Hi there,
FYI, I've experienced a stability issue with the jetson-tk1 NIC since
kernel 5.3 and later.
This is reported upstream at
https://bugzilla.kernel.org/show_bug.cgi?id=206217
To sum-up: under some "MMC and network I/O load" (dnf update or scp of
large file), the pciport receives AER errors that are actually fatal
to the network interface and cannot be recovered unless a reboot.
I've bisected the issue and found the commit that once reverted,
restore a good behaviour:
https://patchwork.ozlabs.org/project/linux-tegra/patch/20200420164304.288...
I haven't experienced any other regression since then.
What I would like to ask is:
1/ Is there any others reproducers for this issue on jetson-tk1 ?
(issue only relevant on tegra124 SOC).
2/ As upstream agreed that a revert would be preferred until more
investigation, can we consider to apply as a downstream patch until
then ?
Thanks for any advices on the topic.
--
-
Nicolas (kwizart)