aboutsummaryrefslogtreecommitdiff
path: root/fs/nfs
AgeCommit message (Collapse)AuthorFilesLines
2015-04-11generic_write_checks(): drop isblk argumentAl Viro1-1/+1
all remaining callers are passing 0; some just obscure that fact. Signed-off-by: Al Viro <[email protected]>
2015-04-11direct_IO: remove rw from a_ops->direct_IO()Omar Sandoval1-2/+1
Now that no one is using rw, remove it completely. Signed-off-by: Omar Sandoval <[email protected]> Signed-off-by: Al Viro <[email protected]>
2015-04-11direct_IO: use iov_iter_rw() instead of rw everywhereOmar Sandoval1-1/+1
The rw parameter to direct_IO is redundant with iov_iter->type, and treated slightly differently just about everywhere it's used: some users do rw & WRITE, and others do rw == WRITE where they should be doing a bitwise check. Simplify this with the new iov_iter_rw() helper, which always returns either READ or WRITE. Signed-off-by: Omar Sandoval <[email protected]> Signed-off-by: Al Viro <[email protected]>
2015-04-11make new_sync_{read,write}() staticAl Viro2-4/+0
All places outside of core VFS that checked ->read and ->write for being NULL or called the methods directly are gone now, so NULL {read,write} with non-NULL {read,write}_iter will do the right thing in all cases. Signed-off-by: Al Viro <[email protected]>
2015-04-11Merge branch 'iocb' into for-nextAl Viro2-3/+2
2015-03-27NFS: Block new writes while syncing data in nfs_getattr()Trond Myklebust1-0/+2
Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1/pnfs: Separate out metadata and data consistency for pNFSTrond Myklebust9-8/+47
The LAYOUTCOMMIT operation means different things to different layout types. For blocks and objects, it is both a data and metadata consistency operation. For files and flexfiles, it is only a metadata consistency operation. This patch separates out the 2 cases, allowing the files/flexfiles layout drivers to optimise away the data consistency calls to layoutcommit. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1/pnfs: Ensure we send layoutcommit before return-on-closeTrond Myklebust1-1/+4
We must not send a close or delegreturn that would result in a return-on-close of the layout without ensuring that we've also sent the necessary layoutcommit. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1/pnfs: Ensure that writes respect the O_SYNC flag when doing O_DIRECTTrond Myklebust3-0/+3
If the caller does not specify the O_SYNC flag, then it is legitimate to return from O_DIRECT without doing a pNFS layoutcommit operation. However if the file is opened O_DIRECT|O_SYNC then we'd better get it right. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4: Truncating file opens should also sync O_DIRECT writesTrond Myklebust2-2/+3
We don't just want to sync out buffered writes, but also O_DIRECT ones. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFS: File unlock needs to be a metadata synchronisation pointTrond Myklebust1-1/+1
File unlock needs to update both data and metadata on the NFS server in order to act as a synchronisation point for other clients. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFS: Add a helper to sync both O_DIRECT and buffered writesTrond Myklebust1-6/+9
Then apply it to nfs_setattr() and nfs_getattr(). Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1/pnfs: Refactor pnfs_set_layoutcommit()Trond Myklebust4-42/+14
pnfs_set_layoutcommit() and pnfs_commit_set_layoutcommit() are 100% identical except for the function arguments. Refactor to eliminate the difference. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1/pnfs: Fix setting of layoutcommit last write byteTrond Myklebust1-9/+8
If the NFS_INO_LAYOUTCOMMIT flag was unset, then we _must_ ensure that we also reset the last write byte (lwb) for that layout. The current code depends on us clearing the lwb when we clear NFS_INO_LAYOUTCOMMIT, which is not the case when we call pnfs_clear_layoutcommit(). Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4: Return the delegation before returning the layout in evict_inode()Trond Myklebust1-2/+3
Minor optimisation for the case where the layout has return-on-close enabled. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4: Allow tracing of NFSv4 fsync callsTrond Myklebust2-0/+8
I appear to have missed this when adding the ftrace probes. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFS: Fix free_deveiceid -> free_deviceidTrond Myklebust2-4/+4
Make it easier to grep for these functions by name. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1: Don't cache deviceids that have no notificationsTrond Myklebust3-0/+13
