summaryrefslogtreecommitdiffstats
path: root/fs
AgeCommit message (Collapse)Author
2008-09-26[XFS] Remove xfs_iext_irec_compact_full()Lachlan McIlroy
Yet another bug was found in xfs_iext_irec_compact_full() and while the source of the bug was found it wasn't an easy task to track it down because the conditions are very difficult to reproduce. A HUGE thank-you goes to Russell Cattelan and Eric Sandeen for their significant effort in tracking down the source of this corruption. xfs_iext_irec_compact_full() and xfs_iext_irec_compact_pages() are almost identical - they both compact indirect extent lists by moving extents from subsequent buffers into earlier ones. xfs_iext_irec_compact_pages() only moves extents if all of the extents in the next buffer will fit into the empty space in the buffer before it. xfs_iext_irec_compact_full() will go a step further and move part of the next buffer if all the extents wont fit. It will then shift the remaining extents in the next buffer up to the start of the buffer. The bug here was that we did not update er_extoff and this caused extent list corruption. It does not appear that this extra functionality gains us much. Calling xfs_iext_irec_compact_pages() instead will do a good enough job at compacting the indirect list and will be quicker too. For the case in xfs_iext_indirect_to_direct() the total number of extents in the indirect list will fit into one buffer so we will never need the extra functionality of xfs_iext_irec_compact_full() there. Also xfs_iext_irec_compact_pages() doesn't need to do a memmove() (the buffers will never overlap) so we don't want the performance hit that can incur. SGI-PV: 987159 SGI-Modid: xfs-linux-melb:xfs-kern:32166a Signed-off-by: Lachlan McIlroy <lachlan@sgi.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
2008-09-26[XFS] Fix extent list corruption in xfs_iext_irec_compact_full().Lachlan McIlroy
If we don't move all the records from the next buffer into the current buffer then we need to update the er_extoff field of the next buffer as we shift the remaining records to the start of the buffer. SGI-PV: 987159 SGI-Modid: xfs-linux-melb:xfs-kern:32165a Signed-off-by: Lachlan McIlroy <lachlan@sgi.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net> Signed-off-by: Russell Cattelan <cattelan@thebarn.com>
2008-09-249p: use an IS_ERR test rather than a NULL testJulien Brunel
In case of error, the function p9_client_walk returns an ERR pointer, but never returns a NULL pointer. So a NULL test that comes after an IS_ERR test should be deleted. The semantic match that finds this problem is as follows: (http://www.emn.fr/x-info/coccinelle/) // <smpl> @match_bad_null_test@ expression x, E; statement S1,S2; @@ x = p9_client_walk(...) ... when != x = E * if (x != NULL) S1 else S2 // </smpl> Signed-off-by: Julien Brunel <brunel@diku.dk> Signed-off-by: Julia Lawall <julia@diku.dk> Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2008-09-24cifs: explicitly revoke SPNEGO key after session setupJeff Layton
cifs: explicitly revoke SPNEGO key after session setup The SPNEGO blob returned by an upcall can only be used once. Explicitly revoke it to make sure that we never pick it up again after session setup exits. This doesn't seem to be that big an issue on more recent kernels, but older kernels seem to link keys into the session keyring by default. That said, explicitly revoking the key seems like a reasonable thing to do here. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-24cifs: Convert cifs to new aops.Nick Piggin
cifs: Convert cifs to new aops. This patch is based on the one originally posted by Nick Piggin. His patch was very close, but had a couple of small bugs. Nick's original comments follow: This is another relatively naive conversion. Always do the read upfront when the page is not uptodate (unless we're in the writethrough path). Fix an uninitialized data exposure where SetPageUptodate was called before the page was uptodate. SetPageUptodate and switch to writeback mode in the case that the full page was dirtied. Acked-by: Shaggy <shaggy@austin.ibm.com> Acked-by: Badari Pulavarty <pbadari@us.ibm.com> Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-24[CIFS] update DOS attributes in cifsInode if we successfully changed themSteve French
Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-24cifs: remove NULL termination from rename target in CIFSSMBRenameOpenFIleJeff Layton
cifs: remove NULL termination from rename target in CIFSSMBRenameOpenFIle The rename destination isn't supposed to be null terminated. Also, change the name string arg to be const. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-24cifs: work around samba returning -ENOENT on SetFileDisposition callJeff Layton
