Whether this bug is worth fixing, we need to consider the whole picture:
- The delay is fundamentally the physical spin-up time of a HDD returning from
STANDBY, which is unavoidable regardless of kernel.
- The only reason it blocks the resume path is that 95hdparm-apm runs hdparm
synchronously inside the sleep hook.
- A system without the hdparm package already resumes immediately and lets the
disk spin up in the background. The hdparm hook is causing the anomaly.
- On this machine the disk is unmounted, so there is no functional impact from
resuming before it is ready.
Possible fix (if we decide to fix it):
Make the resume hook launch hdparm outside the systemd-suspend.service cgroup
so it no longer gates resume completion, with a fallback for non-systemd
systems:
--- a/debian/extra/95hdparm-apm
+++ b/debian/extra/95hdparm-apm
@@ resume_hdparm_apm()
if [ -n "$apm_opt" ]; then
- hdparm $apm_opt $dev
+ # Avoid blocking the suspend service while libata EH holds the
+ # SCSI queue frozen during disk spin-up (LP: #2160187).
+ if command -v systemd-run >/dev/null 2>&1; then
+ systemd-run --no-block --quiet -- hdparm $apm_opt $dev
+ else
+ hdparm $apm_opt $dev
+ fi
fi
@@ resume_hdparm_spindown()
if [ -n "$apm_opt" ]; then
- hdparm $apm_opt $dev
+ if command -v systemd-run >/dev/null 2>&1; then
+ systemd-run --no-block --quiet -- hdparm $apm_opt $dev
+ else
+ hdparm $apm_opt $dev
+ fi
fi
Tested on the affected machine (6.17.0-1029-oem, no nohdparm): 5
consecutive fwts s3 cycles all showed resume 0.44s (down from ~6.9s)
with hardware sleep at 91-93%. The APM level is still correctly applied
to the disk ~6s later once EH completes (verified via journal:
hdparm[...]: APM_level = 254), so drive power management behavior is
preserved.
** Changed in: linux-oem-6.17 (Ubuntu)
Status: New => Triaged
--
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
