summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Merge branch 'for-2.6.38' of git://linux-nfs.org/~bfields/linuxLinus Torvalds2011-01-1438-383/+608
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * 'for-2.6.38' of git://linux-nfs.org/~bfields/linux: (62 commits) nfsd4: fix callback restarting nfsd: break lease on unlink, link, and rename nfsd4: break lease on nfsd setattr nfsd: don't support msnfs export option nfsd4: initialize cb_per_client nfsd4: allow restarting callbacks nfsd4: simplify nfsd4_cb_prepare nfsd4: give out delegations more quickly in 4.1 case nfsd4: add helper function to run callbacks nfsd4: make sure sequence flags are set after destroy_session nfsd4: re-probe callback on connection loss nfsd4: set sequence flag when backchannel is down nfsd4: keep finer-grained callback status rpc: allow xprt_class->setup to return a preexisting xprt rpc: keep backchannel xprt as long as server connection rpc: move sk_bc_xprt to svc_xprt nfsd4: allow backchannel recovery nfsd4: support BIND_CONN_TO_SESSION nfsd4: modify session list under cl_lock Documentation: fl_mylease no longer exists ... Fix up conflicts in fs/nfsd/vfs.c with the vfs-scale work. The vfs-scale work touched some msnfs cases, and this merge removes support for that entirely, so the conflict was trivial to resolve.
| * nfsd4: fix callback restartingJ. Bruce Fields2011-01-141-12/+9
| | | | | | | | | | | | | | Ensure a new callback is added to the client's list of callbacks at most once. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd: break lease on unlink, link, and renameJ. Bruce Fields2011-01-131-4/+23
| | | | | | | | | | | | | | Any change to any of the links pointing to an entry should also break delegations. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: break lease on nfsd setattrJ. Bruce Fields2011-01-131-8/+4
| | | | | | | | | | | | | | Leases (delegations) should really be broken on any metadata change, not just on size change. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd: don't support msnfs export optionJ. Bruce Fields2011-01-133-44/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We've long had these pointless #ifdef MSNFS's sprinkled throughout the code--pointless because MSNFS is always defined (and we give no config option to make that easy to change). So we could just remove the ifdef's and compile the resulting code unconditionally. But as long as we're there: why not just rip out this code entirely? The only purpose is to implement the "msnfs" export option which turns on Windows-like behavior in some cases, and: - the export option isn't documented anywhere; - the userland utilities (which would need to be able to parse "msnfs" in an export file) don't support it; - I don't know how to maintain this, as I don't know what the proper behavior is; and - google shows no evidence that anyone has ever used this. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: initialize cb_per_clientJ. Bruce Fields2011-01-131-0/+1
| | | | | | | | | | | | | | Otherwise a callback that is aborted before it runs will result in a list_del on an uninitialized list head. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: allow restarting callbacksJ. Bruce Fields2011-01-113-6/+32
| | | | | | | | | | | | | | | | | | | | If we lose the backchannel and then the client repairs the problem, resend any callbacks. We use a new cb_done flag to track whether there is still work to be done for the callback or whether it can be destroyed with the rpc. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: simplify nfsd4_cb_prepareJ. Bruce Fields2011-01-111-22/+4
| | | | | | | | | | | | Remove handling for a nonexistant case (status && !-EAGAIN). Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: give out delegations more quickly in 4.1 caseJ. Bruce Fields2011-01-111-2/+15
| | | | | | | | Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: add helper function to run callbacksJ. Bruce Fields2011-01-111-2/+7
| | | | | | | | Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: make sure sequence flags are set after destroy_sessionJ. Bruce Fields2011-01-113-2/+8
| | | | | | | | | | | | | | If this loses any backchannel, make sure we have a chance to notice that and set the sequence flags. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: re-probe callback on connection lossJ. Bruce Fields2011-01-111-1/+1
| | | | | | | | | | | | This makes sure we set the sequence flag when necessary. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: set sequence flag when backchannel is downJ. Bruce Fields2011-01-113-8/+8
| | | | | | | | | | | | Implement the SEQ4_STATUS_CB_PATH_DOWN flag. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: keep finer-grained callback statusJ. Bruce Fields2011-01-113-17/+22
