GitHub/LineageOS/android_kernel_motorola_exynos9610.git
9 years agoocfs2: do not set fs read-only if rec[0] is empty while committing truncate
Xue jiufei [Fri, 4 Sep 2015 22:44:48 +0000 (15:44 -0700)]
ocfs2: do not set fs read-only if rec[0] is empty while committing truncate

While appending an extent to a file, it will call these functions:
ocfs2_insert_extent

  -> call ocfs2_grow_tree() if there's no free rec
     -> ocfs2_add_branch add a new branch to extent tree,
        now rec[0] in the leaf of rightmost path is empty
  -> ocfs2_do_insert_extent
     -> ocfs2_rotate_tree_right
       -> ocfs2_extend_rotate_transaction
          -> jbd2_journal_restart if jbd2_journal_extend fail
     -> ocfs2_insert_path
        -> ocfs2_extend_trans
          -> jbd2_journal_restart if jbd2_journal_extend fail
        -> ocfs2_insert_at_leaf
     -> ocfs2_et_update_clusters
Function jbd2_journal_restart() may be called and it may happened that
buffers dirtied in ocfs2_add_branch() are committed
while buffers dirtied in ocfs2_insert_at_leaf() and
ocfs2_et_update_clusters() are not.
So an empty rec[0] is left in rightmost path which will cause
read-only filesystem when call ocfs2_commit_truncate()
with the error message: "Inode %lu has an empty extent record".

This is not a serious problem, so remove the rightmost path when call
ocfs2_commit_truncate().

Signed-off-by: joyce.xue <xuejiufei@huawei.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: call ocfs2_journal_access_di() before ocfs2_journal_dirty() in ocfs2_write_end...
yangwenfang [Fri, 4 Sep 2015 22:44:45 +0000 (15:44 -0700)]
ocfs2: call ocfs2_journal_access_di() before ocfs2_journal_dirty() in ocfs2_write_end_nolock()

1: After we call ocfs2_journal_access_di() in ocfs2_write_begin(),
   jbd2_journal_restart() may also be called, in this function transaction
   A's t_updates-- and obtains a new transaction B.  If
   jbd2_journal_commit_transaction() is happened to commit transaction A,
   when t_updates==0, it will continue to complete commit and unfile
   buffer.

   So when jbd2_journal_dirty_metadata(), the handle is pointed a new
   transaction B, and the buffer head's journal head is already freed,
   jh->b_transaction == NULL, jh->b_next_transaction == NULL, it returns
   EINVAL, So it triggers the BUG_ON(status).

thread 1                                          jbd2
ocfs2_write_begin                     jbd2_journal_commit_transaction
ocfs2_write_begin_nolock
  ocfs2_start_trans
    jbd2__journal_start(t_updates+1,
                       transaction A)
    ocfs2_journal_access_di
    ocfs2_write_cluster_by_desc
      ocfs2_mark_extent_written
        ocfs2_change_extent_flag
          ocfs2_split_extent
            ocfs2_extend_rotate_transaction
              jbd2_journal_restart
              (t_updates-1,transaction B) t_updates==0
                                        __jbd2_journal_refile_buffer
                                        (jh->b_transaction = NULL)
ocfs2_write_end
ocfs2_write_end_nolock
    ocfs2_journal_dirty
        jbd2_journal_dirty_metadata(bug)
   ocfs2_commit_trans

2.  In ext4, I found that: jbd2_journal_get_write_access() called by
   ext4_write_end.

ext4_write_begin
    ext4_journal_start
        __ext4_journal_start_sb
            ext4_journal_check_start
            jbd2__journal_start

ext4_write_end
    ext4_mark_inode_dirty
        ext4_reserve_inode_write
            ext4_journal_get_write_access
                jbd2_journal_get_write_access
        ext4_mark_iloc_dirty
            ext4_do_update_inode
                ext4_handle_dirty_metadata
                    jbd2_journal_dirty_metadata

3. So I think we should put ocfs2_journal_access_di before
   ocfs2_journal_dirty in the ocfs2_write_end.  and it works well after my
   modification.

Signed-off-by: vicky <vicky.yangwenfang@huawei.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Zhangguanghui <zhang.guanghui@h3c.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: use 64bit variables to track heartbeat time
Tina Ruchandani [Fri, 4 Sep 2015 22:44:43 +0000 (15:44 -0700)]
ocfs2: use 64bit variables to track heartbeat time

o2hb_elapsed_msecs computes the time taken for a disk heartbeat.
'struct timeval' variables are used to store start and end times.  On
32-bit systems, the 'tv_sec' component of 'struct timeval' will overflow
in year 2038 and beyond.

This patch solves the overflow with the following:

1. Replace o2hb_elapsed_msecs using 'ktime_t' values to measure start
   and end time, and built-in function 'ktime_ms_delta' to compute the
   elapsed time.  ktime_get_real() is used since the code prints out the
   wallclock time.

2. Changes format string to print time as a single 64-bit nanoseconds
   value ("%lld") instead of seconds and microseconds.  This simplifies
   the code since converting ktime_t to that format would need expensive
   computation.  However, the debug log string is less readable than the
   previous format.

Signed-off-by: Tina Ruchandani <ruchandani.tina@gmail.com>
Suggested by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: fix race between crashed dio and rm
Joseph Qi [Fri, 4 Sep 2015 22:44:40 +0000 (15:44 -0700)]
ocfs2: fix race between crashed dio and rm

There is a race case between crashed dio and rm, which will lead to
OCFS2_VALID_FL not set read-only.

  N1                              N2
  ------------------------------------------------------------------------
  dd with direct flag
                                  rm file
  crashed with an dio entry left
  in orphan dir
                                  clear OCFS2_VALID_FL in
                                  ocfs2_remove_inode
                                  recover N1 and read the corrupted inode,
                                  and set filesystem read-only

So we skip the inode deletion this time and wait for dio entry recovered
first.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: avoid access invalid address when read o2dlm debug messages
Yiwen Jiang [Fri, 4 Sep 2015 22:44:37 +0000 (15:44 -0700)]
ocfs2: avoid access invalid address when read o2dlm debug messages

The following case will lead to a lockres is freed but is still in use.

cat /sys/kernel/debug/o2dlm/locking_state dlm_thread
lockres_seq_start
    -> lock dlm->track_lock
    -> get resA
                                                resA->refs decrease to 0,
                                                call dlm_lockres_release,
                                                and wait for "cat" unlock.
Although resA->refs is already set to 0,
increase resA->refs, and then unlock
                                                lock dlm->track_lock
                                                    -> list_del_init()
                                                    -> unlock
                                                    -> free resA

In such a race case, invalid address access may occurs.  So we should
delete list res->tracking before resA->refs decrease to 0.

