Skip to contentSkip to conversation

FAILED: patch "[PATCH] NFSD: NFSv4 file creation neglects setting ACL" failed to apply to 6.6-stable tree

PATCH In review [email protected] 2 participants 24 messages Dec 29, 2025 – Jan 12

Subsystems (3)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 6.6-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y
git checkout FETCH_HEAD
git cherry-pick -x 913f7cf77bf14c13cfea70e89bcb6d0b22239562
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2025122941-civic-revered-b250@gregkh' --subject-prefix 'PATCH 6.6.y' HEAD^..

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 913f7cf77bf14c13cfea70e89bcb6d0b22239562 Mon Sep 17 00:00:00 2001
From: Chuck Lever <[email protected]>
Date: Tue, 18 Nov 2025 19:51:19 -0500
Subject: [PATCH] NFSD: NFSv4 file creation neglects setting ACL
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

An NFSv4 client that sets an ACL with a named principal during file creation retrieves the ACL afterwards, and finds that it is only a default ACL (based on the mode bits) and not the ACL that was requested during file creation. This violates RFC 8881 section 6.4.1.3: "the ACL attribute is set as given".

The issue occurs in nfsd_create_setattr(), which calls nfsd_attrs_valid() to determine whether to call nfsd_setattr(). However, nfsd_attrs_valid() checks only for iattr changes and security labels, but not POSIX ACLs. When only an ACL is present, the function returns false, nfsd_setattr() is skipped, and the POSIX ACL is never applied to the inode.

Subsequently, when the client retrieves the ACL, the server finds no POSIX ACL on the inode and returns one generated from the file's mode bits rather than returning the originally-specified ACL.

Fixes c0cbe70742f4 ("NFSD: add posix ACLs to struct nfsd_attrs")
Cc Roland Mainz <[email protected]>
fs/nfsd/vfs.h +2 -1
index fa46f8b5f132..1dd3ae3ceb3a 100644
@@ -67,7 +67,8 @@ static inline bool nfsd_attrs_valid(struct nfsd_attrs *attrs)
 	struct iattr *iap = attrs->na_iattr;
 
 	return (iap->ia_valid || (attrs->na_seclabel &&
-		attrs->na_seclabel->len));
+		attrs->na_seclabel->len) ||
+		attrs->na_pacl || attrs->na_dpacl);
 }
 
 __be32		nfserrno (int errno);
maintainer
#1 lore raw

cover NFSD: NFSv4 file creation neglects setting ACL

From: Chuck Lever <[email protected]>

I received an automated report that patch "NFSD: NFSv4 file creation neglects setting ACL" failed to apply to the 6.6-stable tree. This series is my attempt to address that failure.

- First, applied several pre-requisite patches
- LLM agent review for possible regressions reported no issues
- CI testing reported no regressions

Chuck Lever (4):
nfsd: convert to new timestamp accessors
nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()
nfsd: set security label during create operations
NFSD: NFSv4 file creation neglects setting ACL

 fs/nfsd/nfs3proc.c  | 10 ++++++----
 fs/nfsd/nfs3xdr.c   |  5 +----
 fs/nfsd/nfs4proc.c  | 11 +++++------
 fs/nfsd/nfs4state.c |  2 +-
 fs/nfsd/nfsctl.c    |  2 +-
 fs/nfsd/nfsproc.c   |  6 +++---
 fs/nfsd/vfs.c       | 22 ++++++++++++++--------
 fs/nfsd/vfs.h       | 11 ++++++++++-
 fs/nfsd/xdr3.h      |  2 +-
 9 files changed, 42 insertions(+), 29 deletions(-)
Signature · 1 line
2.52.0
maintainer
#2 lore raw

PATCH 1/4 nfsd: convert to new timestamp accessors

From: Chuck Lever <[email protected]>

[ Upstream commit 335a7be84b526861f3deb4fdd5d5c2a48cf1feef ]

[ cel: adjust to the 6.6.y version of fs/nfsd/blocklayout.c ]
Stable-dep-of: 24d92de9186e ("nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()")

---
 fs/nfsd/nfs3proc.c | 4 ++--
 fs/nfsd/nfs4proc.c | 8 ++++----
 fs/nfsd/nfsctl.c   | 2 +-
 fs/nfsd/vfs.c      | 2 +-
 4 files changed, 8 insertions(+), 8 deletions(-)
fs/nfsd/nfs3proc.c +2 -2
index 268ef57751c4..666bad8182e5 100644
@@ -294,8 +294,8 @@ nfsd3_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS3_CREATE_EXCLUSIVE:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				break;
 			}
fs/nfsd/nfs4proc.c +4 -4
index 886c09267544..37b918e4a53d 100644
@@ -322,8 +322,8 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS4_CREATE_EXCLUSIVE:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				open->op_created = true;
 				break;		/* subtle */
@@ -331,8 +331,8 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS4_CREATE_EXCLUSIVE4_1:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				open->op_created = true;
 				goto set_attr;	/* subtle */
fs/nfsd/nfsctl.c +1 -1
index 887035b74467..81e0b4726567 100644
@@ -1140,7 +1140,7 @@ static struct inode *nfsd_get_inode(struct super_block *sb, umode_t mode)
 	/* Following advice from simple_fill_super documentation: */
 	inode->i_ino = iunique(sb, NFSD_MaxReserved);
 	inode->i_mode = mode;
-	inode->i_atime = inode->i_mtime = inode_set_ctime_current(inode);
+	simple_inode_init_ts(inode);
 	switch (mode & S_IFMT) {
 	case S_IFDIR:
 		inode->i_fop = &simple_dir_operations;
fs/nfsd/vfs.c +1 -1
index 5ee7149ceaa5..1faf65147223 100644
@@ -521,7 +521,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 
 	nfsd_sanitize_attrs(inode, iap);
 
-	if (check_guard && guardtime != inode_get_ctime(inode).tv_sec)
+	if (check_guard && guardtime != inode_get_ctime_sec(inode))
 		return nfserr_notsync;
 
 	/*
Signature · 1 line
2.52.0
author maintainer
#3 lore raw

Re: [PATCH 6.6.y 1/4] nfsd: convert to new timestamp accessors

Quoted 3 lines · Chuck Lever

From: Chuck Lever <[email protected]>

[ Upstream commit 335a7be84b526861f3deb4fdd5d5c2a48cf1feef ]

I don't see this git id anywhere in Linus's tree, are you sure it is correct?

thanks,

greg k-h

maintainer
#4 lore raw

Re: [PATCH 6.6.y 1/4] nfsd: convert to new timestamp accessors

Quoted 11 lines · Greg Kroah-Hartman
Quoted 3 lines · Chuck Lever

From: Chuck Lever <[email protected]>

[ Upstream commit 335a7be84b526861f3deb4fdd5d5c2a48cf1feef ]

I don't see this git id anywhere in Linus's tree, are you sure it is correct?

thanks,

greg k-h

If I start from the current upstream master, I find this instead:

commit 11fec9b9fb04fd1b3330a3b91ab9dcfa81ad5ad3

Author:     Jeff Layton <[email protected]>
AuthorDate: Wed Oct 4 14:52:37 2023 -0400
Commit:     Christian Brauner <[email protected]>
CommitDate: Wed Oct 18 14:08:24 2023 +0200

nfsd: convert to new timestamp accessors

Convert to using the new inode timestamp accessor functions.

Link:
https://lore.kernel.org/r/[email protected]

I picked up 335a7be84b526861f3deb4fdd5d5c2a48cf1feef from an nfsd-related tag by mistake. Do you want to drop this series and I can rework it properly?

Signature · 1 line
Chuck Lever
author maintainer
#5 lore raw

Re: [PATCH 6.6.y 1/4] nfsd: convert to new timestamp accessors

Quoted 25 lines · Chuck Lever
Quoted 11 lines · Greg Kroah-Hartman
Quoted 3 lines · Chuck Lever

From: Chuck Lever <[email protected]>

[ Upstream commit 335a7be84b526861f3deb4fdd5d5c2a48cf1feef ]

I don't see this git id anywhere in Linus's tree, are you sure it is correct?

thanks,

greg k-h

If I start from the current upstream master, I find this instead:

commit 11fec9b9fb04fd1b3330a3b91ab9dcfa81ad5ad3

Author:     Jeff Layton <[email protected]>
AuthorDate: Wed Oct 4 14:52:37 2023 -0400
Commit:     Christian Brauner <[email protected]>
CommitDate: Wed Oct 18 14:08:24 2023 +0200

nfsd: convert to new timestamp accessors

Convert to using the new inode timestamp accessor functions.

Quoted 8 lines

Link:
https://lore.kernel.org/r/[email protected]

I picked up 335a7be84b526861f3deb4fdd5d5c2a48cf1feef from an nfsd-related tag by mistake. Do you want to drop this series and I can rework it properly?

