* [PATCH 1/2] nfsd4: remove BUG_ON @ 2013-03-05 22:17 J. Bruce Fields 2013-03-05 22:17 ` [PATCH 2/2] nfsd4: handle seqid-mutating open errors from xdr decoding J. Bruce Fields 0 siblings, 1 reply; 3+ messages in thread From: J. Bruce Fields @ 2013-03-05 22:17 UTC (permalink / raw) To: linux-nfs From: "J. Bruce Fields" <bfields@redhat.com> This BUG_ON just crashes the thread a little earlier than it would otherwise--it doesn't seem useful. Signed-off-by: J. Bruce Fields <bfields@redhat.com> --- fs/nfsd/nfs4proc.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index ae73175..c7e4e8c 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -1282,12 +1282,9 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, if (op->status) goto encode_op; - if (opdesc->op_func) { - if (opdesc->op_get_currentstateid) - opdesc->op_get_currentstateid(cstate, &op->u); - op->status = opdesc->op_func(rqstp, cstate, &op->u); - } else - BUG_ON(op->status == nfs_ok); + if (opdesc->op_get_currentstateid) + opdesc->op_get_currentstateid(cstate, &op->u); + op->status = opdesc->op_func(rqstp, cstate, &op->u); if (!op->status) { if (opdesc->op_set_currentstateid) -- 1.7.9.5 ^ permalink raw reply related [flat|nested] 3+ messages in thread
* [PATCH 2/2] nfsd4: handle seqid-mutating open errors from xdr decoding 2013-03-05 22:17 [PATCH 1/2] nfsd4: remove BUG_ON J. Bruce Fields @ 2013-03-05 22:17 ` J. Bruce Fields 2013-03-05 22:22 ` J. Bruce Fields 0 siblings, 1 reply; 3+ messages in thread From: J. Bruce Fields @ 2013-03-05 22:17 UTC (permalink / raw) To: linux-nfs From: "J. Bruce Fields" <bfields@redhat.com> If a client sets an owner (or group_owner or acl) attribute on open for create, and the mapping of that owner to an id fails, then we return BAD_OWNER. But BAD_OWNER is a seqid-mutating error, so we can't shortcut the open processing that case: we have to at least look up the owner so we can find the seqid to bump. Signed-off-by: J. Bruce Fields <bfields@redhat.com> --- fs/nfsd/nfs4proc.c | 27 ++++++++++++++++++++++++++- fs/nfsd/nfs4xdr.c | 1 + fs/nfsd/xdr4.h | 1 + 3 files changed, 28 insertions(+), 1 deletion(-) diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index c7e4e8c..42c498c 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -351,6 +351,10 @@ nfsd4_open(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, } if (status) goto out; + if (open->op_xdr_error) { + status = open->op_xdr_error; + goto out; + } status = nfsd4_check_open_attributes(rqstp, cstate, open); if (status) @@ -417,6 +421,24 @@ out: } /* + * OPEN is the only seqid-mutating operation whose decoding can fail + * with a seqid-mutating error (specifically, decoding of user names in + * the attributes). Therefore we have to do some processing to look up + * the stateowner so that we can bump the seqid. + */ +static __be32 nfsd4_open_omfg(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_op *op) +{ + struct nfsd4_open *open = (struct nfsd4_open *)&op->u; + + if (!seqid_mutating_err(ntohl(op->status))) + return op->status; + if (nfsd4_has_session(cstate)) + return op->status; + open->op_xdr_error = op->status; + return nfsd4_open(rqstp, cstate, open); +} + +/* * filehandle-manipulating ops. */ static __be32 @@ -1244,8 +1266,11 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, * for example, if there is a miscellaneous XDR error * it will be set to nfserr_bad_xdr. */ - if (op->status) + if (op->status) { + if (op->opnum == OP_OPEN) + op->status = nfsd4_open_omfg(rqstp, cstate, op); goto encode_op; + } /* We must be able to encode a successful response to * this operation, with enough room left over to encode a diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 0116886..a82df26 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -804,6 +804,7 @@ nfsd4_decode_open(struct nfsd4_compoundargs *argp, struct nfsd4_open *open) open->op_iattr.ia_valid = 0; open->op_openowner = NULL; + open->op_xdr_error = 0; /* seqid, share_access, share_deny, clientid, ownerlen */ READ_BUF(4); READ32(open->op_seqid); diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h index 546f898..be0a79d 100644 --- a/fs/nfsd/xdr4.h +++ b/fs/nfsd/xdr4.h @@ -237,6 +237,7 @@ struct nfsd4_open { u32 op_share_deny; /* request */ u32 op_deleg_want; /* request */ stateid_t op_stateid; /* response */ + __be32 