Signed-off-by: Yiwen Jiang <jiangyiwen@huawei.com>
Reviewed-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Mark Fasheh <mfasheh@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: take inode lock in ocfs2_iop_set/get_acl()
Tariq Saeed [Fri, 4 Sep 2015 22:44:34 +0000 (15:44 -0700)]
ocfs2: take inode lock in ocfs2_iop_set/get_acl()

This bug in mainline code is pointed out by Mark Fasheh.  When
ocfs2_iop_set_acl() and ocfs2_iop_get_acl() are entered from VFS layer,
inode lock is not held.  This seems to be regression from older kernels.
The patch is to fix that.

Orabug: 20189959
Signed-off-by: Tariq Saeed <tariq.x.saeed@oracle.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: fix BUG_ON() in ocfs2_ci_checkpointed()
Tariq Saeed [Fri, 4 Sep 2015 22:44:31 +0000 (15:44 -0700)]
ocfs2: fix BUG_ON() in ocfs2_ci_checkpointed()

PID: 614    TASK: ffff882a739da580  CPU: 3   COMMAND: "ocfs2dc"
  #0 [ffff882ecc3759b0] machine_kexec at ffffffff8103b35d
  #1 [ffff882ecc375a20] crash_kexec at ffffffff810b95b5
  #2 [ffff882ecc375af0] oops_end at ffffffff815091d8
  #3 [ffff882ecc375b20] die at ffffffff8101868b
  #4 [ffff882ecc375b50] do_trap at ffffffff81508bb0
  #5 [ffff882ecc375ba0] do_invalid_op at ffffffff810165e5
  #6 [ffff882ecc375c40] invalid_op at ffffffff815116fb
     [exception RIP: ocfs2_ci_checkpointed+208]
     RIP: ffffffffa0a7e940  RSP: ffff882ecc375cf0  RFLAGS: 00010002
     RAX: 0000000000000001  RBX: 000000000000654b  RCX: ffff8812dc83f1f8
     RDX: 00000000000017d9  RSI: ffff8812dc83f1f8  RDI: ffffffffa0b2c318
     RBP: ffff882ecc375d20   R8: ffff882ef6ecfa60   R9: ffff88301f272200
     R10: 0000000000000000  R11: 0000000000000000  R12: ffffffffffffffff
     R13: ffff8812dc83f4f0  R14: 0000000000000000  R15: ffff8812dc83f1f8
     ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
  #7 [ffff882ecc375d28] ocfs2_check_meta_downconvert at ffffffffa0a7edbd [ocfs2]
  #8 [ffff882ecc375d38] ocfs2_unblock_lock at ffffffffa0a84af8 [ocfs2]
  #9 [ffff882ecc375dc8] ocfs2_process_blocked_lock at ffffffffa0a85285 [ocfs2]
#10 [ffff882ecc375e18] ocfs2_downconvert_thread_do_work at ffffffffa0a85445 [ocfs2]
#11 [ffff882ecc375e68] ocfs2_downconvert_thread at ffffffffa0a854de [ocfs2]
#12 [ffff882ecc375ee8] kthread at ffffffff81090da7
#13 [ffff882ecc375f48] kernel_thread_helper at ffffffff81511884
assert is tripped because the tran is not checkpointed and the lock level is PR.

Some time ago, chmod command had been executed. As result, the following call
chain left the inode cluster lock in PR state, latter on causing the assert.
system_call_fastpath
  -> my_chmod
   -> sys_chmod
    -> sys_fchmodat
     -> notify_change
      -> ocfs2_setattr
       -> posix_acl_chmod
        -> ocfs2_iop_set_acl
         -> ocfs2_set_acl
          -> ocfs2_acl_set_mode
Here is how.
1119 int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
1120 {
1247         ocfs2_inode_unlock(inode, 1); <<< WRONG thing to do.
..
1258         if (!status && attr->ia_valid & ATTR_MODE) {
1259                 status =  posix_acl_chmod(inode, inode->i_mode);

519 posix_acl_chmod(struct inode *inode, umode_t mode)
520 {
..
539         ret = inode->i_op->set_acl(inode, acl, ACL_TYPE_ACCESS);

287 int ocfs2_iop_set_acl(struct inode *inode, struct posix_acl *acl, ...
288 {
289         return ocfs2_set_acl(NULL, inode, NULL, type, acl, NULL, NULL);

224 int ocfs2_set_acl(handle_t *handle,
225                          struct inode *inode, ...
231 {
..
252                                 ret = ocfs2_acl_set_mode(inode, di_bh,
253                                                          handle, mode);

168 static int ocfs2_acl_set_mode(struct inode *inode, struct buffer_head ...
170 {
183         if (handle == NULL) {
                    >>> BUG: inode lock not held in ex at this point <<<
184                 handle = ocfs2_start_trans(OCFS2_SB(inode->i_sb),
185                                            OCFS2_INODE_UPDATE_CREDITS);

ocfs2_setattr.#1247 we unlock and at #1259 call posix_acl_chmod. When we reach
ocfs2_acl_set_mode.#181 and do trans, the inode cluster lock is not held in EX
mode (it should be). How this could have happended?

We are the lock master, were holding lock EX and have released it in
ocfs2_setattr.#1247.  Note that there are no holders of this lock at
this point.  Another node needs the lock in PR, and we downconvert from
EX to PR.  So the inode lock is PR when do the trans in
ocfs2_acl_set_mode.#184.  The trans stays in core (not flushed to disc).
Now another node want the lock in EX, downconvert thread gets kicked
(the one that tripped assert abovt), finds an unflushed trans but the
lock is not EX (it is PR).  If the lock was at EX, it would have flushed
the trans ocfs2_ci_checkpointed -> ocfs2_start_checkpoint before
downconverting (to NULL) for the request.

ocfs2_setattr must not drop inode lock ex in this code path.  If it
does, takes it again before the trans, say in ocfs2_set_acl, another
cluster node can get in between, execute another setattr, overwriting
the one in progress on this node, resulting in a mode acl size combo
that is a mix of the two.

Orabug: 20189959
Signed-off-by: Tariq Saeed <tariq.x.saeed@oracle.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: optimize error handling in dlm_request_join
Norton.Zhu [Fri, 4 Sep 2015 22:44:28 +0000 (15:44 -0700)]
ocfs2: optimize error handling in dlm_request_join

Currently error handling in dlm_request_join is a little obscure, so
optimize it to promote readability.

