On 10/16/20 11:43 AM, Ilya Maximets wrote: > On 10/16/20 11:38 AM, Ilya Maximets wrote: >> On 10/16/20 7:57 AM, Eli Britstein wrote: >>> >>> On 10/16/2020 3:02 AM, Ilya Maximets wrote: >>>> External email: Use caution opening links or attachments >>>> >>>> >>>> On 10/12/20 5:25 PM, Sriharsha Basavapatna via dev wrote: >>>>> On Mon, Oct 12, 2020 at 4:48 PM Eli Britstein <[email protected]> wrote: >>>>>> >>>>>> On 10/12/2020 10:20 AM, Sriharsha Basavapatna wrote: >>>>>>> On Sun, Sep 6, 2020 at 5:51 PM Eli Britstein <[email protected]> wrote: >>>>>>>> ping >>>>>>>> >>>>>>>> On 7/30/2020 1:58 PM, Eli Britstein wrote: >>>>>>>>> From: Lei Wang <[email protected]> >>>>>>>>> >>>>>>>>> Struct match has the tunnel values/masks in >>>>>>>>> match->flow.tunnel/match->wc.masks.tunnel. >>>>>>>>> Load actions such as load:0xa566c10->NXM_NX_TUN_IPV4_DST[], >>>>>>>>> load:0xbba->NXM_NX_TUN_ID[] are utilizing the tunnel masks fields, >>>>>>>>> but those should not be used for matching. >>>>>>>>> Offloading fails if masks is not clear. Clear it if no tunnel used. >>>>>>>>> >>>>>>>>> Signed-off-by: Lei Wang <[email protected]> >>>>>>>>> Reviewed-by: Eli Britstein <[email protected]> >>>>>>>>> Reviewed-by: Gaetan Rivet <[email protected]> >>>>>>>>> Signed-off-by: Eli Britstein <[email protected]> >>>>> Acked-by: Sriharsha Basavapatna <[email protected]> >>>>> >>>>> See my comment below. >>>>> >>>>>>>>> --- >>>>>>>>> lib/netdev-offload-dpdk.c | 4 ++++ >>>>>>>>> 1 file changed, 4 insertions(+) >>>>>>>>> >>>>>>>>> diff --git a/lib/netdev-offload-dpdk.c b/lib/netdev-offload-dpdk.c >>>>>>>>> index de6101e4d..0d23e4879 100644 >>>>>>>>> --- a/lib/netdev-offload-dpdk.c >>>>>>>>> +++ b/lib/netdev-offload-dpdk.c >>>>>>>>> @@ -682,6 +682,10 @@ parse_flow_match(struct flow_patterns *patterns, >>>>>>>>> >>>>>>>>> consumed_masks = &match->wc.masks; >>>>>>>>> >>>>>>>>> + if (!flow_tnl_dst_is_set(&match->flow.tunnel)) { >>>>>>>>> + memset(&match->wc.masks.tunnel, 0, sizeof >>>>>>>>> match->wc.masks.tunnel); >>>> Small nit: For consistency, I think, we should access via 'consumed_masks' >>>> pointer, not directly. What do you think? >>>> I could change that on commit. >>> OK. >>>> >>>> >>>> Another point is that we could actually consider this as a bug fix, but >>>> I'm not sure which commit to use for a 'Fixes' tag. Suggestions? >>>> Technically, the issue exists on all branches down to 2.10, but 2.13 and >>>> lower will require a bit different fix due t major rework of this code. >>> >>> I see the original "validate" code also has it, so maybe e8a2b5bf92bb >>> ("netdev-dpdk: implement flow offload with rte flow") >>> >>> Do you want me to send v2 with this fixes tag and usage of consumed_masks >>> (above), or you can do it when applying? >> >> I can do that.
Applied this version to master and branch-2.14. Best regards, Ilya Maximets. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
