[PATCH] mount.nfs: Ignore vers=4

[PATCH] mount.nfs: Ignore vers=4

[Chuck Lever]

Since mount.nfs looks at the vers= option to determine pmap parameters when
negotiating, it throws EINVAL if "4" is specified, even if the kernel
supports vers=4 for "nfs" type file systems.

Teach mount.nfs to ignore vers=4 and just pass the options to the kernel.

Also fix up umount.nfs to skip the UMNT call for vers=4 mounts.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---

 utils/mount/network.c   |   12 ++++++++----
 utils/mount/nfsumount.c |    4 ++++
 utils/mount/stropts.c   |    6 ++++++
 3 files changed, 18 insertions(+), 4 deletions(-)

diff --git a/utils/mount/network.c b/utils/mount/network.c
index f6fa5fd..420e15a 100644
--- a/utils/mount/network.c
+++ b/utils/mount/network.c
@@ -90,6 +90,7 @@ static const char *nfs_transport_opttbl[] = {
 static const char *nfs_version_opttbl[] = {
 	"v2",
 	"v3",
+	"v4",
 	"vers",
 	"nfsvers",
 	NULL,
@@ -1215,10 +1216,13 @@ nfs_nfs_version(struct mount_options *options, unsigned long *version)
 	case 1: /* v3 */
 		*version = 3;
 		return 1;
-	case 2:	/* vers */
+	case 2: /* v4 */
+		*version = 4;
+		return 1;
+	case 3:	/* vers */
 		switch (po_get_numeric(options, "vers", &tmp)) {
 		case PO_FOUND:
-			if (tmp >= 2 && tmp <= 3) {
+			if (tmp >= 2 && tmp <= 4) {
 				*version = tmp;
 				return 1;
 			}
@@ -1229,10 +1233,10 @@ nfs_nfs_version(struct mount_options *options, unsigned long *version)
 		case PO_BAD_VALUE:
 			return 0;
 		}
-	case 3: /* nfsvers */
+	case 4: /* nfsvers */
 		switch (po_get_numeric(options, "nfsvers", &tmp)) {
 		case PO_FOUND:
-			if (tmp >= 2 && tmp <= 3) {
+			if (tmp >= 2 && tmp <= 4) {
 				*version = tmp;
 				return 1;
 			}
diff --git a/utils/mount/nfsumount.c b/utils/mount/nfsumount.c
index f81db14..c5505b1 100644
--- a/utils/mount/nfsumount.c
+++ b/utils/mount/nfsumount.c
@@ -179,6 +179,10 @@ static int nfs_umount_do_umnt(struct mount_options *options,
 		return EX_FAIL;
 	}
 
+	/* Skip UMNT call for vers=4 mounts */
+	if (nfs_pmap.pm_vers == 4)
+		return EX_SUCCESS;
+
 	*hostname = nfs_umount_hostname(options, *hostname);
 	if (!*hostname) {
 		nfs_error(_("%s: out of memory"), progname);
diff --git a/utils/mount/stropts.c b/utils/mount/stropts.c
index a12ace7..3e88c85 100644
--- a/utils/mount/stropts.c
+++ b/utils/mount/stropts.c
@@ -438,6 +438,7 @@ nfs_rewrite_pmap_mount_options(struct mount_options *options)
 	/*
 	 * Extract just the options needed to contact server.
 	 * Bail now if any of these have bad values.
+	 * Skip probing the server if this is a vers=4 mount
 	 */
 	if (!nfs_extract_server_addresses(options, nfs_saddr, &nfs_salen,
 						mnt_saddr, &mnt_salen)) {
@@ -448,6 +449,11 @@ nfs_rewrite_pmap_mount_options(struct mount_options *options)
 		errno = EINVAL;
 		return 0;
 	}
+	if (nfs_pmap.pm_vers == 4) {
+		if (!nfs_append_clientaddr_option(nfs_saddr, nfs_salen, options))
+			errno = EINVAL;
+		goto out;
+	}
 
 	/*
 	 * The kernel NFS client doesn't support changing the RPC

Re: [PATCH] mount.nfs: Ignore vers=4

[Steve Dickson]

On 08/30/2009 12:37 PM, Chuck Lever wrote:
> Since mount.nfs looks at the vers= option to determine pmap parameters when
> negotiating, it throws EINVAL if "4" is specified, even if the kernel
> supports vers=4 for "nfs" type file systems.
> 
> Teach mount.nfs to ignore vers=4 and just pass the options to the kernel.
How will mount.nfs know that the kernel now supports the "vers=4" option?

steved

Re: [PATCH] mount.nfs: Ignore vers=4

[Chuck Lever]

On Aug 31, 2009, at 8:44 AM, Steve Dickson wrote:
> On 08/30/2009 12:37 PM, Chuck Lever wrote:
>> Since mount.nfs looks at the vers= option to determine pmap  
>> parameters when
>> negotiating, it throws EINVAL if "4" is specified, even if the kernel
>> supports vers=4 for "nfs" type file systems.
>>
>> Teach mount.nfs to ignore vers=4 and just pass the options to the  
>> kernel.
> How will mount.nfs know that the kernel now supports the "vers=4"  
> option?

The same way it knows after we add any new text-based feature... The  
option will be passed down, and if the kernel supports it, it will  
work.  If not, mount(2) will return EINVAL, and the mount.nfs command  
will report the problem.

--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com