Comment 13 for bug 1641322

Revision history for this message
Thomas M Steenholdt (tmus) wrote :

FWIW, this appears to work around the issue:

Create file /etc/udev/rules.d/90-nvme-power.rules :
---
KERNEL=="nvme*[0-9]n*[0-9]", ATTRS{model}=="SAMSUNG MZSLW1T0HMLH-000L1*", ATTR{device/power/control}="on"
---

On my system at least.

I'm sure the kernel should somehow blacklist settings that bakes certain devices behave erratically, but until then, this appears to do the trick. No issue in 4 days!