cifs: work around samba returning -ENOENT on SetFileDisposition call Samba seems to return STATUS_OBJECT_NAME_NOT_FOUND when we try to set the delete on close bit after doing a rename by filehandle. This looks like a samba bug to me, but a lot of servers will do this. For now, pretend an -ENOENT return is a success. Samba does however seem to respect the CREATE_DELETE_ON_CLOSE bit when opening files that already exist. Windows will ignore it, but so adding it to the open flags should be harmless. We're also currently ignoring the return code on the rename by filehandle, so no need to set rc based on it. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-24cifs: fix inverted NULL check after kmallocJeff Layton
cifs: fix inverted NULL check after kmalloc Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-23ext4: Combine proc file handling into a single set of functionsTheodore Ts'o
Previously mballoc created a separate set of functions for each proc file. This combines the tunables into a single set of functions which gets used for all of the per-superblock proc files, saving approximately 2k of compiled object code. Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-09-23[CIFS] clean up upcall handling for dns_resolver keysSteve French
We're given the datalen in the downcall, so there's no need to do any calls to strlen(). Just keep track of the datalen in the key. Finally, add a sanity check of the data in the downcall to make sure that it looks like a real IP address. Signed-off-by: Jeff Layton <jlayton@redhat.com> Acked-by: David Howells <dhowells@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-23[CIFS] fix busy-file renames and refactor cifs_rename logicSteve French
Break out the code that does the actual renaming into a separate function and have cifs_rename call that. That function will attempt a path based rename first and then do a filehandle based one if it looks like the source is busy. The existing logic tried a path based rename first, but if we needed to remove the destination then it only attempted a filehandle based rename afterward. Not all servers support renaming by filehandle, so we need to always attempt path rename first and fall back to filehandle rename if it doesn't work. This also fixes renames of open files on windows servers (at least when the source and destination directories are the same). CC: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-23cifs: add function to set file dispositionJeff Layton
cifs: add function to set file disposition The proper way to set the delete on close bit on an already existing file is to use SET_FILE_INFO with an infolevel of SMB_FILE_DISPOSITION_INFO. Add a function to do that and have the silly-rename code use it. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-23[CIFS] add constants for string lengths of keynames in SPNEGO upcall stringSteve French
Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-23cifs: move rename and delete-on-close logic into helper functionJeff Layton
cifs: move rename and delete-on-close logic into helper function When a file is still open on the server, we attempt to set the DELETE_ON_CLOSE bit and rename it to a new filename. When the last opener closes the file, the server should delete it. This patch moves this mechanism into a helper function and has the two places in cifs_unlink that do this procedure call it. It also fixes the open flags to be correct. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-23ext4: move /proc setup and teardown out of mballoc.cTheodore Ts'o
...and into the core setup/teardown code in fs/ext4/super.c so that other parts of ext4 can define tuning parameters. Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-09-23cifs: have find_writeable_file prefer filehandles opened by same taskJeff Layton
When the CIFS client goes to write out pages, it needs to pick a filehandle to write to. find_writeable_file however just picks the first filehandle that it finds. This can cause problems when a lock is issued against a particular filehandle and we pick a different filehandle to write to. This patch tries to avert this situation by having find_writable_file prefer filehandles that have a pid that matches the current task. This seems to fix lock test 11 from the connectathon test suite when run against a windows server. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-22cifs: don't use GFP_KERNEL with GFP_NOFSPekka Enberg
GFP_KERNEL and GFP_NOFS are mutually exclusive. If you combine them, you end up with plain GFP_KERNEL which can deadlock in cases where you really want GFP_NOFS. Cc: Steve French <sfrench@samba.org> Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-22ext4: Don't use 'struct dentry' for internal lookupsTheodore Ts'o
