Hangbin Liu <liuhangbin@xxxxxxxxx> writes:
> On Thu, Mar 18, 2021 at 03:19:47PM +0100, Toke Høiland-Jørgensen wrote:
>> Hangbin Liu <liuhangbin@xxxxxxxxx> writes:
>>
>> > On Wed, Mar 17, 2021 at 01:03:02PM +0100, Toke Høiland-Jørgensen wrote:
>> >> FYI, this no longer applies to bpf-next due to Björn's refactor in
>> >> commit: ee75aef23afe ("bpf, xdp: Restructure redirect actions")
>> >
>> > Thanks Toke, I need to see how to get the map via map_id, does
>> > bpf_map_get_curr_or_next() works? Should I call bpf_map_put() after
>> > using?
>>
>> I would expect that to be terrible for performance; I think it would be
>> better to just add back the map pointer into struct bpf_redirect_info.
>> If you only set the map pointer when the multicast flag is set, you can
>> just check that pointer to disambiguate between when you need to call
>> dev_map_enqueue() and dev_map_enqueue_multi(), in which case you don't
>> need to add back the flags member...
>
> There are 2 flags, BROADCAST and EXCLUDE_INGRESS. There is no way
> to only check the map pointer and ignore flags..
Ah, right, of course, my bad :)
Well, in that case adding both members back is probably the right thing
to do...
-Toke
