#
#asahi-dev · 2026-11-05 06:42
yukaMy nix store seems to be mostly fine though
#asahi-dev · 2026-11-05 06:41
yukathe logs end at midnight where it ran nix-gc
#asahi-dev · 2026-11-05 06:41
yukaI'm guessing some memory corruption affecting the file system structures
#asahi-dev · 2026-11-05 06:41
yukaLinux pretty sure
#asahi-dev · 2026-11-05 06:40
MichaelReeves[m]like is the actual SSD failing itself or is smth wrong with Linux
#asahi-dev · 2026-11-05 06:40
MichaelReeves[m]does 1TR work?
#asahi-dev · 2026-11-05 06:38
yuka(xfs btw)
#asahi-dev · 2026-11-05 06:37
yukamy filesystem is too fucked
#asahi-dev · 2026-11-05 06:36
yukaLol it no longer boots
#asahi-dev · 2026-11-05 06:36
LarstiQno that was months ago if not more
#asahi-dev · 2026-11-05 06:35
yukaLarstiQ also in the last few days?
#asahi-dev · 2026-11-05 06:35
LarstiQMy one and only .git corruption happened on M1 Pro as well
#asahi-dev · 2026-11-05 06:35
yukaThat would be a bit many at one ;)
#asahi-dev · 2026-11-05 06:35
JamesCalligeros[m]Cosmic rays?
#asahi-dev · 2026-11-05 06:34
yukaguess I'll leave a dmesg -w running on a secondary monitor to catch the logs
#asahi-dev · 2026-11-05 06:33
yukaInput/output error returned on my M1 Pro over night
#asahi-dev · 2026-11-05 01:42
s0ullight[m]But looks like libkrunfw will remove the guest patch :(
#asahi-dev · 2026-11-05 01:38
s0ullight[m]To answer my own question, libkrun, muvm and mesa need changes. Guest kernel, virglrenderer and xwayland seem fine
#asahi-dev · 2026-11-04 21:34
s0ullight[m]does anyone know whatever happened to virtio fence passing? Did it end up becoming a thing. All I found was a rfc from 2023.
#asahi-dev · 2026-11-04 17:42
yukaI was able to recover my linux .git dir and send out the M5 v2 series
#asahi-dev · 2026-11-04 17:42
yukaI mean it seems to be gone now, I'm on 7.2.14-1 though
#asahi-dev · 2026-11-04 17:37
jannaucan you easily boot into an older kernel? or macos?
#asahi-dev · 2026-11-04 17:37
yukanothing stands out
#asahi-dev · 2026-11-04 17:34
jannaudoes `nvme smart-log /dev/nvme0` report anything out of the ordinary?
#asahi-dev · 2026-11-04 17:33
yukafair fair
#asahi-dev · 2026-11-04 17:32
jannauunlikely that it is asahi-7.2.15-1 compared to asahi-7.2.14-1 (which runs here w/o issues) judging by the diff
#asahi-dev · 2026-11-04 17:27
yukanow the question is: is this my machine being broken, or is it some issue in asahi-7.2.15-1
#asahi-dev · 2026-11-04 17:26
yukaafter rebooting there were input output errors again once I started touching the linux .git
#asahi-dev · 2026-11-04 17:26
yukashortly afterwards the machine crashed and I could not log in via ssh
#asahi-dev · 2026-11-04 17:25
yukahad some input/output errors on my M1 Pro internal ssd
#asahi-dev · 2026-11-04 11:24
mazthank you.
#asahi-dev · 2026-11-04 11:23
yukaok!
#asahi-dev · 2026-11-04 11:21
mazyuka: look, we can work together on improving the PCIe driver *and* your confidence that you are actually doing the right thing. but it has to be something that is the result of a discussion. It would be *really* good if you posted what you have on the list so that we can have that discussion and finally make progress on this, because it's literally been years in the making.
#asahi-dev · 2026-11-04 11:09
yukamaz: I'm thankful for your maintainership of the driver. It's just a lot more difficult for a novice like myself to be confident in the, let's say, 50 lines compared to the 15 lines, indeed I was not confident the error handling would work properly, so I left it to someone who can confidently implement it according to your requirements. since then I have collected a bit more C experience but still would prefer not rewriting the driver.
#asahi-dev · 2026-11-04 11:04
jannauand at that point we're duplicating the implementations in pwrctrl
#asahi-dev · 2026-11-04 11:03
jannauand I guess we need separate loops for create_device and power_on_device
#asahi-dev · 2026-11-04 10:57
jannauthe unwind code on error is a little unwiedly but otherwise I don't see an issue with it
#asahi-dev · 2026-11-04 10:55
jannauI would also exclude the documentation from count of changed. it's neither bad nor in pcie-apple
#asahi-dev · 2026-11-04 10:53
mazI'm absolutely serious.
#asahi-dev · 2026-11-04 10:53
mazhere's alternative: you rewrite the whole PCIe driver according to your taste, take it over and maintain it. You have my blessing.
#asahi-dev · 2026-11-04 10:52
mazand these 100 lines and mostly documentation. how is that a problem to document stuff?
#asahi-dev · 2026-11-04 10:52
yukaso I just don't agree with the approach
#asahi-dev · 2026-11-04 10:52
yukawhen the 15 line diff does effectively the same
#asahi-dev · 2026-11-04 10:51
mazif you're not happy with it, then explain what is not good enough.
#asahi-dev · 2026-11-04 10:51
yukait makes a 15 line diff into a 100 line diff
#asahi-dev · 2026-11-04 10:50
yukaiirc this is implemented here but I was not happy/confident with the code and maz said not to bother sending it then
#asahi-dev · 2026-11-04 10:49
yukahttps://cyberchaos.dev/yuka/linux/-/commits/b4/apple-pcie-pwren
#asahi-dev · 2026-11-04 10:49
mazif that doesn't work for any odd reason, I'd like to understand why.
#asahi-dev · 2026-11-04 10:48
mazbetween*
#asahi-dev · 2026-11-04 10:48
maz(at least consensus metween the PCI maintainer and me)

plus ›