cute

i tried reproducing the openzfs comedy bug and when i looked away for a few minutes i came back to it having rebooted from panicking

and now i'm trying to remember how to get mdb to tell me what just happened from the saved core

it seems like it panicked because it decided my pool deadlocked, and since it was an NVMe device in a VM with a couple of cp commands, that's quite the exciting claim.

oh good, there goes my data

just what i always wanted

^1.

[illumos-gate] 16072 Want manuals for libjedec -- Robert Mustacchi <rm⊙fo>

[illumos-gate] 16081 libjedec tests should run as both 32-bit and 64-bit -- Robert Mustacchi <rm⊙fo>

andyf: reinstall my plex vm and some stuff I noticed: 1. the initial boot hang was because I forgot to pass the audio device // 2. BAR failure is indeed a warning // 3. not having a full reset works, but seems to cause more weird difference in GuC firmware loading

So the three i915 driver options I pass def very helpful :) one to stop it trying to ReBAR, one to disable the display output so linux stops trying to setup a display, one to always do a full reset of the hardware

I can also sort of force ASPM to work by using setpci which does seem to drop idle power usage to below what the P1000 was doing.

Anyway, off to rebuilding the VM

[illumos-gate] 16088 asy can overrun TX on interrupt race -- Patrick Mooney <pmooney⊙pc>