I need a bit of help from my fellow W680-ACE owners! (ATX version) -
I've been seeing an erratic issue where SATA link power management appears to cause random timeouts when an access is made to a connected HDD after an idle period. These are recorded as ATA errors in the `journalctl` log that look like:
ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x50000 action 0x6 frozen
ata5: SError: { PHYRdyChg CommWake }
ata5.00: failed command: FLUSH CACHE EXT
ata5.00: cmd ea/00:00:00:00:00/00:00:00:00:00/a0 tag 12
res 40/00:ff:81:00:00/00:00:00:00:00/40 Emask 0x4 (timeout)
ata5.00: status: { DRDY }
ata5: hard resetting link
ata5: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
ata5.00: configured for UDMA/133
ata5.00: retrying FLUSH 0xea Emask 0x4
ata5: EH complete
Details:
SATA link power management was enabled by:
In the UEFI/BIOS settings:
Storage PCH in AHCI mode (the default)
"Aggressive LPM support" enabled
All "Enable SATA link power management for hostX" options set to "Good" in `powertop`
Note, you can't run `powertop --auto-tune` without freezing your system if you're using the onboard NICs as the drivers appear to be bugged. I have these disabled.
HDDs:
Model: ST16000NM001G (Exos X16 16TB)
Tried connecting to both the SATA connectors and the SlimSAS
Besides the link timeouts, there are no issues - SMART data is clear and full `badblocks` passes completed.
PSU is a Corsair RM850x (2021) and is new.
To do a quick check for historical link dropouts on your system, you can run:
journalctl | grep 'failed command:'
Investigation so far:
Replaced SATA cables - Issue occurs
Tried different SATA ports - Issue occurs
Reduced to a single HDD - Issue occurs
Connected an old Samsung HD103UJ HDD to the board - Could not reproduce the issue
Connected one of the HDDs and the OS SSD and booted on my old Z97 system - Could not reproduce the issue
I'm starting to suspect one of:
My PCH/chipset is faulty
There's a bug/incompatibility between the PCH/chipset and these particular HDDs
There's a bug in the kernel's support for this PCH/chipset when SATA link power management is used.
If anyone else is successfully using SATA link power management I'd be interested in whether you see this issue or not.
Thanks in advance!