On 05/12/2020 16.45, Russell King - ARM Linux admin wrote: > On Sat, Dec 05, 2020 at 03:49:02PM +0100, Rasmus Villemoes wrote: >> So, I found out that the problem disappers when I disable >> vlan_filtering, and googling then led me to Russell's patches from >> around March >> (https://patchwork.ozlabs.org/project/netdev/cover/20200218114515.gl18...@shell.armlinux.org.uk/). >> >> But, unlike from what I gather from Russell's description, the problem >> is there whether or not the bridge is created with vlan_filtering >> enabled from the outset or not. > > No. My problem is where the bridge is created _without_ vlan_filtering > enabled, and is subsequently enabled. That caused traffic to stop as > soon as vlan_filtering was enabled. > > Note that if the bridge were created with vlan_filtering enabled from > the start, there would be no problem.
That was what I was trying to say: I see the problem whether or not vlan_filtering is enabled from the beginning (hence the "unlike..."). So since disabling vlan_filtering (or never enabling it in the first place) makes traffic flow, the symptoms look similar to the problem you described (and that was what led me to those patches). >> Also, cherry-picking 517648c8d1 to >> 5.9.12 doesn't help. The problem also exists on 5.4.80, and (somewhat >> naively) backporting 54a0ed0df496 as well as 517648c8d1 doesn't change that. > > I'm not sure what 517648c8d1 is - this isn't a mainline or net-next > commit ID. Sorry, I mistakenly took the sha1 from the 5.4 branch I cherry-picked it to - I did indeed mean your subsequent 1fb74191988f setting the flag for mv88e6xxxx. > You will, however, find that the problem was subsequently fixed by > 1fb74191988f on top of 54a0ed0df496, Unfortunately not. So either it's simply some completely different underlying issue, or there's something else going on in my setup that prevents the configure_vlan_while_not_filtering patches from fixing it. Thanks, Rasmus
