Merge tag 'trace-v7.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-tracegrafted

Pull tracing fixes from Steven Rostedt:

 - Fix error output of boot instance creation failure

   Currently if a boot instance creation fails, instead of printing out
   the name of the instance that failed, it prints "(null)". That is
   because it prints "cur_str" that had already been processed by
   strsep(). Print the saved name instead.

   While at it, print the error code of the failure.

 - Fix use-after-free for same named historgrams

   Histograms can be named so that they can be used in multiple events.
   But if the named histogram has a variable attached, the second event
   that uses the named histogram which duplicates it and needs to free
   the original after duplication leaves the old variable in place and
   still visible. If another histogram uses than variable, it will use
   the stale one which will try to reference the freed duplicate
   histogram and crash the kernel.

   Free the duplicate variables along with the duplicated histogram
   data.

 - Check return value of kthread_run() in event self test

   The events self tests uses a kthread for testing but does not check
   if it succeeded in creating a kthread. If the kthread creation were
   to fail, the code will still try to call kthread_stop() on the error
   returned.

 - Fix race between reading trace_pipe and updating subbuffer size

   If a user is reading the trace_pipe file at the same time they update
   the ring buffer sub-buffer size, can cause the trace_pipe read to
   read stale data. Add trace_access_lock() around updating the ring
   buffer sub-buffer size.

 - Fix eventfs_inode on failure path in creation of the events directory

   In the creation of the "events" directory, if after allocating the
   eventfs_inode a failure is detected, it calls cleanup_ei() which
   calls free_ei(). The free_ei() will test if eventfs_inode being freed
   has no children. It is a bug if it does. But on the failure case of
   the creation of the "events" directory, the children lists have not
   yet been initialized and the free will trigger a warning because
   list_empty() on an uninitialized list returns false.

   Move the initialization into init_ei() where it makes more sense and
   makes sure that a created eventfs_inode has its lists initialized
   upon creation.

 - Check return value of kthread_run() in ftrace direct sample code

   The sample code that shows how to use the ftrace direct calls does
   not test the return of kthread_run() to see if it succeeds. Return a
   failure if the kthread_run() doesn't succeed.

 - Clear user events state on fork in case of alloc failure

   On fork, the child gets a pointer to the parent's user events state.
   It makes a copy of it then updates the child's pointer to it. But if
   the allocation fails, the duplication function leaves the child with
   a pointer to its parent's descriptor. When the child cleans up its
   data, it will free the parent's descriptor while the parent is still
   using it.

   In the duplication function, set the child's user_event_mm to NULL
   before testing if the allocation succeeded, and when it exits it will
   not free the parent's descriptor.

 - Fix retry exhaustion in simple ring buffer reader swap

   simple_ring_buffer_swap_reader_page() starts with retry set to 8 and
   post-decrements it only after a failed link replacement. On the final
   attempt, a successful replacement leaves retry at zero, while a
   failed replacement leaves it at -1.

   But the check for success expects the retry value to be non-zero and
   exits with an error on zero. This is the opposite result. Fix it.

 - Fail nicely when the remote swap_reader_page() returns an error

   Currently, if the swap_reader_page() of a remote buffer fails, it
   triggers a WARN_ON_ONCE() and continues normally. Instead, have it
   exit with an error and a pr_warn() print instead of a full WARNING.

* tag 'trace-v7.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  ring-buffer: Stop remote reader update when page swap fails
  tracing: Fix retry exhaustion in simple ring buffer reader swap
  tracing/user_events: Clear copied tracing state before fork duplication
  samples/ftrace: Fix kthread_stop() on ERR_PTR in ftrace-direct-multi-modify
  samples/ftrace: Fix kthread_stop() on ERR_PTR in ftrace-direct-modify
  eventfs: Initialize ei->children and ei->list in init_ei()
  tracing: Fix use-after-free in trace_pipe read on sub-buffer order change
  tracing: Fix crash passing ERR_PTR to kthread_stop()
  tracing: Fix use-after-free with same-name named triggers
  tracing: Fix logged instance name on creation failure