This is a port of a patch from Linus which fixes a 200+ byte stack usage problem in ext4_get_parent(). It's more efficient to pass down only the actual parts of the dentry that matter: the parent inode and the name, instead of allocating a struct dentry on the stack. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-10-06ext4/jbd2: Avoid WARN() messages when failing to write to the superblockTheodore Ts'o
This fixes some very common warnings reported by kerneloops.org Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-09-18GFS2: high time to take some time over atimeSteven Whitehouse
Until now, we've used the same scheme as GFS1 for atime. This has failed since atime is a per vfsmnt flag, not a per fs flag and as such the "noatime" flag was not getting passed down to the filesystems. This patch removes all the "special casing" around atime updates and we simply use the VFS's atime code. The net result is that GFS2 will now support all the same atime related mount options of any other filesystem on a per-vfsmnt basis. We do lose the "lazy atime" updates, but we gain "relatime". We could add lazy atime to the VFS at a later date, if there is a requirement for that variant still - I suspect relatime will be enough. Also we lose about 100 lines of code after this patch has been applied, and I have a suspicion that it will speed things up a bit, even when atime is "on". So it seems like a nice clean up as well. From a user perspective, everything stays the same except the loss of the per-fs atime quantum tweekable (ought to be per-vfsmnt at the very least, and to be honest I don't think anybody ever used it) and that a number of options which were ignored before now work correctly. Please let me know if you've got any comments. I'm pushing this out early so that you can all see what my plans are. Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2008-09-18GFS2: The war on bloatSteven Whitehouse
The following patch shrinks the gfs2_args structure which is embedded in every GFS2 superblock. It cuts down the size of the options to a single unsigned int (the 13 bits of bitfields will be rounded up to that size by the compiler) from the current 11 unsigned ints. So on x86 thats 44 bytes shrinking to 4 bytes, in each and every GFS2 superblock. Signed-off-by: Steven Whitehouse <swhitho@redhat.com>
2008-09-18UBIFS: fix printk format warningsAlexander Beregalov
fs/ubifs/dir.c:428: warning: format '%llu' expects type 'long long unsigned int', but argument 5 has type 'long unsigned int' fs/ubifs/debug.c:541: warning: format '%llu' expects type 'long long unsigned int', but argument 2 has type 'long unsigned int' Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
2008-09-17UBIFS: remove incorrect assertAdrian Hunter
The assert was not valid because one of the variables 'taken_empty_lebs' has transient values out of sync with the other variables. Signed-off-by: Adrian Hunter <ext-adrian.hunter@nokia.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
2008-09-17UBIFS: TNC / GC race fixesAdrian Hunter
- update GC sequence number if any nodes may have been moved even if GC did not finish the LEB - don't ignore error return when reading Signed-off-by: Adrian Hunter <ext-adrian.hunter@nokia.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
2008-09-17UBIFS: create the name of the background thread in every caseSebastian Siewior
If the ubifs partition is mounted RO and then remounted RW we end up with no thread name in ubifs_remount_rw() and the thread appears nameless. Signed-off-by: Sebastian Siewior <bigeasy@linutronix.de> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
2008-09-17[XFS] Don't do I/O beyond eof when unreserving spaceLachlan McIlroy
When unreserving space with boundaries that are not block aligned we round up the start and round down the end boundaries and then use this function, xfs_zero_remaining_bytes(), to zero the parts of the blocks that got dropped during the rounding. The problem is we don't consider if these blocks are beyond eof. Worse still is if we encounter delayed allocations beyond eof we will try to use the magic delayed allocation block number as a real block number. If the file size is ever extended to expose these blocks then we'll go through xfs_zero_eof() to zero them anyway. SGI-PV: 983683 SGI-Modid: xfs-linux-melb:xfs-kern:32055a Signed-off-by: Lachlan McIlroy <lachlan@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org>
2008-09-17[XFS] Fix use-after-free with buffersLachlan McIlroy
We have a use-after-free issue where log completions access buffers via the buffer log item and the buffer has already been freed. Fix this by taking a reference on the buffer when attaching the buffer log item and release the hold when the buffer log item is detached and we no longer need the buffer. Also create a new function xfs_buf_item_free() to combine some common code. SGI-PV: 985757 SGI-Modid: xfs-linux-melb:xfs-kern:32025a Signed-off-by: Lachlan McIlroy <lachlan@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org>