If packet.code is invalid, reset it to JOIN_DISALLOW to keep it
meaningful.  It only influences the log printing.

Signed-off-by: Norton.Zhu <norton.zhu@huawei.com>
Cc: Srinivas Eeda <srinivas.eeda@oracle.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: fix a tiny case that inode can not removed
Yiwen Jiang [Fri, 4 Sep 2015 22:44:25 +0000 (15:44 -0700)]
ocfs2: fix a tiny case that inode can not removed

When running dirop_fileop_racer we found a case that inode
can not removed.

Two nodes, say Node A and Node B, mount the same ocfs2 volume.  Create
two dirs /race/1/ and /race/2/ in the filesystem.

  Node A                            Node B
  rm -r /race/2/
                                    mv /race/1/ /race/2/
  call ocfs2_unlink(), get
  the EX mode of /race/2/
                                    wait for B unlock /race/2/
  decrease i_nlink of /race/2/ to 0,
  and add inode of /race/2/ into
  orphan dir, unlock /race/2/
                                    got EX mode of /race/2/. because
                                    /race/1/ is dir, so inc i_nlink
                                    of /race/2/ and update into disk,
                                    unlock /race/2/
  because i_nlink of /race/2/
  is not zero, this inode will
  always remain in orphan dir

This patch fixes this case by test whether i_nlink of new dir is zero.

Signed-off-by: Yiwen Jiang <jiangyiwen@huawei.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Cc: Xue jiufei <xuejiufei@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: add ip_alloc_sem in direct IO to protect allocation changes
WeiWei Wang [Fri, 4 Sep 2015 22:44:23 +0000 (15:44 -0700)]
ocfs2: add ip_alloc_sem in direct IO to protect allocation changes

In ocfs2, ip_alloc_sem is used to protect allocation changes on the
node.  In direct IO, we add ip_alloc_sem to protect date consistent
between direct-io and ocfs2_truncate_file race (buffer io use
ip_alloc_sem already).  Although inode->i_mutex lock is used to avoid
concurrency of above situation, i think ip_alloc_sem is still needed
because protect allocation changes is significant.

Other filesystem like ext4 also uses rw_semaphore to protect data
consistent between get_block-vs-truncate race by other means, So
ip_alloc_sem in ocfs2 direct io is needed.

Signed-off-by: Weiwei Wang <wangww631@huawei.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: clear the rest of the buffers on error
Goldwyn Rodrigues [Fri, 4 Sep 2015 22:44:20 +0000 (15:44 -0700)]
ocfs2: clear the rest of the buffers on error

In case a validation fails, clear the rest of the buffers and return the
error to the calling function.

This also facilitates bubbling up the error originating from ocfs2_error
to calling functions.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: acknowledge return value of ocfs2_error()
Goldwyn Rodrigues [Fri, 4 Sep 2015 22:44:17 +0000 (15:44 -0700)]
ocfs2: acknowledge return value of ocfs2_error()

Caveat: This may return -EROFS for a read case, which seems wrong.  This
is happening even without this patch series though.  Should we convert
EROFS to EIO?

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: add errors=continue
Goldwyn Rodrigues [Fri, 4 Sep 2015 22:44:11 +0000 (15:44 -0700)]
ocfs2: add errors=continue

OCFS2 is often used in high-availaibility systems.  However, ocfs2
converts the filesystem to read-only at the drop of the hat.  This may
not be necessary, since turning the filesystem read-only would affect
other running processes as well, decreasing availability.

This attempt is to add errors=continue, which would return the EIO to
the calling process and terminate furhter processing so that the
filesystem is not corrupted further.  However, the filesystem is not
converted to read-only.

As a future plan, I intend to create a small utility or extend
fsck.ocfs2 to fix small errors such as in the inode.  The input to the
utility such as the inode can come from the kernel logs so we don't have
to schedule a downtime for fixing small-enough errors.

The patch changes the ocfs2_error to return an error.  The error
returned depends on the mount option set.  If none is set, the default
is to turn the filesystem read-only.

Perhaps errors=continue is not the best option name.  Historically it is
used for making an attempt to progress in the current process itself.
Should we call it errors=eio? or errors=killproc? Suggestions/Comments
welcome.

Sources are available at:
  https://github.com/goldwynr/linux/tree/error-cont

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: flush inode data to disk and free inode when i_count becomes zero
Xue jiufei [Fri, 4 Sep 2015 22:44:11 +0000 (15:44 -0700)]
ocfs2: flush inode data to disk and free inode when i_count becomes zero

Disk inode deletion may be heavily delayed when one node unlink a file
after the same dentry is freed on another node(say N1) because of memory
shrink but inode is left in memory.  This inode can only be freed while
N1 doing the orphan scan work.

However, N1 may skip orphan scan for several times because other nodes
may do the work earlier.  In our tests, it may take 1 hour on 4 nodes
cluster and it hurts the user experience.  So we think the inode should
be freed after the data flushed to disk when i_count becomes zero to
avoid such circumstances.

Signed-off-by: Joyce.xue <xuejiufei@huawei.com>
Cc: Joel Becker <jlbec@evilplan.org>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: trusted xattr missing CAP_SYS_ADMIN check
Sanidhya Kashyap [Fri, 4 Sep 2015 22:44:08 +0000 (15:44 -0700)]
ocfs2: trusted xattr missing CAP_SYS_ADMIN check

The trusted extended attributes are only visible to the process which
hvae CAP_SYS_ADMIN capability but the check is missing in ocfs2
xattr_handler trusted list.  The check is important because this will be
used for implementing mechanisms in the userspace for which other
ordinary processes should not have access to.

Signed-off-by: Sanidhya Kashyap <sanidhya.gatech@gmail.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Taesoo kim <taesoo@gatech.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: set filesytem read-only when ocfs2_delete_entry failed.
jiangyiwen [Fri, 4 Sep 2015 22:44:06 +0000 (15:44 -0700)]
ocfs2: set filesytem read-only when ocfs2_delete_entry failed.

In ocfs2_rename, it will lead to an inode with two entried(old and new) if
ocfs2_delete_entry(old) failed.  Thus, filesystem will be inconsistent.

The case is described below:

ocfs2_rename
    -> ocfs2_start_trans
    -> ocfs2_add_entry(new)
    -> ocfs2_delete_entry(old)
        -> __ocfs2_journal_access *failed* because of -ENOMEM
    -> ocfs2_commit_trans

So filesystem should be set to read-only at the moment.