Now dropped, please rework, thanks!

greg k-h

author maintainer
#6 lore raw

Re: [PATCH 6.6.y 1/4] nfsd: convert to new timestamp accessors

In reply to Chuck Lever
Quoted 8 lines · Chuck Lever

From: Chuck Lever <[email protected]>

[ Upstream commit 335a7be84b526861f3deb4fdd5d5c2a48cf1feef ]

[ cel: adjust to the 6.6.y version of fs/nfsd/blocklayout.c ]
Stable-dep-of: 24d92de9186e ("nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()")

---

Is this really commit 11fec9b9fb04 ("nfsd: convert to new timestamp accessors")? if so, what happened to the original authorship and other signed-off-by lines?

thanks,

greg k-h

maintainer
#7 lore raw

PATCH 2/4 nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()

In reply to Chuck Lever

From: Chuck Lever <[email protected]>

[ Upstream commit 24d92de9186ebc340687caf7356e1070773e67bc ]

nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()

The main point of the guarded SETATTR is to prevent races with other WRITE and SETATTR calls. That requires that the check of the guard time against the inode ctime be done after taking the inode lock.

Furthermore, we need to take into account the 32-bit nature of timestamps in NFSv3, and the possibility that files may change at a faster rate than once a second.

Stable-dep-of: 442d27ff09a2 ("nfsd: set security label during create operations")

---
 fs/nfsd/nfs3proc.c  |  6 ++++--
 fs/nfsd/nfs3xdr.c   |  5 +----
 fs/nfsd/nfs4proc.c  |  3 +--
 fs/nfsd/nfs4state.c |  2 +-
 fs/nfsd/nfsproc.c   |  6 +++---
 fs/nfsd/vfs.c       | 20 +++++++++++++-------
 fs/nfsd/vfs.h       |  2 +-
 fs/nfsd/xdr3.h      |  2 +-
 8 files changed, 25 insertions(+), 21 deletions(-)
fs/nfsd/nfs3proc.c +4 -2
index 666bad8182e5..f4ccbb1f49ba 100644
@@ -71,13 +71,15 @@ nfsd3_proc_setattr(struct svc_rqst *rqstp)
 	struct nfsd_attrs attrs = {
 		.na_iattr	= &argp->attrs,
 	};
+	const struct timespec64 *guardtime = NULL;
 
 	dprintk("nfsd: SETATTR(3)  %s\n",
 				SVCFH_fmt(&argp->fh));
 
 	fh_copy(&resp->fh, &argp->fh);
-	resp->status = nfsd_setattr(rqstp, &resp->fh, &attrs,
-				    argp->check_guard, argp->guardtime);
+	if (argp->check_guard)
+		guardtime = &argp->guardtime;
+	resp->status = nfsd_setattr(rqstp, &resp->fh, &attrs, guardtime);
 	return rpc_success;
 }
 
fs/nfsd/nfs3xdr.c +1 -4
index f32128955ec8..a7a07470c1f8 100644
@@ -295,17 +295,14 @@ svcxdr_decode_sattr3(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 static bool
 svcxdr_decode_sattrguard3(struct xdr_stream *xdr, struct nfsd3_sattrargs *args)
 {
-	__be32 *p;
 	u32 check;
 
 	if (xdr_stream_decode_bool(xdr, &check) < 0)
 		return false;
 	if (check) {
-		p = xdr_inline_decode(xdr, XDR_UNIT * 2);
-		if (!p)
+		if (!svcxdr_decode_nfstime3(xdr, &args->guardtime))
 			return false;
 		args->check_guard = 1;
-		args->guardtime = be32_to_cpup(p);
 	} else
 		args->check_guard = 0;
 
fs/nfsd/nfs4proc.c +1 -2
index 37b918e4a53d..c9ae789bb045 100644
@@ -1160,8 +1160,7 @@ nfsd4_setattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		goto out;
 	save_no_wcc = cstate->current_fh.fh_no_wcc;
 	cstate->current_fh.fh_no_wcc = true;
-	status = nfsd_setattr(rqstp, &cstate->current_fh, &attrs,
-				0, (time64_t)0);
+	status = nfsd_setattr(rqstp, &cstate->current_fh, &attrs, NULL);
 	cstate->current_fh.fh_no_wcc = save_no_wcc;
 	if (!status)
 		status = nfserrno(attrs.na_labelerr);
fs/nfsd/nfs4state.c +1 -1
index 57f1f6aa19c9..511fedc37008 100644
@@ -5223,7 +5223,7 @@ nfsd4_truncate(struct svc_rqst *rqstp, struct svc_fh *fh,
 		return 0;
 	if (!(open->op_share_access & NFS4_SHARE_ACCESS_WRITE))
 		return nfserr_inval;
-	return nfsd_setattr(rqstp, fh, &attrs, 0, (time64_t)0);
+	return nfsd_setattr(rqstp, fh, &attrs, NULL);
 }
 
 static __be32 nfs4_get_vfs_file(struct svc_rqst *rqstp, struct nfs4_file *fp,
fs/nfsd/nfsproc.c +3 -3
index a7315928a760..36370b957b63 100644
@@ -103,7 +103,7 @@ nfsd_proc_setattr(struct svc_rqst *rqstp)
 		}
 	}
 
-	resp->status = nfsd_setattr(rqstp, fhp, &attrs, 0, (time64_t)0);
+	resp->status = nfsd_setattr(rqstp, fhp, &attrs, NULL);
 	if (resp->status != nfs_ok)
 		goto out;
 
@@ -390,8 +390,8 @@ nfsd_proc_create(struct svc_rqst *rqstp)
 		 */
 		attr->ia_valid &= ATTR_SIZE;
 		if (attr->ia_valid)
-			resp->status = nfsd_setattr(rqstp, newfhp, &attrs, 0,
-						    (time64_t)0);
+			resp->status = nfsd_setattr(rqstp, newfhp, &attrs,
+						    NULL);
 	}
 
 out_unlock:
fs/nfsd/vfs.c +13 -7
index 1faf65147223..51c2ad3847c4 100644
@@ -459,7 +459,6 @@ static int __nfsd_setattr(struct dentry *dentry, struct iattr *iap)
  * @rqstp: controlling RPC transaction
  * @fhp: filehandle of target
  * @attr: attributes to set
- * @check_guard: set to 1 if guardtime is a valid timestamp
  * @guardtime: do not act if ctime.tv_sec does not match this timestamp
  *
  * This call may adjust the contents of @attr (in particular, this
@@ -471,8 +470,7 @@ static int __nfsd_setattr(struct dentry *dentry, struct iattr *iap)
  */
 __be32
 nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
-	     struct nfsd_attrs *attr,
-	     int check_guard, time64_t guardtime)
+	     struct nfsd_attrs *attr, const struct timespec64 *guardtime)
 {
 	struct dentry	*dentry;
 	struct inode	*inode;
@@ -521,9 +519,6 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 
 	nfsd_sanitize_attrs(inode, iap);
 
-	if (check_guard && guardtime != inode_get_ctime_sec(inode))
-		return nfserr_notsync;
-
 	/*
 	 * The size case is special, it changes the file in addition to the
 	 * attributes, and file systems don't expect it to be mixed with
@@ -541,6 +536,16 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	err = fh_fill_pre_attrs(fhp);
 	if (err)
 		goto out_unlock;
+
+	if (guardtime) {
+		struct timespec64 ctime = inode_get_ctime(inode);
+		if ((u32)guardtime->tv_sec != (u32)ctime.tv_sec ||
+		    guardtime->tv_nsec != ctime.tv_nsec) {
+			err = nfserr_notsync;
+			goto out_fill_attrs;
+		}
+	}
+
 	for (retries = 1;;) {
 		struct iattr attrs;
 
@@ -568,6 +573,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 		attr->na_aclerr = set_posix_acl(&nop_mnt_idmap,
 						dentry, ACL_TYPE_DEFAULT,
 						attr->na_dpacl);
+out_fill_attrs:
 	fh_fill_post_attrs(fhp);
 out_unlock:
 	inode_unlock(inode);
@@ -1374,7 +1380,7 @@ nfsd_create_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	 * if the attributes have not changed.
 	 */
 	if (iap->ia_valid)
-		status = nfsd_setattr(rqstp, resfhp, attrs, 0, (time64_t)0);
+		status = nfsd_setattr(rqstp, resfhp, attrs, NULL);
 	else
 		status = nfserrno(commit_metadata(resfhp));
 
fs/nfsd/vfs.h +1 -1
index e3c29596f4df..b476028e020b 100644
@@ -69,7 +69,7 @@ __be32		 nfsd_lookup_dentry(struct svc_rqst *, struct svc_fh *,
 				const char *, unsigned int,
 				struct svc_export **, struct dentry **);
 __be32		nfsd_setattr(struct svc_rqst *, struct svc_fh *,
-				struct nfsd_attrs *, int, time64_t);
+			     struct nfsd_attrs *, const struct timespec64 *);
 int nfsd_mountpoint(struct dentry *, struct svc_export *);
 #ifdef CONFIG_NFSD_V4
 __be32		nfsd4_vfs_fallocate(struct svc_rqst *, struct svc_fh *,
fs/nfsd/xdr3.h +1 -1
index 03fe4e21306c..522067b7fd75 100644
@@ -14,7 +14,7 @@ struct nfsd3_sattrargs {
 	struct svc_fh		fh;
 	struct iattr		attrs;
 	int			check_guard;
-	time64_t		guardtime;
+	struct timespec64	guardtime;
 };
 
 struct nfsd3_diropargs {
Signature · 1 line
2.52.0
maintainer
#8 lore raw

PATCH 3/4 nfsd: set security label during create operations

In reply to Chuck Lever

From: Chuck Lever <[email protected]>

[ Upstream commit 442d27ff09a218b61020ab56387dbc508ad6bfa6 ]

When security labeling is enabled, the client can pass a file security label as part of a create operation for the new file, similar to mode and other attributes. At present, the security label is received by nfsd and passed down to nfsd_create_setattr(), but nfsd_setattr() is never called and therefore the label is never set on the new file. This bug may have been introduced on or around commit d6a97d3f589a ("NFSD:
add security label to struct nfsd_attrs"). Looking at nfsd_setattr() I am uncertain as to whether the same issue presents for file ACLs and therefore requires a similar fix for those.