2008-09-17[XFS] Prevent lockdep false positives when locking two inodes.David Chinner
If we call xfs_lock_two_inodes() to grab both the iolock and the ilock, then drop the ilocks on both inodes, then grab them again (as xfs_swap_extents() does) then lockdep will report a locking order problem. This is a false positive. To avoid this, disallow xfs_lock_two_inodes() fom locking both inode locks at once - force calers to make two separate calls. This means that nested dropping and regaining of the ilocks will retain the same lockdep subclass and so lockdep will not see anything wrong with this code. SGI-PV: 986238 SGI-Modid: xfs-linux-melb:xfs-kern:31999a Signed-off-by: David Chinner <david@fromorbit.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Peter Leckie <pleckie@sgi.com> Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
2008-09-17[XFS] Fix barrier status change detection.David Chinner
The current code in xlog_iodone() uses the wrong macro to check if the barrier has been cleared due to an EOPNOTSUPP error form the lower layer. SGI-PV: 986143 SGI-Modid: xfs-linux-melb:xfs-kern:31984a Signed-off-by: David Chinner <david@fromorbit.com> Signed-off-by: Nathaniel W. Turner <nate@houseofnate.net> Signed-off-by: Peter Leckie <pleckie@sgi.com> Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
2008-09-17[XFS] Prevent direct I/O from mapping extents beyond eofLachlan McIlroy
With the help from some tracing I found that we try to map extents beyond eof when doing a direct I/O read. It appears that the way to inform the generic direct I/O path (ie do_direct_IO()) that we have breached eof is to return an unmapped buffer from xfs_get_blocks_direct(). This will cause do_direct_IO() to jump to the hole handling code where is will check for eof and then abort. This problem was found because a direct I/O read was trying to map beyond eof and was encountering delayed allocations. The delayed allocations beyond eof are speculative allocations and they didn't get converted when the direct I/O flushed the file because there was only enough space in the current AG to convert and write out the dirty pages within eof. Note that xfs_iomap_write_allocate() wont necessarily convert all the delayed allocation passed to it - it will return after allocating the first extent - so if the delayed allocation extends beyond eof then it will stay that way. SGI-PV: 983683 SGI-Modid: xfs-linux-melb:xfs-kern:31929a Signed-off-by: Lachlan McIlroy <lachlan@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org>
2008-09-17[XFS] Fix regression introduced by remount fixupChristoph Hellwig
Logically we would return an error in xfs_fs_remount code to prevent users from believing they might have changed mount options using remount which can't be changed. But unfortunately mount(8) adds all options from mtab and fstab to the mount arguments in some cases so we can't blindly reject options, but have to check for each specified option if it actually differs from the currently set option and only reject it if that's the case. Until that is implemented we return success for every remount request, and silently ignore all options that we can't actually change. SGI-PV: 985710 SGI-Modid: xfs-linux-melb:xfs-kern:31908a Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
2008-09-17[XFS] Move memory allocations for log tracing out of the critical pathLachlan McIlroy
Memory allocations for log->l_grant_trace and iclog->ic_trace are done on demand when the first event is logged. In xlog_state_get_iclog_space() we call xlog_trace_iclog() under a spinlock and allocating memory here can cause us to sleep with a spinlock held and deadlock the system. For the log grant tracing we use KM_NOSLEEP but that means we can lose trace entries. Since there is no locking to serialize the log grant tracing we could race and have multiple allocations and leak memory. So move the allocations to where we initialize the log/iclog structures. Use KM_NOFS to avoid recursing into the filesystem and drop log->l_trace since it's not even used. SGI-PV: 983738 SGI-Modid: xfs-linux-melb:xfs-kern:31896a Signed-off-by: Lachlan McIlroy <lachlan@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org>
2008-09-16[CIFS] use common code for turning off ATTR_READONLY in cifs_unlinkSteve French
We already have a cifs_set_file_info function that can flip DOS attribute bits. Have cifs_unlink call it to handle turning ATTR_HIDDEN on and ATTR_READONLY off when an unlink attempt returns -EACCES. This also removes a level of indentation from cifs_unlink. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-16cifs: clean up variables in cifs_unlinkJeff Layton