1 files changed, 343 insertions, 0 deletions

diff --git a/fs/affs/Changes b/fs/affs/Changes
new file mode 100644
index 000000000..b41c2c979
--- /dev/null
+++ b/fs/affs/Changes
@@ -0,0 +1,343 @@
+(Note: I consider version numbers as cheap. That means
+that I do not like numbers like 0.1 and the like for
+things that can be used since quite some time. But
+then, 3.1 doesn't mean 'perfectly stable', too.)
+
+Known bugs:
+-----------
+
+- Doesn't work on the alpha. The only 64/32-bit
+  problem that I'm aware of (pointer/int conversion
+  in readdir()) gives compiler warnings but is
+  apparently not causing the failure, as directory
+  reads basically work (but all files are of size 0).
+  Alas, I've got no alpha to debug. :-(
+
+- The partition checker (drivers/block/genhd.c)
+  doesn't work with devices which have 256 byte
+  blocks (some very old SCSI drives). 
+
+- The feature to automatically make the fs clean
+  might leave a trashed file system with the
+  bitmap flag set valid.
+
+- When a file is truncated to a size that is not
+  a multiple of the blocksize, the rest of the
+  last allocated block is not cleared. Well,
+  this fs never claimed to be Posix conformant.
+
+Please direct bug reports to: zippel@linux-m68k.org
+
+Version 3.20
+------------
+- kill kernel lock
+- fix for a possible bitmap corruption
+
+Version 3.19
+------------
+
+- sizeof changes from Kernel Janitor Project
+- several bug fixes found with fsx
+
+Version 3.18
+------------
+
+- change to global min macro + warning fixes
+- add module tags
+
+Version 3.17
+------------
+
+- locking fixes
+- wrong sign in __affs_hash_dentry
+- remove unnecessary check in affs_new_inode
+- enable international mode for dircache fs
+
+Version 3.16
+------------
+
+- use mark_buffer_dirty_inode instead of mark_buffer_dirty.
+- introduce affs_lock_{link|dir|ext}.
+
+Version 3.15
+------------
+
+- disable link to directories until we can properly support them.
+- locking fixes for link creation/removal.
+
+Version 3.14
+------------
+
+- correctly cut off long file names for compares
+- correctly initialize s_last_bmap
+
+Version 3.13
+------------
+
+Major cleanup for 2.4 [Roman Zippel]
+- new extended block handling
+- new bitmap allocation functions
+- locking should be safe for the future
+- cleanup of some interfaces
+
+Version 3.12
+------------
+
+more 2.4 fixes: [Roman Zippel]
+- s_lock changes
+- increased getblock mess
+- clear meta blocks 
+
+Version 3.11
+------------
+
+- Converted to use 2.3.x page cache [Dave Jones]
+- Corruption in truncate() bugfix [Ken Tyler <kent@werple.net.au>]
+
+Version 3.10
+------------
+
+- Changed partition checker to allow devices
+  with physical blocks != 512 bytes.
+
+- The partition checker now also ignores the
+  word at 0xd0 that Windows likes to write to.
+
+Version 3.9
+-----------
+
+- Moved cleanup from release_file() to put_inode().
+  This makes the first one obsolete.
+
+- truncate() zeroes the unused remainder of a
+  partially used last block when a file is truncated.
+  It also marks the inode dirty now (which is not
+  really necessary as notify_change() will do
+  it anyway).
+
+- Added a few comments, fixed some typos (and
+  introduced some new ones), made the debug messages
+  more consistent. Changed a bad example in the
+  doc file (affs.txt).
+
+- Sets the NOEXEC flag in read_super() for old file
+  systems, since you can't run programs on them.
+
+Version 3.8
+-----------
+Bill Hawes kindly reviewed the affs and sent me the
+patches he did. They're marked (BH). Thanks, Bill!
+
+- Cleanup of error handling in read_super().
+  Didn't release all resources in case of an
+  error. (BH)
+
+- put_inode() releases the ext cache only if it's
+  no longer needed. (BH)
+
+- One set of dentry callbacks is enough. (BH)
+
+- Cleanup of error handling in namei.c. (BH)
+
+- Cleanup of error handling in file.c. (BH)
+
+- The original blocksize of the device is
+  restored when the fs is unmounted. (BH)
+
+- getblock() did not invalidate the key cache
+  when it allocated a new block.
+
+- Removed some unnecessary locks as Bill
+  suggested.
+
+- Simplified match_name(), changed all hashing
+  and case insensitive name comparisons to use
+  uppercase. This makes the tolower() routines
+  obsolete.
+
+- Added mount option 'mufs' to force muFS
+  uid/gid interpretation.
+
+- File mode changes were not updated on disk.
+  This was fixed before, but somehow got lost.
+
+Version 3.7
+-----------
+
+- Added dentry callbacks to allow the dcache to
+  operate case insensitive and length ignorant
+  like the affs itself.
+
+- getblock() didn't update the lastblock field in the
+  inode if the fs was not an OFS. This bug only shows
+  up if a file was enlarged via truncate() and there
+  was not enough space.
+
+- Remove some more superfluous code left over from
+  the old link days ...
+
+- Fixed some oversights which were in patch 2.1.78.
