On Mon, 31 Aug 2026 18:44:29 -0700 Long Li wrote:
quoted
MANA reconfigures a port by destroying its queues and building them
again.
quoted
mana_detach() tears the whole queue set down, mana_attach() allocates
a new one, and every path that changes a queue property does this: the
ethtool channel, ring and private-flag setters, the MTU change and XDP
attach.
Given the sheer number of reports I'm going to assume that at least one of
the Medium+ ones is right and respin is needed. Please LMK if that's not the
case.
If you're respinning please take a pass and delete the more pointless
comments your LLM has added.
--
pw-bot: cr
Thank you for looking at this.
Sashiko flagged that two fixes should be sent with the "Fixes" tag. I have sent those two separate patches to net:
https://lore.kernel.org/all/20260904202640.3900685-1-longli@microsoft.com/ (local)
https://lore.kernel.org/all/20260905004401.3937066-1-longli@microsoft.com/ (local)
But I can also argue that those fixes are related to this patch series and should be included in this series instead. Sending them together reduces code churns.
Please let me know if those net patches are appropriate, or they should be sent part of the net-next series.
I'll respin the patches with all the Sashiko reviews addressed.
Long