An alternative approach would be to introduce a new LSM hook to set the "create SID" of the current task prior to the actual file creation, which would atomically label the new inode at creation time. This would be better for SELinux and a similar approach has been used previously (see security_dentry_create_files_as) but perhaps not usable by other LSMs.

Reproducer:
1. Install a Linux distro with SELinux - Fedora is easiest
2. git clone https://github.com/SELinuxProject/selinux-testsuite
3. Install the requisite dependencies per selinux-testsuite/README.md
4. Run something like the following script:
MOUNT=$HOME/selinux-testsuite
sudo systemctl start nfs-server
sudo exportfs -o rw,no_root_squash,security_label localhost:$MOUNT sudo mkdir -p /mnt/selinux-testsuite
sudo mount -t nfs -o vers=4.2 localhost:$MOUNT /mnt/selinux-testsuite pushd /mnt/selinux-testsuite/
sudo make -C policy load
pushd tests/filesystem
sudo runcon -t test_filesystem_t ./create_file -f trans_test_file \
-e test_filesystem_filetranscon_t -v
sudo rm -f trans_test_file
popd
sudo make -C policy unload
popd
sudo umount /mnt/selinux-testsuite
sudo exportfs -u localhost:$MOUNT
sudo rmdir /mnt/selinux-testsuite
sudo systemctl stop nfs-server

Expected output:
<eliding noise from commands run prior to or after the test itself> Process context:

        unconfined_u:unconfined_r:test_filesystem_t:s0-s0:c0.c1023

Created file: trans_test_file
File context: unconfined_u:object_r:test_filesystem_filetranscon_t:s0 File context is correct

Actual output:
<eliding noise from commands run prior to or after the test itself> Process context:

        unconfined_u:unconfined_r:test_filesystem_t:s0-s0:c0.c1023

Created file: trans_test_file
File context: system_u:object_r:test_file_t:s0
File context error, expected:
test_filesystem_filetranscon_t
got:
test_file_t

Stable-dep-of: 913f7cf77bf1 ("NFSD: NFSv4 file creation neglects setting ACL")

---
 fs/nfsd/vfs.c | 2 +-
 fs/nfsd/vfs.h | 8 ++++++++
 2 files changed, 9 insertions(+), 1 deletion(-)
fs/nfsd/vfs.c +1 -1
index 51c2ad3847c4..d2ca0739ae81 100644
@@ -1379,7 +1379,7 @@ nfsd_create_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	 * Callers expect new file metadata to be committed even
 	 * if the attributes have not changed.
 	 */
-	if (iap->ia_valid)
+	if (nfsd_attrs_valid(attrs))
 		status = nfsd_setattr(rqstp, resfhp, attrs, NULL);
 	else
 		status = nfserrno(commit_metadata(resfhp));
fs/nfsd/vfs.h +8 -0
index b476028e020b..df9baaee052e 100644
@@ -60,6 +60,14 @@ static inline void nfsd_attrs_free(struct nfsd_attrs *attrs)
 	posix_acl_release(attrs->na_dpacl);
 }
 
+static inline bool nfsd_attrs_valid(struct nfsd_attrs *attrs)
+{
+	struct iattr *iap = attrs->na_iattr;
+
+	return (iap->ia_valid || (attrs->na_seclabel &&
+		attrs->na_seclabel->len));
+}
+
 __be32		nfserrno (int errno);
 int		nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
 		                struct svc_export **expp);
Signature · 1 line
2.52.0
maintainer
#9 lore raw

PATCH 4/4 NFSD: NFSv4 file creation neglects setting ACL

In reply to Chuck Lever

From: Chuck Lever <[email protected]>

[ Upstream commit 913f7cf77bf14c13cfea70e89bcb6d0b22239562 ]

An NFSv4 client that sets an ACL with a named principal during file creation retrieves the ACL afterwards, and finds that it is only a default ACL (based on the mode bits) and not the ACL that was requested during file creation. This violates RFC 8881 section 6.4.1.3: "the ACL attribute is set as given".

The issue occurs in nfsd_create_setattr(), which calls nfsd_attrs_valid() to determine whether to call nfsd_setattr(). However, nfsd_attrs_valid() checks only for iattr changes and security labels, but not POSIX ACLs. When only an ACL is present, the function returns false, nfsd_setattr() is skipped, and the POSIX ACL is never applied to the inode.

Subsequently, when the client retrieves the ACL, the server finds no POSIX ACL on the inode and returns one generated from the file's mode bits rather than returning the originally-specified ACL.

Fixes c0cbe70742f4 ("NFSD: add posix ACLs to struct nfsd_attrs")
Cc Roland Mainz <[email protected]>
---
 fs/nfsd/vfs.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/nfsd/vfs.h +2 -1
index df9baaee052e..6f059c5ac22b 100644
@@ -65,7 +65,8 @@ static inline bool nfsd_attrs_valid(struct nfsd_attrs *attrs)
 	struct iattr *iap = attrs->na_iattr;
 
 	return (iap->ia_valid || (attrs->na_seclabel &&
-		attrs->na_seclabel->len));
+		attrs->na_seclabel->len) ||
+		attrs->na_pacl || attrs->na_dpacl);
 }
 
 __be32		nfserrno (int errno);
Signature · 1 line
2.52.0
maintainer
#10 lore raw

v2 cover NFSD: NFSv4 file creation neglects setting ACL

In reply to [email protected]

From: Chuck Lever <[email protected]>

I received an automated report that patch "NFSD: NFSv4 file creation neglects setting ACL" failed to apply to the 6.6-stable tree. This series is my attempt to address that failure.

- First, applied several pre-requisite patches
- LLM agent review for possible regressions reported no issues
- CI testing reported no regressions

Changes since v1:
- Replace 1/4 with the upstreamed version of that commit

Chuck Lever (1):
NFSD: NFSv4 file creation neglects setting ACL

Jeff Layton (1):
nfsd: convert to new timestamp accessors

Stephen Smalley (1):
nfsd: set security label during create operations

Trond Myklebust (1):
nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()

 fs/nfsd/blocklayout.c |  4 +++-
 fs/nfsd/nfs3proc.c    | 10 ++++++----
 fs/nfsd/nfs3xdr.c     |  5 +----
 fs/nfsd/nfs4proc.c    | 11 +++++------
 fs/nfsd/nfs4state.c   |  2 +-
 fs/nfsd/nfsctl.c      |  2 +-
 fs/nfsd/nfsproc.c     |  6 +++---
 fs/nfsd/vfs.c         | 22 ++++++++++++++--------
 fs/nfsd/vfs.h         | 11 ++++++++++-
 fs/nfsd/xdr3.h        |  2 +-
 10 files changed, 45 insertions(+), 30 deletions(-)
Signature · 1 line
2.52.0
maintainer
#11 lore raw

PATCH v2 1/4 nfsd: convert to new timestamp accessors

In reply to [email protected]

From: Jeff Layton <[email protected]>

[ Upstream commit 11fec9b9fb04fd1b3330a3b91ab9dcfa81ad5ad3 ]

Convert to using the new inode timestamp accessor functions.

