Description of problem:

My Dell Precision 3530 experiences ethernet network instability with centos9 stream kernels released after 5.14.0-308. The kernel messages report

e1000e 0000:00:1f.6 eno1: Detected Hardware Unit Hang:#012 TDH <8e>#012 TDT <9d>#012 next_to_use <9d>#012 next_to_clean <8c>#012buffer_info[next_to_clean]:#012 time_stamp <fffc9bd8>#012 next_to_watch <8e>#012 jiffies <fffcb340>#012 next_to_watch.status <0>#012MAC Status <40080083>#012PHY Status <796d>#012PHY 1000BASE-T Status <3c00>#012PHY Extended Status <3000>#012PCI Status <10>

Looking at the kernel changelog, it appears to be related to this change, but I have not done a bisect to be sure.

• Thu May 04 2023 Jan Stancek <jstancek@redhat.com> [5.14.0-308.el9]

• e1000e: Remove redundant pci_enable_pcie_error_reporting() (Ken Cox) [2153331]
• e1000e: Enable Link Partner Advertised Support (Ken Cox) [2153331]
• e1000e: Fix TX dispatch condition (Ken Cox) [2153331]
• e1000e: Remove unnecessary use of kmap_atomic() (Ken Cox) [2153331]
• e1000e: Add e1000e trace module (Ken Cox) [2153331]
• e1000e: Add support for the next LOM generation (Ken Cox) [2153331]
• e1000e: Separate MTP board type from ADP (Ken Cox) [2153331]

How reproducible: 100%

Steps to Reproduce:
1. Boot into centos9 stream kernel after 5.14.0-308
2. Attempt to do some network related work, until the e1000e hangs start happening and a reboot is required