- 30 Oct, 2008 40 commits
-
-
David Chinner authored
Normally dquots are written back via delayed write mechanisms. They are flushed to their backing buffer by xfssyncd, which is then pushed out by either AIL or xfsbufd flushing. The flush from the xfssyncd is supposed to be non-blocking, but xfs_qm_dqflush() always waits for pinned duots, which means that it will block for the length of time it takes to do a synchronous log force. This causes unnecessary extra log I/O to be issued whenever we try to flush a busy dquot. Avoid the log forces and blocking xfssyncd by making xfs_qm_dqflush() pay attention to what type of sync it is doing when it sees a pinned dquot and not waiting when doing non-blocking flushes. SGI-PV: 988147 SGI-Modid: xfs-linux-melb:xfs-kern:32287a Signed-off-by:
David Chinner <david@fromorbit.com> Signed-off-by:
Peter Leckie <pleckie@sgi.com> Signed-off-by:
Lachlan McIlroy <lachlan@sgi.com>
-
David Chinner authored
xfs_iflush_all() walks the m_inodes list to find inodes that need reclaiming. We already have such a list - the m_del_inodes list. Replace xfs_iflush_all() with a call to xfs_finish_reclaim_all() and clean up xfs_finish_reclaim_all() to handle the different flush modes now needed. Originally based on a patch from Christoph Hellwig. Version 3 o rediff against new linux-2.6/xfs_sync.c code Version 2 o revert xfs_syncsub() inode reclaim behaviour back to original code o xfs_quiesce_fs() should use XFS_IFLUSH_DELWRI_ELSE_ASYNC, not XFS_IFLUSH_ASYNC, to prevent change of behaviour. SGI-PV: 988139 SGI-Modid: xfs-linux-melb:xfs-kern:32284a Signed-off-by:
Christoph Hellwig <hch@infradead.org>
-
David Chinner authored
Move all the xfssyncd code to the new xfs_sync.c file. This places it closer to the actual code that it interacts with, rather than just being associated with high level VFS code. SGI-PV: 988139 SGI-Modid: xfs-linux-melb:xfs-kern:32283a Signed-off-by:
-
David Chinner authored
The sync code in XFS is spread around several files. While it used to make sense to have such a distribution, the code is about to be cleaned up and so centralising it in one spot as the first step makes sense. SGI-PV: 988139 SGI-Modid: xfs-linux-melb:xfs-kern:32282a Signed-off-by:
-
Barry Naujok authored
SGI-PV: 987246 SGI-Modid: xfs-linux-melb:xfs-kern:32233a Signed-off-by:
Barry Naujok <bnaujok@sgi.com> Signed-off-by:
Eric Sandeen <sandeen@sandeen.net> Signed-off-by:
-
Barry Naujok authored
SGI-PV: 987683 SGI-Modid: xfs-linux-melb:xfs-kern:32232a Signed-off-by:
-
Barry Naujok authored
SGI-PV: 986558 SGI-Modid: xfs-linux-melb:xfs-kern:32231a Signed-off-by:
-
Lachlan McIlroy authored
SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32224a Signed-off-by:
-
Peter Leckie authored
This is a fix for patch " Clean up dquot pincount code" which introduced a build breakage due to a missing & in xfs_qm_dquot_logitem_pin. SGI-PV: 986789 SGI-Modid: xfs-linux-melb:xfs-kern:32221a Signed-off-by:
Donald Douwsma <donaldd@sgi.com> Signed-off-by:
-
Peter Leckie authored
This is a code cleanup and optimization that removes a per mount point spinlock from the quota code and cleans up the code. The patch changes the pincount from being an int protected by a spinlock to an atomic_t allowing the pincount to be manipulated without holding the spinlock. This cleanup also protects against random wakup's of both the aild and xfssyncd by reevaluating the pincount after been woken. Two latter patches will address the Spurious wakeups. SGI-PV: 986789 SGI-Modid: xfs-linux-melb:xfs-kern:32215a Signed-off-by:
-
Lachlan McIlroy authored
It's possible to have outstanding xfs_ioend_t's queued when the file size is zero. This can happen in the direct I/O path when a direct I/O write fails due to ENOSPC. In this case the xfs_ioend_t will still be queued (ie xfs_end_io_direct() does not know that the I/O failed so can't force the xfs_ioend_t to be flushed synchronously). When we truncate a file on unlink we don't know to wait for these xfs_ioend_ts and we can have a use-after-free situation if the inode is reclaimed before the xfs_ioend_t is finally processed. As was suggested by Dave Chinner lets wait for all I/Os to complete when truncating the file size to zero. SGI-PV: 981668 SGI-Modid: xfs-linux-melb:xfs-kern:32216a Signed-off-by:
-
Christoph Hellwig authored
Make the existing bmap btree tracing generic so that it applies to all btree types. Some fragments lifted from a patch by Dave Chinner. This adds two files that were missed from the previous btree tracing checkin. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32210a Signed-off-by:
Bill O'Donnell <billodo@sgi.com> Signed-off-by:
-
Christoph Hellwig authored
Lots of functionality in xfs_btree.c isn't needed by callers outside of this file anymore, so mark these functions static. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32209a Signed-off-by:
-
Christoph Hellwig authored
Add methods to check whether two keys/records are in the righ order. This replaces the xfs_btree_check_key and xfs_btree_check_rec methods. For the callers from xfs_bmap.c just opencode the bmbt-specific asserts. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32208a Signed-off-by:
-
Christoph Hellwig authored
