NVMe drive is not detected after multiple hot plug operations

This document (000019625) is provided subject to the disclaimer at the end of this document.

Environment

SUSE Linux Enterprise Server 15 SP1

Situation

NVMe drive is not detected when multiple hot plug operations are performed.

This issue may occur on the following platforms:
Dell EMC PowerEdge R6515
Dell EMC PowerEdge R6525
Dell EMC PowerEdge C6515
Dell EMC PowerEdge R7515
Dell EMC PowerEdge R7525

Resolution

Install Security Update for the Linux Kernel kernel-default-4.12.14-197.40.1 (April 29 2020)

Cause

The issue is caused by a race condition related to MSI interrupts and their handling in pciehp driver interrupt service routine.

Additional Information

The problem was resolved with :
[PATCH v4] PCI: pciehp: Fix MSI interrupt race
https://lore.kernel.org/linux-pci/78b4ced5072bfe6e369d20e8b47c279b8c7af12e.1582121613.git.lukas@wunner.de/T/#u

The issue could be worked around by following steps.

1. Read the Slot Status Register in the PCI Express Capability structure by running the following command:

setpci -s e0:03.2 CAP_EXP+0x1a.w
Output: 0148 (value returned)

2. Clear the event bits that are impacted by running the following command:

setpci -s e0:03.2 CAP_EXP+0x1a.w=0x0108

3. Re-read the Slot Status Register to confirm that event bits are cleared by running the following command:

setpci -s e0:03.2 CAP_EXP+0x1a.w
Output: 0040 (value returned)

4. Unplug the drive and then plug-in the drive after clearing the event bits.

Note: The output of each command could vary based on your system configuration.

Disclaimer

This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.