Signed-off-by: Yiwen Jiang <jiangyiwen@huawei.com>
Cc: Joseph Qi <joseph.qi@huawei.com>
Cc: Joel Becker <jlbec@evilplan.org>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2/dlm: use list_for_each_entry instead of list_for_each
Joseph Qi [Fri, 4 Sep 2015 22:44:03 +0000 (15:44 -0700)]
ocfs2/dlm: use list_for_each_entry instead of list_for_each

Use list_for_each_entry instead of list_for_each to simplify code.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: remove unneeded code in dlm_register_domain_handlers
Joseph Qi [Fri, 4 Sep 2015 22:44:00 +0000 (15:44 -0700)]
ocfs2: remove unneeded code in dlm_register_domain_handlers

The last goto statement is unneeded, so remove it.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: fix BUG when o2hb_register_callback fails
Joseph Qi [Fri, 4 Sep 2015 22:43:57 +0000 (15:43 -0700)]
ocfs2: fix BUG when o2hb_register_callback fails

In dlm_register_domain_handlers, if o2hb_register_callback fails, it
will call dlm_unregister_domain_handlers to unregister.  This will
trigger the BUG_ON in o2hb_unregister_callback because hc_magic is 0.
So we should call o2hb_setup_callback to initialize hc first.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: remove unneeded code in ocfs2_dlm_init
Joseph Qi [Fri, 4 Sep 2015 22:43:54 +0000 (15:43 -0700)]
ocfs2: remove unneeded code in ocfs2_dlm_init

status is already initialized and it will only be 0 or negatives in the
code flow.  So remove the unneeded assignment after the lable 'local'.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: adjust code to match locking/unlocking order
Joseph Qi [Fri, 4 Sep 2015 22:43:52 +0000 (15:43 -0700)]
ocfs2: adjust code to match locking/unlocking order

Unlocking order in ocfs2_unlink and ocfs2_rename mismatches the
corresponding locking order, although it won't cause issues, adjust the
code so that it looks more reasonable.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: clean up unused local variables in ocfs2_file_write_iter
Joseph Qi [Fri, 4 Sep 2015 22:43:49 +0000 (15:43 -0700)]
ocfs2: clean up unused local variables in ocfs2_file_write_iter

Since commit 86b9c6f3f891 ("ocfs2: remove filesize checks for sync I/O
journal commit") removes filesize checks for sync I/O journal commit,
variables old_size and old_clusters are not actually used any more.  So
clean them up.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: do not log twice error messages
Christophe JAILLET [Fri, 4 Sep 2015 22:43:46 +0000 (15:43 -0700)]
ocfs2: do not log twice error messages

'o2hb_map_slot_data' and 'o2hb_populate_slot_data' are called from only
one place, in 'o2hb_region_dev_write'.  Return value is checked and
'mlog_errno' is called to log a message if it is not 0.

So there is no need to call 'mlog_errno' directly within these functions.
This would result on logging the message twice.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: do not BUG if buffer not uptodate in __ocfs2_journal_access
Joseph Qi [Fri, 4 Sep 2015 22:43:43 +0000 (15:43 -0700)]
ocfs2: do not BUG if buffer not uptodate in __ocfs2_journal_access

When storage network is unstable, it may trigger the BUG in
__ocfs2_journal_access because of buffer not uptodate.  We can retry the
write in this case or return error instead of BUG.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Reported-by: Zhangguanghui <zhang.guanghui@h3c.com>
Tested-by: Zhangguanghui <zhang.guanghui@h3c.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: fix several issues of append dio
Joseph Qi [Fri, 4 Sep 2015 22:43:40 +0000 (15:43 -0700)]
ocfs2: fix several issues of append dio

1) Take rw EX lock in case of append dio.
2) Explicitly treat the error code -EIOCBQUEUED as normal.
3) Set di_bh to NULL after brelse if it may be used again later.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Yiwen Jiang <jiangyiwen@huawei.com>
Cc: Weiwei Wang <wangww631@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: fix race between dio and recover orphan
Joseph Qi [Fri, 4 Sep 2015 22:43:37 +0000 (15:43 -0700)]
ocfs2: fix race between dio and recover orphan

During direct io the inode will be added to orphan first and then
deleted from orphan.  There is a race window that the orphan entry will
be deleted twice and thus trigger the BUG when validating
OCFS2_DIO_ORPHANED_FL in ocfs2_del_inode_from_orphan.

ocfs2_direct_IO_write
    ...
    ocfs2_add_inode_to_orphan
    >>>>>>>> race window.
             1) another node may rm the file and then down, this node
             take care of orphan recovery and clear flag
             OCFS2_DIO_ORPHANED_FL.
             2) since rw lock is unlocked, it may race with another
             orphan recovery and append dio.
    ocfs2_del_inode_from_orphan

So take inode mutex lock when recovering orphans and make rw unlock at the
end of aio write in case of append dio.

Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Reported-by: Yiwen Jiang <jiangyiwen@huawei.com>
Cc: Weiwei Wang <wangww631@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agosh: use PFN_DOWN macro
Alexander Kuleshov [Fri, 4 Sep 2015 22:43:35 +0000 (15:43 -0700)]
sh: use PFN_DOWN macro

Replace ((x) >> PAGE_SHIFT) with the predefined PFN_DOWN macro.

Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
Acked-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agontfs: delete unnecessary checks before calling iput()
SF Markus Elfring [Fri, 4 Sep 2015 22:43:32 +0000 (15:43 -0700)]
ntfs: delete unnecessary checks before calling iput()

iput() tests whether its argument is NULL and then returns immediately.
Thus the test around the call is not needed.

This issue was detected by using the Coccinelle software.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Cc: Julia Lawall <julia.lawall@lip6.fr>
Reviewed-by: Anton Altaparmakov <anton@tuxera.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoscripts/spelling.txt: add some typo-words
Zhao Lei [Fri, 4 Sep 2015 22:43:29 +0000 (15:43 -0700)]
scripts/spelling.txt: add some typo-words

I wrote a small script to show word-pair from all linux spelling-typo
commits, and get following result by sort | uniq -c:

    181 occured -> occurred
     78 transfered -> transferred
     67 recieved -> received
     65 dependant -> dependent
     58 wether -> whether
     56 accomodate -> accommodate
     54 occured -> occurred
     51 recieve -> receive
     47 cant -> can't
     40 sucessfully -> successfully
     ...

Some of them are not in spelling.txt, this patch adds the most common
word-pairs into spelling.txt.

Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoscripts: decode_stacktrace: fix ARM architecture decoding
Robert Jarzmik [Fri, 4 Sep 2015 22:43:26 +0000 (15:43 -0700)]
scripts: decode_stacktrace: fix ARM architecture decoding

Fix the stack decoder for the ARM architecture.
An ARM stack is designed as :

[   81.547704] [<c023eb04>] (bucket_find_contain) from [<c023ec88>] (check_sync+0x40/0x4f8)
[   81.559668] [<c023ec88>] (check_sync) from [<c023f8c4>] (debug_dma_sync_sg_for_cpu+0x128/0x194)
[   81.571583] [<c023f8c4>] (debug_dma_sync_sg_for_cpu) from [<c0327dec>] (__videobuf_s

The current script doesn't expect the symbols to be bound by
parenthesis, and triggers the following errors :

  awk: cmd. line:1: error: Unmatched ( or \(: / (check_sync$/
  [   81.547704] (bucket_find_contain) from (check_sync+0x40/0x4f8)

Fix it by chopping starting and ending parenthesis from the each symbol
name.

As a side note, this probably comes from the function
dump_backtrace_entry(), which is implemented differently for each
architecture.  That makes a single decoding script a bit a challenge.

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Michal Marek <mmarek@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoscripts/Lindent: handle missing indent gracefully
Jean Delvare [Fri, 4 Sep 2015 22:43:24 +0000 (15:43 -0700)]
scripts/Lindent: handle missing indent gracefully

If indent is not found, bail out immediately instead of spitting random
shell script error messages.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agokerneldoc: Convert error messages to GNU error message format
Bart Van Assche [Fri, 4 Sep 2015 22:43:21 +0000 (15:43 -0700)]
kerneldoc: Convert error messages to GNU error message format

Editors like emacs and vi recognize a number of error message formats.
The format used by the kerneldoc tool is not recognized by emacs.

Change the kerneldoc error message format to the GNU style such that the
emacs prev-error and next-error commands can be used to navigate through
kerneldoc error messages.  For more information about the GNU error
message format, see also
  https://www.gnu.org/prep/standards/html_node/Errors.html.

This patch has been generated via the following sed command:

  sed -i.orig 's/Error(\${file}:\$.):/\${file}:\$.: error:/g;s/Warning(\${file}:\$.):/\${file}:\$.: warning:/g;s/Warning(\${file}):/\${file}:1: warning:/g;s/Info(\${file}:\$.):/\${file}:\$.: info:/g' scripts/kernel-doc

Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Johannes Berg <johannes.berg@intel.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoscripts/spelling.txt: spelling of uninitialized
Sudip Mukherjee [Fri, 4 Sep 2015 22:43:18 +0000 (15:43 -0700)]
scripts/spelling.txt: spelling of uninitialized

I just did a spelling mistake of uninitialized and wrote that as
unintialized.  Fortunately I noticed it in my final review.

Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoscripts/spelling.txt: add misspelled words for check
Maninder Singh [Fri, 4 Sep 2015 22:43:15 +0000 (15:43 -0700)]
scripts/spelling.txt: add misspelled words for check

misspelled words for check:-
 chcek
 chck
 cehck

I myself did these spell mistakes in changelog for patches, Thus
suggesting to add in spelling.txt, so that checkpatch.pl warns it
earlier.  References:-

./arch/powerpc/kernel/exceptions-64e.S:456: . . . make sure you chcek
https://lkml.org/lkml/2015/6/25/289
./arch/x86/mm/pageattr.c:1368: * No need to cehck in that case

[akpm@linux-foundation.org: add whcih->which, whcih I always get wrong]
Signed-off-by: Maninder Singh <maninder1.s@samsung.com>
Acked-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agofsnotify: get rid of fsnotify_destroy_mark_locked()
Jan Kara [Fri, 4 Sep 2015 22:43:12 +0000 (15:43 -0700)]
fsnotify: get rid of fsnotify_destroy_mark_locked()

fsnotify_destroy_mark_locked() is subtle to use because it temporarily
releases group->mark_mutex.  To avoid future problems with this
function, split it into two.

fsnotify_detach_mark() is the part that needs group->mark_mutex and
fsnotify_free_mark() is the part that must be called outside of
group->mark_mutex.  This way it's much clearer what's going on and we
also avoid some pointless acquisitions of group->mark_mutex.

Signed-off-by: Jan Kara <jack@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agofsnotify: remove mark->free_list
Jan Kara [Fri, 4 Sep 2015 22:43:09 +0000 (15:43 -0700)]
fsnotify: remove mark->free_list

Free list is used when all marks on given inode / mount should be
destroyed when inode / mount is going away.  However we can free all of
the marks without using a special list with some care.

Signed-off-by: Jan Kara <jack@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agofsnotify: document mark locking
Jan Kara [Fri, 4 Sep 2015 22:43:06 +0000 (15:43 -0700)]
fsnotify: document mark locking

Signed-off-by: Jan Kara <jack@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agofsnotify: fix check in inotify fdinfo printing
Jan Kara [Fri, 4 Sep 2015 22:43:03 +0000 (15:43 -0700)]
fsnotify: fix check in inotify fdinfo printing

A check in inotify_fdinfo() checking whether mark is valid was always
true due to a bug.  Luckily we can never get to invalidated marks since
we hold mark_mutex and invalidated marks get removed from the group list
when they are invalidated under that mutex.

Anyway fix the check to make code more future proof.

Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agofs/notify: optimize inotify/fsnotify code for unwatched files
Dave Hansen [Fri, 4 Sep 2015 22:43:01 +0000 (15:43 -0700)]
fs/notify: optimize inotify/fsnotify code for unwatched files

I have a _tiny_ microbenchmark that sits in a loop and writes single
bytes to a file.  Writing one byte to a tmpfs file is around 2x slower
than reading one byte from a file, which is a _bit_ more than I expecte.
This is a dumb benchmark, but I think it's hard to deny that write() is
a hot path and we should avoid unnecessary overhead there.

I did a 'perf record' of 30-second samples of read and write.  The top
item in a diffprofile is srcu_read_lock() from fsnotify().  There are
active inotify fd's from systemd, but nothing is actually listening to
the file or its part of the filesystem.

I *think* we can avoid taking the srcu_read_lock() for the common case
where there are no actual marks on the file.  This means that there will
both be nothing to notify for *and* implies that there is no need for
clearing the ignore mask.

This patch gave a 13.1% speedup in writes/second on my test, which is an
improvement from the 10.8% that I saw with the last version.