These are equivalent to the xfs_btree_* versions, and the only remaining caller can be switched to the generic one after they are exported. Also remove some now dead infrastructure in xfs_bmap_btree.c. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32207a Signed-off-by:
-
Christoph Hellwig authored
Not really much reason to make it generic given that it's so small, but this is the last non-method in xfs_alloc_btree.c and xfs_ialloc_btree.c, so it makes the whole btree implementation more structured. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32206a Signed-off-by:
-
Christoph Hellwig authored
Make the btree delete code generic. Based on a patch from David Chinner with lots of changes to follow the original btree implementations more closely. While this loses some of the generic helper routines for inserting/moving/removing records it also solves some of the one off bugs in the original code and makes it easier to verify. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32205a Signed-off-by:
-
Christoph Hellwig authored
xfs_bmbt_killroot is a mostly generic implementation of moving from a real block based root to an inode based root. So move it to xfs_btree.c where it can use all the nice infrastructure there and make it pointer size agnostic The new name for it is xfs_btree_kill_iroot, following the old naming but making it clear we're dealing with the root in inode case here, and to avoid confusion with xfs_btree_new_root which is used for the not inode rooted case. I've also added a comment describing what it does and why it's named the way it is. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32203a Signed-off-by:
-
Christoph Hellwig authored
Make the btree insert code generic. Based on a patch from David Chinner with lots of changes to follow the original btree implementations more closely. While this loses some of the generic helper routines for inserting/moving/removing records it also solves some of the one off bugs in the original code and makes it easier to verify. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32202a Signed-off-by:
-
Christoph Hellwig authored
xfs_bmbt_newroot is a mostly generic implementation of moving from an inode root to a real block based root. So move it to xfs_btree.c where it can use all the nice infrastructure there and make it pointer size agnostic The new name for it is xfs_btree_new_iroot, following the old naming but making it clear we're dealing with the root in inode case here, and to avoid confusion with xfs_btree_new_root which is used for the not inode rooted case. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32201a Signed-off-by:
-
Christoph Hellwig authored
From: Dave Chinner <dgc@sgi.com> Add a xfs_btree_new_root helper for the alloc and ialloc btrees. The bmap btree needs it's own version and is not converted. [hch: split out from bigger patch and minor adaptions] SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32200a Signed-off-by:
-
Christoph Hellwig authored
Make the btree split code generic. Based on a patch from David Chinner with lots of changes to follow the original btree implementations more closely. While this loses some of the generic helper routines for inserting/moving/removing records it also solves some of the one off bugs in the original code and makes it easier to verify. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32198a Signed-off-by:
-
Christoph Hellwig authored
Make the btree left shift code generic. Based on a patch from David Chinner with lots of changes to follow the original btree implementations more closely. While this loses some of the generic helper routines for inserting/moving/removing records it also solves some of the one off bugs in the original code and makes it easier to verify. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32197a Signed-off-by:
-
Christoph Hellwig authored
Make the btree right shift code generic. Based on a patch from David Chinner with lots of changes to follow the original btree implementations more closely. While this loses some of the generic helper routines for inserting/moving/removing records it also solves some of the one off bugs in the original code and makes it easier to verify. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32196a Signed-off-by:
-
Christoph Hellwig authored
From: Dave Chinner <dgc@sgi.com> The most complicated part here is the lastrec tracking for the alloc btree. Most logic is in the update_lastrec method which has to do some hopefully good enough dirty magic to maintain it. [hch: split out from bigger patch and a rework of the lastrec logic] SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32194a Signed-off-by:
-
Christoph Hellwig authored
From: Dave Chinner <dgc@sgi.com> Note that there are many > 80 char lines introduced due to the xfs_btree_key casts. But the places where this happens is throw-away code once the whole btree code gets merged into a common implementation. The same is true for the temporary xfs_alloc_log_keys define to the new name. All old users will be gone after a few patches. [hch: split out from bigger patch and minor adaptions] SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32193a Signed-off-by:
-
Christoph Hellwig authored