op_xdr_error; /* see nfsd4_open_omfg() */ u32 op_recall; /* recall */ struct nfsd4_change_info op_cinfo; /* response */ u32 op_rflags; /* response */ -- 1.7.9.5 ^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH 2/2] nfsd4: handle seqid-mutating open errors from xdr decoding 2013-03-05 22:17 ` [PATCH 2/2] nfsd4: handle seqid-mutating open errors from xdr decoding J. Bruce Fields @ 2013-03-05 22:22 ` J. Bruce Fields 0 siblings, 0 replies; 3+ messages in thread From: J. Bruce Fields @ 2013-03-05 22:22 UTC (permalink / raw) To: linux-nfs On Tue, Mar 05, 2013 at 05:17:59PM -0500, J. Bruce Fields wrote: > From: "J. Bruce Fields" <bfields@redhat.com> > > If a client sets an owner (or group_owner or acl) attribute on open for > create, and the mapping of that owner to an id fails, then we return > BAD_OWNER. But BAD_OWNER is a seqid-mutating error, so we can't > shortcut the open processing that case: we have to at least look up the > owner so we can find the seqid to bump. This seems like an awful hack. Better might be to move idmapping out of xdr decoding, but that's a bigger change and I wonder if it would break anything else. --b. > > Signed-off-by: J. Bruce Fields <bfields@redhat.com> > --- > fs/nfsd/nfs4proc.c | 27 ++++++++++++++++++++++++++- > fs/nfsd/nfs4xdr.c | 1 + > fs/nfsd/xdr4.h | 1 + > 3 files changed, 28 insertions(+), 1 deletion(-) > > diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c > index c7e4e8c..42c498c 100644 > --- a/fs/nfsd/nfs4proc.c > +++ b/fs/nfsd/nfs4proc.c > @@ -351,6 +351,10 @@ nfsd4_open(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, > } > if (status) > goto out; > + if (open->op_xdr_error) { > + status = open->op_xdr_error; > + goto out; > + } > > status = nfsd4_check_open_attributes(rqstp, cstate, open); > if (status) > @@ -417,6 +421,24 @@ out: > } > > /* > + * OPEN is the only seqid-mutating operation whose decoding can fail > + * with a seqid-mutating error (specifically, decoding of user names in > + * the attributes). Therefore we have to do some processing to look up > + * the stateowner so that we can bump the seqid. > + */ > +static __be32 nfsd4_open_omfg(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_op *op) > +{ > + struct nfsd4_open *open = (struct nfsd4_open *)&op->u; > + > + if (!seqid_mutating_err(ntohl(op->status))) > + return op->status; > + if (nfsd4_has_session(cstate)) > + return op->status; > + open->op_xdr_error = op->status; > + return nfsd4_open(rqstp, cstate, open); > +} > + > +/* > * filehandle-manipulating ops. > */ > static __be32 > @@ -1244,8 +1266,11 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, > * for example, if there is a miscellaneous XDR error > * it will be set to nfserr_bad_xdr. > */ > - if (op->status) > + if (op->status) { > + if (op->opnum == OP_OPEN) > + op->status = nfsd4_open_omfg(rqstp, cstate, op); > goto encode_op; > + } > > /* We must be able to encode a successful response to > * this operation, with enough room left over to encode a > diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c > index 0116886..a82df26 100644 > --- a/fs/nfsd/nfs4xdr.c > +++ b/fs/nfsd/nfs4xdr.c > @@ -804,6 +804,7 @@ nfsd4_decode_open(struct nfsd4_compoundargs *argp, struct nfsd4_open *open) > open->op_iattr.ia_valid = 0; > open->op_openowner = NULL; > > + open->op_xdr_error = 0; > /* seqid, share_access, share_deny, clientid, ownerlen */ > READ_BUF(4); > READ32(open->op_seqid); > diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h > index 546f898..be0a79d 100644 > --- a/fs/nfsd/xdr4.h > +++ b/fs/nfsd/xdr4.h > @@ -237,6 +237,7 @@ struct nfsd4_open { > u32 op_share_deny; /* request */ > u32 op_deleg_want; /* request */ > stateid_t op_stateid; /* response */ > + __be32 op_xdr_error; /* see nfsd4_open_omfg() */ > u32 op_recall; /* recall */ > struct nfsd4_change_info op_cinfo; /* response */ > u32 op_rflags; /* response */ > -- > 1.7.9.5 > ^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2013-03-05 22:22 UTC | newest] Thread overview: 3+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2013-03-05 22:17 [PATCH 1/2] nfsd4: remove BUG_ON J. Bruce Fields 2013-03-05 22:17 ` [PATCH 2/2] nfsd4: handle seqid-mutating open errors from xdr decoding J. Bruce Fields 2013-03-05 22:22 ` J. Bruce Fields
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).