Hello,
Just looking for some input here (hoping somebody has run across something similar).
Background:
Finally got around to updating my desktop to nixos-21.05. All went well (always kudos to the NixOS crew). However, I noticed when I hibernate my system (LUKS, full-disk outside of boot) and resume the CD/disk tray opens and closes before my cryptsetup prompt comes up (stage-1).
Debugging:
Knowing that this occurs at stage-1 drastically reduced my search area (has to be udev/systemd/kernel etc.). But nothing stuck out at me with a quick browse of the nixpkgs tree log. So I went down the full bi-sect route. Around 5000 commits split later I knocked it down to a kernel update (I normally run linuxPackages_latest). So I backed it off to the LTS release, problem still occurred. I despise when these things happens, as this points to a bad fix back-ported.
Little while later picking off the kernel minor number (5.10.47 - good, 5.10.48 - bad) and a shuffle through the release log and I’ve got my bad commit:
scsi: sr: Return appropriate error code when disk is ejected
- else if (med->media_event_code == 3)
return DISK_EVENT_EJECT_REQUEST;
Now my programming background was in networking, so I’m not as well versed with the SCSI driver (I also haven’t committed any kernel work in >10 years). So I’m wondering if i should go about reporting this upstream or go about looking through the disk-access stack and seeing whats picking up the event and acting on it. Thoughts?