Skip to contentSkip to conversation

mountd: allow high ports on all pseudofs exports

PATCH In review J. Bruce Fields 4 participants 8 messages Dec 2, 2020 – Dec 26, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/2 mountd: allow high ports on all pseudofs exports

From: "J. Bruce Fields" <[email protected]>

We originally tried to grant permissions on the v4 pseudoroot filesystem that were the absolute minimum required for a client to reach a given export. This turns out to be complicated, and we've never gotten it quite right. Also, the tradition from the MNT protocol was to allow anyone to browse the list of exports.

So, do as we already did with security flavors and just allow clients from high ports to access the whole pseudofilesystem.

---
 utils/mountd/v4root.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)
utils/mountd/v4root.c +2 -5
index a9ea167a07e0..2ac4e87898c0 100644
@@ -36,7 +36,7 @@ static nfs_export pseudo_root = {
 		.e_path = "/",
 		.e_flags = NFSEXP_READONLY | NFSEXP_ROOTSQUASH
 				| NFSEXP_NOSUBTREECHECK | NFSEXP_FSID
-				| NFSEXP_V4ROOT,
+				| NFSEXP_V4ROOT | NFSEXP_INSECURE_PORT,
 		.e_anonuid = 65534,
 		.e_anongid = 65534,
 		.e_squids = NULL,
@@ -60,8 +60,6 @@ set_pseudofs_security(struct exportent *pseudo, int flags)
 	struct flav_info *flav;
 	int i;
 
-	if (flags & NFSEXP_INSECURE_PORT)
-		pseudo->e_flags |= NFSEXP_INSECURE_PORT;
 	if ((flags & NFSEXP_ROOTSQUASH) == 0)
 		pseudo->e_flags &= ~NFSEXP_ROOTSQUASH;
 	for (flav = flav_map; flav < flav_map + flav_map_size; flav++) {
@@ -70,8 +68,7 @@ set_pseudofs_security(struct exportent *pseudo, int flags)
 		i = secinfo_addflavor(flav, pseudo);
 		new = &pseudo->e_secinfo[i];
 
-		if (flags & NFSEXP_INSECURE_PORT)
-			new->flags |= NFSEXP_INSECURE_PORT;
+		new->flags |= NFSEXP_INSECURE_PORT;
 	}
 }
 
Signature · 1 line
2.28.0
#1 lore raw

PATCH 2/2 mountd: always root squash on the pseudofs

From: "J. Bruce Fields" <[email protected]>

As with security flavors and "secure" ports, we tried to code this so that pseudofs directories would inherit root squashing from their children, but it doesn't really work as coded and I'm not sure it's useful.

Just root squash always. If it turns out somebody's exporting directories that are only readable by root, I guess we can try to do something else here, but frankly that sounds like a pretty weird configuration.

---
 utils/mountd/v4root.c | 2 --
 1 file changed, 2 deletions(-)
utils/mountd/v4root.c +0 -2
index 2ac4e87898c0..36543401f296 100644
@@ -60,8 +60,6 @@ set_pseudofs_security(struct exportent *pseudo, int flags)
 	struct flav_info *flav;
 	int i;
 
-	if ((flags & NFSEXP_ROOTSQUASH) == 0)
-		pseudo->e_flags &= ~NFSEXP_ROOTSQUASH;
 	for (flav = flav_map; flav < flav_map + flav_map_size; flav++) {
 		struct sec_entry *new;
 
Signature · 1 line
2.28.0
#7 lore raw
In reply to J. Bruce Fields
Quoted 10 lines · J. Bruce Fields

From: "J. Bruce Fields" <[email protected]>

We originally tried to grant permissions on the v4 pseudoroot filesystem that were the absolute minimum required for a client to reach a given export. This turns out to be complicated, and we've never gotten it quite right. Also, the tradition from the MNT protocol was to allow anyone to browse the list of exports.

So, do as we already did with security flavors and just allow clients from high ports to access the whole pseudofilesystem.

Oh, except then we may as well also remove this "flags" parameter.

--b.

utils/mountd/v4root.c +3 -3
index 36543401f296..f6eb126660f3 100644
@@ -55,7 +55,7 @@ static nfs_export pseudo_root = {
 };
 
 static void
-set_pseudofs_security(struct exportent *pseudo, int flags)
+set_pseudofs_security(struct exportent *pseudo)
 {
 	struct flav_info *flav;
 	int i;
@@ -85,7 +85,7 @@ v4root_create(char *path, nfs_export *export)
 	strncpy(eep.e_path, path, sizeof(eep.e_path)-1);
 	if (strcmp(path, "/") != 0)
 		eep.e_flags &= ~NFSEXP_FSID;
-	set_pseudofs_security(&eep, curexp->e_flags);
+	set_pseudofs_security(&eep);
 	exp = export_create(&eep, 0);
 	if (exp == NULL)
 		return NULL;
@@ -133,7 +133,7 @@ pseudofs_update(char *hostname, char *path, nfs_export *source)
 		return 0;
 	}
 	/* Update an existing V4ROOT export: */
-	set_pseudofs_security(&exp->m_export, source->m_export.e_flags);
+	set_pseudofs_security(&exp->m_export);
 	return 0;
 }
 
#2 lore raw

Re: [PATCH 2/2] mountd: always root squash on the pseudofs

In reply to J. Bruce Fields
Quoted 25 lines · J. Bruce Fields

From: "J. Bruce Fields" <[email protected]>

As with security flavors and "secure" ports, we tried to code this so that pseudofs directories would inherit root squashing from their children, but it doesn't really work as coded and I'm not sure it's useful.

