[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20230721-bz2223560-v3-2-bb57e302bab7@kernel.org>
Date: Fri, 21 Jul 2023 10:29:11 -0400
From: Jeff Layton <jlayton@...nel.org>
To: Chuck Lever <chuck.lever@...cle.com>, Neil Brown <neilb@...e.de>,
Olga Kornievskaia <kolga@...app.com>,
Dai Ngo <Dai.Ngo@...cle.com>, Tom Talpey <tom@...pey.com>
Cc: linux-nfs@...r.kernel.org, linux-kernel@...r.kernel.org,
Jeff Layton <jlayton@...nel.org>, Boyang Xue <bxue@...hat.com>
Subject: [PATCH v3 2/2] nfsd: remove unsafe BUG_ON from set_change_info
At one time, nfsd would scrape inode information directly out of struct
inode in order to populate the change_info4. At that time, the BUG_ON in
set_change_info made some sense, since having it unset meant a coding
error.
More recently, it calls vfs_getattr to get this information, which can
fail. If that fails, fh_pre_saved can end up not being set. While this
situation is unfortunate, we don't need to crash the box.
Move set_change_info to nfs4proc.c since all of the callers are there.
Revise the condition for setting "atomic" to also check for
fh_pre_saved. Drop the BUG_ON and and just have it zero out both
change_attr4s when this occurs.
Reported-by: Boyang Xue <bxue@...hat.com>
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=2223560
Signed-off-by: Jeff Layton <jlayton@...nel.org>
---
fs/nfsd/nfs4proc.c | 32 ++++++++++++++++++++++++++++++++
fs/nfsd/xdr4.h | 11 -----------
2 files changed, 32 insertions(+), 11 deletions(-)
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index 9285e1eab4d5..3f6710c9c5c9 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -382,6 +382,38 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
return status;
}
+/**
+ * set_change_info - set up the change_info4 for a reply
+ * @cinfo: pointer to nfsd4_change_info to be populated
+ * @fhp: pointer to svc_fh to use as source
+ *
+ * Many operations in NFSv4 require change_info4 in the reply. This function
+ * populates that from the info that we (should!) have already collected. In
+ * the event that we didn't get any pre-attrs, just zero out both.
+ */
+static void
+set_change_info(struct nfsd4_change_info *cinfo, struct svc_fh *fhp)
+{
+ cinfo->atomic = (u32)(fhp->fh_pre_saved && fhp->fh_post_saved && !fhp->fh_no_atomic_attr);
+ cinfo->before_change = fhp->fh_pre_change;
+ cinfo->after_change = fhp->fh_post_change;
+
+ /*
+ * If fetching the pre-change attributes failed, then we should
+ * have already failed the whole operation. We could have still
+ * failed to fetch post-change attributes however.
+ *
+ * If we didn't get post-op attrs, just zero-out the after
+ * field since we don't know what it should be. If the pre_saved
+ * field isn't set for some reason, throw warning and just copy
+ * whatever is in the after field.
+ */
+ if (WARN_ON_ONCE(!fhp->fh_pre_saved))
+ cinfo->before_change = 0;
+ if (!fhp->fh_post_saved)
+ cinfo->after_change = 0;
+}
+
static __be32
do_open_lookup(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_open *open, struct svc_fh **resfh)
{
diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h
index b2931fdf53be..9e67f63c5f4d 100644
--- a/fs/nfsd/xdr4.h
+++ b/fs/nfsd/xdr4.h
@@ -775,17 +775,6 @@ void warn_on_nonidempotent_op(struct nfsd4_op *op);
#define NFS4_SVC_XDRSIZE sizeof(struct nfsd4_compoundargs)
-static inline void
-set_change_info(struct nfsd4_change_info *cinfo, struct svc_fh *fhp)
-{
- BUG_ON(!fhp->fh_pre_saved);
- cinfo->atomic = (u32)(fhp->fh_post_saved && !fhp->fh_no_atomic_attr);
-
- cinfo->before_change = fhp->fh_pre_change;
- cinfo->after_change = fhp->fh_post_change;
-}
-
-
bool nfsd4_mach_creds_match(struct nfs4_client *cl, struct svc_rqst *rqstp);
bool nfs4svc_decode_compoundargs(struct svc_rqst *rqstp, struct xdr_stream *xdr);
bool nfs4svc_encode_compoundres(struct svc_rqst *rqstp, struct xdr_stream *xdr);
--
2.41.0
Powered by blists - more mailing lists