- 20 Jun, 2009 2 commits
-
-
Ricardo Labiaga authored
xprt_alloc_bc_request() is always called in soft interrupt context. Grab the spin_lock instead of the bottom half spin_lock. Softirqs do not preempt other softirqs running on the same processor, so there is no need to disable bottom halves. Signed-off-by:
Ricardo Labiaga <Ricardo.Labiaga@netapp.com> Signed-off-by:
Benny Halevy <bhalevy@panasas.com> Signed-off-by:
Trond Myklebust <Trond.Myklebust@netapp.com>
-
Benny Halevy authored
nfs4_open_recover_helper clears opendata->o_res before calling nfs4_init_opendata_res, thus causing NFSv4.0 OPEN operations to be sent rather than nfsv4.1. Signed-off-by:
-
- 19 Jun, 2009 1 commit
-
-
Trond Myklebust authored
Conflicts: fs/nfs/client.c fs/nfs/super.c
-
- 18 Jun, 2009 18 commits
-
-
James Morris authored
Unless I'm mistaken, NFS_INO_INVALID_ACL is being checked twice during getacl calls (i.e. first via nfs_revalidate_inode() and then by each all site). Signed-off-by: -
Chuck Lever authored
Specifying "port=-5" with the kernel's current mount option parser generates "unrecognized mount option". If "sloppy" is set, this causes the mount to succeed and use the default values; the desired behavior is that, since this is a valid option with an invalid value, the mount should fail, even with "sloppy." To properly handle "sloppy" parsing, we need to distinguish between correct options with invalid values, and incorrect options. We will need to parse integer values by hand, therefore, and not rely on match_token(). For instance, these must all fail with "invalid value": port=12345678 port=-5 port=samuel and not with "unrecognized option," as they do currently. Thus, for the sake of match_token() we need to treat the values for these options as strings, and do the conversion to integers using strict_strtol(). This is basically the same solution we used for the earlier "retry=" fix (commit ecbb3845), except in this case the kernel actually has to parse the value, rather than ignore it. Signed-off-by:
Chuck Lever <chuck.lever@oracle.com> Signed-off-by:
-
Chuck Lever authored
Ian Kent reports: "I've noticed a couple of other regressions with the options vers and proto option of mount.nfs(8). The commands: mount -t nfs -o vers=<invalid version> <server>:/<path> /<mountpoint> mount -t nfs -o proto=<invalid proto> <server>:/<path> /<mountpoint> both immediately fail. But if the "-s" option is also used they both succeed with the mount falling back to defaults (by the look of it). In the past these failed even when the sloppy option was given, as I think they should. I believe the sloppy option is meant to allow the mount command to still function for mount options (for example in shared autofs maps) that exist on other Unix implementations but aren't present in the Linux mount.nfs(8). So, an invalid value specified for a known mount option is different to an unknown mount option and should fail appropriately." See RH bugzilla 486266. Signed-off-by:
-
Chuck Lever authored
Clean up: Remove xdr_decode_fhstatus() and xdr_decode_fhstatus3(), now that they are unused. Signed-off-by:
-
Chuck Lever authored
Solder xdr_stream-based XDR decoding functions into the in-kernel mountd client that are more careful about checking data types and watching for buffer overflows. The new MNT3 decoder includes support for auth-flavor list decoding. The "_sz" macro for MNT3 replies was missing the size of the file handle. I've added this back, and included the size of the auth flavor array. Signed-off-by:
-
Chuck Lever authored
Introduce an xdr_stream-based XDR decoder that can unpack the auth- flavor list returned in a MNT3 reply. The nfs_mount() function's caller allocates an array, and passes the size and a pointer to it. The decoder decodes all the flavors it can into the array, and returns the number of decoded flavors. If the caller is not interested in the auth flavors, it can pass a value of zero as the size of the pre-allocated array. Signed-off-by:
-
Chuck Lever authored
Introduce xdr_stream-based XDR file handle decoders to the in-kernel mountd client. These are more careful than the existing decoder functions about buffer overflows and data type and range checking. Signed-off-by:
-
Chuck Lever authored
