[PATCH v2 0/2][SRU][Jammy] e1000e report hardware hang
AceLan Kao
acelan.kao at canonical.com
Wed Aug 3 06:11:56 UTC 2022
From: "Chia-Lin Kao (AceLan)" <acelan.kao at canonical.com>
BugLink: https://bugs.launchpad.net/bugs/1973104
[Impact]
e1000e hardware unit hang after s2idle
May 06 15:34:49 ubuntu kernel: e1000e 0000:00:1f.6 enp0s31f6: Detected Hardware Unit Hang:
TDH <1>
TDT <5>
next_to_use <5>
next_to_clean <1>
buffer_info[next_to_clean]:
time_stamp <1000587f0>
next_to_watch <1>
jiffies <1000589c0>
next_to_watch.status <0>
MAC Status <40080283>
PHY Status <796d>
PHY 1000BASE-T Status <3800>
PHY Extended Status <3000>
PCI Status <10>
[Fix]
Below commits fix this issue which have been included in v5.19-rc8
https://patchwork.ozlabs.org/project/intel-wired-lan/patch/20220508070905.1878172-1-sasha.neftin@intel.com/
6cfa45361d3e Revert "e1000e: Fix possible HW unit hang after an s0ix exit"
b49feacbeffc e1000e: Enable GPT clock before sending message to CSME
[Test]
Verified on the target machine, and confirmed the issue is gone.
[Where problems could occur]
The impact is low, it keeps the GPT clock enabled for CSME when
e1000e_pm_resume() is called. With the new fix, the old workaround could
be removed. I can't see this may introduce any regression.
Sasha Neftin (2):
e1000e: Enable GPT clock before sending message to CSME
Revert "e1000e: Fix possible HW unit hang after an s0ix exit"
drivers/net/ethernet/intel/e1000e/hw.h | 1 -
drivers/net/ethernet/intel/e1000e/ich8lan.c | 4 ---
drivers/net/ethernet/intel/e1000e/ich8lan.h | 1 -
drivers/net/ethernet/intel/e1000e/netdev.c | 30 +++------------------
4 files changed, 4 insertions(+), 32 deletions(-)
--
2.25.1
More information about the kernel-team
mailing list