Revert "nfsd4: a client's own opens needn't prevent delegations"
authorJ. Bruce Fields <bfields@redhat.com>
Mon, 8 Mar 2021 15:52:29 +0000 (10:52 -0500)
committerSalvatore Bonaccorso <carnil@debian.org>
Fri, 19 Mar 2021 18:20:52 +0000 (18:20 +0000)
commitde43b4156416e7bd81fc05fc399eeda293cb7266
tree7a4573d22da903fcbfa5fbdddf3cb581deab5cac
parent358f7c5456258fcacbb6716da4174a5665b03c67
Revert "nfsd4: a client's own opens needn't prevent delegations"

commit 6ee65a773096ab3f39d9b00311ac983be5bdeb7c upstream.

This reverts commit 94415b06eb8aed13481646026dc995f04a3a534a.

That commit claimed to allow a client to get a read delegation when it
was the only writer.  Actually it allowed a client to get a read
delegation when *any* client has a write open!

The main problem is that it's depending on nfs4_clnt_odstate structures
that are actually only maintained for pnfs exports.

This causes clients to miss writes performed by other clients, even when
there have been intervening closes and opens, violating close-to-open
cache consistency.

We can do this a different way, but first we should just revert this.

I've added pynfs 4.1 test DELEG19 to test for this, as I should have
done originally!

Cc: stable@vger.kernel.org
Reported-by: Timo Rothenpieler <timo@rothenpieler.org>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Gbp-Pq: Topic bugfix/all
Gbp-Pq: Name revert-nfsd4-a-client-s-own-opens-needn-t-prevent-delegations.patch
fs/locks.c
fs/nfsd/nfs4state.c