- 02 Mar, 2012 11 commits
-
-
Chuck Lever authored
For debugging, introduce a simplistic function to print NFS file handles on the system console. The main function is hooked into the dprintk debugging facility, but you can directly call the helper, _nfs_display_fhandle(), if you want to print a handle unconditionally. Signed-off-by:
Chuck Lever <chuck.lever@oracle.com> Signed-off-by:
Trond Myklebust <Trond.Myklebust@netapp.com>
-
Chuck Lever authored
For NFSv4 mounts, the clientaddr= mount option has always been required. Now we have rpc_localaddr() in the kernel, which was modeled after the same logic in the mount.nfs command that constructs the clientaddr= mount option. If user space doesn't provide a clientaddr= mount option, the kernel can now construct its own. Signed-off-by:
-
Chuck Lever authored
NFSv4.0 clients must send endpoint information for their callback service to NFSv4.0 servers during their first contact with a server. Traditionally on Linux, user space provides the callback endpoint IP address via the "clientaddr=" mount option. During an NFSv4 migration event, it is possible that an FSID may be migrated to a destination server that is accessible via a different source IP address than the source server was. The client must update callback endpoint information on the destination server so that it can maintain leases and allow delegation. Without a new "clientaddr=" option from user space, however, the kernel itself must construct an appropriate IP address for the callback update. Provide an API in the RPC client for upper layer RPC consumers to acquire a source address for a remote. The mechanism used by the mount.nfs command is copied: set up a connected UDP socket to the designated remote, then scrape the source address off the socket. We are careful to select the correct network namespace when setting up the temporary UDP socket. Signed-off-by:
-
Trond Myklebust authored
When the cl_xprt field is updated, the cl_server field will also have to change. Since the contents of cl_server follow the remote endpoint of cl_xprt, just move that field to the rpc_xprt. Signed-off-by:
-
Trond Myklebust authored
A migration event will replace the rpc_xprt used by an rpc_clnt. To ensure this can be done safely, all references to cl_xprt must now use a form of rcu_dereference(). Special care is taken with rpc_peeraddr2str(), which returns a pointer to memory whose lifetime is the same as the rpc_xprt. Signed-off-by:
-
Chuck Lever authored
CLOSE is new with NFSv4. Sometimes it's important to know the timing of this operation compared to things like lease renewal. Signed-off-by:
-
Chuck Lever authored
I noticed recently that decode_attr_fs_locations() is not generating very pretty debugging output. The pathname components each appear on a separate line of output, though that does not appear to be the intended display behavior. The preferred way to generate continued lines of output on the console is to use pr_cont(). Note that incoming pathname4 components contain a string that is not necessarily NUL-terminated. I did actually see some trailing garbage on the console. In addition to correcting the line continuation problem, add a string precision format specifier to ensure that each component string is displayed properly, and that vsnprintf() does not Oops. Someone pointed out that allowing incoming network data to possibly generate a console line of unbounded length may not be such a good idea. Since this output will rarely be enabled, and there is a hard upper bound (NFS4_PATHNAME_MAXCOMPONENTS) in our implementation, this is probably not a major concern. It might be useful to additionally sanity-check the length of each incoming component, however. RFC 3530bis15 does not suggest a maximum number of UTF-8 characters per component for either the pathname4 or component4 types. However, we could invent one that is appropriate for our implementation. Another possibility is to scrap all of this and print these pathnames in upper layers after a reasonable amount of sanity checking in the XDR layer. This would give us an opportunity to allocate a full buffer so that the whole pathname would be output via a single dprintk. Introduced by commit 7aaa0b3b: "NFSv4: convert fs-locations-components to conform to RFC3530," (June 9, 2006). Signed-off-by:
-
Chuck Lever authored
Eliminate a number of implicit type casts in comparisons, and these compiler warnings: fs/nfs/dir.c: In function ‘nfs_readdir_clear_array’: fs/nfs/dir.c:264:16: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] fs/nfs/dir.c: In function ‘nfs_readdir_search_for_cookie’: fs/nfs/dir.c:352:16: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] fs/nfs/dir.c: In function ‘nfs_do_filldir’: fs/nfs/dir.c:769:38: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] fs/nfs/dir.c:780:9: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] Signed-off-by:
-
Trond Myklebust authored
Signed-off-by: -
Trond Myklebust authored
The 'minorversion' mount option is now deprecated, so we need to display the minor version number in the 'vers=' format. Signed-off-by: -
Trond Myklebust authored
Allow the user to mount an NFSv4.0 or NFSv4.1 partition using a standard syntax of '-overs=4.0', or '-overs=4.1' rather than the more cumbersome '-overs=4,minorversion=1'. See also the earlier patch by Dros Adamson, which added the Linux-specific syntax '-ov4.0', '-ov4.1'. Signed-off-by:
-
- 01 Mar, 2012 7 commits
-
-
Weston Andros Adamson authored
Shows the implementation ids in /proc/self/mountstats. This doesn't break the nfs-utils mountstats tool. Signed-off-by:
Weston Andros Adamson <dros@netapp.com> Signed-off-by:
-
Weston Andros Adamson authored
server_scope would never be freed if nfs4_check_cl_exchange_flags() returned non-zero Signed-off-by:
-
Weston Andros Adamson authored
Send the nfs implementation id in EXCHANGE_ID requests unless the module parameter nfs.send_implementation_id is 0. This adds a CONFIG variable for the nii_domain that defaults to "kernel.org". Signed-off-by:
-
Bryan Schumaker authored
This patch removes the old hashmap-based caching and instead uses a "request key actor" to place an upcall to the legacy idmapper rather than going through /sbin/request-key. This will only be used as a fallback if /etc/request-key.conf isn't configured to use nfsidmap. Signed-off-by:
Bryan Schumaker <bjschuma@netapp.com> Signed-off-by:
-
Bryan Schumaker authored
The keyctl_set_timeout function isn't exported to other parts of the kernel, but I want to use it for the NFS idmapper. I already have the key, but I wanted a generic way to set the timeout. Signed-off-by:
David Howells <dhowells@redhat.com> Signed-off-by:
-
Trond Myklebust authored
The NFS4CLNT_LAYOUTRECALL bit is a long-term impediment to scalability. It basically stops all other recalls by a given server once any layout recall is requested. If the recall is for a different file, then we don't care. If the recall applies to the same file, then we're in one of two situations: Either we are in the case of a replay of an existing request, in which case the session is supposed to deal with matters, or we are dealing with a completely different request, in which case we should just try to process it. Signed-off-by: -
Trond Myklebust authored
The NFS4CLNT_LAYOUTRECALL tests in pnfs_layout_process and pnfs_update_layout are redundant. In the case of a bulk layout recall, we're always testing for the NFS_LAYOUT_BULK_RECALL flay anyway. In the case of a file or segment recall, the call to pnfs_set_layout_stateid() updates the layout_header 'barrier' sequence id, which triggers the test in pnfs_layoutgets_blocked() and is less race-prone than NFS4CLNT_LAYOUTRECALL anyway. Signed-off-by:
-
- 27 Feb, 2012 4 commits
-
-
Stanislav Kinsbursky authored
Currently, wait queue, used for polling of RPC pipe changes from user-space, is a part of RPC pipe. But the pipe data itself can be released on NFS umount prior to dentry-inode pair, connected to it (is case of this pair is open by some process). This is not a problem for almost all pipe users, because all PipeFS file operations checks pipe reference prior to using it. Except evenfd. This thing registers itself with "poll" file operation and thus has a reference to pipe wait queue. This leads to oopses on destroying eventfd after NFS umount (like rpc_idmapd do) since not pipe data left to the point already. The solution is to wait queue from pipe data to internal RPC inode data. This looks more logical, because this wiat queue used only for user-space processes, which already holds inode reference. Note: upcalls have to get pipe->dentry prior to dereferecing wait queue to make sure, that mount point won't disappear from underneath us. Signed-off-by:
Stanislav Kinsbursky <skinsbursky@parallels.com> Signed-off-by:
-
Stanislav Kinsbursky authored
There are 2 tightly bound objects: pipe data (created for kernel needs, has reference to dentry, which depends on PipeFS mount/umount) and PipeFS dentry/inode pair (created on mount for user-space needs). They both independently may have or have not a valid reference to each other. This means, that we have to make sure, that pipe->dentry reference is valid on upcalls, and dentry->pipe reference is valid on downcalls. The latter check is absent - my fault. IOW, PipeFS dentry can be opened by some process (rpc.idmapd for example), but it's pipe data can belong to NFS mount, which was unmounted already and thus pipe data was destroyed. To fix this, pipe reference have to be set to NULL on rpc_unlink() and checked on PipeFS file operations instead of pipe->dentry check. Note: PipeFS "poll" file operation will be updated in next patch, because it's logic is more complicated. Signed-off-by:
-
Stanislav Kinsbursky authored
v3: 1) Lookup for client is performed from the beginning of the list on each PipeFS event handling operation. Lockdep is sad otherwise, because inode mutex is taken on PipeFS dentry creation, which can be called on mount notification, where this per-net client lock is taken on clients list walk. Signed-off-by:
-
Stanislav Kinsbursky authored
v3: 1) Lookup for client is performed from the beginning of the list on each PipeFS event handling operation. Lockdep is sad otherwise, because inode mutex is taken on PipeFS dentry creation, which can be called on mount notification, where this per-net client lock is taken on clients list walk. Signed-off-by:
-
- 26 Feb, 2012 1 commit
-
-
Trond Myklebust authored
Signed-off-by:
-
- 19 Feb, 2012 2 commits
-
-
Trond Myklebust authored
Otherwise we have no guarantee that the net namespace won't just disappear from underneath us once the task that created it is destroyed. Signed-off-by: -
Trond Myklebust authored
Currently, the nfs_parsed_mount_data->net field is initialised in the nfs_parse_mount_options() function, which means that it only gets set if we're using text based mounts. The legacy binary mount interface is therefore broken. Fix is to initialise the ->net field in nfs_alloc_parsed_mount_data. Signed-off-by:
-
- 17 Feb, 2012 2 commits
-
-
Weston Andros Adamson authored
Include RPC statistics from all data servers in /proc/self/mountstats for pNFS filelayout mounts. Signed-off-by:
-
Andy Adamson authored
Signed-off-by:
Andy Adamson <andros@netapp.com> Signed-off-by:
-
- 16 Feb, 2012 4 commits
-
-
Chuck Lever authored
Clean up: Fix a debugging message which had an obsolete function name in it (nfs_follow_mountpoint). Introduced by commit 36d43a43 "NFS: Use d_automount() rather than abusing follow_link()" (January 14, 2011) Signed-off-by:
-
Chuck Lever authored
Our dprintk() debugging facility doesn't specify any verbosity level for it's printk() calls, but it should. The default verbosity for printk's is KERN_DEFAULT. You might argue that these are debugging printk's and thus the verbosity should be KERN_DEBUG. That would mean that to see NFS and SUNRPC debugging output an admin would also have to boost the syslog verbosity, which would be insufferably noisy. Signed-off-by:
-
Andy Adamson authored
With static RPC slots, the xprt backlog queue stats were useful in showing when the transport (TCP) was starved by lack of RPC slots. The new dynamic RPC slot code, commit d9ba131d, always provides an RPC slot and so only uses the xprt backlog queue when the tcp_max_slot_table_entries value has been hit or when an allocation error occurs. All requests are now placed on the xprt sending or pending queue which need to be monitored for debugging. The max_slot stat shows the maximum number of dynamic RPC slots reached which is useful when debugging performance issues. Add the new fields at the end of the mountstats xprt stanza so that mountstats outputs the previous correct values and ignores the new fields. Bump NFS_IOSTATS_VERS. Signed-off-by:
-
Stanislav Kinsbursky authored
Signed-off-by:
-
- 15 Feb, 2012 9 commits
-
-
Vitaliy Gusev authored
ca_maxoperations: For the backchannel, the server MUST NOT change the value the client offers. For the fore channel, the server MAY change the requested value. ca_maxrequests: For the backchannel, the server MUST NOT change the value the client offers. For the fore channel, the server MAY change the requested value. Signed-off-by:Vitaliy Gusev <gusev.vitaliy@nexenta.com> Signed-off-by:
-
Weston Andros Adamson authored
Don't allow invalid 'vers' and 'minorversion' combinations in mount options, such as "vers=3,minorversion=1". Signed-off-by:
-
Trond Myklebust authored
The tracepoint code relies on the queue->name being defined in order to be able to display the name of the waitqueue on which an RPC task is sleeping. Reported-by:
Randy Dunlap <rdunlap@xenotime.net> Reported-by:
Steven Rostedt <rostedt@goodmis.org> Signed-off-by:
-
Trond Myklebust authored
Don't allocate the legacy idmapper tables until we actually need them. Signed-off-by:
Jeff Layton <jlayton@redhat.com>
-
Trond Myklebust authored
Add the appropriate 'select KEYS' to the NFSv4 Kconfig entry. Signed-off-by: -
Trond Myklebust authored
Instead of pre-allocating the storage for all the strings, we can significantly reduce the size of that table by doing the allocation when we do the downcall. Signed-off-by:
-
Weston Andros Adamson authored
Signed-off-by:
-
Stanislav Kinsbursky authored
It's not compilable in case of CONFIG_NFS_V4_1 is not set. Signed-off-by:
-
Trond Myklebust authored
Ensure that we initialise the nfs_net->nfs_client_lock spinlock. Also ensure that nfs_server_remove_lists() doesn't try to dereference server->nfs_client before that is initialised. Signed-off-by:
-
