#1091805 nvme: Does your device have a faulty power saving mode enabled? on Samsung 990 Pro 4TB

#1091805#5
Date:
2024-12-31 17:19:10 UTC
From:
To:
On a Dell Precision 3650 Tower (Firmware Version: 1.35.0),
we have two Samsung SSD 990 PRO with Heatsink 4TB with different
firmware versions shutting down randomly with the following message:

```
Dec 30 08:05:11 [redacted] kernel: nvme nvme1: Disabling device after
reset failure: -19
Dec 30 08:05:11 [redacted] kernel: nvme 0000:02:00.0: Unable to change
power state from D3cold to D0, device inaccessible
Dec 30 08:05:11 [redacted] kernel: nvme nvme1: Try
"nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off"
and report a bug
Dec 30 08:05:11 [redacted] kernel: nvme nvme1: Does your device have a
faulty power saving mode enabled?
Dec 30 08:05:11 [redacted] kernel: nvme nvme1: controller is down; will
reset: CSTS=0xffffffff, PCI_STATUS=0xffff
```

Device is thereafter not responsive to any command using the `nvme`
command. Starting the system with the suggested kernel parameters yields
no change (the previous log snippet is taken from a system started with
those arguments).

Disks as described by `lshw`:
```
[...]
            *-nvme
                 description: NVMe device
                 product: Samsung SSD 990 PRO with Heatsink 4TB
                 vendor: Samsung Electronics Co Ltd
                 physical id: 0
                 bus info: pci@0000:01:00.0
                 logical name: /dev/nvme0
                 version: 4B2QJXD7
                 serial: S7HRNJ0X201019H
                 width: 64 bits
                 clock: 33MHz
                 capabilities: nvme pm msi pciexpress msix nvm_express
bus_master cap_list
                 configuration: driver=nvme latency=0
nqn=nqn.1994-11.com.samsung:nvme:990PRO:M.2:S7HRNJ0X201019H state=live
                 resources: irq:16 memory:71400000-71403fff
               *-namespace:0
                    description: NVMe disk
                    physical id: 0
                    logical name: hwmon0
               *-namespace:1
                    description: NVMe disk
                    physical id: 2
                    logical name: /dev/ng0n1
               *-namespace:2
                    description: NVMe disk
                    physical id: 1
                    bus info: nvme@0:1
                    logical name: /dev/nvme0n1
                    serial: [redacted]
                    size: 3726GiB
                    capacity: 3726GiB
                    capabilities: lvm2
                    configuration: logicalsectorsize=512 sectorsize=512
wwid=eui.0025384241403599
[...]
            *-generic
                 description: NVMe device
                 product: Samsung SSD 990 PRO with Heatsink 4TB
                 vendor: Samsung Electronics Co Ltd
                 physical id: 0
                 bus info: pci@0000:02:00.0
                 logical name: /dev/nvme1
                 version: 0B2QJXG7
                 serial: S7HRNJ0WC05772H
                 width: 32 bits
                 clock: 66MHz
                 capabilities: bus_master vga_palette cap_list
                 configuration: driver=nvme latency=255 maxlatency=255
mingnt=255 nqn=nqn.1994-11.com.samsung:nvme:990PRO:M.2:S7HRNJ0WC05772H
state=dead
                 resources: irq:16 memory:71300000-71303fff
               *-namespace:0
                    description: NVMe disk
                    physical id: 0
                    logical name: hwmon1
               *-namespace:1
                    description: NVMe disk
                    physical id: 2
                    logical name: /dev/ng1n1
               *-namespace:2
                    description: NVMe disk
                    physical id: 1
                    bus info: nvme@1:1
                    logical name: /dev/nvme1n1
                    configuration: logicalsectorsize=512 sectorsize=512
wwid=eui.0025384c3144c366
[...]
```