Tested against an OEM client. When returning the max-access blob on a file we only check for DELETE access in the ACL, we also need to check for DELETE_CHILD access in the containing directory. Patch to follow (tested by Ira and the OEM vendor at the SNIA event).
Also needs patching for 4.0.0rc.next.
Created attachment 7933 [details] git-am fix for 4.0.0rc Patches for 3.6.x to follow.
Created attachment 7934 [details] git-am fix for 3.6.x
Comment on attachment 7933 [details] git-am fix for 4.0.0rc Looks good to me
Comment on attachment 7934 [details] git-am fix for 3.6.x Does this pass 'make test'?
Karolin, please pick the 4.0.0 patch
(In reply to comment #5) > Comment on attachment 7934 [details] > git-am fix for 3.6.x > > Does this pass 'make test'? Hmmm. This is what Ira tested against the client, but I haven't run the 3.6.x "make test" though. I'll do that now and report back for the 3.6.x patch. Jeremy.
FYI. It passes the small subset of smb2 tests we already have in 3.6.x. I'll also run the smb2 torture tests in master against the 3.6.x build to confirm. Jeremy.
Tested using master smbtorture. It passes the smb2.create.gentest call which requests io.in.query_maximal_access, so yes I'd say the 3.6.x patch is good. Jeremy.
Comment on attachment 7934 [details] git-am fix for 3.6.x Looks good.
Karolin, please pick for the releases
Pushed to autobuild-v4-0-test and v3-6-test. Closing out bug report. Thanks!