| | | | | | | | | | | | | | | | | | Distinguish between when the callback channel is known to be down, and when it is not yet confirmed. This will be useful in the 4.1 case. Also, we don't seem to be using the fact that this field is atomic. Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
| * rpc: allow xprt_class->setup to return a preexisting xprtJ. Bruce Fields2011-01-113-9/+13
| | | | | | | | | | | | | | This allows us to reuse the xprt associated with a server connection if one has already been set up. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * rpc: keep backchannel xprt as long as server connectionJ. Bruce Fields2011-01-113-11/+29
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Multiple backchannels can share the same tcp connection; from rfc 5661 section 2.10.3.1: A connection's association with a session is not exclusive. A connection associated with the channel(s) of one session may be simultaneously associated with the channel(s) of other sessions including sessions associated with other client IDs. However, multiple backchannels share a connection, they must all share the same xid stream (hence the same rpc_xprt); the only way we have to match replies with calls at the rpc layer is using the xid. So, keep the rpc_xprt around as long as the connection lasts, in case we're asked to use the connection as a backchannel again. Requests to create new backchannel clients over a given server connection should results in creating new clients that reuse the existing rpc_xprt. But to start, just reject attempts to associate multiple rpc_xprt's with the same underlying bc_xprt. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * rpc: move sk_bc_xprt to svc_xprtJ. Bruce Fields2011-01-114-6/+8
| | | | | | | | | | | | | | This seems obviously transport-level information even if it's currently used only by the server socket code. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: allow backchannel recoveryJ. Bruce Fields2011-01-112-9/+43
| | | | | | | | | | | | | | | | | | Now that we have a list of connections to choose from, we can teach the callback code to just pick a suitable connection and use that, instead of insisting on forever using the connection that the first create_session was sent with. Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
| * nfsd4: support BIND_CONN_TO_SESSIONJ. Bruce Fields2011-01-116-10/+102
| | | | | | | | | | | | | | Basic xdr and processing for BIND_CONN_TO_SESSION. This adds a connection to the list of connections associated with a session. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * nfsd4: modify session list under cl_lockJ. Bruce Fields2011-01-111-0/+6
| | | | | | | | | | | | We want to traverse this from the callback code. Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
| * Documentation: fl_mylease no longer existsJ. Bruce Fields2011-01-111-2/+0
| | | | | | | | Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * Merge commit 'v2.6.37' into for-2.6.38-incomingJ. Bruce Fields2011-01-11424-3094/+4332
| |\ | | | | | | | | | | | | I made a slight mess of Documentation/filesystems/Locking; resolve conflicts with upstream before fixing it up.
| * | Documentation: don't remove lock manager fl_release_privateJ. Bruce Fields2011-01-111-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | I thought I'd removed the last user of this, but missed fs/locks.c:lease_release_private_callback(). Thanks to Christoph for pointing this out. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfs4: set source address when callback is generatedTakuma Umeya2011-01-043-3/+21
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | when callback is generated in NFSv4 server, it doesn't set the source address. When an alias IP is utilized on NFSv4 server and suppose the client is accessing via that alias IP (e.g. eth0:0), the client invokes the callback to the IP address that is set on the original device (e.g. eth0). This behavior results in timeout of xprt. The patch sets the IP address that the client should invoke callback to. Signed-off-by: Takuma Umeya <tumeya@redhat.com> [bfields@redhat.com: Simplify gen_callback arguments, use helper function] Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: return nfs errno from name_to_id functionsJ. Bruce Fields2011-01-044-14/+14
| | | | | | | | | | | | | | | | | | This avoids the need for the confusing ESRCH mapping. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: remove outdated pathname-commentsJ. Bruce Fields2011-01-042-4/+0
| | | | | | | | | | | | Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: move idmap and acl header files into fs/nfsdJ. Bruce Fields2011-01-047-7/+8