The spec says that once all layouts that reference a given deviceid have been returned, then we are only allowed to continue to cache the deviceid if the metadata server supports notifications. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1: Allow getdeviceinfo to return notification info back to callerTrond Myklebust2-9/+10
We are only allowed to cache deviceinfo if the server supports notifications and actually promises to call us back when changes occur. Right now, we request those notifications, but then we don't check the server's reply. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1: Cleanup - don't opencode nfs4_put_deviceid_node()Trond Myklebust1-4/+2
There really is no reason to do so. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4.1: Convert pNFS deviceid to use kfree_rcu()Trond Myklebust6-8/+7
Use of synchronize_rcu() when unmounting and potentially freeing a lot of deviceids is problematic. There really is no reason why we can't just use kfree_rcu() here. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-27NFSv4: Return delegations synchronously in evict_inodeTrond Myklebust1-1/+1
Kinglong Mee reports that asynchronous delegations are being killed by the call to rpc_shutdown_client() when unmounting. This can lead to state leakage on the server until the client lease expires. Reported-by: Kinglong Mee <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2015-03-25fs: move struct kiocb to fs.hChristoph Hellwig1-1/+0
struct kiocb now is a generic I/O container, so move it to fs.h. Also do a #include diet for aio.h while we're at it. Signed-off-by: Christoph Hellwig <[email protected]> Signed-off-by: Al Viro <[email protected]>
2015-03-13fs: split generic and aio kiocbChristoph Hellwig1-1/+1
Most callers in the kernel want to perform synchronous file I/O, but still have to bloat the stack with a full struct kiocb. Split out the parts needed in filesystem code from those in the aio code, and only allocate those needed to pass down argument on the stack. The aio code embedds the generic iocb in the one it allocates and can easily get back to it by using container_of. Also add a ->ki_complete method to struct kiocb, this is used to call into the aio code and thus removes the dependency on aio for filesystems impementing asynchronous operations. It will also allow other callers to substitute their own completion callback. We also add a new ->ki_flags field to work around the nasty layering violation recently introduced in commit 5e33f6 ("usb: gadget: ffs: add eventfd notification about ffs events"). Signed-off-by: Christoph Hellwig <[email protected]> Signed-off-by: Al Viro <[email protected]>
2015-03-13nfs: clean up nfs_direct_IOPeng Tao1-7/+0
This follows up "nfs: fix dio deadlock when O_DIRECT flag is flipped" and removes the unnecessary CONFIG_NFS_SWAP switch. Signed-off-by: Peng Tao <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2015-03-12fs: remove ki_nbytesChristoph Hellwig1-1/+1
There is no need to pass the total request length in the kiocb, as we already get passed in through the iov_iter argument. Signed-off-by: Christoph Hellwig <[email protected]> Signed-off-by: Al Viro <[email protected]>
2015-03-12NFSv4: Append delegations to the per-client list instead of prependingTrond Myklebust1-1/+1
Do so on the assumption that for most use cases, that list will turn into a more or less LRU-ordered list, and so the list traversals in nfs_client_return_marked_delegations() are likely to be shorter before hitting a candidate to return. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-12NFS: remount with security change should return EINVALBenjamin Coddington1-2/+1
A remount that alters security flavors can appear to succeed when it should instead return -EINVAL. Check to see if the current security flavor exists within the flavors specified in the remount options, and if not fail the remount. Signed-off-by: Benjamin Coddington <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2015-03-12nfs: do not export discarded symbolsArnd Bergmann1-1/+1
The function nfs4_pnfs_v3_ds_connect_unload is exported so it can be used by other modules, but is also marked '__exit' and will be discarded when built into the kernel, as pointed out by this linker error: `nfs4_pnfs_v3_ds_connect_unload' referenced in section `___ksymtab_gpl+nfs4_pnfs_v3_ds_connect_unload' of fs/built-in.o: defined in discarded section `.exit.text' of fs/built-in.o This removes the __exit annotation to make it safe to call this function. Signed-off-by: Arnd Bergmann <[email protected]> Fixes: 5f01d9539496 ("nfs41: create NFSv3 DS connection if specified") Signed-off-by: Trond Myklebust <[email protected]>