Change parameters to cifs_unlink to match the ones used in the generic VFS. Add some local variables to cut down on the amount of struct dereferencing that needs to be done, and eliminate some unneeded NULL pointer checks on the parent directory inode. Finally, rename pTcon to "tcon" to more closely match standard kernel coding style. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2008-09-15GFS2: GFS2 will panic if you misspell any mount optionsAbhijith Das
The gfs2 superblock pointer is NULL after a failed mount. When control eventually goes to gfs2_kill_sb, we dereference this NULL pointer. This patch ensures that the gfs2 superblock pointer is not NULL before being dereferenced in gfs2_kill_sb. Signed-off-by: Abhijith Das <adas@redhat.com> Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2008-09-15GFS2: Direct IO write at end of file errorBob Peterson
This patch fixes a problem whereby a direct_io write doesn't fall back to buffered write properly at end of file. Signed-off-by: Bob Peterson <rpeterso@redhat.com> Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2008-09-14timers: fix itimer/many thread hangFrank Mayhar
Overview This patch reworks the handling of POSIX CPU timers, including the ITIMER_PROF, ITIMER_VIRT timers and rlimit handling. It was put together with the help of Roland McGrath, the owner and original writer of this code. The problem we ran into, and the reason for this rework, has to do with using a profiling timer in a process with a large number of threads. It appears that the performance of the old implementation of run_posix_cpu_timers() was at least O(n*3) (where "n" is the number of threads in a process) or worse. Everything is fine with an increasing number of threads until the time taken for that routine to run becomes the same as or greater than the tick time, at which point things degrade rather quickly. This patch fixes bug 9906, "Weird hang with NPTL and SIGPROF." Code Changes This rework corrects the implementation of run_posix_cpu_timers() to make it run in constant time for a particular machine. (Performance may vary between one machine and another depending upon whether the kernel is built as single- or multiprocessor and, in the latter case, depending upon the number of running processors.) To do this, at each tick we now update fields in signal_struct as well as task_struct. The run_posix_cpu_timers() function uses those fields to make its decisions. We define a new structure, "task_cputime," to contain user, system and scheduler times and use these in appropriate places: struct task_cputime { cputime_t utime; cputime_t stime; unsigned long long sum_exec_runtime; }; This is included in the structure "thread_group_cputime," which is a new substructure of signal_struct and which varies for uniprocessor versus multiprocessor kernels. For uniprocessor kernels, it uses "task_cputime" as a simple substructure, while for multiprocessor kernels it is a pointer: struct thread_group_cputime { struct task_cputime totals; }; struct thread_group_cputime { struct task_cputime *totals; }; We also add a new task_cputime substructure directly to signal_struct, to cache the earliest expiration of process-wide timers, and task_cputime also replaces the it_*_expires fields of task_struct (used for earliest expiration of thread timers). The "thread_group_cputime" structure contains process-wide timers that are updated via account_user_time() and friends. In the non-SMP case the structure is a simple aggregator; unfortunately in the SMP case that simplicity was not achievable due to cache-line contention between CPUs (in one measured case performance was actually _worse_ on a 16-cpu system than the same test on a 4-cpu system, due to this contention). For SMP, the thread_group_cputime counters are maintained as a per-cpu structure allocated using alloc_percpu(). The timer functions update only the timer field in the structure corresponding to the running CPU, obtained using per_cpu_ptr(). We define a set of inline functions in sched.h that we use to maintain the thread_group_cputime structure and hide the differences between UP and SMP implementations from the rest of the kernel. The thread_group_cputime_init() function initializes the thread_group_cputime structure for the given task. The thread_group_cputime_alloc() is a no-op for UP; for SMP it calls the out-of-line function thread_group_cputime_alloc_smp() to allocate and fill in the per-cpu structures and fields. The thread_group_cputime_free() function, also a no-op for UP, in SMP frees the per-cpu structures. The thread_group_cputime_clone_thread() function (also a UP no-op) for SMP calls thread_group_cputime_alloc() if the per-cpu structures haven't yet been allocated. The thread_group_cputime() function fills the task_cputime structure it is passed with the contents of the thread_group_cputime fields; in UP it's that simple but in SMP it must also safely check that tsk->signal is non-NULL (if it is it just uses the appropriate fields of task_struct) and, if so, sums the per-cpu values for each online CPU. Finally, the three functions account_group_user_time(), account_group_system_time() and account_group_exec_runtime() are used by timer functions to update the respective fields of the thread_group_cputime structure. Non-SMP operation is trivial and will not be mentioned further. The per-cpu structure is always allocated when a task creates its first new thread, via a call to thread_group_cputime_clone_thread() from copy_signal(). It is freed at process exit via a call to thread_group_cputime_free() from cleanup_signal(). All functions that formerly summed utime/stime/sum_sched_runtime values from from all threads in the thread group now use thread_group_cputime() to snapshot the values in the thread_group_cputime structure or the values in the task structure itself if the per-cpu structure hasn't been allocated. Finally, the code in kernel/posix-cpu-timers.c has changed quite a bit. The run_posix_cpu_timers() function has been split into a fast path and a slow path; the former safely checks whether there are any expired thread timers and, if not, just returns, while the slow path does the heavy lifting. With the dedicated thread group fields, timers are no longer "rebalanced" and the process_timer_rebalance() function and related code has gone away. All summing loops are gone and all code that used them now uses the thread_group_cputime() inline. When process-wide timers are set, the new task_cputime structure in signal_struct is used to cache the earliest expiration; this is checked in the fast path. Performance The fix appears not to add significant overhead to existing operations. It generally performs the same as the current code except in two cases, one in which it performs slightly worse (Case 5 below) and one in which it performs very significantly better (Case 2 below). Overall it's a wash except in those two cases. I've since done somewhat more involved testing on a dual-core Opteron system. Case 1: With no itimer running, for a test with 100,000 threads, the fixed kernel took 1428.5 seconds, 513 seconds more than the unfixed system, all of which was spent in the system. There were twice as many voluntary context switches with the fix as without it. Case 2: With an itimer running at .01 second ticks and 4000 threads (the most an unmodified kernel can handle), the fixed kernel ran the test in eight percent of the time (5.8 seconds as opposed to 70 seconds) and had better tick accuracy (.012 seconds per tick as opposed to .023 seconds per tick). Case 3: A 4000-thread test with an initial timer tick of .01 second and an interval of 10,000 seconds (i.e. a timer that ticks only once) had very nearly the same performance in both cases: 6.3 seconds elapsed for the fixed kernel versus 5.5 seconds for the unfixed kernel. With fewer threads (eight in these tests), the Case 1 test ran in essentially the same time on both the modified and unmodified kernels (5.2 seconds versus 5.8 seconds). The Case 2 test ran in about the same time as well, 5.9 seconds versus 5.4 seconds but again with much better tick accuracy, .013 seconds per tick versus .025 seconds per tick for the unmodified kernel. Since the fix affected the rlimit code, I also tested soft and hard CPU limits. Case 4: With a hard CPU limit of 20 seconds and eight threads (and an itimer running), the modified kernel was very slightly favored in that while it killed the process in 19.997 seconds of CPU time (5.002 seconds of wall time), only .003 seconds of that was system time, the rest was user time. The unmodified kernel killed the process in 20.001 seconds of CPU (5.014 seconds of wall time) of which .016 seconds was system time. Really, though, the results were too close to call. The results were essentially the same with no itimer running. Case 5: With a soft limit of 20 seconds and a hard limit of 2000 seconds (where the hard limit would never be reached) and an itimer running, the modified kernel exhibited worse tick accuracy than the unmodified kernel: .050 seconds/tick versus .028 seconds/tick. Otherwise, performance was almost indistinguishable. With no itimer running this test exhibited virtually identical behavior and times in both cases. In times past I did some limited performance testing. those results are below. On a four-cpu Opteron system without