Stable-dep-of: 24d92de9186e ("nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()")

---
 fs/nfsd/blocklayout.c | 4 +++-
 fs/nfsd/nfs3proc.c    | 4 ++--
 fs/nfsd/nfs4proc.c    | 8 ++++----
 fs/nfsd/nfsctl.c      | 2 +-
 fs/nfsd/vfs.c         | 2 +-
 5 files changed, 11 insertions(+), 9 deletions(-)
fs/nfsd/blocklayout.c +3 -1
index 59f119cce3dc..db4b67523934 100644
@@ -117,11 +117,13 @@ static __be32
 nfsd4_block_commit_blocks(struct inode *inode, struct nfsd4_layoutcommit *lcp,
 		struct iomap *iomaps, int nr_iomaps)
 {
+	struct timespec64 mtime = inode_get_mtime(inode);
+	loff_t new_size = lcp->lc_last_wr + 1;
 	struct iattr iattr = { .ia_valid = 0 };
 	int error;
 
 	if (lcp->lc_mtime.tv_nsec == UTIME_NOW ||
-	    timespec64_compare(&lcp->lc_mtime, &inode->i_mtime) < 0)
+	    timespec64_compare(&lcp->lc_mtime, &mtime) < 0)
 		lcp->lc_mtime = current_time(inode);
 	iattr.ia_valid |= ATTR_ATIME | ATTR_CTIME | ATTR_MTIME;
 	iattr.ia_atime = iattr.ia_ctime = iattr.ia_mtime = lcp->lc_mtime;
fs/nfsd/nfs3proc.c +2 -2
index 268ef57751c4..666bad8182e5 100644
@@ -294,8 +294,8 @@ nfsd3_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS3_CREATE_EXCLUSIVE:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				break;
 			}
fs/nfsd/nfs4proc.c +4 -4
index 886c09267544..37b918e4a53d 100644
@@ -322,8 +322,8 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS4_CREATE_EXCLUSIVE:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				open->op_created = true;
 				break;		/* subtle */
@@ -331,8 +331,8 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS4_CREATE_EXCLUSIVE4_1:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				open->op_created = true;
 				goto set_attr;	/* subtle */
fs/nfsd/nfsctl.c +1 -1
index 887035b74467..81e0b4726567 100644
@@ -1140,7 +1140,7 @@ static struct inode *nfsd_get_inode(struct super_block *sb, umode_t mode)
 	/* Following advice from simple_fill_super documentation: */
 	inode->i_ino = iunique(sb, NFSD_MaxReserved);
 	inode->i_mode = mode;
-	inode->i_atime = inode->i_mtime = inode_set_ctime_current(inode);
+	simple_inode_init_ts(inode);
 	switch (mode & S_IFMT) {
 	case S_IFDIR:
 		inode->i_fop = &simple_dir_operations;
fs/nfsd/vfs.c +1 -1
index 5ee7149ceaa5..1faf65147223 100644
@@ -521,7 +521,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 
 	nfsd_sanitize_attrs(inode, iap);
 
-	if (check_guard && guardtime != inode_get_ctime(inode).tv_sec)
+	if (check_guard && guardtime != inode_get_ctime_sec(inode))
 		return nfserr_notsync;
 
 	/*
Signature · 1 line
2.52.0
maintainer
#12 lore raw

Re: [PATCH 6.6.y v2 1/4] nfsd: convert to new timestamp accessors

Quoted 10 lines · Chuck Lever

From: Jeff Layton <[email protected]>

[ Upstream commit 11fec9b9fb04fd1b3330a3b91ab9dcfa81ad5ad3 ]

Convert to using the new inode timestamp accessor functions.

Stable-dep-of: 24d92de9186e ("nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()")

Quoted 101 lines
---
 fs/nfsd/blocklayout.c | 4 +++-
 fs/nfsd/nfs3proc.c    | 4 ++--
 fs/nfsd/nfs4proc.c    | 8 ++++----
 fs/nfsd/nfsctl.c      | 2 +-
 fs/nfsd/vfs.c         | 2 +-
 5 files changed, 11 insertions(+), 9 deletions(-)
fs/nfsd/blocklayout.c +3 -1
index 59f119cce3dc..db4b67523934 100644
@@ -117,11 +117,13 @@ static __be32
 nfsd4_block_commit_blocks(struct inode *inode, struct nfsd4_layoutcommit *lcp,
 		struct iomap *iomaps, int nr_iomaps)
 {
+	struct timespec64 mtime = inode_get_mtime(inode);
+	loff_t new_size = lcp->lc_last_wr + 1;
 	struct iattr iattr = { .ia_valid = 0 };
 	int error;
 
 	if (lcp->lc_mtime.tv_nsec == UTIME_NOW ||
-	    timespec64_compare(&lcp->lc_mtime, &inode->i_mtime) < 0)
+	    timespec64_compare(&lcp->lc_mtime, &mtime) < 0)
 		lcp->lc_mtime = current_time(inode);
 	iattr.ia_valid |= ATTR_ATIME | ATTR_CTIME | ATTR_MTIME;
 	iattr.ia_atime = iattr.ia_ctime = iattr.ia_mtime = lcp->lc_mtime;
fs/nfsd/nfs3proc.c +2 -2
index 268ef57751c4..666bad8182e5 100644
@@ -294,8 +294,8 @@ nfsd3_create_file(struct svc_rqst *rqstp, struct 
 svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS3_CREATE_EXCLUSIVE:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				break;
fs/nfsd/nfs3proc.c +0 -0
 			}
fs/nfsd/nfs4proc.c +2 -2
index 886c09267544..37b918e4a53d 100644
@@ -322,8 +322,8 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct 
 svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS4_CREATE_EXCLUSIVE:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				open->op_created = true;
fs/nfsd/nfs4proc.c +2 -2
 				break;		/* subtle */
@@ -331,8 +331,8 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct 
 svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS4_CREATE_EXCLUSIVE4_1:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				open->op_created = true;
fs/nfsd/nfs4proc.c +0 -0
 				goto set_attr;	/* subtle */
fs/nfsd/nfsctl.c +1 -1
index 887035b74467..81e0b4726567 100644
@@ -1140,7 +1140,7 @@ static struct inode *nfsd_get_inode(struct 
 super_block *sb, umode_t mode)
 	/* Following advice from simple_fill_super documentation: */
 	inode->i_ino = iunique(sb, NFSD_MaxReserved);
 	inode->i_mode = mode;
-	inode->i_atime = inode->i_mtime = inode_set_ctime_current(inode);
+	simple_inode_init_ts(inode);
 	switch (mode & S_IFMT) {
 	case S_IFDIR:
fs/nfsd/nfsctl.c +0 -0
 		inode->i_fop = &simple_dir_operations;
fs/nfsd/vfs.c +1 -1
index 5ee7149ceaa5..1faf65147223 100644
@@ -521,7 +521,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh 
 *fhp,
 
 	nfsd_sanitize_attrs(inode, iap);
 
-	if (check_guard && guardtime != inode_get_ctime(inode).tv_sec)
+	if (check_guard && guardtime != inode_get_ctime_sec(inode))
 		return nfserr_notsync;
 

/*

Signature · 1 line
2.52.0
Signature · 1 line
Chuck Lever
author maintainer
#13 lore raw

Re: [PATCH 6.6.y v2 1/4] nfsd: convert to new timestamp accessors

In reply to Chuck Lever
Quoted 11 lines · Chuck Lever

From: Jeff Layton <[email protected]>

[ Upstream commit 11fec9b9fb04fd1b3330a3b91ab9dcfa81ad5ad3 ]

Convert to using the new inode timestamp accessor functions.

Stable-dep-of: 24d92de9186e ("nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()")

Quoted 7 lines
---
 fs/nfsd/blocklayout.c | 4 +++-
 fs/nfsd/nfs3proc.c    | 4 ++--
 fs/nfsd/nfs4proc.c    | 8 ++++----
 fs/nfsd/nfsctl.c      | 2 +-
 fs/nfsd/vfs.c         | 2 +-
 5 files changed, 11 insertions(+), 9 deletions(-)

Adds a build warning, which breaks the build:

fs/nfsd/blocklayout.c: In function ‘nfsd4_block_commit_blocks’:
fs/nfsd/blocklayout.c:123:16: error: unused variable ‘new_size’ [-Werror=unused-variable]

  123 |         loff_t new_size = lcp->lc_last_wr + 1;
      |                ^~~~~~~~

cc1: all warnings being treated as errors

try a 3rd version?

thanks,

greg k-h

maintainer
#14 lore raw

Re: [PATCH 6.6.y v2 1/4] nfsd: convert to new timestamp accessors

Quoted 27 lines · Greg Kroah-Hartman
Quoted 11 lines · Chuck Lever

From: Jeff Layton <[email protected]>

[ Upstream commit 11fec9b9fb04fd1b3330a3b91ab9dcfa81ad5ad3 ]

Convert to using the new inode timestamp accessor functions.

Stable-dep-of: 24d92de9186e ("nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()")

Quoted 7 lines
---
 fs/nfsd/blocklayout.c | 4 +++-
 fs/nfsd/nfs3proc.c    | 4 ++--
 fs/nfsd/nfs4proc.c    | 8 ++++----
 fs/nfsd/nfsctl.c      | 2 +-
 fs/nfsd/vfs.c         | 2 +-
 5 files changed, 11 insertions(+), 9 deletions(-)

Adds a build warning, which breaks the build:

fs/nfsd/blocklayout.c: In function ‘nfsd4_block_commit_blocks’:
fs/nfsd/blocklayout.c:123:16: error: unused variable ‘new_size’ [-Werror=unused-variable]

  123 |         loff_t new_size = lcp->lc_last_wr + 1;
      |                ^~~~~~~~
Quoted 7 lines

cc1: all warnings being treated as errors

try a 3rd version?

thanks,

greg k-h

Harumph. I didn't see any warnings before I posted yesterday.

/me trudges back to his drawing board.