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jan Kara <jack@suse.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Eric Paris <eparis@redhat.com>
Cc: John McCutchan <john@johnmccutchan.com>
Cc: Robert Love <rlove@rlove.org>
Cc: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agodrivers/video/concole: add negative dependency for VGA_CONSOLE on ARC
Yuriy Kolerov [Fri, 4 Sep 2015 22:42:58 +0000 (15:42 -0700)]
drivers/video/concole: add negative dependency for VGA_CONSOLE on ARC

Architectures which support VGA console must define screen_info
structurture from "uapi/linux/screen_info.h".  Otherwise undefined
symbol error occurs.  Usually it's defined in "setup.c" for each
architecture.

If an architecture does not support VGA console (ARC's case) there are 2
ways: define a dummy instance of screen_info or add a negative
dependency for VGA_CONSOLE in to prevent selecting this option.

I've implemented the second way.  However the best solution is to add
HAVE_VGA_CONSOLE option for targets which support VGA console.  Then
turn off VGA_CONSOLE by default and add dependency to HAVE_VGA_CONSOLE.
But right now it's better to just add a negative dependency for ARC and
then consider how to collaborate about this issue with maintainers of
other architectures.

Signed-off-by: Yuriy Kolerov <yuriy.kolerov@synopsys.com>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Jaya Kumar <jayalk@intworks.biz>
Cc: Vineet Gupta <vgupta@synopsys.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agocapabilities: add a securebit to disable PR_CAP_AMBIENT_RAISE
Andy Lutomirski [Fri, 4 Sep 2015 22:42:51 +0000 (15:42 -0700)]
capabilities: add a securebit to disable PR_CAP_AMBIENT_RAISE

Per Andrew Morgan's request, add a securebit to allow admins to disable
PR_CAP_AMBIENT_RAISE.  This securebit will prevent processes from adding
capabilities to their ambient set.

For simplicity, this disables PR_CAP_AMBIENT_RAISE entirely rather than
just disabling setting previously cleared bits.

Signed-off-by: Andy Lutomirski <luto@kernel.org>
Acked-by: Andrew G. Morgan <morgan@kernel.org>
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Serge Hallyn <serge.hallyn@canonical.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Aaron Jones <aaronmdjones@gmail.com>
Cc: Ted Ts'o <tytso@mit.edu>
Cc: Andrew G. Morgan <morgan@kernel.org>
Cc: Mimi Zohar <zohar@linux.vnet.ibm.com>
Cc: Austin S Hemmelgarn <ahferroin7@gmail.com>
Cc: Markku Savela <msa@moth.iki.fi>
Cc: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: James Morris <james.l.morris@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoselftests/capabilities: Add tests for capability evolution
Andy Lutomirski [Fri, 4 Sep 2015 22:42:48 +0000 (15:42 -0700)]
selftests/capabilities: Add tests for capability evolution

This test focuses on ambient capabilities.  It requires either root or
the ability to create user namespaces.  Some of the test cases will be
skipped for nonroot users.

Signed-off-by: Andy Lutomirski <luto@kernel.org>
Acked-by: Kees Cook <keescook@chromium.org>
Cc: Christoph Lameter <cl@linux.com> # Original author
Cc: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Shuah Khan <shuahkh@osg.samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agocapabilities: ambient capabilities
Andy Lutomirski [Fri, 4 Sep 2015 22:42:45 +0000 (15:42 -0700)]
capabilities: ambient capabilities

Credit where credit is due: this idea comes from Christoph Lameter with
a lot of valuable input from Serge Hallyn.  This patch is heavily based
on Christoph's patch.

===== The status quo =====

On Linux, there are a number of capabilities defined by the kernel.  To
perform various privileged tasks, processes can wield capabilities that
they hold.

Each task has four capability masks: effective (pE), permitted (pP),
inheritable (pI), and a bounding set (X).  When the kernel checks for a
capability, it checks pE.  The other capability masks serve to modify
what capabilities can be in pE.

Any task can remove capabilities from pE, pP, or pI at any time.  If a
task has a capability in pP, it can add that capability to pE and/or pI.
If a task has CAP_SETPCAP, then it can add any capability to pI, and it
can remove capabilities from X.

Tasks are not the only things that can have capabilities; files can also
have capabilities.  A file can have no capabilty information at all [1].
If a file has capability information, then it has a permitted mask (fP)
and an inheritable mask (fI) as well as a single effective bit (fE) [2].
File capabilities modify the capabilities of tasks that execve(2) them.

A task that successfully calls execve has its capabilities modified for
the file ultimately being excecuted (i.e.  the binary itself if that
binary is ELF or for the interpreter if the binary is a script.) [3] In
the capability evolution rules, for each mask Z, pZ represents the old
value and pZ' represents the new value.  The rules are:

  pP' = (X & fP) | (pI & fI)
  pI' = pI
  pE' = (fE ? pP' : 0)
  X is unchanged

For setuid binaries, fP, fI, and fE are modified by a moderately
complicated set of rules that emulate POSIX behavior.  Similarly, if
euid == 0 or ruid == 0, then fP, fI, and fE are modified differently
(primary, fP and fI usually end up being the full set).  For nonroot
users executing binaries with neither setuid nor file caps, fI and fP
are empty and fE is false.

As an extra complication, if you execute a process as nonroot and fE is
set, then the "secure exec" rules are in effect: AT_SECURE gets set,
LD_PRELOAD doesn't work, etc.

This is rather messy.  We've learned that making any changes is
dangerous, though: if a new kernel version allows an unprivileged
program to change its security state in a way that persists cross
execution of a setuid program or a program with file caps, this
persistent state is surprisingly likely to allow setuid or file-capped
programs to be exploited for privilege escalation.

===== The problem =====

Capability inheritance is basically useless.

If you aren't root and you execute an ordinary binary, fI is zero, so
your capabilities have no effect whatsoever on pP'.  This means that you
can't usefully execute a helper process or a shell command with elevated
capabilities if you aren't root.

On current kernels, you can sort of work around this by setting fI to
the full set for most or all non-setuid executable files.  This causes
pP' = pI for nonroot, and inheritance works.  No one does this because
it's a PITA and it isn't even supported on most filesystems.

If you try this, you'll discover that every nonroot program ends up with
secure exec rules, breaking many things.

This is a problem that has bitten many people who have tried to use
capabilities for anything useful.

===== The proposed change =====

This patch adds a fifth capability mask called the ambient mask (pA).
pA does what most people expect pI to do.

pA obeys the invariant that no bit can ever be set in pA if it is not
set in both pP and pI.  Dropping a bit from pP or pI drops that bit from
pA.  This ensures that existing programs that try to drop capabilities
still do so, with a complication.  Because capability inheritance is so
broken, setting KEEPCAPS, using setresuid to switch to nonroot uids, and
then calling execve effectively drops capabilities.  Therefore,
setresuid from root to nonroot conditionally clears pA unless
SECBIT_NO_SETUID_FIXUP is set.  Processes that don't like this can
re-add bits to pA afterwards.

