nfsd4: don't allow owner override on 4.1 CLAIM_FH opens

The Linux client is using CLAIM_FH to implement regular opens, not just
recovery cases, so it depends on the server to check permissions
correctly.

Therefore the owner override, which may make sense in the delegation
recovery case, isn't right in the CLAIM_FH case.

Symptoms: on a client with 49f9a0fafd
"NFSv4.1: Enable open-by-filehandle", Bryan noticed this:

	touch test.txt
	chmod 000 test.txt
	echo test > test.txt

succeeding.

Cc: stable@kernel.org
Reported-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
This commit is contained in:
J. Bruce Fields 2013-05-03 16:09:09 -04:00
parent 676e4ebd5f
commit 9f415eb255
1 changed files with 13 additions and 2 deletions

View File

@ -279,6 +279,7 @@ do_open_fhandle(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, str
{
struct svc_fh *current_fh = &cstate->current_fh;
__be32 status;
int accmode = 0;
/* We don't know the target directory, and therefore can not
* set the change info
@ -290,9 +291,19 @@ do_open_fhandle(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, str
open->op_truncate = (open->op_iattr.ia_valid & ATTR_SIZE) &&
(open->op_iattr.ia_size == 0);
/*
* In the delegation case, the client is telling us about an
* open that it *already* performed locally, some time ago. We
* should let it succeed now if possible.
*
* In the case of a CLAIM_FH open, on the other hand, the client
* may be counting on us to enforce permissions (the Linux 4.1
* client uses this for normal opens, for example).
*/
if (open->op_claim_type == NFS4_OPEN_CLAIM_DELEG_CUR_FH)
accmode = NFSD_MAY_OWNER_OVERRIDE;
status = do_open_permission(rqstp, current_fh, open,
NFSD_MAY_OWNER_OVERRIDE);
status = do_open_permission(rqstp, current_fh, open, accmode);
return status;
}