- 07 Apr, 2009 40 commits
-
-
Ryusuke Konishi authored
Nilfs creates checkpoints even for garbage collection or metadata updates such as checkpoint mode change. So, user often sees checkpoints created only by such internal operations. This is inconvenient in some situations. For example, application that monitors checkpoints and changes them to snapshots, will fall into an infinite loop because it cannot distinguish internally created checkpoints. This patch solves this sort of problem by adding a flag to checkpoint for identification. Signed-off-by:
Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Ryusuke Konishi authored
The sketch file is a file to mark checkpoints with user data. It was experimentally introduced in the original implementation, and now obsolete. The file was handled differently with regular files; the file size got truncated when a checkpoint was created. This stops the special treatment and will treat it as a regular file. Most users are not affected because mkfs.nilfs2 no longer makes this file. Signed-off-by:
-
Ryusuke Konishi authored
This adds a missing endian conversion of checksum field in the super block. This fixes compatibility issue on big endian machines which will come to surface after supporting recovery of super block. Signed-off-by:
-
Ryusuke Konishi authored
Pekka Enberg advised me: > It would be nice if BUG(), BUG_ON(), and panic() calls would be > converted to proper error handling using WARN_ON() calls. The BUG() > call in nilfs_cpfile_delete_checkpoints(), for example, looks to be > triggerable from user-space via the ioctl() system call. This will follow the comment and keep them to a minimum. Acked-by:
Pekka Enberg <penberg@cs.helsinki.fi> Signed-off-by:
-
Ryusuke Konishi authored
This adds a new argument to the nilfs_sustat structure. The extended field allows to delete volatile active state of segments, which was needed to protect freshly-created segments from garbage collection but has confused code dealing with segments. This extension alleviates the mess and gives room for further simplifications. The volatile active flag is not persistent, so it's eliminable on this occasion without affecting compatibility other than the ioctl change. Signed-off-by:
-
Ryusuke Konishi authored
Pekka Enberg suggested converting ->ioctl operations to use ->unlocked_ioctl to avoid BKL. The conversion was verified to be safe, so I will take it on this occasion. Cc: Pekka Enberg <penberg@cs.helsinki.fi> Signed-off-by:
-
Ryusuke Konishi authored
This removes compat code from the nilfs ioctls and applies the same function for both .ioctl and .compat_ioctl file operations. Signed-off-by:
-
Ryusuke Konishi authored
Nilfs ioctl had structures not having fixed sized types such as: struct nilfs_argv { void *v_base; size_t v_nmembs; size_t v_size; int v_index; int v_flags; }; Further, some of them are wrongly aligned: e.g. struct nilfs_cpmode { __u64 cm_cno; int cm_mode; }; The size of wrongly aligned structures varies depending on architectures, and it breaks the identity of ioctl commands, which leads to arch dependent errors. Previously, these are compensated by using compat_ioctl. This fixes these problems and allows removal of compat ioctl. Since this will change sizes of those structures, binary compatibility for the past utilities will once break; new utilities have to be used instead. However, it would be helpful to avoid platform dependent problems in the long term. Signed-off-by: -
Ryusuke Konishi authored
This removes NILFS_IOCTL_TIMEDWAIT command from ioctl interface along with the related flags and wait queue. The command is terrible because it just sleeps in the ioctl. I prefer to avoid this by devising means of event polling in userland program. By reconsidering the userland GC daemon, I found this is possible without changing behaviour of the daemon and sacrificing efficiency. Signed-off-by:
-
Ryusuke Konishi authored
