Il 26/06/2013 16:09, Michael R. Hines ha scritto:
*This requires some steps:*
1. First, maintain a new data structure: something like
"These memory ranges are 'being unpinned'" - block all potential writes
to these addresses until the unpinning completes.
2. Once the source unpin completes, send the asynchronous control
channel message
to the other side for unpinning.
2. Mark the data structure and return and allow the migration to continue
with the next RDMA write.
3. Upon completion of the unpinning on the destination,
respond to the source that it was finished.
4. Source then clears the data structure for the successfully unpinned
memory ranges.
5. At this point, one or more writes may (or may not) be blocking on the
unpinned memory areas and will poll the data structure and find that
the unpinning has completed.
6. Then issue the new writes and proceed as normal.
7. Repeat step 1.
After more discussion on IRC I think I understand this. It's not
trivial, but not super-complex either... it would really be nice to add
it to the protocol already in 1.6.
Paolo