2015-03-12NFSv4.1: don't export static symbolJulia Lawall1-2/+0
The semantic patch that fixes this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @r@ type T; identifier f; @@ static T f (...) { ... } @@ identifier r.f; declarer name EXPORT_SYMBOL_GPL; @@ -EXPORT_SYMBOL_GPL(f); // </smpl> Signed-off-by: Julia Lawall <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2015-03-03NFSv4.1: Clear the old state by our client id before establishing a new leaseTrond Myklebust3-5/+17
If the call to exchange-id returns with the EXCHGID4_FLAG_CONFIRMED_R flag set, then that means our lease was established by a previous mount instance. Ensure that we detect this situation, and that we clear the state held by that mount. Reported-by: Jorge Mora <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2015-03-03NFSv4: Fix a race in NFSv4.1 server trunking discoveryTrond Myklebust3-8/+17
We do not want to allow a race with another NFS mount to cause nfs41_walk_client_list() to establish a lease on our nfs_client before we're done checking for trunking. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-03NFS: Don't write enable new pages while an invalidation is proceedingTrond Myklebust2-0/+4
nfs_vm_page_mkwrite() should wait until the page cache invalidation is finished. This is the second patch in a 2 patch series to deprecate the NFS client's reliance on nfs_release_page() in the context of nfs_invalidate_mapping(). Signed-off-by: Trond Myklebust <[email protected]>
2015-03-03NFS: Fix a regression in the read() syscallTrond Myklebust2-5/+36
When invalidating the page cache for a regular file, we want to first sync all dirty data to disk and then call invalidate_inode_pages2(). The latter relies on nfs_launder_page() and nfs_release_page() to deal respectively with dirty pages, and unstable written pages. When commit 9590544694bec ("NFS: avoid deadlocks with loop-back mounted NFS filesystems.") changed the behaviour of nfs_release_page(), then it made it possible for invalidate_inode_pages2() to fail with an EBUSY. Unfortunately, that error is then propagated back to read(). Let's therefore work around the problem for now by protecting the call to sync the data and invalidate_inode_pages2() so that they are atomic w.r.t. the addition of new writes. Later on, we can revisit whether or not we still need nfs_launder_page() and nfs_release_page(). Signed-off-by: Trond Myklebust <[email protected]>
2015-03-02NFSv4: Ensure we skip delegations that are already being returnedTrond Myklebust1-0/+6
In nfs_client_return_marked_delegations() and nfs_delegation_reap_unclaimed() we want to optimise the loop traversal by skipping delegations that are already in the process of being returned. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-02NFSv4: Pin the superblock while we're returning the delegationTrond Myklebust1-4/+16
This patch ensures that the superblock doesn't go ahead and disappear underneath us while the state manager thread is returning delegations. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-02NFSv4: Ensure we honour NFS_DELEGATION_RETURNING in nfs_inode_set_delegation()Trond Myklebust1-1/+4
Ensure that nfs_inode_set_delegation() doesn't inadvertently detach a delegation that is already in the process of being returned. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-02NFSv4: Ensure that we don't reap a delegation that is being returnedTrond Myklebust1-5/+7
Signed-off-by: Trond Myklebust <[email protected]>
2015-03-02NFS: Fix stateid used for NFS v4 closesAnna Schumaker1-2/+2
After 566fcec60 the client uses the "current stateid" from the nfs4_state structure to close a file. This could potentially contain a delegation stateid, which is disallowed by the protocol and causes servers to return NFS4ERR_BAD_STATEID. This patch restores the (correct) behavior of sending the open stateid to close a file. Reported-by: Olga Kornievskaia <[email protected]> Fixes: 566fcec60 (NFSv4: Fix an atomicity problem in CLOSE) Signed-off-by: Anna Schumaker <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2015-03-01NFSv4: Don't call put_rpccred() under the rcu_read_lock()Trond Myklebust1-1/+1
put_rpccred() can sleep. Fixes: 8f649c3762547 ("NFSv4: Fix the locking in nfs_inode_reclaim_delegation()") Cc: [email protected] # 2.6.35+ Signed-off-by: Trond Myklebust <[email protected]>
2015-03-01NFS: Don't require a filehandle to refresh the inode in nfs_prime_dcache()Trond Myklebust1-3/+13