this fix, a sixteen-thread test executed in 3569.991 seconds, of which user was 3568.435s and system was 1.556s. On the same system with the fix, user and elapsed time were about the same, but system time dropped to 0.007 seconds. Performance with eight, four and one thread were comparable. Interestingly, the timer ticks with the fix seemed more accurate: The sixteen-thread test with the fix received 149543 ticks for 0.024 seconds per tick, while the same test without the fix received 58720 for 0.061 seconds per tick. Both cases were configured for an interval of 0.01 seconds. Again, the other tests were comparable. Each thread in this test computed the primes up to 25,000,000. I also did a test with a large number of threads, 100,000 threads, which is impossible without the fix. In this case each thread computed the primes only up to 10,000 (to make the runtime manageable). System time dominated, at 1546.968 seconds out of a total 2176.906 seconds (giving a user time of 629.938s). It received 147651 ticks for 0.015 seconds per tick, still quite accurate. There is obviously no comparable test without the fix. Signed-off-by: Frank Mayhar <fmayhar@google.com> Cc: Roland McGrath <roland@redhat.com> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-09-13rescan_partitions(): make device capacity errors non-fatalAndrew Morton
Herton Krzesinski reports that the error-checking changes in 04ebd4aee52b06a2c38127d9208546e5b96f3a19 ("block/ioctl.c and fs/partition/check.c: check value returned by add_partition") cause his buggy USB camera to no longer mount. "The camera is an Olympus X-840. The original issue comes from the camera itself: its format program creates a partition with an off by one error". Buggy devices happen. It is better for the kernel to warn and to proceed with the mount. Reported-by: Herton Ronaldo Krzesinski <herton@mandriva.com.br> Cc: Abdel Benamrouche <draconux@gmail.com> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: Alan Stern <stern@rowland.harvard.edu> Cc: David Brownell <david-b@pacbell.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-09-13mm: ifdef Quicklists in /proc/meminfoHugh Dickins
A "Quicklists: 0 kB" line has just started appearing in /proc/meminfo, but most architectures (including x86) don't have them configured, so #ifdef it, like the highmem lines. And those architectures which do have quicklists configured are using them for page tables: so let's place it next to PageTables. Signed-off-by: Hugh Dickins <hugh@veritas.com> Acked-by: Christoph Lameter <cl@linux-foundation.org> Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-09-13bfs: fix Lockdep warningEric Sesterhenn
This fixes: ============================================= [ INFO: possible recursive locking detected ] 2.6.27-rc5-00283-g70bb089 #68 --------------------------------------------- touch/6855 is trying to acquire lock: (&info->bfs_lock){--..}, at: [<c02262f5>] bfs_delete_inode+0x9e/0x18c but task is already holding lock: (&info->bfs_lock){--..}, at: [<c0226c00>] bfs_create+0x45/0x187 other info that might help us debug this: 2 locks held by touch/6855: #0: (&type->i_mutex_dir_key#5){--..}, at: [<c018ad13>] do_filp_open+0x10b/0x62f #1: (&info->bfs_lock){--..}, at: [<c0226c00>] bfs_create+0x45/0x187 stack backtrace: Pid: 6855, comm: touch Not tainted 2.6.27-rc5-00283-g70bb089 #68 [<c013e769>] validate_chain+0x458/0x9f4 [<c013bece>] ? trace_hardirqs_off+0xb/0xd [<c013f36b>] __lock_acquire+0x666/0x6e0 [<c013f440>] lock_acquire+0x5b/0x77 [<c02262f5>] ? bfs_delete_inode+0x9e/0x18c [<c06aab74>] mutex_lock_nested+0xbc/0x234 [<c02262f5>] ? bfs_delete_inode+0x9e/0x18c [<c02262f5>] ? bfs_delete_inode+0x9e/0x18c [<c02262f5>] bfs_delete_inode+0x9e/0x18c [<c0226257>] ? bfs_delete_inode+0x0/0x18c [<c01925e1>] generic_delete_inode+0x94/0xfe [<c019265d>] generic_drop_inode+0x12/0x12f [<c0191b7e>] iput+0x4b/0x4e [<c0226d1e>] bfs_create+0x163/0x187 [<c0188b42>] vfs_create+0xa6/0x114 [<c018adb5>] do_filp_open+0x1ad/0x62f [<c0107cdc>] ? native_sched_clock+0x82/0x96 [<c06ac309>] ? _spin_unlock+0x27/0x3c [<c019379e>] ? alloc_fd+0xbf/0xc9 [<c06ae2f4>] ? sub_preempt_count+0x9d/0xab [<c019379e>] ? alloc_fd+0xbf/0xc9 [<c0180391>] do_sys_open+0x42/0xb8 [<c041d564>] ? trace_hardirqs_on_thunk+0xc/0x10 [<c0180449>] sys_open+0x1e/0x26 [<c01038bd>] sysenter_do_call+0x12/0x31 ======================= The problem is that we don't unlock the bfs->lock mutex before calling iput (we do in the other cases). Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de> Cc: Tigran Aivazian <tigran@aivazian.fsnet.co.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-09-13proc: more debugging for "already registered" caseAlexey Dobriyan
