On Sun, Oct 18, 2026 at 03:54:34AM +0200, Julius Bairaktaris wrote: > On Sun, Oct 18, 2026 at 03:42:58AM +0200, Pablo Neira Ayuso wrote: > > This series are a no-go in this approach. > > Understood, I drop it. > > > Please, given we are discussing one series of you quite recently... > > let's handle one thing at a time. > > Agreed, the fdb series first. > > Some context on why there are so many: they all come from the OpenWrt > PR that adds flowtable offload for the Qualcomm IPQ807x PPE: > > https://github.com/openwrt/openwrt/pull/24806 > > That PR carries the kernel changes it needs as patches. I am sending > them upstream so that OpenWrt can carry backports instead of local > patches. > > I know the rule that there must be an in-tree user. The PPE driver is > not upstream, so I only send changes that stand on their own: fixes to > the software path, or features that existing drivers like airoha and > mtk use.
Thanks for the context. As said, let's go one by one after each issue/scenario. We will eventually get to all of your described scenarios.