If the server does not return a valid set of attributes that we can use to either create a file or refresh the inode, then there is no value in calling nfs_prime_dcache(). However if we're just refreshing the inode using the attributes that the server returned, then it shouldn't matter whether or not we have a filehandle, as long as we check the fsid+fileid combination. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-01NFSv3: Use the readdir fileid as the mounted-on-fileidTrond Myklebust1-0/+5
When we call readdirplus, set the fileid normally returned by readdir as the mounted-on-fileid, since that is commonly the case if there is a mountpoint. To ensure that we get it right, we only set the flag if the readdir fileid differs from the one returned in the readdirplus attributes. This again means that we can avoid the issues described in commit 2ef47eb1aee17 ("NFS: Fix use of nfs_attr_use_mounted_on_fileid()"), which only fixed NFSv4. Signed-off-by: Trond Myklebust <[email protected]>
2015-03-01NFS: Don't invalidate a submounted dentry in nfs_prime_dcache()Trond Myklebust1-0/+6
If we're traversing a directory which contains a submounted filesystem, or one that has a referral, the NFS server that is processing the READDIR request will often return information for the underlying (mounted-on) directory. It may, or may not, also return filehandle information. If this happens, and the lookup in nfs_prime_dcache() returns the dentry for the submounted directory, the filehandle comparison will fail, and we call d_invalidate(). Post-commit 8ed936b5671bf ("vfs: Lazily remove mounts on unlinked files and directories."), this means the entire subtree is unmounted. The following minimal patch addresses this problem by punting on the invalidation if there is a submount. Kudos to Neil Brown <[email protected]> for having tracked down this issue (see link). Reported-by: Nix <[email protected]> Link: http://lkml.kernel.org/r/[email protected] Cc: [email protected] # 3.18+ Signed-off-by: Trond Myklebust <[email protected]>
2015-03-01NFSv4: Set a barrier in the update_changeattr() helperTrond Myklebust2-0/+2
Ensure that we don't regress the changes that were made to the directory. Signed-off-by: Trond Myklebust <[email protected]> Tested-by: Chuck Lever <[email protected]>
2015-03-01NFS: Fix nfs_post_op_update_inode() to set an attribute barrierTrond Myklebust1-0/+1
nfs_post_op_update_inode() is called after a self-induced attribute update. Ensure that it also sets the barrier. Signed-off-by: Trond Myklebust <[email protected]> Tested-by: Chuck Lever <[email protected]>
2015-03-01NFS: Remove size hack in nfs_inode_attrs_need_update()Trond Myklebust1-8/+0
Prior to this patch, we used to always OK attribute updates that extended the file size on the assumption that we might be performing writeback. Now that we have attribute barriers to protect the writeback related updates, we should remove this hack, as it can cause truncate() operations to apparently be reverted if/when a readahead or getattr RPC call races with our on-the-wire SETATTR. Signed-off-by: Trond Myklebust <[email protected]> Tested-by: Chuck Lever <[email protected]>
2015-03-01NFSv4: Add attribute update barriers to delegreturn and pNFS layoutcommitTrond Myklebust1-0/+1
Ensure that other operations that race with delegreturn and layoutcommit cannot revert the attribute updates that were made on the server. Signed-off-by: Trond Myklebust <[email protected]> Tested-by: Chuck Lever <[email protected]>
2015-03-01NFS: Add attribute update barriers to NFS writebacksTrond Myklebust6-8/+56
Ensure that other operations that race with our write RPC calls cannot revert the file size updates that were made on the server. Signed-off-by: Trond Myklebust <[email protected]> Tested-by: Chuck Lever <[email protected]>
2015-03-01NFS: Set an attribute barrier on all updatesTrond Myklebust1-0/+4
Ensure that we update the attribute barrier even if there were no invalidations, provided that this value is newer than the old one. Signed-off-by: Trond Myklebust <[email protected]> Tested-by: Chuck Lever <[email protected]>
2015-03-01NFS: Add attribute update barriers to nfs_setattr_update_inode()Trond Myklebust4-10/+17
Ensure that other operations which raced with our setattr RPC call cannot revert the file attribute changes that were made on the server. To do so, we artificially bump the attribute generation counter on the inode so that all calls to nfs_fattr_init() that precede ours will be dropped. The motivation for the patch came from Chuck Lever's reports of readaheads racing with truncate operations and causing the file size to be reverted. Reported-by: Chuck Lever <[email protected]> Signed-off-by: Trond Myklebust <[email protected]> Tested-by: Chuck Lever <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.