Signature · 1 line
Chuck Lever
maintainer
#15 lore raw

PATCH v2 2/4 nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()

In reply to [email protected]

From: Trond Myklebust <[email protected]>

[ Upstream commit 24d92de9186ebc340687caf7356e1070773e67bc ]

nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()

The main point of the guarded SETATTR is to prevent races with other WRITE and SETATTR calls. That requires that the check of the guard time against the inode ctime be done after taking the inode lock.

Furthermore, we need to take into account the 32-bit nature of timestamps in NFSv3, and the possibility that files may change at a faster rate than once a second.

Stable-dep-of: 442d27ff09a2 ("nfsd: set security label during create operations")

---
 fs/nfsd/nfs3proc.c  |  6 ++++--
 fs/nfsd/nfs3xdr.c   |  5 +----
 fs/nfsd/nfs4proc.c  |  3 +--
 fs/nfsd/nfs4state.c |  2 +-
 fs/nfsd/nfsproc.c   |  6 +++---
 fs/nfsd/vfs.c       | 20 +++++++++++++-------
 fs/nfsd/vfs.h       |  2 +-
 fs/nfsd/xdr3.h      |  2 +-
 8 files changed, 25 insertions(+), 21 deletions(-)
fs/nfsd/nfs3proc.c +4 -2
index 666bad8182e5..f4ccbb1f49ba 100644
@@ -71,13 +71,15 @@ nfsd3_proc_setattr(struct svc_rqst *rqstp)
 	struct nfsd_attrs attrs = {
 		.na_iattr	= &argp->attrs,
 	};
+	const struct timespec64 *guardtime = NULL;
 
 	dprintk("nfsd: SETATTR(3)  %s\n",
 				SVCFH_fmt(&argp->fh));
 
 	fh_copy(&resp->fh, &argp->fh);
-	resp->status = nfsd_setattr(rqstp, &resp->fh, &attrs,
-				    argp->check_guard, argp->guardtime);
+	if (argp->check_guard)
+		guardtime = &argp->guardtime;
+	resp->status = nfsd_setattr(rqstp, &resp->fh, &attrs, guardtime);
 	return rpc_success;
 }
 
fs/nfsd/nfs3xdr.c +1 -4
index f32128955ec8..a7a07470c1f8 100644
@@ -295,17 +295,14 @@ svcxdr_decode_sattr3(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 static bool
 svcxdr_decode_sattrguard3(struct xdr_stream *xdr, struct nfsd3_sattrargs *args)
 {
-	__be32 *p;
 	u32 check;
 
 	if (xdr_stream_decode_bool(xdr, &check) < 0)
 		return false;
 	if (check) {
-		p = xdr_inline_decode(xdr, XDR_UNIT * 2);
-		if (!p)
+		if (!svcxdr_decode_nfstime3(xdr, &args->guardtime))
 			return false;
 		args->check_guard = 1;
-		args->guardtime = be32_to_cpup(p);
 	} else
 		args->check_guard = 0;
 
fs/nfsd/nfs4proc.c +1 -2
index 37b918e4a53d..c9ae789bb045 100644
@@ -1160,8 +1160,7 @@ nfsd4_setattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		goto out;
 	save_no_wcc = cstate->current_fh.fh_no_wcc;
 	cstate->current_fh.fh_no_wcc = true;
-	status = nfsd_setattr(rqstp, &cstate->current_fh, &attrs,
-				0, (time64_t)0);
+	status = nfsd_setattr(rqstp, &cstate->current_fh, &attrs, NULL);
 	cstate->current_fh.fh_no_wcc = save_no_wcc;
 	if (!status)
 		status = nfserrno(attrs.na_labelerr);
fs/nfsd/nfs4state.c +1 -1
index 57f1f6aa19c9..511fedc37008 100644
@@ -5223,7 +5223,7 @@ nfsd4_truncate(struct svc_rqst *rqstp, struct svc_fh *fh,
 		return 0;
 	if (!(open->op_share_access & NFS4_SHARE_ACCESS_WRITE))
 		return nfserr_inval;
-	return nfsd_setattr(rqstp, fh, &attrs, 0, (time64_t)0);
+	return nfsd_setattr(rqstp, fh, &attrs, NULL);
 }
 
 static __be32 nfs4_get_vfs_file(struct svc_rqst *rqstp, struct nfs4_file *fp,
fs/nfsd/nfsproc.c +3 -3
index a7315928a760..36370b957b63 100644
@@ -103,7 +103,7 @@ nfsd_proc_setattr(struct svc_rqst *rqstp)
 		}
 	}
 
-	resp->status = nfsd_setattr(rqstp, fhp, &attrs, 0, (time64_t)0);
+	resp->status = nfsd_setattr(rqstp, fhp, &attrs, NULL);
 	if (resp->status != nfs_ok)
 		goto out;
 
@@ -390,8 +390,8 @@ nfsd_proc_create(struct svc_rqst *rqstp)
 		 */
 		attr->ia_valid &= ATTR_SIZE;
 		if (attr->ia_valid)
-			resp->status = nfsd_setattr(rqstp, newfhp, &attrs, 0,
-						    (time64_t)0);
+			resp->status = nfsd_setattr(rqstp, newfhp, &attrs,
+						    NULL);
 	}
 
 out_unlock:
fs/nfsd/vfs.c +13 -7
index 1faf65147223..51c2ad3847c4 100644
@@ -459,7 +459,6 @@ static int __nfsd_setattr(struct dentry *dentry, struct iattr *iap)
  * @rqstp: controlling RPC transaction
  * @fhp: filehandle of target
  * @attr: attributes to set
- * @check_guard: set to 1 if guardtime is a valid timestamp
  * @guardtime: do not act if ctime.tv_sec does not match this timestamp
  *
  * This call may adjust the contents of @attr (in particular, this
@@ -471,8 +470,7 @@ static int __nfsd_setattr(struct dentry *dentry, struct iattr *iap)
  */
 __be32
 nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
-	     struct nfsd_attrs *attr,
-	     int check_guard, time64_t guardtime)
+	     struct nfsd_attrs *attr, const struct timespec64 *guardtime)
 {
 	struct dentry	*dentry;
 	struct inode	*inode;
@@ -521,9 +519,6 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 
 	nfsd_sanitize_attrs(inode, iap);
 
-	if (check_guard && guardtime != inode_get_ctime_sec(inode))
-		return nfserr_notsync;
-
 	/*
 	 * The size case is special, it changes the file in addition to the
 	 * attributes, and file systems don't expect it to be mixed with
@@ -541,6 +536,16 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	err = fh_fill_pre_attrs(fhp);
 	if (err)
 		goto out_unlock;
+
+	if (guardtime) {
+		struct timespec64 ctime = inode_get_ctime(inode);
+		if ((u32)guardtime->tv_sec != (u32)ctime.tv_sec ||
+		    guardtime->tv_nsec != ctime.tv_nsec) {
+			err = nfserr_notsync;
+			goto out_fill_attrs;
+		}
+	}
+
 	for (retries = 1;;) {
 		struct iattr attrs;
 
@@ -568,6 +573,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 		attr->na_aclerr = set_posix_acl(&nop_mnt_idmap,
 						dentry, ACL_TYPE_DEFAULT,
 						attr->na_dpacl);
+out_fill_attrs:
 	fh_fill_post_attrs(fhp);
 out_unlock:
 	inode_unlock(inode);
@@ -1374,7 +1380,7 @@ nfsd_create_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	 * if the attributes have not changed.
 	 */
 	if (iap->ia_valid)
-		status = nfsd_setattr(rqstp, resfhp, attrs, 0, (time64_t)0);
+		status = nfsd_setattr(rqstp, resfhp, attrs, NULL);
 	else
 		status = nfserrno(commit_metadata(resfhp));
 
fs/nfsd/vfs.h +1 -1
index e3c29596f4df..b476028e020b 100644
@@ -69,7 +69,7 @@ __be32		 nfsd_lookup_dentry(struct svc_rqst *, struct svc_fh *,
 				const char *, unsigned int,
 				struct svc_export **, struct dentry **);
 __be32		nfsd_setattr(struct svc_rqst *, struct svc_fh *,
-				struct nfsd_attrs *, int, time64_t);
+			     struct nfsd_attrs *, const struct timespec64 *);
 int nfsd_mountpoint(struct dentry *, struct svc_export *);
 #ifdef CONFIG_NFSD_V4
 __be32		nfsd4_vfs_fallocate(struct svc_rqst *, struct svc_fh *,
fs/nfsd/xdr3.h +1 -1
index 03fe4e21306c..522067b7fd75 100644
@@ -14,7 +14,7 @@ struct nfsd3_sattrargs {
 	struct svc_fh		fh;
 	struct iattr		attrs;
 	int			check_guard;
-	time64_t		guardtime;
+	struct timespec64	guardtime;
 };
 
 struct nfsd3_diropargs {
Signature · 1 line
2.52.0
maintainer
#16 lore raw

PATCH v2 3/4 nfsd: set security label during create operations

In reply to [email protected]

From: Stephen Smalley <[email protected]>

[ Upstream commit 442d27ff09a218b61020ab56387dbc508ad6bfa6 ]

When security labeling is enabled, the client can pass a file security label as part of a create operation for the new file, similar to mode and other attributes. At present, the security label is received by nfsd and passed down to nfsd_create_setattr(), but nfsd_setattr() is never called and therefore the label is never set on the new file. This bug may have been introduced on or around commit d6a97d3f589a ("NFSD:
add security label to struct nfsd_attrs"). Looking at nfsd_setattr() I am uncertain as to whether the same issue presents for file ACLs and therefore requires a similar fix for those.

