nfsd4: simplify reading of opnum
authorJ. Bruce Fields <bfields@redhat.com>
Sat, 17 Nov 2012 03:28:38 +0000 (22:28 -0500)
committerJ. Bruce Fields <bfields@redhat.com>
Mon, 26 Nov 2012 14:08:13 +0000 (09:08 -0500)
commit8a61b18c9b13987310d0f3ba13aa04af51f02a1c
tree0a2b520aed04683bcd3fde519106ed25b5533de2
parent447bfcc936ce28636833e89c4b82f424a291dde9
nfsd4: simplify reading of opnum

The comment here is totally bogus:
- OP_WRITE + 1 is RELEASE_LOCKOWNER.  Maybe there was some older
  version of the spec in which that served as a sort of
  OP_ILLEGAL?  No idea, but it's clearly wrong now.
- In any case, I can't see that the spec says anything about
  what to do if the client sends us less ops than promised.
  It's clearly nutty client behavior, and we should do
  whatever's easiest: returning an xdr error (even though it
  won't be consistent with the error on the last op returned)
  seems fine to me.

Signed-off-by: J. Bruce Fields <bfields@redhat.com>
fs/nfsd/nfs4xdr.c