The capability evolution rules are changed:

  pA' = (file caps or setuid or setgid ? 0 : pA)
  pP' = (X & fP) | (pI & fI) | pA'
  pI' = pI
  pE' = (fE ? pP' : pA')
  X is unchanged

If you are nonroot but you have a capability, you can add it to pA.  If
you do so, your children get that capability in pA, pP, and pE.  For
example, you can set pA = CAP_NET_BIND_SERVICE, and your children can
automatically bind low-numbered ports.  Hallelujah!

Unprivileged users can create user namespaces, map themselves to a
nonzero uid, and create both privileged (relative to their namespace)
and unprivileged process trees.  This is currently more or less
impossible.  Hallelujah!

You cannot use pA to try to subvert a setuid, setgid, or file-capped
program: if you execute any such program, pA gets cleared and the
resulting evolution rules are unchanged by this patch.

Users with nonzero pA are unlikely to unintentionally leak that
capability.  If they run programs that try to drop privileges, dropping
privileges will still work.

It's worth noting that the degree of paranoia in this patch could
possibly be reduced without causing serious problems.  Specifically, if
we allowed pA to persist across executing non-pA-aware setuid binaries
and across setresuid, then, naively, the only capabilities that could
leak as a result would be the capabilities in pA, and any attacker
*already* has those capabilities.  This would make me nervous, though --
setuid binaries that tried to privilege-separate might fail to do so,
and putting CAP_DAC_READ_SEARCH or CAP_DAC_OVERRIDE into pA could have
unexpected side effects.  (Whether these unexpected side effects would
be exploitable is an open question.) I've therefore taken the more
paranoid route.  We can revisit this later.

An alternative would be to require PR_SET_NO_NEW_PRIVS before setting
ambient capabilities.  I think that this would be annoying and would
make granting otherwise unprivileged users minor ambient capabilities
(CAP_NET_BIND_SERVICE or CAP_NET_RAW for example) much less useful than
it is with this patch.

===== Footnotes =====

[1] Files that are missing the "security.capability" xattr or that have
unrecognized values for that xattr end up with has_cap set to false.
The code that does that appears to be complicated for no good reason.

[2] The libcap capability mask parsers and formatters are dangerously
misleading and the documentation is flat-out wrong.  fE is *not* a mask;
it's a single bit.  This has probably confused every single person who
has tried to use file capabilities.

[3] Linux very confusingly processes both the script and the interpreter
if applicable, for reasons that elude me.  The results from thinking
about a script's file capabilities and/or setuid bits are mostly
discarded.

Preliminary userspace code is here, but it needs updating:
https://git.kernel.org/cgit/linux/kernel/git/luto/util-linux-playground.git/commit/?h=cap_ambient&id=7f5afbd175d2

Here is a test program that can be used to verify the functionality
(from Christoph):

/*
 * Test program for the ambient capabilities. This program spawns a shell
 * that allows running processes with a defined set of capabilities.
 *
 * (C) 2015 Christoph Lameter <cl@linux.com>
 * Released under: GPL v3 or later.
 *
 *
 * Compile using:
 *
 * gcc -o ambient_test ambient_test.o -lcap-ng
 *
 * This program must have the following capabilities to run properly:
 * Permissions for CAP_NET_RAW, CAP_NET_ADMIN, CAP_SYS_NICE
 *
 * A command to equip the binary with the right caps is:
 *
 * setcap cap_net_raw,cap_net_admin,cap_sys_nice+p ambient_test
 *
 *
 * To get a shell with additional caps that can be inherited by other processes:
 *
 * ./ambient_test /bin/bash
 *
 *
 * Verifying that it works:
 *
 * From the bash spawed by ambient_test run
 *
 * cat /proc/$$/status
 *
 * and have a look at the capabilities.
 */

#include <stdlib.h>
#include <stdio.h>
#include <errno.h>
#include <cap-ng.h>
#include <sys/prctl.h>
#include <linux/capability.h>

/*
 * Definitions from the kernel header files. These are going to be removed
 * when the /usr/include files have these defined.
 */
#define PR_CAP_AMBIENT 47
#define PR_CAP_AMBIENT_IS_SET 1
#define PR_CAP_AMBIENT_RAISE 2
#define PR_CAP_AMBIENT_LOWER 3
#define PR_CAP_AMBIENT_CLEAR_ALL 4

static void set_ambient_cap(int cap)
{
int rc;

capng_get_caps_process();
rc = capng_update(CAPNG_ADD, CAPNG_INHERITABLE, cap);
if (rc) {
printf("Cannot add inheritable cap\n");
exit(2);
}
capng_apply(CAPNG_SELECT_CAPS);

/* Note the two 0s at the end. Kernel checks for these */
if (prctl(PR_CAP_AMBIENT, PR_CAP_AMBIENT_RAISE, cap, 0, 0)) {
perror("Cannot set cap");
exit(1);
}
}

int main(int argc, char **argv)
{
int rc;

set_ambient_cap(CAP_NET_RAW);
set_ambient_cap(CAP_NET_ADMIN);
set_ambient_cap(CAP_SYS_NICE);

printf("Ambient_test forking shell\n");
if (execv(argv[1], argv + 1))
perror("Cannot exec");

return 0;
}

Signed-off-by: Christoph Lameter <cl@linux.com> # Original author
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Acked-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Acked-by: Kees Cook <keescook@chromium.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Aaron Jones <aaronmdjones@gmail.com>
Cc: Ted Ts'o <tytso@mit.edu>
Cc: Andrew G. Morgan <morgan@kernel.org>
Cc: Mimi Zohar <zohar@linux.vnet.ibm.com>
Cc: Austin S Hemmelgarn <ahferroin7@gmail.com>
Cc: Markku Savela <msa@moth.iki.fi>
Cc: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: James Morris <james.l.morris@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agokernel/kthread.c:kthread_create_on_node(): clarify documentation
Andrew Morton [Fri, 4 Sep 2015 22:42:42 +0000 (15:42 -0700)]
kernel/kthread.c:kthread_create_on_node(): clarify documentation

- Make it clear that the `node' arg refers to memory allocations only:
  kthread_create_on_node() does not pin the new thread to that node's
  CPUs.

- Encourage the use of NUMA_NO_NODE.