An alternative approach would be to introduce a new LSM hook to set the "create SID" of the current task prior to the actual file creation, which would atomically label the new inode at creation time. This would be better for SELinux and a similar approach has been used previously (see security_dentry_create_files_as) but perhaps not usable by other LSMs.

Reproducer:
1. Install a Linux distro with SELinux - Fedora is easiest
2. git clone https://github.com/SELinuxProject/selinux-testsuite
3. Install the requisite dependencies per selinux-testsuite/README.md
4. Run something like the following script:
MOUNT=$HOME/selinux-testsuite
sudo systemctl start nfs-server
sudo exportfs -o rw,no_root_squash,security_label localhost:$MOUNT sudo mkdir -p /mnt/selinux-testsuite
sudo mount -t nfs -o vers=4.2 localhost:$MOUNT /mnt/selinux-testsuite pushd /mnt/selinux-testsuite/
sudo make -C policy load
pushd tests/filesystem
sudo runcon -t test_filesystem_t ./create_file -f trans_test_file \
-e test_filesystem_filetranscon_t -v
sudo rm -f trans_test_file
popd
sudo make -C policy unload
popd
sudo umount /mnt/selinux-testsuite
sudo exportfs -u localhost:$MOUNT
sudo rmdir /mnt/selinux-testsuite
sudo systemctl stop nfs-server

Expected output:
<eliding noise from commands run prior to or after the test itself> Process context:

        unconfined_u:unconfined_r:test_filesystem_t:s0-s0:c0.c1023

Created file: trans_test_file
File context: unconfined_u:object_r:test_filesystem_filetranscon_t:s0 File context is correct

Actual output:
<eliding noise from commands run prior to or after the test itself> Process context:

        unconfined_u:unconfined_r:test_filesystem_t:s0-s0:c0.c1023

Created file: trans_test_file
File context: system_u:object_r:test_file_t:s0
File context error, expected:
test_filesystem_filetranscon_t
got:
test_file_t

Stable-dep-of: 913f7cf77bf1 ("NFSD: NFSv4 file creation neglects setting ACL")

---
 fs/nfsd/vfs.c | 2 +-
 fs/nfsd/vfs.h | 8 ++++++++
 2 files changed, 9 insertions(+), 1 deletion(-)
fs/nfsd/vfs.c +1 -1
index 51c2ad3847c4..d2ca0739ae81 100644
@@ -1379,7 +1379,7 @@ nfsd_create_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	 * Callers expect new file metadata to be committed even
 	 * if the attributes have not changed.
 	 */
-	if (iap->ia_valid)
+	if (nfsd_attrs_valid(attrs))
 		status = nfsd_setattr(rqstp, resfhp, attrs, NULL);
 	else
 		status = nfserrno(commit_metadata(resfhp));
fs/nfsd/vfs.h +8 -0
index b476028e020b..df9baaee052e 100644
@@ -60,6 +60,14 @@ static inline void nfsd_attrs_free(struct nfsd_attrs *attrs)
 	posix_acl_release(attrs->na_dpacl);
 }
 
+static inline bool nfsd_attrs_valid(struct nfsd_attrs *attrs)
+{
+	struct iattr *iap = attrs->na_iattr;
+
+	return (iap->ia_valid || (attrs->na_seclabel &&
+		attrs->na_seclabel->len));
+}
+
 __be32		nfserrno (int errno);
 int		nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
 		                struct svc_export **expp);
Signature · 1 line
2.52.0
maintainer
#17 lore raw

PATCH v2 4/4 NFSD: NFSv4 file creation neglects setting ACL

In reply to [email protected]

From: Chuck Lever <[email protected]>

[ Upstream commit 913f7cf77bf14c13cfea70e89bcb6d0b22239562 ]

An NFSv4 client that sets an ACL with a named principal during file creation retrieves the ACL afterwards, and finds that it is only a default ACL (based on the mode bits) and not the ACL that was requested during file creation. This violates RFC 8881 section 6.4.1.3: "the ACL attribute is set as given".

The issue occurs in nfsd_create_setattr(), which calls nfsd_attrs_valid() to determine whether to call nfsd_setattr(). However, nfsd_attrs_valid() checks only for iattr changes and security labels, but not POSIX ACLs. When only an ACL is present, the function returns false, nfsd_setattr() is skipped, and the POSIX ACL is never applied to the inode.

Subsequently, when the client retrieves the ACL, the server finds no POSIX ACL on the inode and returns one generated from the file's mode bits rather than returning the originally-specified ACL.

Fixes c0cbe70742f4 ("NFSD: add posix ACLs to struct nfsd_attrs")
Cc Roland Mainz <[email protected]>
---
 fs/nfsd/vfs.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/nfsd/vfs.h +2 -1
index df9baaee052e..6f059c5ac22b 100644
@@ -65,7 +65,8 @@ static inline bool nfsd_attrs_valid(struct nfsd_attrs *attrs)
 	struct iattr *iap = attrs->na_iattr;
 
 	return (iap->ia_valid || (attrs->na_seclabel &&
-		attrs->na_seclabel->len));
+		attrs->na_seclabel->len) ||
+		attrs->na_pacl || attrs->na_dpacl);
 }
 
 __be32		nfserrno (int errno);
Signature · 1 line
2.52.0
maintainer
#18 lore raw

v3 cover NFSD: NFSv4 file creation neglects setting ACL

In reply to [email protected]

From: Chuck Lever <[email protected]>

I received an automated report that patch "NFSD: NFSv4 file creation neglects setting ACL" failed to apply to the 6.6-stable tree. This series is my attempt to address that failure.

- First, applied several pre-requisite patches
- LLM agent review for possible regressions reported no issues
- CI testing reported no regressions

Changes since v2:
- Add a Signed-off-by to 1/4
- Address a build warning introduced in 1/4
- Fix the In-Reply-To header

Changes since v1:
- Replace 1/4 with the upstreamed version of that commit

Chuck Lever (1):
NFSD: NFSv4 file creation neglects setting ACL

Jeff Layton (1):
nfsd: convert to new timestamp accessors

Stephen Smalley (1):
nfsd: set security label during create operations

Trond Myklebust (1):
nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()

 fs/nfsd/blocklayout.c |  3 ++-
 fs/nfsd/nfs3proc.c    | 10 ++++++----
 fs/nfsd/nfs3xdr.c     |  5 +----
 fs/nfsd/nfs4proc.c    | 11 +++++------
 fs/nfsd/nfs4state.c   |  2 +-
 fs/nfsd/nfsctl.c      |  2 +-
 fs/nfsd/nfsproc.c     |  6 +++---
 fs/nfsd/vfs.c         | 22 ++++++++++++++--------
 fs/nfsd/vfs.h         | 11 ++++++++++-
 fs/nfsd/xdr3.h        |  2 +-
 10 files changed, 44 insertions(+), 30 deletions(-)
Signature · 1 line
2.52.0
author maintainer
#19 lore raw

Re: [PATCH 6.6.y v3 0/4] NFSD: NFSv4 file creation neglects setting ACL

Quoted 10 lines · Chuck Lever

From: Chuck Lever <[email protected]>

I received an automated report that patch "NFSD: NFSv4 file creation neglects setting ACL" failed to apply to the 6.6-stable tree. This series is my attempt to address that failure.

- First, applied several pre-requisite patches
- LLM agent review for possible regressions reported no issues
- CI testing reported no regressions

Quoted 7 lines

Changes since v2:
- Add a Signed-off-by to 1/4
- Address a build warning introduced in 1/4
- Fix the In-Reply-To header

Changes since v1:
- Replace 1/4 with the upstreamed version of that commit

Nice, this worked, thanks!

greg k-h

maintainer
#20 lore raw

PATCH v3 1/4 nfsd: convert to new timestamp accessors

In reply to [email protected]

From: Jeff Layton <[email protected]>

[ Upstream commit 11fec9b9fb04fd1b3330a3b91ab9dcfa81ad5ad3 ]

Convert to using the new inode timestamp accessor functions.

