[Bug 2029502] Re: udev ata_id environment keys not imported for sata hdd

[Nick Rosbrook]

I don't know enough about these interactions to know what is best here.
On the one hand, I don't exactly know the motivation for only setting
ID_SERIAL if it is unset, but that seems like a reasonable thing to do.
And systemd/udev cannot control other packages installing udev rules
that will be run earlier and try to do similar things (as sg3-utils is
here).

I know you asked this question on systemd-devel too, but you might have
better luck with a github issue/PR trying to make the change. Otherwise,
I would suggest using a local udev rule that meets your needs.

If you do open an upstream issue or something, please link it here.

** Also affects: sg3-utils (Ubuntu)
   Importance: Undecided
       Status: New

** Changed in: systemd (Ubuntu)
       Status: New => Incomplete

** Changed in: sg3-utils (Ubuntu)
       Status: New => Incomplete

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/2029502

Title:
  udev ata_id environment keys not imported for sata hdd

Status in sg3-utils package in Ubuntu:
  Incomplete
Status in systemd package in Ubuntu:
  Incomplete

Bug description:
  Hello,

  While debugging why `pm-utils` hdparam udev rules were not picked up by my
  SATA HDD, I found that it depends on the presence of the `ID_ATA_FEATURE_SET_APM`
  env key to set the drive advanced power management. This key is returned by
  `ata_id` correctly but, among other values, they are never imported as
  environment keys by udev.

  They seem to be imported in `/lib/udev/rules.d/60-persistent-
  storage.rules`

  # ATA
  KERNEL=="sd*[!0-9]|sr*", ENV{ID_SERIAL}!="?*", SUBSYSTEMS=="scsi", \
   ATTRS{vendor}=="ATA", IMPORT{program}="ata_id --export $devnode"

  But this rule is not applied since `sg3_utils` already set an
  `ID_SERIAL` in
  https://github.com/hreinecke/sg3_utils/blob/master/scripts/55-scsi-
  sg3_id.rules#L100C61-L100C61

  I workaround this by introducing a new rule that imports the `ata_id` output on that
  device, without the ID_SERIAL condition, which makes hdparm to detect APM correctly.

  So, regardless of how hdparm is detecting APM capabilities, I'm not sure the current
  behavior is expected or not. I would expect that the `ata_id` env properties are
  always imported if the device has `ID_ATA=1`

  # lsb_release -rd
  Description:	Ubuntu 22.04.2 LTS
  Release:	22.04
  # apt-cache policy systemd
  systemd:
    Installed: 249.11-0ubuntu3.9

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/sg3-utils/+bug/2029502/+subscriptions