aboutsummaryrefslogtreecommitdiff
path: root/fs/nfs/write.c
AgeCommit message (Collapse)AuthorFilesLines
2011-03-24NFSv4.1 convert layoutcommit sync to booleanAndy Adamson1-3/+5
Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-23NFSv4.1: layoutcommitAndy Adamson1-1/+14
The filelayout driver sends LAYOUTCOMMIT only when COMMIT goes to the data server (as opposed to the MDS) and the data server WRITE is not NFS_FILE_SYNC. Only whole file layout support means that there is only one IOMODE_RW layout segment. Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Alexandros Batsakis <[email protected]> Signed-off-by: Boaz Harrosh <[email protected]> Signed-off-by: Dean Hildebrand <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Mingyang Guo <[email protected]> Signed-off-by: Tao Guo <[email protected]> Signed-off-by: Zhang Jingwang <[email protected]> Tested-by: Boaz Harrosh <[email protected]> Signed-off-by: Benny Halevy <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-23NFSv4.1: filelayout driver specific code for COMMITFred Isaman1-7/+14
Implement all the hooks created in the previous patches. This requires exporting quite a few functions and adding a few structure fields. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-23NFSv4.1: remove GETATTR from ds commitsFred Isaman1-2/+5
Any COMMIT compound directed to a data server needs to have the GETATTR calls suppressed. We here, make sure the field we are testing (data->lseg) is set and refcounted correctly. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-23NFSv4.1: add generic layer hooks for pnfs COMMITFred Isaman1-15/+26
We create three major hooks for the pnfs code. pnfs_mark_request_commit() is called during writeback_done from nfs_mark_request_commit, which gives the driver an opportunity to claim it wants control over commiting a particular req. pnfs_choose_commit_list() is called from nfs_scan_list to choose which list a given req should be added to, based on where we intend to send it for COMMIT. It is up to the driver to have preallocated list headers for each destination it may need. pnfs_commit_list() is how the driver actually takes control, it is used instead of nfs_commit_list(). In order to pass information between the above functions, we create a union in nfs_page to hold a lseg (which is possible because the req is not on any list while in transition), and add some flags to indicate if we need to use the pnfs code. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-23NFSv4.1: pull out code from nfs_commit_releaseFred Isaman1-3/+9
Create a separate support function for later use by data server commit code. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-23NFSv4.1: pull error handling out of nfs_commit_listFred Isaman1-10/+16
Create a separate support function for later use by data server commit code. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-23NFSv4.1: rearrange nfs_commit_rpcsetupFred Isaman1-27/+33
Reorder nfs_commit_rpcsetup, preparing for a pnfs entry point. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-23NFSv4.1: don't send COMMIT to ds for data sync writesFred Isaman1-1/+4
Based on consensus reached in Feb 2011 interim IETF meeting regarding use of LAYOUTCOMMIT, it has been decided that a NFS_DATA_SYNC return from a WRITE to data server should not initiate a COMMIT. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-21NFS: Fix a hang/infinite loop in nfs_wb_page()Trond Myklebust1-12/+19
When one of the two waits in nfs_commit_inode() is interrupted, it returns a non-negative value, which causes nfs_wb_page() to think that the operation was successful causing it to busy-loop rather than exiting. It also causes nfs_file_fsync() to incorrectly report the file as being successfully committed to disk. This patch fixes both problems by ensuring that we return an error if the attempts to wait fail. Signed-off-by: Trond Myklebust <[email protected]> Cc: [email protected]
2011-03-21FS: Use stable writes when not doing a bulk flushTrond Myklebust1-3/+13
If we're only doing a single write, and there are no other unstable writes being queued up, we might want to just flip to using a stable write RPC call. Reviewed-by: NeilBrown <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: Clear lseg pointer in ->doio functionFred Isaman1-0/+2
Now that we have access to the pointer, clear it immediately after the put, instead of in caller. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: rearrange ->doio argsFred Isaman1-12/+16
This will make it possible to clear the lseg pointer in the same function as it is put, instead of in the caller nfs_pageio_doio(). Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: pnfs filelayout driver writeFred Isaman1-1/+4