From: Dave Chinner <dgc@sgi.com> [hch: split out from bigger patch and minor adaptions] SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32192a Signed-off-by:
-
Christoph Hellwig authored
From: Dave Chinner <dgc@sgi.com> [hch: split out from bigger patch and minor adaptions] SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32191a Signed-off-by:
-
Christoph Hellwig authored
From: Dave Chinner <dgc@sgi.com> Because this is the first major generic btree routine this patch includes some infrastrucure, first a few routines to deal with a btree block that can be either in short or long form, second xfs_btree_read_buf_block, which is the new central routine to read a btree block given a cursor, and third the new xfs_btree_ptr_addr routine to calculate the address for a given btree pointer record. [hch: split out from bigger patch and minor adaptions] SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32190a Signed-off-by:
-
Christoph Hellwig authored
Add new helpers in xfs_btree.c to find the record, key and block pointer entries inside a btree block. To implement this genericly the ->get_maxrecs methods and two new xfs_btree_ops entries for the key and record sizes are used. Also add a big comment describing how the addressing inside a btree block works. Note that these helpers are unused until users are introduced in the next patches and this patch will thus cause some harmless compiler warnings. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32189a Signed-off-by:
-
Christoph Hellwig authored
Factor xfs_btree_maxrecs into a per-btree operation. The get_maxrecs method is based on a patch from Dave Chinner. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32188a Signed-off-by:
-
Christoph Hellwig authored
Make the existing bmap btree tracing generic so that it applies to all btree types. Some fragments lifted from a patch by Dave Chinner. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32187a Signed-off-by:
-
David Chinner authored
From: Dave Chinner <dgc@sgi.com> Introduce statistics coverage of all the btrees and cover all the btree operations, not just some. Invaluable for determining test code coverage of all the btree operations.... SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32184a Signed-off-by:
-
Christoph Hellwig authored
Move the various btree validation helpers around in xfs_btree.c so that they are close to each other and in common #ifdef DEBUG sections. Also add a new xfs_btree_check_ptr helper to check a btree ptr that can be either long or short form. Split out from a bigger patch from Dave Chinner with various small changes applied by me. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32183a Signed-off-by:
-
Christoph Hellwig authored
From: Dave Chinner <dgc@sgi.com> Refactor xfs_btree_readahead to make it more readable: (a) remove the inline xfs_btree_readahead wrapper and move all checks out of line into the main routine. (b) factor out helpers for short/long form btrees (c) move check for root in inodes from the callers into xfs_btree_readahead [hch: split out from a big patch and minor cleanups] SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32182a Signed-off-by:
-
Christoph Hellwig authored
Add a flag to the xfs btree cursor when using long (64bit) block pointers instead of checking btnum == XFS_BTNUM_BMAP. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32181a Signed-off-by:
-
Christoph Hellwig authored
The bmap btree is rooted in the inode and not in a disk block. Make the support for this feature more generic by adding a btree flag to for this feature instead of relying on the XFS_BTNUM_BMAP btnum check. Also clean up xfs_btree_get_block where this new flag is used. Based upon a patch from Dave Chinner. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32180a Signed-off-by:
-
Christoph Hellwig authored
Add generic union types for btree pointers, keys and records. The generic btree pointer contains either a 32 and 64bit big endian scalar for short and long form btrees, and the key and record contain the relevant type for each possible btree. Split out from a bigger patch from Dave Chinner and simplified a little further. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32178a Signed-off-by:
-
Christoph Hellwig authored
xfs_btree_init_cursor contains close to little shared code for the different btrees and will get even more non-common code in the future. Split it up into one routine per btree type. Because xfs_btree_dup_cursor needs to call the init routine for a generic btree cursor add a new btree operation vector that contains a dup_cursor method that initializes a new cursor based on an existing one. The btree operations vector is based on an idea and code from Dave Chinner and will grow more entries later during this series. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32176a Signed-off-by:
-
Christoph Hellwig authored
This type is only embedded in struct xfs_btree_block and never used directly. By moving the fields directly into struct xfs_btree_block a lot of the macros for struct xfs_btree_sblock and struct xfs_btree_lblock can be used for struct xfs_btree_block too now which helps greatly with some of the migrations during implementing the generic btree code. SGI-PV: 985583 SGI-Modid: xfs-linux-melb:xfs-kern:32174a Signed-off-by:
-