Stable-dep-of: 24d92de9186e ("nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()")

[ cel: d68886bae76a has already been applied ]
---
 fs/nfsd/blocklayout.c | 3 ++-
 fs/nfsd/nfs3proc.c    | 4 ++--
 fs/nfsd/nfs4proc.c    | 8 ++++----
 fs/nfsd/nfsctl.c      | 2 +-
 fs/nfsd/vfs.c         | 2 +-
 5 files changed, 10 insertions(+), 9 deletions(-)
fs/nfsd/blocklayout.c +2 -1
index 59f119cce3dc..b6b4e389a901 100644
@@ -117,11 +117,12 @@ static __be32
 nfsd4_block_commit_blocks(struct inode *inode, struct nfsd4_layoutcommit *lcp,
 		struct iomap *iomaps, int nr_iomaps)
 {
+	struct timespec64 mtime = inode_get_mtime(inode);
 	struct iattr iattr = { .ia_valid = 0 };
 	int error;
 
 	if (lcp->lc_mtime.tv_nsec == UTIME_NOW ||
-	    timespec64_compare(&lcp->lc_mtime, &inode->i_mtime) < 0)
+	    timespec64_compare(&lcp->lc_mtime, &mtime) < 0)
 		lcp->lc_mtime = current_time(inode);
 	iattr.ia_valid |= ATTR_ATIME | ATTR_CTIME | ATTR_MTIME;
 	iattr.ia_atime = iattr.ia_ctime = iattr.ia_mtime = lcp->lc_mtime;
fs/nfsd/nfs3proc.c +2 -2
index 268ef57751c4..666bad8182e5 100644
@@ -294,8 +294,8 @@ nfsd3_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS3_CREATE_EXCLUSIVE:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				break;
 			}
fs/nfsd/nfs4proc.c +4 -4
index 886c09267544..37b918e4a53d 100644
@@ -322,8 +322,8 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS4_CREATE_EXCLUSIVE:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				open->op_created = true;
 				break;		/* subtle */
@@ -331,8 +331,8 @@ nfsd4_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 			status = nfserr_exist;
 			break;
 		case NFS4_CREATE_EXCLUSIVE4_1:
-			if (d_inode(child)->i_mtime.tv_sec == v_mtime &&
-			    d_inode(child)->i_atime.tv_sec == v_atime &&
+			if (inode_get_mtime_sec(d_inode(child)) == v_mtime &&
+			    inode_get_atime_sec(d_inode(child)) == v_atime &&
 			    d_inode(child)->i_size == 0) {
 				open->op_created = true;
 				goto set_attr;	/* subtle */
fs/nfsd/nfsctl.c +1 -1
index 887035b74467..81e0b4726567 100644
@@ -1140,7 +1140,7 @@ static struct inode *nfsd_get_inode(struct super_block *sb, umode_t mode)
 	/* Following advice from simple_fill_super documentation: */
 	inode->i_ino = iunique(sb, NFSD_MaxReserved);
 	inode->i_mode = mode;
-	inode->i_atime = inode->i_mtime = inode_set_ctime_current(inode);
+	simple_inode_init_ts(inode);
 	switch (mode & S_IFMT) {
 	case S_IFDIR:
 		inode->i_fop = &simple_dir_operations;
fs/nfsd/vfs.c +1 -1
index 5ee7149ceaa5..1faf65147223 100644
@@ -521,7 +521,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 
 	nfsd_sanitize_attrs(inode, iap);
 
-	if (check_guard && guardtime != inode_get_ctime(inode).tv_sec)
+	if (check_guard && guardtime != inode_get_ctime_sec(inode))
 		return nfserr_notsync;
 
 	/*
Signature · 1 line
2.52.0
maintainer
#21 lore raw

PATCH v3 2/4 nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()

In reply to [email protected]

From: Trond Myklebust <[email protected]>

[ Upstream commit 24d92de9186ebc340687caf7356e1070773e67bc ]

nfsd: Fix NFSv3 atomicity bugs in nfsd_setattr()

The main point of the guarded SETATTR is to prevent races with other WRITE and SETATTR calls. That requires that the check of the guard time against the inode ctime be done after taking the inode lock.

Furthermore, we need to take into account the 32-bit nature of timestamps in NFSv3, and the possibility that files may change at a faster rate than once a second.

Stable-dep-of: 442d27ff09a2 ("nfsd: set security label during create operations")

---
 fs/nfsd/nfs3proc.c  |  6 ++++--
 fs/nfsd/nfs3xdr.c   |  5 +----
 fs/nfsd/nfs4proc.c  |  3 +--
 fs/nfsd/nfs4state.c |  2 +-
 fs/nfsd/nfsproc.c   |  6 +++---
 fs/nfsd/vfs.c       | 20 +++++++++++++-------
 fs/nfsd/vfs.h       |  2 +-
 fs/nfsd/xdr3.h      |  2 +-
 8 files changed, 25 insertions(+), 21 deletions(-)
fs/nfsd/nfs3proc.c +4 -2
index 666bad8182e5..f4ccbb1f49ba 100644
@@ -71,13 +71,15 @@ nfsd3_proc_setattr(struct svc_rqst *rqstp)
 	struct nfsd_attrs attrs = {
 		.na_iattr	= &argp->attrs,
 	};
+	const struct timespec64 *guardtime = NULL;
 
 	dprintk("nfsd: SETATTR(3)  %s\n",
 				SVCFH_fmt(&argp->fh));
 
 	fh_copy(&resp->fh, &argp->fh);
-	resp->status = nfsd_setattr(rqstp, &resp->fh, &attrs,
-				    argp->check_guard, argp->guardtime);
+	if (argp->check_guard)
+		guardtime = &argp->guardtime;
+	resp->status = nfsd_setattr(rqstp, &resp->fh, &attrs, guardtime);
 	return rpc_success;
 }
 
fs/nfsd/nfs3xdr.c +1 -4
index f32128955ec8..a7a07470c1f8 100644
@@ -295,17 +295,14 @@ svcxdr_decode_sattr3(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 static bool
 svcxdr_decode_sattrguard3(struct xdr_stream *xdr, struct nfsd3_sattrargs *args)
 {
-	__be32 *p;
 	u32 check;
 
 	if (xdr_stream_decode_bool(xdr, &check) < 0)
 		return false;
 	if (check) {
-		p = xdr_inline_decode(xdr, XDR_UNIT * 2);
-		if (!p)
+		if (!svcxdr_decode_nfstime3(xdr, &args->guardtime))
 			return false;
 		args->check_guard = 1;
-		args->guardtime = be32_to_cpup(p);
 	} else
 		args->check_guard = 0;
 
fs/nfsd/nfs4proc.c +1 -2
index 37b918e4a53d..c9ae789bb045 100644
@@ -1160,8 +1160,7 @@ nfsd4_setattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		goto out;
 	save_no_wcc = cstate->current_fh.fh_no_wcc;
 	cstate->current_fh.fh_no_wcc = true;
-	status = nfsd_setattr(rqstp, &cstate->current_fh, &attrs,
-				0, (time64_t)0);
+	status = nfsd_setattr(rqstp, &cstate->current_fh, &attrs, NULL);
 	cstate->current_fh.fh_no_wcc = save_no_wcc;
 	if (!status)
 		status = nfserrno(attrs.na_labelerr);
fs/nfsd/nfs4state.c +1 -1
index 57f1f6aa19c9..511fedc37008 100644
@@ -5223,7 +5223,7 @@ nfsd4_truncate(struct svc_rqst *rqstp, struct svc_fh *fh,
 		return 0;
 	if (!(open->op_share_access & NFS4_SHARE_ACCESS_WRITE))
 		return nfserr_inval;
-	return nfsd_setattr(rqstp, fh, &attrs, 0, (time64_t)0);
+	return nfsd_setattr(rqstp, fh, &attrs, NULL);
 }
 
 static __be32 nfs4_get_vfs_file(struct svc_rqst *rqstp, struct nfs4_file *fp,
fs/nfsd/nfsproc.c +3 -3
index a7315928a760..36370b957b63 100644
@@ -103,7 +103,7 @@ nfsd_proc_setattr(struct svc_rqst *rqstp)
 		}
 	}
 
-	resp->status = nfsd_setattr(rqstp, fhp, &attrs, 0, (time64_t)0);
+	resp->status = nfsd_setattr(rqstp, fhp, &attrs, NULL);
 	if (resp->status != nfs_ok)
 		goto out;
 
@@ -390,8 +390,8 @@ nfsd_proc_create(struct svc_rqst *rqstp)
 		 */
 		attr->ia_valid &= ATTR_SIZE;
 		if (attr->ia_valid)
-			resp->status = nfsd_setattr(rqstp, newfhp, &attrs, 0,
-						    (time64_t)0);
+			resp->status = nfsd_setattr(rqstp, newfhp, &attrs,
+						    NULL);
 	}
 
 out_unlock:
fs/nfsd/vfs.c +13 -7
index 1faf65147223..51c2ad3847c4 100644
@@ -459,7 +459,6 @@ static int __nfsd_setattr(struct dentry *dentry, struct iattr *iap)
  * @rqstp: controlling RPC transaction
  * @fhp: filehandle of target
  * @attr: attributes to set
- * @check_guard: set to 1 if guardtime is a valid timestamp
  * @guardtime: do not act if ctime.tv_sec does not match this timestamp
  *
  * This call may adjust the contents of @attr (in particular, this
@@ -471,8 +470,7 @@ static int __nfsd_setattr(struct dentry *dentry, struct iattr *iap)
  */
 __be32
 nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
-	     struct nfsd_attrs *attr,
-	     int check_guard, time64_t guardtime)
+	     struct nfsd_attrs *attr, const struct timespec64 *guardtime)
 {
 	struct dentry	*dentry;
 	struct inode	*inode;
@@ -521,9 +519,6 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 
 	nfsd_sanitize_attrs(inode, iap);
 