Allows the pnfs filelayout driver to write to the data servers. Note that COMMIT to data servers will be implemented in a future patch. To avoid improper behavior, for the moment any WRITE to a data server that would also require a COMMIT to the data server is sent NFS_FILE_SYNC. Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Dean Hildebrand <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Mingyang Guo <[email protected]> Signed-off-by: Oleg Drokin <[email protected]> Signed-off-by: Ricardo Labiaga <[email protected]> Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Benny Halevy <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: implement generic pnfs layer write switchAndy Adamson1-0/+4
Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Boaz Harrosh <[email protected]> Signed-off-by: Dean Hildebrand <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: J. Bruce Fields <[email protected]> Signed-off-by: Mike Sager <[email protected]> Signed-off-by: Ricardo Labiaga <[email protected]> Signed-off-by: Tao Guo <[email protected]> Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Benny Halevy <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: trigger LAYOUTGET for writesFred Isaman1-12/+20
Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: Send lseg down into nfs_write_rpcsetupFred Isaman1-2/+5
We grab the lseg sent in from the doio function and attach it to each struct nfs_write_data created. This is how the lseg will be sent to the layout driver. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: rearrange nfs_write_rpcsetupAndy Adamson1-36/+46
Reorder nfs_write_rpcsetup, preparing for a pnfs entry point. Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: shift pnfs_update_layout locationsFred Isaman1-2/+2
Move the pnfs_update_layout call location to nfs_pageio_do_add_request(). Grab the lseg sent in the doio function to nfs_read_rpcsetup and attach it to each nfs_read_data so it can be sent to the layout driver. Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Dean Hildebrand <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Benny Halevy <[email protected]> Signed-off-by: Boaz Harrosh <[email protected]> Signed-off-by: Oleg Drokin <[email protected]> Signed-off-by: Tao Guo <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFSv4.1: coelesce across layout stripesFred Isaman1-0/+3
Add a pg_test layout driver hook which is used to avoid coelescing I/O across layout stripes. Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Andy Adamson <[email protected]> Signed-off-by: Dean Hildebrand <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Benny Halevy <[email protected]> Signed-off-by: Boaz Harrosh <[email protected]> Signed-off-by: Oleg Drokin <[email protected]> Signed-off-by: Tao Guo <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-11NFS: change nfs_writeback_done to return voidFred Isaman1-4/+4
The return values are not used by any callers. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-03-10nfs: close NFSv4 COMMIT vs. CLOSE raceJeff Layton1-0/+2
I've been adding in more artificial delays in the NFSv4 commit and close codepaths to uncover races. The kernel I'm testing has the patch to close the race in __rpc_wait_for_completion_task that's in Trond's cthon2011 branch. The reproducer I've been using does this in a loop: mkdir("DIR"); fd = open("DIR/FILE", O_WRONLY|O_CREAT|O_EXCL, 0644); write(fd, "abcdefg", 7); close(fd); unlink("DIR/FILE"); rmdir("DIR"); The above reproducer shouldn't result in any silly-renaming. However, when I add a "msleep(100)" just after the nfs_commit_clear_lock call in nfs_commit_release, I can almost always force one to occur. If I can force it to occur with that, then it can happen without that delay given the right timing. nfs_commit_inode waits for the NFS_INO_COMMIT bit to clear when called with FLUSH_SYNC set. nfs_commit_rpcsetup on the other hand does not wait for the task to complete before putting its reference to it, so the last reference get put in rpc_release task and gets queued to a workqueue. In this situation, the last open context reference may be put by the COMMIT release instead of the close() syscall. The close() syscall returns too quickly and the unlink runs while the d_count is still high since the COMMIT release hasn't put its dentry reference yet. Fix this by having rpc_commit_rpcsetup wait for the RPC call to complete before putting the task reference when FLUSH_SYNC is set. With this, the last reference is put by the process that's initiating the FLUSH_SYNC commit and the race is closed. Signed-off-by: Jeff Layton <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2011-01-19NFS: fix handling of malloc failure during nfs_flush_multi()Fred Isaman1-1/+1
