024-09-21 00:49
Kitlithweird question: is there a way from MacOS to inspect the contents of non-partitioned space/recreate a partition that was previously deleted? diskutil when told to not format says that it'll zero the space.#asahi · 2024-01-05 05:55
kitlithI think I'm done for the night, but I am somewhat interested in digging deeper at some point. will check logs tomorrow#asahi · 2024-01-05 04:34
kitlithAccording to DM compat, apple_nvme is apple,nvme-ans2 compatible#asahi · 2024-01-05 04:31
kitlithhm. what's the best way to check that?#asahi · 2024-01-05 04:22
kitlithno device specified#asahi · 2024-01-05 04:22
kitlithit returns 0 on both partitions 5 and 9, and 1 on partition 7 and 1.