Just root squash always.  If it turns out somebody's exporting directories that are only readable by root, I guess we can try to do something else here, but frankly that sounds like a pretty weird configuration.

---
 utils/mountd/v4root.c | 2 --

 1 file changed, 2 deletions(-)
utils/mountd/v4root.c +0 -0
index 2ac4e87898c0..36543401f296 100644
@@ -60,8 +60,6 @@ set_pseudofs_security(struct exportent *pseudo, int
 flags)
         struct flav_info *flav;
         int i;
Quoted 7 lines


-       if ((flags & NFSEXP_ROOTSQUASH) == 0)
-               pseudo->e_flags &= ~NFSEXP_ROOTSQUASH; for (flav = flav_map; flav < flav_map + flav_map_size; flav++) {
                struct sec_entry *new;

Hmm... What is the harm in allowing root to be unsquashed here? Isn't this really all about respecting lookup permissions, or could a user actually modify something in the pseudofs? If the latter, then that sounds like a bug (the pseudofs should always be read-only).

The consequence of not being able to look up a directory in the pseudofs is that the NFSv4 client will be completely unable to mount that subtree, so squashing root could make a major difference.

Signature · 3 lines
Trond Myklebust
Linux NFS client maintainer, Hammerspace
[email protected]
#3 lore raw

Re: [PATCH 2/2] mountd: always root squash on the pseudofs

Quoted 30 lines · Trond Myklebust
Quoted 29 lines · J. Bruce Fields

From: "J. Bruce Fields" <[email protected]>

As with security flavors and "secure" ports, we tried to code this so that pseudofs directories would inherit root squashing from their children, but it doesn't really work as coded and I'm not sure it's useful.

Just root squash always.  If it turns out somebody's exporting directories that are only readable by root, I guess we can try to do something else here, but frankly that sounds like a pretty weird configuration.

---
 utils/mountd/v4root.c | 2 --

 1 file changed, 2 deletions(-)
utils/mountd/v4root.c +0 -0
index 2ac4e87898c0..36543401f296 100644
@@ -60,8 +60,6 @@ set_pseudofs_security(struct exportent *pseudo, int
 flags)
         struct flav_info *flav;
         int i;


-       if ((flags & NFSEXP_ROOTSQUASH) == 0)
-               pseudo->e_flags &= ~NFSEXP_ROOTSQUASH; for (flav = flav_map; flav < flav_map + flav_map_size;

Quoted 8 lines
Quoted 3 lines

flav++) {
                struct sec_entry *new;

Hmm... What is the harm in allowing root to be unsquashed here? Isn't this really all about respecting lookup permissions, or could a user actually modify something in the pseudofs? If the latter, then that sounds like a bug (the pseudofs should always be read-only).

Yeah, it should only be read-only.

Quoted 3 lines

The consequence of not being able to look up a directory in the pseudofs is that the NFSv4 client will be completely unable to mount that subtree, so squashing root could make a major difference.

Fair enough, I'll resend.

--b.

#4 lore raw

PATCH 1/2 mountd: allow high ports on all pseudofs exports

From: "J. Bruce Fields" <[email protected]>

We originally tried to grant permissions on the v4 pseudoroot filesystem that were the absolute minimum required for a client to reach a given export. This turns out to be complicated, and we've never gotten it quite right. Also, the tradition from the MNT protocol was to allow anyone to browse the list of exports.

So, do as we already did with security flavors and just allow clients from high ports to access the whole pseudofilesystem.

---
 utils/mountd/v4root.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)
utils/mountd/v4root.c +5 -8
index a9ea167a07e0..39dd87a94e59 100644
@@ -36,7 +36,7 @@ static nfs_export pseudo_root = {
 		.e_path = "/",
 		.e_flags = NFSEXP_READONLY | NFSEXP_ROOTSQUASH
 				| NFSEXP_NOSUBTREECHECK | NFSEXP_FSID
-				| NFSEXP_V4ROOT,
+				| NFSEXP_V4ROOT | NFSEXP_INSECURE_PORT,
 		.e_anonuid = 65534,
 		.e_anongid = 65534,
 		.e_squids = NULL,
@@ -55,13 +55,11 @@ static nfs_export pseudo_root = {
 };
 
 static void
-set_pseudofs_security(struct exportent *pseudo, int flags)
+set_pseudofs_security(struct exportent *pseudo)
 {
 	struct flav_info *flav;
 	int i;
 
