[Chromebook] UFS disk device type "unknown"

[micmalti]

[x] I've read and accepted the Bug Reporting Howto [x] I've provided all required tlp-stat outputs via Gist (see below)

Describe the bug

Hi, I've been happily using tlp for 5+ years and never had any issues until I recently switched over to a new laptop. Now, I keep getting this kernel message every 5 seconds at random intervals, both on battery and AC: sd 0:0:0:0: [sda] Starting disk

From what I understand, a similar message is reported whenever tlp endlessly tries to stop/start the laptop disk, so I checked powertop and confirmed that the Runtime PM for disk sda option was marked as 'good'. The recommendation for disabling this option is to uncomment SATA_LINKPWR_DENYLIST="hostX", yet when I ran tlp-stat -d to determine the host, it seems as though the disk isn't being recognized by tlp: tlp-stat

Is this perhaps due to a general lack of support for UFS drives, which this particular model is making use of (SDINFDO4-128G)?

[micmalti]

It isn't tlp that's spamming the log but a bad kernel commit. It's still unclear though why tlp isn't properly identifying the disk.

[linrunner]

Hi, I have no idea what a "UFS drive" is supposed to be and what type of controller it uses, obviously it's not SATA/AHCI.

Please show the output of

ls -l /dev/disk/by-id

[micmalti]

Here's the output (main and EFI):

total 0
lrwxrwxrwx 1 root root  9 Aug  8 14:24 scsi-0WDC_SDINFDO4-128G_WDC -> ../../sda
lrwxrwxrwx 1 root root 10 Aug  8 14:24 scsi-0WDC_SDINFDO4-128G_WDC-part1 -> ../../sda1
lrwxrwxrwx 1 root root 10 Aug  8 14:24 scsi-0WDC_SDINFDO4-128G_WDC-part2 -> ../../sda2

[linrunner]

@micmalti Sorry for the late reply. I'm currently busy with the 1.7 beta.

I checked the kernel docs for UFS

• https://docs.kernel.org/scsi/ufs.html
• https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-driver-ufs

but couldn't figure out how UFS pm works. Without having the hardware, I can neither investigate further nor develop or even test anything.

Your output shows the controller as PCIe device, so at least runtime pm will be enabled ("auto") on battery power :

/sys/bus/pci/devices/0000:00:12.7/power/control = on (0x010901, Universal Flash Storage controller, ufshcd)

[micmalti]

That's fine by me. It's just a reporting issue then.