Bgp give some ordering increases #16774
Conversation
| @@ -4091,7 +4091,10 @@ void bgp_process(struct bgp *bgp, struct bgp_dest *dest, | |||
|
|
|||
| /* can't be enqueued twice */ | |||
| assert(STAILQ_NEXT(dest, pq) == NULL); | |||
| STAILQ_INSERT_TAIL(&pqnode->pqueue, dest, pq); | |||
| if (early_process) | |||
There was a problem hiding this comment.
it's never true at the moment. I'm just adding the ability to use this. It will be useful down the line, I am trying to coordinate with some other people that is all
| @@ -806,7 +806,8 @@ extern void bgp_withdraw(struct peer *peer, const struct prefix *p, | |||
|
|
|||
| /* for bgp_nexthop and bgp_damp */ | |||
| extern void bgp_process(struct bgp *bgp, struct bgp_dest *dest, | |||
| struct bgp_path_info *pi, afi_t afi, safi_t safi); | |||
There was a problem hiding this comment.
instead of changing this function, maybe add a "process_early" version, that you can use in the specific places where you want the "early" behavior?
There was a problem hiding this comment.
adding "early," though, allows you to control the behavior on a route-by-route basis, perhaps even (eventually) using a route-map or some policy to let some routes be processed before others (?) ...
There was a problem hiding this comment.
I agree that there needs to be an API, and that it needs to be called - I'm just thinking it'd be better to have a dedicated api, that would only involve changing the places where it's needed. This change requires many callers to be changed, so that they can say "false"...
There was a problem hiding this comment.
I added a bgp_process_early I believe it solves Marks issues
There was a problem hiding this comment.
that's great, thanks.
so ... could we have a comment, even a one-liner, to say that the "early" version puts the dest at the front of the list?
donaldsharp
force-pushed
the
bgp_give_some_ordering_increases
branch
from
ed70531 to
304ebd8
Compare
| @@ -806,7 +806,8 @@ extern void bgp_withdraw(struct peer *peer, const struct prefix *p, | |||
|
|
|||
| /* for bgp_nexthop and bgp_damp */ | |||
| extern void bgp_process(struct bgp *bgp, struct bgp_dest *dest, | |||
| struct bgp_path_info *pi, afi_t afi, safi_t safi); | |||
There was a problem hiding this comment.
that's great, thanks.
so ... could we have a comment, even a one-liner, to say that the "early" version puts the dest at the front of the list?
There is a need to be able to process certain bgp routes earlier than others. Especially when there is major trauma going on in the network. Start the ability for this to happen. Signed-off-by: Donald Sharp <[email protected]>
The check for an equivalent bgp pointer makes no sense in the context of the workqueue as that we have a work queue per bgp process, as such the bgp pointer will always be the same as the pqnode. Signed-off-by: Donald Sharp <[email protected]>
donaldsharp
force-pushed
the
bgp_give_some_ordering_increases
branch
from
304ebd8 to
220b4ef
Compare
|
yes no-one is using it yet. |
a) Allow bgp_process to push a item to the front of the list
b) Remove a redundant check