-	if (flags & NFSEXP_INSECURE_PORT)
-		pseudo->e_flags |= NFSEXP_INSECURE_PORT;
 	if ((flags & NFSEXP_ROOTSQUASH) == 0)
 		pseudo->e_flags &= ~NFSEXP_ROOTSQUASH;
 	for (flav = flav_map; flav < flav_map + flav_map_size; flav++) {
@@ -70,8 +68,7 @@ set_pseudofs_security(struct exportent *pseudo, int flags)
 		i = secinfo_addflavor(flav, pseudo);
 		new = &pseudo->e_secinfo[i];
 
-		if (flags & NFSEXP_INSECURE_PORT)
-			new->flags |= NFSEXP_INSECURE_PORT;
+		new->flags |= NFSEXP_INSECURE_PORT;
 	}
 }
 
@@ -90,7 +87,7 @@ v4root_create(char *path, nfs_export *export)
 	strncpy(eep.e_path, path, sizeof(eep.e_path)-1);
 	if (strcmp(path, "/") != 0)
 		eep.e_flags &= ~NFSEXP_FSID;
-	set_pseudofs_security(&eep, curexp->e_flags);
+	set_pseudofs_security(&eep);
 	exp = export_create(&eep, 0);
 	if (exp == NULL)
 		return NULL;
@@ -138,7 +135,7 @@ pseudofs_update(char *hostname, char *path, nfs_export *source)
 		return 0;
 	}
 	/* Update an existing V4ROOT export: */
-	set_pseudofs_security(&exp->m_export, source->m_export.e_flags);
+	set_pseudofs_security(&exp->m_export);
 	return 0;
 }
 
Signature · 1 line
2.28.0
#5 lore raw

PATCH 2/2 mountd: never root squash on the pseudofs

From: "J. Bruce Fields" <[email protected]>

As with security flavors and "secure" ports, we tried to code this so that pseudofs directories would inherit root squashing from their children, but it doesn't really work as coded and I'm not sure it's useful.

Let's just not root squash. The risk is pretty low since the pseudofs is readonly, and we'd rather not risk failing a mount unnecessarily.

---
 utils/mountd/v4root.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
utils/mountd/v4root.c +1 -3
index 39dd87a94e59..c42ba72380ea 100644
@@ -34,7 +34,7 @@ static nfs_export pseudo_root = {
 	.m_export = {
 		.e_hostname = "*",
 		.e_path = "/",
-		.e_flags = NFSEXP_READONLY | NFSEXP_ROOTSQUASH
+		.e_flags = NFSEXP_READONLY
 				| NFSEXP_NOSUBTREECHECK | NFSEXP_FSID
 				| NFSEXP_V4ROOT | NFSEXP_INSECURE_PORT,
 		.e_anonuid = 65534,
@@ -60,8 +60,6 @@ set_pseudofs_security(struct exportent *pseudo)
 	struct flav_info *flav;
 	int i;
 
-	if ((flags & NFSEXP_ROOTSQUASH) == 0)
-		pseudo->e_flags &= ~NFSEXP_ROOTSQUASH;
 	for (flav = flav_map; flav < flav_map + flav_map_size; flav++) {
 		struct sec_entry *new;
 
Signature · 1 line
2.28.0
#6 lore raw

Re: [PATCH 2/2] mountd: never root squash on the pseudofs

Quoted 11 lines · On 12/2/20 8:14 PM, [email protected] wrote:

From: "J. Bruce Fields" <[email protected]>

As with security flavors and "secure" ports, we tried to code this so that pseudofs directories would inherit root squashing from their children, but it doesn't really work as coded and I'm not sure it's useful.

Let's just not root squash. The risk is pretty low since the pseudofs is readonly, and we'd rather not risk failing a mount unnecessarily.

My apologies for taking so long to get to this... I lost it in the weeds ;-)

Both patches Committed!

steved.

Quoted 27 lines
---
 utils/mountd/v4root.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
utils/mountd/v4root.c +1 -3
index 39dd87a94e59..c42ba72380ea 100644
@@ -34,7 +34,7 @@ static nfs_export pseudo_root = {
 	.m_export = {
 		.e_hostname = "*",
 		.e_path = "/",
-		.e_flags = NFSEXP_READONLY | NFSEXP_ROOTSQUASH
+		.e_flags = NFSEXP_READONLY
 				| NFSEXP_NOSUBTREECHECK | NFSEXP_FSID
 				| NFSEXP_V4ROOT | NFSEXP_INSECURE_PORT,
 		.e_anonuid = 65534,
@@ -60,8 +60,6 @@ set_pseudofs_security(struct exportent *pseudo)
 	struct flav_info *flav;
 	int i;
 
-	if ((flags & NFSEXP_ROOTSQUASH) == 0)
-		pseudo->e_flags &= ~NFSEXP_ROOTSQUASH;
 	for (flav = flav_map; flav < flav_map + flav_map_size; flav++) {
 		struct sec_entry *new;