This will fix the weird behavior of lscp command in listing continuously created checkpoints; the output of lscp is rewinded regularly for the recent nilfs. As a result of debugging, a defect was found in nilfs_cpfile_do_get_cpinfo() function. Though the function can be repeatedly called to enumerate checkpoints and it can skip invalid checkpoint entries, the index value was not carried between successive calls. The bug has long been present, and came to surface after applying a bugfix nilfs2-fix-problems-of-memory-allocation-in-ioctl.patch, which increased frequency of calling the function. The similar bugfix was already applied for ``snapshots'' by nilfs2-fix-gc-failure-on-volumes-keeping-numerous-snapshots.patch. This fixes the problem by making the index argument bidirectional on the function. Signed-off-by:
-
Pekka Enberg authored
This cleans up the strange indirect function calling convention used in nilfs to follow the normal kernel coding style. Signed-off-by:
-
Ryusuke Konishi authored
A few tool developers gave me requests for fixing inconvenient return value of nilfs_get_cpinfo() ioctl; if the requested mode is NILFS_SNAPSHOT and the specified start entry is not a snapshot, the ioctl unnaturally returns one as the number of acquired snapshot item. In addition, the ioctl function returns an ENOENT error for checkpoints within blocks deleted by garbage collection. These behaviors require corrections for programs which enumerate snapshots. This resolves the inconvenience by changing the return values to zero for the above cases. Signed-off-by:
-
Ryusuke Konishi authored
This resolves the following failure of nilfs2 cleaner daemon: nilfs_cleanerd[20670]: cannot clean segments: No such file or directory nilfs_cleanerd[20670]: shutdown When creating thousands of snapshots, the cleaner daemon had rarely died as above due to an error returned from the kernel code. After applying the recent patch which fixed memory allocation problems in ioctl (Message-Id: <20081215.155840.105124170.ryusuke@osrg.net>), the problem gets more frequent. It turned out to be a bug of nilfs_ioctl_wrap_copy function and one of its callback routines to read out information of snapshots; if the nilfs_ioctl_wrap_copy function divided a large read request into multiple requests, the second and later requests have failed since a restart position on snapshot meta data was not properly set forward. It's a deficiency of the callback interface that cannot pass the restart position among multiple requests. This patch fixes the issue by allowing nilfs_ioctl_wrap_copy and snapshot read functions to exchange a position argument. Signed-off-by:
-
Ryusuke Konishi authored
Signed-off-by:
-
Ryusuke Konishi authored
The file gcinode.c gives buffer cache functions for on-disk blocks moved in garbage collection. Joern Engel has suggested inserting its explanations in the source file (Message-ID: <20080917144146.GD8750@logfs.org> and <20080917224953.GB14644@logfs.org>). This follows the comment. Cc: Joern Engel <joern@logfs.org> Signed-off-by:
-
Ryusuke Konishi authored
Pekka Enberg pointed out that double error handlings found after nilfs_transaction_end() can be avoided by separating abort operation: OK, I don't understand this. The only way nilfs_transaction_end() can fail is if we have NILFS_TI_SYNC set and we fail to construct the segment. But why do we want to construct a segment if we don't commit? I guess what I'm asking is why don't we have a separate nilfs_transaction_abort() function that can't fail for the erroneous case to avoid this double error value tracking thing? This does the separation and renames nilfs_transaction_end() to nilfs_transaction_commit() for clarification. Since, some calls of these functions were used just for exclusion control against the segment constructor, they are replaced with semaphore operations. Acked-by:
-
Ryusuke Konishi authored
This will remove the following unnecessary locks and cleanup code in nilfs_clear_inode(): - unnecessary protection using nilfs_transaction_begin() and nilfs_transaction_end(). - cleanup code of i_dirty list field which is never chained when this function is called. - spinlock used when releasing i_bh field. Signed-off-by:
-
Ryusuke Konishi authored
