nfs41: return correct errors on callback replays version 3 Responded to Tronds comments, fixed bugs in cb_recall_slot, and resize slot table on session reset. The first three patches clean up callback processing 0001-nfs41-fix-wrong-error-on-callback-header-xdr-overflo.patch 0002-nfs41-directly-encode-back-channel-error.patch 0003-nfs41-remove-uneeded-checks-in-callback-processing.patch These next two implement correct error returns for v4.1 callback replays. Since our back channel has a ca_maxrequestsize_cached = 0, a replay with cachethis set to true results in a NFS4ERR_TOO_BIG_TO_CACHE error. This code is set up to do a real DRC. A replay with cachethis set to false returns a NFS4ERR_RETRY_UNCACHED_REP error. 0004-nfs41-prepare-for-back-channel-drc.patch 0005-nfs41-back-channel-drc-minimal-implementation.patch ADDED: 0006-nfs41-implement-cb_recall_slot.patch The fore channel session is drained, then the new highest_slot is set. 0007-nfs41-resize-slot-table-in-reset.patch The slot table size is renegotiated on session reset. TODO: The callback code currently returns NFS4ERR_RESOURCE on all xdr overflows. This is correct for v4.0, incorrect for v4.1. Testing: Modified nfsv4.1 pynfs server tested cb_recall replays with the cb_sequence cachethis set to False and to True. Modified nfs4.1 pyNFS server sends CB_RECALL_SLOT call during the Connectathon basic big file test which uses all available slots. Connectathon tests pass. -->Andy -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
