On 11/16/26 8:56 PM, Rosemarie O'Riorden via dev wrote: > On 11/16/26 5:09 AM, Xavier Simonart wrote: >> Hi Rosemarie >> >> Thanks for the feedback. >> >> On Thu, Nov 12, 2026 at 11:45 PM Rosemarie O'Riorden >> <[email protected] <mailto:[email protected]>> wrote: >> >> Hi Xavier, thanks for the fix! >> >> My one question is: should the Fixes tag be what it is? I found commits >> 505ceec8a and 280bef8b6 which specifically handle load balancer changes >> for logical switches and routers, and added the code blocks you edited, >> so they seem very relevant. >> >> Very good question! >> I ran git bisect before adding the Fixes tag, and it points precisely to >> the commit where the test starts failing: "lflow-mgr: Remove the >> ovn_datapath from ovn_lflow." >> I agree that the two patches you mention (handling LB/LB-group changes >> for switches and routers) made the incremental lb_data tracking fragile. >> But the test passes cleanly on top of them alone. >> It's the later, seemingly unrelated, od->dp refactor that actually >> exposes the missing "changed = true" and triggers the assert. >> So I think we should keep the Fixes tag as it is now. WDYT? Maybe a >> maintainer can give us their perspective too on what the Fixes tag >> should be in this case. > > Hi Xavier, I appreciate the response. > > Ok, that is interesting. Well, I suppose it can be left as is then. I > just wasn't sure how they were related, but if that's exactly where the > test starts failing, then it makes sense :) > > Everything looks good to me, so: > > Acked-by: Rosemarie O'Riorden <[email protected]>
Hi Xavier, Rosemarie, Thanks for the fix and review, applied to main, 26.09 and 26.03. Regards, Dumitru _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