Cleanup of the allocated list entries should not call put_nfs_open_context() on each entry, as the context will always be NULL, causing an oops. Signed-off-by: Fred Isaman <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2010-12-07nfs: remove extraneous and problematic calls to nfs_clear_requestTrond Myklebust1-1/+2
When a nfs_page is freed, nfs_free_request is called which also calls nfs_clear_request to clean out the lock and open contexts and free the pagecache page. However, a couple of places in the nfs code call nfs_clear_request themselves. What happens here if the refcount on the request is still high? We'll be releasing contexts and freeing pointers while the request is possibly still in use. Remove those bare calls to nfs_clear_context. That should only be done when the request is being freed. Note that when doing this, we need to watch out for tests of req->wb_page. Previously, nfs_set_page_tag_locked() and nfs_clear_page_tag_locked() would check the value of req->wb_page to figure out if the page is mapped into the nfsi->nfs_page_tree. We now indicate the page is mapped using the new bit PG_MAPPED in req->wb_flags . Reported-by: Jeff Layton <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2010-10-26writeback: remove nonblocking/encountered_congestion referencesWu Fengguang1-3/+1
This removes more dead code that was somehow missed by commit 0d99519efef (writeback: remove unused nonblocking and congestion checks). There are no behavior change except for the removal of two entries from one of the ext4 tracing interface. The nonblocking checks in ->writepages are no longer used because the flusher now prefer to block on get_request_wait() than to skip inodes on IO congestion. The latter will lead to more seeky IO. The nonblocking checks in ->writepage are no longer used because it's redundant with the WB_SYNC_NONE check. We no long set ->nonblocking in VM page out and page migration, because a) it's effectively redundant with WB_SYNC_NONE in current code b) it's old semantic of "Don't get stuck on request queues" is mis-behavior: that would skip some dirty inodes on congestion and page out others, which is unfair in terms of LRU age. Inspired by Christoph Hellwig. Thanks! Signed-off-by: Wu Fengguang <[email protected]> Cc: Theodore Ts'o <[email protected]> Cc: David Howells <[email protected]> Cc: Sage Weil <[email protected]> Cc: Steve French <[email protected]> Cc: Chris Mason <[email protected]> Cc: Jens Axboe <[email protected]> Cc: Christoph Hellwig <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2010-09-29NFS: don't use FLUSH_SYNC on WB_SYNC_NONE COMMIT calls (try #2)Jeff Layton1-7/+9
WB_SYNC_NONE is supposed to mean "don't wait on anything". That should also include not waiting for COMMIT calls to complete. WB_SYNC_NONE is also implied when wbc->nonblocking and wbc->for_background are set, so we can replace those checks in nfs_commit_unstable_pages with a check for WB_SYNC_NONE. Signed-off-by: Jeff Layton <[email protected]> Reviewed-by: Wu Fengguang <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2010-09-24NFSv4.1: keep seq_res.sr_slot as pointer rather than an indexBenny Halevy1-2/+0
Having to explicitly initialize sr_slotid to NFS4_MAX_SLOT_TABLE resulted in numerous bugs. Keeping the current slot as a pointer to the slot table is more straight forward and robust as it's implicitly set up to NULL wherever the seq_res member is initialized to zeroes. Signed-off-by: Benny Halevy <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2010-08-07Merge branch 'nfs-for-2.6.36' of ↵Linus Torvalds1-3/+6
git://git.linux-nfs.org/projects/trondmy/nfs-2.6 * 'nfs-for-2.6.36' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6: (42 commits) NFS: NFSv4.1 is no longer a "developer only" feature NFS: NFS_V4 is no longer an EXPERIMENTAL feature NFS: Fix /proc/mount for legacy binary interface NFS: Fix the locking in nfs4_callback_getattr SUNRPC: Defer deleting the security context until gss_do_free_ctx() SUNRPC: prevent task_cleanup running on freed xprt SUNRPC: Reduce asynchronous RPC task stack usage SUNRPC: Move the bound cred to struct rpc_rqst SUNRPC: Clean up of rpc_bindcred() SUNRPC: Move remaining RPC client related task initialisation into clnt.c SUNRPC: Ensure that rpc_exit() always wakes up a sleeping task SUNRPC: Make the credential cache hashtable size configurable SUNRPC: Store the hashtable size in struct rpc_cred_cache NFS: Ensure the AUTH_UNIX credcache is allocated dynamically NFS: Fix the NFS users of rpc_restart_call() SUNRPC: The function rpc_restart_call() should return success/failure NFSv4: Get rid of the bogus RPC_ASSASSINATED(task) checks NFSv4: Clean up the process of renewing the NFSv4 lease NFSv4.1: Handle NFS4ERR_DELAY on SEQUENCE correctly NFS: nfs_rename() should not have to flush out writebacks ...