This is another patch for fixing the following problems of a memory copy function in nilfs2 ioctl: (1) It tries to allocate 128KB size of memory even for small objects. (2) Though the function repeatedly tries large memory allocations while reducing the size, GFP_NOWAIT flag is not specified. This increases the possibility of system memory shortage. (3) During the retries of (2), verbose warnings are printed because _GFP_NOWARN flag is not used for the kmalloc calls. The first patch was still doing large allocations by kmalloc which are repeatedly tried while reducing the size. Andi Kleen told me that using copy_from_user for large memory is not good from the viewpoint of preempt latency: On Fri, 12 Dec 2008 21:24:11 +0100, Andi Kleen <andi@firstfloor.org> wrote: > > In the current interface, each data item is copied twice: one is to > > the allocated memory from user space (via copy_from_user), and another > > For such large copies it is better to use multiple smaller (e.g. 4K) > copy user, that gives better real time preempt latencies. Each cfu has a > cond_resched(), but only one, not multiple times in the inner loop. He also advised me that: On Sun, 14 Dec 2008 16:13:27 +0100, Andi Kleen <andi@firstfloor.org> wrote: > Better would be if you could go to PAGE_SIZE. order 0 allocations > are typically the fastest / least likely to stall. > > Also in this case it's a good idea to use __get_free_pages() > directly, kmalloc tends to be become less efficient at larger > sizes. For the function in question, the size of buffer memory can be reduced since the buffer is repeatedly used for a number of small objects. On the other hand, it may incur large preempt latencies for larger buffer because a copy_from_user (and a copy_to_user) was applied only once each cycle. With that, this revision uses the order 0 allocations with __get_free_pages() to fix the original problems. Cc: Andi Kleen <andi@firstfloor.org> Signed-off-by: -
Ryusuke Konishi authored
This adds a Makefile for the nilfs2 file system, and updates the makefile and Kconfig file in the file system directory. Signed-off-by:
-
Koji Sato authored
This adds userland interface implemented with ioctl. Signed-off-by:
Koji Sato <sato.koji@lab.ntt.co.jp> Signed-off-by:
-
Ryusuke Konishi authored
This adds the cache of on-disk blocks to be moved in garbage collection. The disk blocks are held with dummy inodes (called gcinodes), and this file provides lookup function of the dummy inodes, and their buffer read function. Signed-off-by:
Seiji Kihara <kihara.seiji@lab.ntt.co.jp> Signed-off-by:
Yoshiji Amagai <amagai.yoshiji@lab.ntt.co.jp> Signed-off-by:
-
Ryusuke Konishi authored
NILFS2 uses another DAT inode during garbage collection to ensure atomicity and consistency of the DAT in the transient state. This twin inode is called GCDAT. This adds functions to initialize the GCDAT and to switch page caches and B-tree node caches between these two inodes. Signed-off-by:
-
Ryusuke Konishi authored
This adds recovery function on mount. Usually the recovery is achieved by just finding the latest super root. When logs without checkpoints were appended for data sync operations after the latest super root, the recovery function will perform roll forwarding and reconstruct new log(s) with a super root. Signed-off-by:
-
Ryusuke Konishi authored
Chris Mason pointed out that there is a missed sync issue in nilfs_writepages(): On Wed, 17 Dec 2008 21:52:55 -0500, Chris Mason wrote: > It looks like nilfs_writepage ignores WB_SYNC_NONE, which is used by > do_sync_mapping_range(). where WB_SYNC_NONE in do_sync_mapping_range() was replaced with WB_SYNC_ALL by Nick's patch (commit: ee53a891). This fixes the problem by letting nilfs_writepages() write out the log of file data within the range if sync_mode is WB_SYNC_ALL. This involves removal of nilfs_file_aio_write() which was previously needed to ensure O_SYNC sync writes. Cc: Chris Mason <chris.mason@oracle.com> Signed-off-by:
-
Ryusuke Konishi authored
This adds the segment constructor (also called log writer). The segment constructor collects dirty buffers for every dirty inode, makes summaries of the buffers, assigns disk block addresses to the buffers, and then submits BIOs for the buffers. Signed-off-by:
-
Ryusuke Konishi authored
This adds the segment buffer which is used to constuct logs. [akpm@linux-foundation.org: BIO_RW_SYNC got removed] Signed-off-by:
