hi,all when a writer adds or deletes some elements in a bucket of a hash table, the bihash algorithm tries to making a working copy of the bucket and the reader may use the working copy to search. However, it is possible for the reader to use the bucket to search before the writer makes the working copy and the writer may then modify the bucket even if the reader is still using it. Another possibility is that when a reader is using the working copy and a writer may change the working copy for another bucket modification. In any case, the reader may can't find a element already in the hash table or get a wrong value of a element. I don't verify the analysis above at a running environment and all the analysis is based on the code. Please help me to check out whether the analysis is right. Regards, chu.penghong
_______________________________________________ vpp-dev mailing list vpp-dev@lists.fd.io https://lists.fd.io/mailman/listinfo/vpp-dev