-	if (check_guard && guardtime != inode_get_ctime_sec(inode))
-		return nfserr_notsync;
-
 	/*
 	 * The size case is special, it changes the file in addition to the
 	 * attributes, and file systems don't expect it to be mixed with
@@ -541,6 +536,16 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	err = fh_fill_pre_attrs(fhp);
 	if (err)
 		goto out_unlock;
+
+	if (guardtime) {
+		struct timespec64 ctime = inode_get_ctime(inode);
+		if ((u32)guardtime->tv_sec != (u32)ctime.tv_sec ||
+		    guardtime->tv_nsec != ctime.tv_nsec) {
+			err = nfserr_notsync;
+			goto out_fill_attrs;
+		}
+	}
+
 	for (retries = 1;;) {
 		struct iattr attrs;
 
@@ -568,6 +573,7 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 		attr->na_aclerr = set_posix_acl(&nop_mnt_idmap,
 						dentry, ACL_TYPE_DEFAULT,
 						attr->na_dpacl);
+out_fill_attrs:
 	fh_fill_post_attrs(fhp);
 out_unlock:
 	inode_unlock(inode);
@@ -1374,7 +1380,7 @@ nfsd_create_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	 * if the attributes have not changed.
 	 */
 	if (iap->ia_valid)
-		status = nfsd_setattr(rqstp, resfhp, attrs, 0, (time64_t)0);
+		status = nfsd_setattr(rqstp, resfhp, attrs, NULL);
 	else
 		status = nfserrno(commit_metadata(resfhp));
 
fs/nfsd/vfs.h +1 -1
index e3c29596f4df..b476028e020b 100644
@@ -69,7 +69,7 @@ __be32		 nfsd_lookup_dentry(struct svc_rqst *, struct svc_fh *,
 				const char *, unsigned int,
 				struct svc_export **, struct dentry **);
 __be32		nfsd_setattr(struct svc_rqst *, struct svc_fh *,
-				struct nfsd_attrs *, int, time64_t);
+			     struct nfsd_attrs *, const struct timespec64 *);
 int nfsd_mountpoint(struct dentry *, struct svc_export *);
 #ifdef CONFIG_NFSD_V4
 __be32		nfsd4_vfs_fallocate(struct svc_rqst *, struct svc_fh *,
fs/nfsd/xdr3.h +1 -1
index 03fe4e21306c..522067b7fd75 100644
@@ -14,7 +14,7 @@ struct nfsd3_sattrargs {
 	struct svc_fh		fh;
 	struct iattr		attrs;
 	int			check_guard;
-	time64_t		guardtime;
+	struct timespec64	guardtime;
 };
 
 struct nfsd3_diropargs {
Signature · 1 line
2.52.0
maintainer
#22 lore raw

PATCH v3 3/4 nfsd: set security label during create operations

In reply to [email protected]

From: Stephen Smalley <[email protected]>

[ Upstream commit 442d27ff09a218b61020ab56387dbc508ad6bfa6 ]

When security labeling is enabled, the client can pass a file security label as part of a create operation for the new file, similar to mode and other attributes. At present, the security label is received by nfsd and passed down to nfsd_create_setattr(), but nfsd_setattr() is never called and therefore the label is never set on the new file. This bug may have been introduced on or around commit d6a97d3f589a ("NFSD:
add security label to struct nfsd_attrs"). Looking at nfsd_setattr() I am uncertain as to whether the same issue presents for file ACLs and therefore requires a similar fix for those.

An alternative approach would be to introduce a new LSM hook to set the "create SID" of the current task prior to the actual file creation, which would atomically label the new inode at creation time. This would be better for SELinux and a similar approach has been used previously (see security_dentry_create_files_as) but perhaps not usable by other LSMs.

Reproducer:
1. Install a Linux distro with SELinux - Fedora is easiest
2. git clone https://github.com/SELinuxProject/selinux-testsuite
3. Install the requisite dependencies per selinux-testsuite/README.md
4. Run something like the following script:
MOUNT=$HOME/selinux-testsuite
sudo systemctl start nfs-server
sudo exportfs -o rw,no_root_squash,security_label localhost:$MOUNT sudo mkdir -p /mnt/selinux-testsuite
sudo mount -t nfs -o vers=4.2 localhost:$MOUNT /mnt/selinux-testsuite pushd /mnt/selinux-testsuite/
sudo make -C policy load
pushd tests/filesystem
sudo runcon -t test_filesystem_t ./create_file -f trans_test_file \
-e test_filesystem_filetranscon_t -v
sudo rm -f trans_test_file
popd
sudo make -C policy unload
popd
sudo umount /mnt/selinux-testsuite
sudo exportfs -u localhost:$MOUNT
sudo rmdir /mnt/selinux-testsuite
sudo systemctl stop nfs-server

Expected output:
<eliding noise from commands run prior to or after the test itself> Process context:

        unconfined_u:unconfined_r:test_filesystem_t:s0-s0:c0.c1023

Created file: trans_test_file
File context: unconfined_u:object_r:test_filesystem_filetranscon_t:s0 File context is correct

Actual output:
<eliding noise from commands run prior to or after the test itself> Process context:

        unconfined_u:unconfined_r:test_filesystem_t:s0-s0:c0.c1023

Created file: trans_test_file
File context: system_u:object_r:test_file_t:s0
File context error, expected:
test_filesystem_filetranscon_t
got:
test_file_t

Stable-dep-of: 913f7cf77bf1 ("NFSD: NFSv4 file creation neglects setting ACL")

---
 fs/nfsd/vfs.c | 2 +-
 fs/nfsd/vfs.h | 8 ++++++++
 2 files changed, 9 insertions(+), 1 deletion(-)
fs/nfsd/vfs.c +1 -1
index 51c2ad3847c4..d2ca0739ae81 100644
@@ -1379,7 +1379,7 @@ nfsd_create_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	 * Callers expect new file metadata to be committed even
 	 * if the attributes have not changed.
 	 */
-	if (iap->ia_valid)
+	if (nfsd_attrs_valid(attrs))
 		status = nfsd_setattr(rqstp, resfhp, attrs, NULL);
 	else
 		status = nfserrno(commit_metadata(resfhp));
fs/nfsd/vfs.h +8 -0
index b476028e020b..df9baaee052e 100644
@@ -60,6 +60,14 @@ static inline void nfsd_attrs_free(struct nfsd_attrs *attrs)
 	posix_acl_release(attrs->na_dpacl);
 }
 
+static inline bool nfsd_attrs_valid(struct nfsd_attrs *attrs)
+{
+	struct iattr *iap = attrs->na_iattr;
+
+	return (iap->ia_valid || (attrs->na_seclabel &&
+		attrs->na_seclabel->len));
+}
+
 __be32		nfserrno (int errno);
 int		nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
 		                struct svc_export **expp);
Signature · 1 line
2.52.0
maintainer
#23 lore raw

PATCH v3 4/4 NFSD: NFSv4 file creation neglects setting ACL

In reply to [email protected]

From: Chuck Lever <[email protected]>

[ Upstream commit 913f7cf77bf14c13cfea70e89bcb6d0b22239562 ]

An NFSv4 client that sets an ACL with a named principal during file creation retrieves the ACL afterwards, and finds that it is only a default ACL (based on the mode bits) and not the ACL that was requested during file creation. This violates RFC 8881 section 6.4.1.3: "the ACL attribute is set as given".

The issue occurs in nfsd_create_setattr(), which calls nfsd_attrs_valid() to determine whether to call nfsd_setattr(). However, nfsd_attrs_valid() checks only for iattr changes and security labels, but not POSIX ACLs. When only an ACL is present, the function returns false, nfsd_setattr() is skipped, and the POSIX ACL is never applied to the inode.

Subsequently, when the client retrieves the ACL, the server finds no POSIX ACL on the inode and returns one generated from the file's mode bits rather than returning the originally-specified ACL.

Fixes c0cbe70742f4 ("NFSD: add posix ACLs to struct nfsd_attrs")
Cc Roland Mainz <[email protected]>
---
 fs/nfsd/vfs.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/nfsd/vfs.h +2 -1
index df9baaee052e..6f059c5ac22b 100644
@@ -65,7 +65,8 @@ static inline bool nfsd_attrs_valid(struct nfsd_attrs *attrs)
 	struct iattr *iap = attrs->na_iattr;
 
 	return (iap->ia_valid || (attrs->na_seclabel &&
-		attrs->na_seclabel->len));
+		attrs->na_seclabel->len) ||
+		attrs->na_pacl || attrs->na_dpacl);
 }
 
 __be32		nfserrno (int errno);
Signature · 1 line
2.52.0