+
+- Fixed a few typos.
+
+Version 3.6
+-----------
+
+- dentry changes. (Thanks to Jes Sorensen for his help.)
+
+- Fixed bug in balloc(): Superblock was not set dirty after
+  the bitmap was changed, so the bitmap wasn't sync'd.
+
+- Fixed nasty bug in find_new_zone(): If the current
+  zone number was zero, the loop didn't terminate,
+  causing a solid lock-up.
+
+- Removed support for old-style directory reads.
+
+- Fixed bug in add_entry(): When doing a sorted insert,
+  the pointer to the next entry in the hash chain wasn't
+  correctly byte-swapped. Since most of the users of the
+  affs use it on a 68k, they didn't notice. But why did
+  I not find this during my tests?
+
+- Fixed some oversights (version wasn't updated on some
+  directory changes).
+
+- Handling of hard links rewritten. To the VFS
+  they appear now as normal Unix links. They are
+  now resolved only once in lookup(). The backside
+  is that unlink(), rename() and rmdir() have to
+  be smart about them, but the result is worth the
+  effort. This also led to some code cleanup.
+
+- Changed name type to unsigned char; the test for
+  invalid filenames didn't work correctly.
+  (Thanks to Michael Krause for pointing at this.)
+
+- Changed mapping of executable flag.
+
+- Changed all network byte-order macros to the
+  recommended ones.
+
+- Added a remount function, so attempts to remount
+  a dircache filesystem or one with errors read/write
+  can be trapped. Previously, ro remounts didn't
+  flush the super block, and rw remounts didn't
+  create allocation zones ...
+
+- Call shrink_dcache_parent() in rmdir().
+  (Thanks to Bill Hawes.)
+
+- Permission checks in unlink().
+
+- Allow mounting of volumes with superfluous
+  bitmap pointers read only, also allows them
+  to be remounted read/write.
+
+- Owner/Group defaults now to the fs user (i.e.
+  the one that mounted it) instead of root. This
+  obsoletes the mount options uid and gid.
+
+- Argument to volume option could overflow the
+  name buffer. It is now silently truncated to
+  30 characters. (Damn it! This kind of bug
+  is too embarrassing.)
+
+- Split inode.c into 2 files, the superblock
+  routines desperately wanted their own file.
+
+- truncate() didn't allocate an extension block
+  cache. If a file was extended by means of
+  truncate(), this led to an Oops.
+
+- fsuser is now checked last.
+
+- rename() will not ignore changes in filename
+  casing any more (though mv(1) still won't allow
+  you to do "mv oldname OldName").
+
+Version 3.5
+-----------
+
+- Extension block caches are now allocated on
+  demand instead of when a file is opened, as
+  files can be read and written without opening
+  them (e. g. the loopback device does this).
+
+- Removed an unused function.
+
+Version 3.4
+-----------
+
+- Hash chains are now sorted by block numbers.
+  (Thanks to Kars de Jong for finding this.)
+- Removed all unnecessary external symbols.
+
+Version 3.3
+-----------
+
+- Tried to make all types 'correct' and consistent.
+- Errors and warnings are now reported via a
+  function. They are all prefixed by a severity
+  and have the same appearance:
+    "AFFS: <function>: <error message>"
+  (There's one exception to this, as in that function
+  is no pointer to the super block available.)
+- The filesystem is remounted read-only after an
+  error.
+- The names of newly created filesystem objects are
+  now checked for validity.
+- Minor cleanups in comments.
+- Added this Changes file. At last!
+
+Version 3.2
+-----------
+
+- Extension block cache: Reading/writing of huge files
+  (several MB) is much faster (of course the added
+  overhead slows down opening, but this is hardly
+  noticeable).
+- The same get_block()-routine can now be used for
+  both OFS and FFS.
+- The super block is now searched in the block that
+  was calculated and in the one following. This
+  should remedy the round-off error introduced by
+  the 1-k blocks that Linux uses.
+- Minor changes to adhere to the new VFS interface.
+- The number of used blocks is now also calculated
+  if the filesystem is mounted read-only.
+- Prefixed some constants with AFFS_ to avoid name
+  clashes.
+- Removed 'EXPERIMENTAL' status.
+
+Version 3.1
+-----------
+
+- Fixed a nasty bug which didn't allow read-only
+  mounts.
+- Allow dir-cache filesystems to be mounted
+  read only.
+- OFS support.
+- Several other changes I just cannot remember
+  any more.
+
+Version 3.0
+-----------
+
+- Almost complete rewrite for the new VFS
+  interface in Linux 1.3.
+- Write support.
+- Support for hard and symbolic links.
+- Lots of things I remember even less ...
+
+Version 2.0
+-----------
+
+- Fixed a few things to get it compiled.
+- Automatic root block calculation.
+- Partition checker for genhd.c
+
+========================================
+
+Let's just call Ray Burr's original affs
+'Version 1.0'.