-
Ryusuke Konishi authored
This adds super block operations for the nilfs2 file system. Signed-off-by:
-
Ryusuke Konishi authored
This adds functions on the_nilfs object, which keeps shared resources and states among a read/write mount and snapshots mounts going individually. the_nilfs is allocated per block device; it is created when user first mount a snapshot or a read/write mount on the device, then it is reused for successive mounts. It will be freed when all mount instances on the device are detached. Signed-off-by:
-
Ryusuke Konishi authored
This adds pathname operations, most of which comes from the ext2 file system. Signed-off-by:
-
Yoshiji Amagai authored
This adds directory handling functions, most of which comes from the ext2 file system. Signed-off-by:
-
Ryusuke Konishi authored
This adds primitives for regular file handling. Signed-off-by:
-
Ryusuke Konishi authored
This adds inode level operations of the nilfs2 file system. Signed-off-by:
-
Koji Sato authored
This adds a meta data file which stores the allocation state of segments. [konishi.ryusuke@lab.ntt.co.jp: fix wrong counting of checkpoints and dirty segments] Signed-off-by:
-
Koji Sato authored
This adds a meta data file which holds checkpoint entries in its data blocks. Signed-off-by:
-
Ryusuke Konishi authored
This adds a meta data file which stores on-disk inodes in its data blocks. Signed-off-by:
-
Koji Sato authored
This adds the disk address translation file (DAT) whose primary function is to convert virtual disk block numbers to actual disk block numbers. The virtual block numbers of NILFS are associated with checkpoint generation numbers, and this file also provides functions to manage the lifetime information of each virtual block number. Signed-off-by:
-
Ryusuke Konishi authored
This adds common functions to allocate or deallocate entries with bitmaps on a meta data file. This feature is used by the DAT and ifile. Signed-off-by:
-
Ryusuke Konishi authored
This adds the meta data file, which serves common buffer functions to the DAT, sufile, cpfile, ifile, and so forth. Signed-off-by:
-
Ryusuke Konishi authored
This adds common routines for buffer/page operations used in B-tree node caches, meta data files, or segment constructor (log writer). NILFS uses copy functions for buffers and pages due to the following reasons: 1) Relocation required for COW Since NILFS changes address of on-disk blocks, moving buffers in page cache is needed for the buffers which are not addressed by a file offset. If buffer size is smaller than page size, this involves partial copy of pages. 2) Freezing mmapped pages NILFS calculates checksums for each log to ensure its validity. If page data changes after the checksum calculation, this validity check will not work correctly. To avoid this failure for mmaped pages, NILFS freezes their data by copying. 3) Copy-on-write for DAT pages NILFS makes clones of DAT page caches in a copy-on-write manner during GC processes, and this ensures atomicity and consistency of the DAT in the transient state. In addition, NILFS uses two obsolete functions, nilfs_mark_buffer_dirty() and nilfs_clear_page_dirty() respectively. * nilfs_mark_buffer_dirty() was required to avoid NULL pointer dereference faults: Since the page cache of B-tree node pages or data page cache of pseudo inodes does not have a valid mapping->host, calling mark_buffer_dirty() for their buffers causes the fault; it calls __mark_inode_dirty(NULL) through __set_page_dirty(). * nilfs_clear_page_dirty() was needed in the two cases: 1) For B-tree node pages and data pages of the dat/gcdat, NILFS2 clears page dirty flags when it copies back pages from the cloned cache (gcdat->{i_mapping,i_btnode_cache}) to its original cache (dat->{i_mapping,i_btnode_cache}). 2) Some B-tree operations like insertion or deletion may dispose buffers in dirty state, and this needs to cancel the dirty state of their pages. clear_page_dirty_for_io() caused faults because it does not clear the dirty tag on the page cache. Signed-off-by: -
Ryusuke Konishi authored
This adds routines for B-tree node buffers. Signed-off-by:
-