| | | | | | | | | | | | | | | | | | These are internal nfsd interfaces. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: name->id mapping should fail with BADOWNER not BADNAMEJ. Bruce Fields2011-01-043-3/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | According to rfc 3530 BADNAME is for strings that represent paths; BADOWNER is for user/group names that don't map. And the too-long name should probably be BADOWNER as well; it's effectively the same as if we couldn't map it. Cc: stable@kernel.org Reported-by: Trond Myklebust <Trond.Myklebust@netapp.com> Reported-by: Simon Kirby <sim@hostway.ca> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | locks: minor setlease cleanupJ. Bruce Fields2011-01-041-1/+1
| | | | | | | | | | | | Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | locks: eliminate fl_mylease callbackJ. Bruce Fields2011-01-043-28/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The nfs server only supports read delegations for now, so we don't care how conflicts are determined. All we care is that unlocks are recognized as matching the leases they are meant to remove. After the last patch, a comparison of struct files will work for that purpose. So we no longer need this callback. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: use a single struct file for delegationsJ. Bruce Fields2011-01-042-5/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | When we converted to sharing struct filess between nfs4 opens I went too far and also used the same mechanism for delegations. But keeping a reference to the struct file ensures it will outlast the lease, and allows us to remove the lease with the same file as we added it. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: eliminate lease delete callbackJ. Bruce Fields2011-01-042-20/+0
| | | | | | | | | | | | | | | | | | | | | nfsd controls the lifetime of the lease, not the lock code, so there's no need for this callback on lease destruction. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | svcrpc: ensure cache_check caller sees updated entryJ. Bruce Fields2011-01-041-1/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Supposes cache_check runs simultaneously with an update on a different CPU: cache_check task doing update ^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^ 1. test for CACHE_VALID 1'. set entry->data & !CACHE_NEGATIVE 2. use entry->data 2'. set CACHE_VALID If the two memory writes performed in step 1' and 2' appear misordered with respect to the reads in step 1 and 2, then the caller could get stale data at step 2 even though it saw CACHE_VALID set on the cache entry. Add memory barriers to prevent this. Reviewed-by: NeilBrown <neilb@suse.de> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | svcrpc: take lock on turning entry NEGATIVE in cache_checkJ. Bruce Fields2011-01-041-7/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We attempt to turn a cache entry negative in place. But that entry may already have been filled in by some other task since we last checked whether it was valid, so we could be modifying an already-valid entry. If nothing else there's a likely leak in such a case when the entry is eventually put() and contents are not freed because it has CACHE_NEGATIVE set. So, take the cache_lock just as sunrpc_cache_update() does. Reviewed-by: NeilBrown <neilb@suse.de> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd: remove some unnecessary dropit handlingJ. Bruce Fields2011-01-044-12/+0
| | | | | | | | | | | | | | | | | | We no longer need a few of these special cases. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd: stop translating EAGAIN to nfserr_dropitJ. Bruce Fields2011-01-041-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | We no longer need this. Also, EWOULDBLOCK is generally a synonym for EAGAIN, but that may not be true on all architectures, so map it as well. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | svcrpc: simpler request droppingJ. Bruce Fields2011-01-044-2/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently we use -EAGAIN returns to determine when to drop a deferred request. On its own, that is error-prone, as it makes us treat -EAGAIN returns from other functions specially to prevent inadvertent dropping. So, use a flag on the request instead. Returning an error on request deferral is still required, to prevent further processing, but we no longer need worry that an error return on its own could result in a drop. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | svcrpc: avoid double reply caused by deferral raceJ. Bruce Fields2011-01-041-7/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit d29068c431599fa "sunrpc: Simplify cache_defer_req and related functions." asserted that cache_check() could determine success or failure of cache_defer_req() by checking the CACHE_PENDING bit. This isn't quite right. We need to know whether cache_defer_req() created a deferred request, in which case sending an rpc reply has become the responsibility of the deferred request, and it is important that we not send our own reply, resulting in two different replies to the same request. And the CACHE_PENDING bit doesn't tell us that; we could have succesfully created a deferred request at the same time as another thread cleared the CACHE_PENDING bit. So, partially revert that commit, to ensure that cache_check() returns -EAGAIN if and only if a deferred request has been created. Signed-off-by: J. Bruce Fields <bfields@redhat.com> Acked-by: NeilBrown <neilb@suse.de>