Introduce data structures and xdr_stream-based decoding functions for unmarshalling mountd status codes properly. Mountd version 3 uses specific standard error return codes that are not errno values and not NFS3ERR_ values. These have a well-defined standard mapping to local errno values. Introduce data structures and a decoder function that map these status codes to local errno values properly. This is new functionality (but not used yet). Version 1 mountd status values are defined by RFC 1094 as UNIX error values (errno values). Errno values on heterogeneous systems do not necessarily match each other. To avoid exposing possibly incorrect errno values to upper layers, the current XDR decoder converts all non-zero MNT version 1 status codes to -EACCES. The OpenGroup XNFS standard provides a mapping similar to but smaller than the version 3 error codes. Implement a decoder that uses the XNFS error codes, replacing the current decoder. For both mountd protocol versions, map unrecognized errors to -EACCES. Finally we introduce a replacement data structure for mnt_fhstatus at this time, which is used by the new XDR decoders. In addition to documenting that the status value returned by the XDR decoders is always an errno, this new structure will be expanded in subsequent patches. Signed-off-by:
-
Chuck Lever authored
Clean up: remove xdr_encode_dirpath() now that it has been replaced. Signed-off-by:
-
Chuck Lever authored
Check the length of the supplied dirpath, and see that it fits properly in the RPC buffer. Signed-off-by:
-
Chuck Lever authored
Clean up: Relocate MNT program procedure number definitions to the only file that uses them. Relocate the version number definitions, which are shared, to nfs.h. Remove duplicate program number definitions. Signed-off-by:
-
Chuck Lever authored
Cut NSM upcall RPC traffic in half -- don't do a NULL call first. The cases where a ping would be helpful are rare. Signed-off-by:
-
Chuck Lever authored
When rpc.statd starts up in user space at boot time, it attempts to write the latest NSM local state number into /proc/sys/fs/nfs/nsm_local_state. If lockd.ko isn't loaded yet (as is the case in most configurations), that file doesn't exist, thus the kernel's NSM state remains set to its initial value of zero during lockd operation. This is a problem because rpc.statd and lockd use the NSM state number to prevent repeated lock recovery on rebooted hosts. If lockd sends a zero NSM state, but then a delayed SM_NOTIFY with a real NSM state number is received, there is no way for lockd or rpc.statd to distinguish that stale SM_NOTIFY from an actual reboot. Thus lock recovery could be performed after the rebooted host has already started reclaiming locks, and those locks will be lost. We could change /etc/init.d/nfslock so it always modprobes lockd.ko before starting rpc.statd. However, if lockd.ko is ever unloaded and reloaded, we are back at square one, since the NSM state is not preserved across an unload/reload cycle. This may happen frequently on clients that use automounter. A period of NFS inactivity causes lockd.ko to be unloaded, and the kernel loses its NSM state setting. Instead, let's use the fact that rpc.statd plants the local system's NSM state in every SM_MON (and SM_UNMON) reply. lockd performs a synchronous SM_MON upcall to the local rpc.statd _before_ sending its first NLM request to a new remote. This would permit rpc.statd to provide the current NSM state to lockd, even after lockd.ko had been unloaded and reloaded. Note that NLMPROC_LOCK arguments are constructed before the nsm_monitor() call, so we have to rearrange argument construction very slightly to make this all work out. And, the kernel appears to treat NSM state as a u32 (see struct nlm_args and nsm_res). Make nsm_local_state a u32 as well, to ensure we don't get bogus comparison results. Signed-off-by:
-
Chuck Lever authored
Clear "ret" if the error return from svc_create_xprt(AF_INET6) was -EAFNOSUPORT. Otherwise, callback start-up will succeed, but nfs_callback_up() will return -EAFNOSUPPORT anyway, and the first NFSv4 mount attempt after a reboot will fail. Bug introduced by commit f738f517 in 2.6.30-rc1. Signed-off-by:
-
Chuck Lever authored
If the kernel cannot start the NFSv4 callback service during a mount request, it returns -ENOMEM to user space, resulting in this message: mount.nfs4: Cannot allocate memory Adjust nfs_alloc_client() and nfs_get_client() to pass NFSv4 callback start-up errors back to user space so a less mysterious error message can be displayed by the mount command. Signed-off-by:
-
Chuck Lever authored
The "intr" mount option has been deprecated for a while, but /proc/mounts continues to display "nointr" whether "intr" or "nointr" has been specified for a mount point. Since these options do not have any effect, simply do not display them. Signed-off-by:
-
Suresh Jayaraman authored
