On 2/15/2022 6:40 PM, Kenneth R. Crudup wrote:
External email: Use caution opening links or attachments
On Mon, 7 Feb 2022, Bjorn Helgaas wrote:
If the problem really was fixed by some change elsewhere, it *should*
still happen on v5.11-rc1. I think we should verify that and try to
figure out what the other change was.
Yeah, not a bad idea. I'm a little busy now, but I'll cut a kernel at
4257f7e0, remake, and see what happens. Gimme a day or so.
So I did this- checked out a branch hard-reset to commit 4257f7e0, used my
current _defconfig on it, and built and tested it- NOW it doesn't exhibit the
NVMe-device-killing behavior I'd seen a year earlier; I'm assuming Dell must
have fixed something in the BIOS?
Anyway, I'd done an "lspci -vvvnn" before and after hibernation, if you're
interested, and a diff of same; I'll send those in subsequent E-mails.
Thanks Kenny for the dump.
I see that the ASPM-L1 state of Realtek NIC which was in disabled state
before hibernate got enabled after hibernate. This is weird. This patch
doesn't do anything to LnkCtl register which has control for ASPM L1
state. Could you please check why ASPM L1 got enabled post hibernation?
-Kenny
--
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange County CA