| * | nfsd: don't drop requests on -ENOMEMJ. Bruce Fields2011-01-041-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | We never want to drop a request if we could return a JUKEBOX/DELAY error instead; so, convert to nfserr_jukebox and let nfsd_dispatch() convert that to a dropit error as a last resort if JUKEBOX/DELAY is unavailable (as in the NFSv2 case). Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd: declare several functions of nfs4callback as staticKirill A. Shutemov2011-01-041-3/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | setup_callback_client(), nfsd4_release_cb() and nfsd4_process_cb_update() do not have users outside the translation unit. Let's declare it as static. Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | SUNRPC: Remove more code when NFSD_DEPRECATED is not configuredJ. Bruce Fields2011-01-042-0/+15
| | | | | | | | | | | | | | | | | | Signed-off-by: NeilBrown <neilb@suse.de> [bfields@redhat.com: moved svcauth_unix_purge outside ifdef's.] Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | svcrpc: modifying valid sunrpc cache entries is racyJ. Bruce Fields2011-01-041-2/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Once a sunrpc cache entry is VALID, we should be replacing it (and allowing any concurrent users to destroy it on last put) instead of trying to update it in place. Otherwise someone referencing the ip_map we're modifying here could try to use the m_client just as we're putting the last reference. The bug should only be seen by users of the legacy nfsd interfaces. (Thanks to Neil for suggestion to use sunrpc_invalidate.) Reviewed-by: NeilBrown <neilb@suse.de> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: fix oops on secinfo_no_name result encodingMi Jinlong2010-12-291-4/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The secinfo_no_name code oopses on encoding with BUG: unable to handle kernel NULL pointer dereference at 00000044 IP: [<e2bd239a>] nfsd4_encode_secinfo+0x1c/0x1c1 [nfsd] We should implement a nfsd4_encode_secinfo_no_name() instead using nfsd4_encode_secinfo(). Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: implement secinfo_no_nameJ. Bruce Fields2010-12-174-2/+48
| | | | | | | | | | | | | | | | | | Implementation of this operation is mandatory for NFSv4.1. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: move guts of nfsd4_lookupp into helperJ. Bruce Fields2010-12-171-6/+10
| | | | | | | | | | | | | | | | | | We'll reuse this code in secinfo_no_name. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfsd4: 4.1 SECINFO should consume filehandleJ. Bruce Fields2010-12-171-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | See the referenced spec language; an attempt by a 4.1 client to use the current filehandle after a secinfo call should result in a NOFILEHANDLE error. Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | net/sunrpc/auth_gss/gss_krb5_crypto.c: Use normal negative error value returnJoe Perches2010-12-171-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | And remove unnecessary double semicolon too. No effect to code, as test is != 0. Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | net: sunrpc: kill unused macrosShan Wei2010-12-173-5/+0
| | | | | | | | | | | | | | | | | | | | | These macros never be used for several years. Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | nfs: add missed CONFIG_NFSD_DEPRECATEDbookjovi@gmail.com2010-12-171-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | these pieces of code only make sense when CONFIG_NFSD_DEPRECATED enabled Signed-off-by: Jovi Zhang <bookjovi@gmail.com> fs/nfsd/nfsctl.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) Signed-off-by: J. Bruce Fields <bfields@redhat.com>
| * | sunrpc: svc_sock_names should hold ref to socket being closed.NeilBrown2010-12-171-4/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently svc_sock_names calls svc_close_xprt on a svc_sock to which it does not own a reference. As soon as svc_close_xprt sets XPT_CLOSE, the socket could be freed by a separate thread (though this is a very unlikely race). It is safer to hold a reference while calling svc_close_xprt. Signed-off-by: NeilBrown <neilb@suse.de> Signed-off-by: J. Bruce Fields <bfields@redhat.com>
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-06 18:31:02 +0000