Actions
Optimization #4652
closedGAP handling improvements seem expensive
Status:
Rejected
Priority:
Normal
Assignee:
-
Target version:
-
Effort:
Difficulty:
Label:
Description
(originally reported y Eric Leblond)
We got a big performance increase with disabling the GAP handling (GAP handling commits in 6.x vs 5.x ). As we are also leaking memory in stream TCP reassembly ( https://redmine.openinfosecfoundation.org/issues/4650 ) , this could also augment the pressure on memory because Suricata is not stopping the reassembly when it hits the first gap.
Actions