2010-08-01NFS: Fix a typo in include/linux/nfs_fs.hTrond Myklebust1-5/+0
nfs_commit_inode() needs to be defined irrespectively of whether or not we are supporting NFSv3 and NFSv4. Allow the compiler to optimise away code in the NFSv2-only case by converting it into an inlined stub function. Reported-and-tested-by: Ingo Molnar <[email protected]> Signed-off-by: Trond Myklebust <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2010-07-30NFS: Ensure that writepage respects the nonblock flagTrond Myklebust1-6/+17
Signed-off-by: Trond Myklebust <[email protected]>
2010-07-30NFS: kswapd must not block in nfs_release_pageTrond Myklebust1-2/+2
See https://bugzilla.kernel.org/show_bug.cgi?id=16056 If other processes are blocked waiting for kswapd to free up some memory so that they can make progress, then we cannot allow kswapd to block on those processes. Signed-off-by: Trond Myklebust <[email protected]> Cc: [email protected]
2010-07-30NFSv4: Ensure that we track the NFSv4 lock state in read/write requests.Trond Myklebust1-1/+4
This patch fixes bugzilla entry 14501: https://bugzilla.kernel.org/show_bug.cgi?id=14501 Signed-off-by: Trond Myklebust <[email protected]>
2010-06-22NFSv4.1: Make nfs4_setup_sequence take a nfs_server argumentTrond Myklebust1-2/+2
In anticipation of the day when we have per-filesystem sessions, and also in order to allow the session to change in the event of a filesystem migration event. Signed-off-by: Trond Myklebust <[email protected]>
2010-05-26NFS: Fix another nfs_wb_page() deadlockTrond Myklebust1-2/+5
J.R. Okajima reports that the call to sync_inode() in nfs_wb_page() can deadlock with other writeback flush calls. It boils down to the fact that we cannot ever call writeback_single_inode() while holding a page lock (even if we do set nr_to_write to zero) since another process may already be waiting in the call to do_writepages(), and so will deny us the I_SYNC lock. Signed-off-by: Trond Myklebust <[email protected]>
2010-05-26NFS: Ensure that we mark the inode as dirty if we exit early from commitTrond Myklebust1-2/+11
If we exit from nfs_commit_inode() without ensuring that the COMMIT rpc call has been completed, we must re-mark the inode as dirty. Otherwise, future calls to sync_inode() with the WB_SYNC_ALL flag set will fail to ensure that the data is on the disk. Signed-off-by: Trond Myklebust <[email protected]>
2010-04-27NFS: Ensure that nfs_wb_page() waits for Pg_writeback to clearTrond Myklebust1-15/+4
Neil Brown reports that he is seeing the BUG_ON(ret == 0) trigger in nfs_page_async_flush. According to the trace in https://bugzilla.novell.com/show_bug.cgi?id=599628 the problem appears to be due to nfs_wb_page() not waiting for the PG_writeback flag to clear. There is a ditto problem in nfs_wb_page_cancel() Signed-off-by: Trond Myklebust <[email protected]>
2010-04-22NFS: Fix an unstable write data integrity raceTrond Myklebust1-4/+32
Commit 2c61be0a9478258f77b66208a0c4b1f5f8161c3c (NFS: Ensure that the WRITE and COMMIT RPC calls are always uninterruptible) exposed a race on file close. In order to ensure correct close-to-open behaviour, we want to wait for all outstanding background commit operations to complete. This patch adds an inode flag that indicates if a commit operation is under way, and provides a mechanism to allow ->write_inode() to wait for its completion if this is a data integrity flush. Signed-off-by: Trond Myklebust <[email protected]>
2010-04-09NFS: Ensure that the WRITE and COMMIT RPC calls are always uninterruptibleTrond Myklebust1-7/+16
We always want to ensure that WRITE and COMMIT completes, whether or not the user presses ^C. Do this by making the call asynchronous, and allowing the user to do an interruptible wait for rpc_task completion. Signed-off-by: Trond Myklebust <[email protected]>