Adds support for splice writes. It effectively calls generic_file_splice_write() to do the writes. We need not worry about O_APPEND case as the combination of splice() writes and O_APPEND is disallowed. This patch propagates NFS write errors back to the caller. The number of bytes written via splice are being added to NFSIO_NORMALWRITTENBYTES as these are effectively cached writes. Signed-off-by:
Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by:
-
Trond Myklebust authored
-
- 17 Jun, 2009 19 commits
-
-
Ricardo Labiaga authored
Validates the callback's sessionID, the slot number, and the sequence ID. Increments the slot's sequence. Detects replays, but simply prints a debug message (if debugging is enabled since we don't yet implement a duplicate request cache for the backchannel. This should not present a problem, since only idempotent callbacks are currently implemented. Signed-off-by:
-
Ricardo Labiaga authored
Finds the 'struct nfs_client' that matches the server's address, major version number, and session ID. Signed-off-by:
-
Ricardo Labiaga authored
Defines a new 'struct nfs4_slot_table' in the 'struct nfs4_session' for use by the backchannel. Initializes, resets, and destroys the backchannel slot table in the same manner the forechannel slot table is initialized, reset, and destroyed. The sequenceid for each slot in the backchannel slot table is initialized to 0, whereas the forechannel slotid's sequenceid is set to 1. Signed-off-by:
-
Ricardo Labiaga authored
Generalize nfs4_init_slot_table() so it can be used to initialize the backchannel slot table in addition to the forechannel slot table. Signed-off-by:
-
Ricardo Labiaga authored
Generalize nfs4_reset_slot_table() so it can be used to reset the backchannel slot table in addition to the forechannel slot table. Signed-off-by:
-
Ricardo Labiaga authored
Change the type of cs_addr and csr_status to 'struct sockaddr' and '__be32' since the cb_sequence processing function will use existing functionality that expects these types. Signed-off-by:
-
Ricardo Labiaga authored
The 'rq_received' member of 'struct rpc_rqst' is used to track when we have received a reply to our request. With v4.1, the backchannel can now accept callback requests over the existing connection. Rename this field to make it clear that it is only used for tracking reply bytes and not all bytes received on the connection. Signed-off-by:
-
Benny Halevy authored
CB_SEQUENCE must appear first in the callback compound RPC. If it is not the first operation NFS4ERR_SEQUENCE_POS must be returned. If the first operation ni the CB_COMPOUND is not CB_SEQUENCE then NFS4ERR_OP_NOT_IN_SESSION must be returned. Signed-off-by:
-
Benny Halevy authored
[nfs41: get rid of READMEM and COPYMEM for callback_xdr.c] Signed-off-by:
-
Benny Halevy authored
Currently, just free up any referring calls information. Signed-off-by:
-
Benny Halevy authored
Signed-off-by: -
Benny Halevy authored
Note that this patch changes the nfsv4.0 behavior also when CONFIG_NFS_V4_1 is not defined where NFS4ERR_MINOR_VERS_MISMATCH will be returned if the client received a CB_COMPOUND with minorversion != 0. Previously, it would have returned NFS4ERR_OP_ILLEGAL for CB_SEQUENCE. (or if the server is broken and sent OP_CB_GETATTR or OP_CB_RECALL with minorversion!=0, they would have been processed normally. Signed-off-by:
-
Benny Halevy authored
Signed-off-by: -
Benny Halevy authored
decode cb_compound header conforming to http://tools.ietf.org/html/draft-ietf-nfsv4-minorversion1-26 Get rid of cb_compound_hdr_arg.callback_ident callback_ident is not used anywhere so we shouldn't waste any memory to store it. Signed-off-by:
-
Benny Halevy authored
Signed-off-by: -
Andy Adamson authored
Frees the preallocated backchannel resources that are associated with this session when the session is destroyed. A backchannel is currently created once per session. Destroy the backchannel only when the session is destroyed. Signed-off-by:
-
Rahul Iyer authored
Obtain the rpc_xprt from the rpc_rqst so that calls and callback replies can both use the same code path. A client needs the rpc_xprt in order to reply to a callback. Signed-off-by:
Rahul Iyer <iyer@netapp.com> Signed-off-by:
-
Andy Adamson authored
Set the SESSION4_BACK_CHAN flag to indicate the client supports a backchannel. Signed-off-by:
-
Andy Adamson authored
The NFS v4.1 callback service has already been setup, and rpc_xprt->serv points to the svc_serv structure describing it. Invoke the xprt_setup_backchannel() initialization to pre- allocate the necessary backchannel structures. Signed-off-by:
Alexandros Batsakis <Alexandros.Batsakis@netapp.com> Signed-off-by:
Andy Adamson <andros@netapp.com> Signed-off-by:
-
