1. Why older kernels / earlier tests didn't fail:

6.17 actually resumes faster. On 6.11, PM: resume of devices complete
took ~390ms. On 6.17, it takes ~109ms. That speedup changes a timing
race: on 6.17 the libata EH thread is already running (queue frozen) by
the time the hdparm hook fires, so hdparm consistently blocks. On 6.11
the slower PM device phase means EH tends to start after the hook has
run, so hdparm sees an unfrozen queue and returns instantly. The libata
EH and hdparm code are identical between 6.11 and 6.17; only the timing
shifted.

2. The disk involved is not in use:

The disk causing the delay is /dev/sdb, a Seagate SATA HDD. On this system it 
is not mounted anywhere:
- / (root) and /swap.img are on NVMe
- fstab has no reference to sdb
- sdb1 exists as a partition but is unmounted secondary storage

The 5s physical spin-up is being paid for a disk nothing on the system
is using at resume time.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2160187

Title:
  Suspend resume time is longer than the threshold on Dell Pro Tower
  QCT1255.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux-oem-6.17/+bug/2160187/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to