2010-04-09NFS: Fix a race with the new commit codeTrond Myklebust1-7/+10
This patch fixes a race which occurs due to the fact that we release the PG_writeback flag while still holding the nfs_page locked. Signed-off-by: Trond Myklebust <[email protected]>
2010-04-09NFS: Ensure that writeback_single_inode() calls write_inode() when syncingTrond Myklebust1-2/+2
Since writeback_single_inode() checks the inode->i_state flags _before_ it flushes out the data, we need to ensure that the I_DIRTY_DATASYNC flag is already set. Otherwise we risk not seeing a call to write_inode(), which again means that we break fsync() et al... Signed-off-by: Trond Myklebust <[email protected]>
2010-03-05NFS: Simplify nfs_wb_page()Trond Myklebust1-97/+23
Signed-off-by: Trond Myklebust <[email protected]>
2010-03-05NFS: Replace __nfs_write_mapping with sync_inode()Trond Myklebust1-37/+5
Now that we have correct COMMIT semantics in writeback_single_inode, we can reduce and simplify nfs_wb_all(). Also replace nfs_wb_nocommit() with a call to filemap_write_and_wait(), which doesn't need to hold the inode->i_mutex. With that done, we can eliminate nfs_write_mapping() altogether. Signed-off-by: Trond Myklebust <[email protected]>
2010-03-05NFS: Simplify nfs_wb_page_cancel()Trond Myklebust1-38/+1
In all cases we should be able to just remove the request and call cancel_dirty_page(). Signed-off-by: Trond Myklebust <[email protected]>
2010-03-05NFS: Ensure inode is always marked I_DIRTY_DATASYNC, if it has unstable pagesTrond Myklebust1-0/+2
Since nfs_scan_list() doesn't wait for locked pages, we have a race in which it is possible to end up with an inode that needs to send a COMMIT, but which does not have the I_DIRTY_DATASYNC flag set. Signed-off-by: Trond Myklebust <[email protected]>
2010-03-05NFS: Run COMMIT as an asynchronous RPC call when wbc->for_background is setTrond Myklebust1-1/+1
Signed-off-by: Trond Myklebust <[email protected]> Acked-by: Peter Zijlstra <[email protected]> Acked-by: Wu Fengguang <[email protected]>
2010-03-05NFS: Reduce the number of unnecessary COMMIT callsTrond Myklebust1-4/+22
If the caller is doing a non-blocking flush, and there are still writebacks pending on the wire, we can usually defer the COMMIT call until those writes are done. Also ensure that we honour the wbc->nonblocking flag. Signed-off-by: Trond Myklebust <[email protected]>
2010-03-05NFS: Add a count of the number of unstable writes carried by an inodeTrond Myklebust1-4/+10
In order to know when we should do opportunistic commits of the unstable writes, when the VM is doing a background flush, we add a field to count the number of unstable writes. Signed-off-by: Trond Myklebust <[email protected]>
2010-03-05NFS: Cleanup - move nfs_write_inode() into fs/nfs/write.cTrond Myklebust1-1/+23
The sole purpose of nfs_write_inode is to commit unstable writes, so move it into fs/nfs/write.c, and make nfs_commit_inode static. Signed-off-by: Trond Myklebust <[email protected]>
2010-02-09NFS: Remove a redundant check for PageFsCache in nfs_migrate_page()Trond Myklebust1-2/+1
Signed-off-by: Trond Myklebust <[email protected]> Acked-by: David Howells <[email protected]>
2010-01-26NFS: Make nfs_commitdata_release staticH Hartley Sweeten1-1/+1
The symbol nfs_commitdata_release is only used locally in this file. Make it static to prevent the following sparse warning: warning: symbol 'nfs_commitdata_release' was not declared. Should it be static? Signed-off-by: H Hartley Sweeten <[email protected]> Cc: Trond Myklebust <[email protected]> Signed-off-by: Trond Myklebust <[email protected]> Reviewed-by: Chuck Lever <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.