Hi- > On Dec 18, 2021, at 8:37 PM, trondmy@xxxxxxxxxx wrote: > > From: Trond Myklebust <trond.myklebust@xxxxxxxxxxxxxxx> > > The following patchset is mainly for improving support for re-exporting > NFSv4 as NFSv3. However it also includes one generic bugfix for NFSv3 to > allow zero length writes. It also improves the writeback performance by > replacing the rwsem with a lock-free errseq_t-based method. > > > - v2: Split patch adding WCC support > v2: Rebase onto v5.16-rc5 I've replaced the following patch. Thank you for the bug report. > nfsd: NFSv3 should allow zero length writes I've provisionally applied the following five to for-next while tests are in progress: > nfsd: map EBADF > nfsd: Add errno mapping for EREMOTEIO > nfsd: Retry once in nfsd_open on an -EOPENSTALE return > nfsd: Add a tracepoint for errors in nfsd4_clone_file_range() > nfsd: Replace use of rwsem with errseq_t The following two are still open for discussion. I get where you want to go, just a quibble about exactly how to get there. > nfsd: Distinguish between required and optional NFSv3 post-op > attributes > nfs: Add export support for weak cache consistency attributes The following two are deferred. IMHO they are not ready. > nfsd: allow lockd to be forcibly disabled > nfsd: Ignore rpcbind errors on nfsd startup When reposting, you can leave out the patches that have already been applied. -- Chuck Lever