Print parent directory name as well. The aim is to catch non-creation of parent directory when proc_mkdir will return NULL and all subsequent registrations go directly in /proc instead of intended directory. Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> [ Fixed insane printk string while at it. - Linus ] Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-09-13ext4: use percpu data structures for lg_prealloc_listEric Sandeen
lg_prealloc_list seems to cry out for a per-cpu data structure; on a large smp system I think this should be better. I've lightly tested this change on a 4-cpu system. Signed-off-by: Eric Sandeen <sandeen@redhat.com> Acked-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-09-13ext4: Renumber EXT4_IOC_MIGRATETheodore Ts'o
Pick an ioctl number for EXT4_IOC_MIGRATE that won't conflict with other ext4 ioctl's. Since there haven't been any major userspace users of this ioctl, we can afford to change this now, to avoid potential problems later. Also, reorder the ioctl numbers in ext4.h to avoid this sort of mistake in the future. Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-10-08ext4: hook the ext3 migration interface to the EXT4_IOC_SETFLAGS ioctlAneesh Kumar K.V
This patch hooks the ext3 to ext4 migrate interface to EXT4_IOC_SETFLAGS ioctl. The userspace interface is via chattr +e. We only allow setting extent flags. Clearing extent flag (migrating from ext4 to ext3) is not supported. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-09-13ext4: elevate write count for migrate ioctlAneesh Kumar K.V
The migrate ioctl writes to the filsystem, so we need to elevate the write count. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-09-11Merge branch 'for_linus' of ↵Linus Torvalds
git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-udf-2.6 * 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-udf-2.6: udf: add llseek method udf: Fix error paths in udf_new_inode() udf: Fix lock inversion between iprune_mutex and alloc_mutex (v2)
2008-09-10ocfs2: Fix a bug in direct IO read.Tao Ma
ocfs2 will become read-only if we try to read the bytes which pass the end of i_size. This can be easily reproduced by following steps: 1. mkfs a ocfs2 volume with bs=4k cs=4k and nosparse. 2. create a small file(say less than 100 bytes) and we will create the file which is allocated 1 cluster. 3. read 8196 bytes from the kernel using O_DIRECT which exceeds the limit. 4. The ocfs2 volume becomes read-only and dmesg shows: OCFS2: ERROR (device sda13): ocfs2_direct_IO_get_blocks: Inode 66010 has a hole at block 1 File system is now read-only due to the potential of on-disk corruption. Please run fsck.ocfs2 once the file system is unmounted. So suppress the ERROR message. Signed-off-by: Tao Ma <tao.ma@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
2008-09-09Merge branch 'linux-next' of git://git.infradead.org/~dedekind/ubifs-2.6Linus Torvalds
* 'linux-next' of git://git.infradead.org/~dedekind/ubifs-2.6: UBIFS: make minimum fanout 3 UBIFS: fix division by zero UBIFS: amend f_fsid UBIFS: fill f_fsid UBIFS: improve statfs reporting even more UBIFS: introduce LEB overhead UBIFS: add forgotten gc_idx_lebs component UBIFS: fix assertion UBIFS: improve statfs reporting UBIFS: remove incorrect index space check UBIFS: push empty flash hack down UBIFS: do not update min_idx_lebs in stafs UBIFS: allow for racing between GC and TNC UBIFS: always read hashed-key nodes under TNC mutex UBIFS: fix zero-length truncations
2008-09-08NFS: Restore missing hunk in NFS mount option parserChuck Lever
Automounter maps can contain mount options valid for other NFS implementations but not for Linux. The Linux automounter uses the mount command's "-s" command line option ("s" for "sloppy") so that mount requests containing such options are not rejected. Commit f45663ce5fb30f76a3414ab3ac69f4dd320e760a attempted to address a known regression with text-based NFS mount option parsing. Unrecognized mount options would cause mount requests to fail, even if the "-s" option was used on the mount command line. Unfortunately, this commit was not complete as submitted. It adds a new mount option, "sloppy". But it is missing a hunk, so it now allows NFS mounts with unrecognized mount options, even if the "sloppy" option is not present. This could be a problem if a required critical mount option such as "sync" is misspelled, for example, and is considered a regression from 2.6.26. This patch restores the missing hunk. Now, the default behavior of text-based NFS mount options is as before: any unrecognized mount option will cause the mount to fail. Please include this in 2.6.27-rc. Thanks to Neil Brown for reporting this. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Acked-by: J. Bruce Fields <bfields@citi.umich.edu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>