[nzimmer@sgi.com: use NUMA_NO_NODE in kthread_create() also]
Cc: Nathan Zimmer <nzimmer@sgi.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agomm: check if section present during memory block registering
Yinghai Lu [Fri, 4 Sep 2015 22:42:39 +0000 (15:42 -0700)]
mm: check if section present during memory block registering

Tony Luck found on his setup, if memory block size 512M will cause crash
during booting.

  BUG: unable to handle kernel paging request at ffffea0074000020
  IP: get_nid_for_pfn+0x17/0x40
  PGD 128ffcb067 PUD 128ffc9067 PMD 0
  Oops: 0000 [#1] SMP
  Modules linked in:
  CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.2.0-rc8 #1
  ...
  Call Trace:
     ? register_mem_sect_under_node+0x66/0xe0
     register_one_node+0x17b/0x240
     ? pci_iommu_alloc+0x6e/0x6e
     topology_init+0x3c/0x95
     do_one_initcall+0xcd/0x1f0

The system has non continuous RAM address:
 BIOS-e820: [mem 0x0000001300000000-0x0000001cffffffff] usable
 BIOS-e820: [mem 0x0000001d70000000-0x0000001ec7ffefff] usable
 BIOS-e820: [mem 0x0000001f00000000-0x0000002bffffffff] usable
 BIOS-e820: [mem 0x0000002c18000000-0x0000002d6fffefff] usable
 BIOS-e820: [mem 0x0000002e00000000-0x00000039ffffffff] usable

So there are start sections in memory block not present.  For example:

    memory block : [0x2c18000000, 0x2c20000000) 512M

first three sections are not present.

The current register_mem_sect_under_node() assume first section is
present, but memory block section number range [start_section_nr,
end_section_nr] would include not present section.

For arch that support vmemmap, we don't setup memmap for struct page
area within not present sections area.

So skip the pfn range that belong to absent section.

[akpm@linux-foundation.org: simplification]
[rientjes@google.com: more simplification]
Fixes: bdee237c0343 ("x86: mm: Use 2GB memory block size on large memory x86-64 systems")
Fixes: 982792c782ef ("x86, mm: probe memory block size for generic x86 64bit")
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: David Rientjes <rientjes@google.com>
Reported-by: Tony Luck <tony.luck@intel.com>
Tested-by: Tony Luck <tony.luck@intel.com>
Cc: Greg KH <greg@kroah.com>
Cc: Ingo Molnar <mingo@elte.hu>
Tested-by: David Rientjes <rientjes@google.com>
Cc: <stable@vger.kernel.org> [3.15+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoocfs2: direct write will call ocfs2_rw_unlock() twice when doing aio+dio
Ryan Ding [Fri, 4 Sep 2015 22:42:36 +0000 (15:42 -0700)]
ocfs2: direct write will call ocfs2_rw_unlock() twice when doing aio+dio

ocfs2_file_write_iter() is usng the wrong return value ('written').  This
will cause ocfs2_rw_unlock() be called both in write_iter & end_io,
triggering a BUG_ON.

This issue was introduced by commit 7da839c47589 ("ocfs2: use
__generic_file_write_iter()").

Orabug: 21612107
Fixes: 7da839c47589 ("ocfs2: use __generic_file_write_iter()")
Signed-off-by: Ryan Ding <ryan.ding@oracle.com>
Reviewed-by: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agomemory-hotplug: add hot-added memory ranges to memblock before allocate node_data...
Tang Chen [Fri, 4 Sep 2015 22:42:32 +0000 (15:42 -0700)]
memory-hotplug: add hot-added memory ranges to memblock before allocate node_data for a node.

Commit f9126ab9241f ("memory-hotplug: fix wrong edge when hot add a new
node") hot-added memory range to memblock, after creating pgdat for new
node.

But there is a problem:

  add_memory()
  |--> hotadd_new_pgdat()
       |--> free_area_init_node()
            |--> get_pfn_range_for_nid()
                 |--> find start_pfn and end_pfn in memblock
  |--> ......
  |--> memblock_add_node(start, size, nid)    --------    Here, just too late.

get_pfn_range_for_nid() will find that start_pfn and end_pfn are both 0.
As a result, when adding memory, dmesg will give the following wrong
message.

  Initmem setup node 5 [mem 0x0000000000000000-0xffffffffffffffff]
  On node 5 totalpages: 0
  Built 5 zonelists in Node order, mobility grouping on.  Total pages: 32588823
  Policy zone: Normal
  init_memory_mapping: [mem 0x60000000000-0x607ffffffff]

The solution is simple, just add the memory range to memblock a little
earlier, before hotadd_new_pgdat().

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Xishi Qiu <qiuxishi@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Gu Zheng <guz.fnst@cn.fujitsu.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: David Rientjes <rientjes@google.com>
Cc: <stable@vger.kernel.org> [4.2.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoMerge tag 'pinctrl-v4.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Fri, 4 Sep 2015 17:22:09 +0000 (10:22 -0700)]
Merge tag 'pinctrl-v4.3-1' of git://git./linux/kernel/git/linusw/linux-pinctrl

Pull pin control updates from Linus Walleij:
 "This is the bulk of pin control changes for the v4.3 development
  cycle.

  Like with GPIO it's a lot of stuff.  If my subsystems are any sign of
  the overall tempo of the kernel v4.3 will be a gigantic diff.

[ It looks like 4.3 is calmer than 4.2 in most other subsystems, but
  we'll see - Linus ]

  Core changes:

   - It is possible configure groups in debugfs.

   - Consolidation of chained IRQ handler install/remove replacing all
     call sites where irq_set_handler_data() and
     irq_set_chained_handler() were done in succession with a combined
     call to irq_set_chained_handler_and_data().  This series was
     created by Thomas Gleixner after the problem was observed by
     Russell King.

   - Tglx also made another series of patches switching
     __irq_set_handler_locked() for irq_set_handler_locked() which is
     way cleaner.

   - Tglx also wrote a good bunch of patches to make use of
     irq_desc_get_xxx() accessors and avoid looking up irq_descs from
     IRQ numbers.  The goal is to get rid of the irq number from the
     handlers in the IRQ flow which is nice.

  Driver feature enhancements:

   - Power management support for the SiRF SoC Atlas 7.

   - Power down support for the Qualcomm driver.

   - Intel Cherryview and Baytrail: switch drivers to use raw spinlocks
     in IRQ handlers to play nice with the realtime patch set.

   - Rework and new modes handling for Qualcomm SPMI-MPP.

   - Pinconf power source config for SH PFC.

  New drivers and subdrivers:

   - A new driver for Conexant Digicolor CX92755.

   - A new driver for UniPhier PH1-LD4, PH1-Pro4, PH1-sLD8, PH1-Pro5,
     ProXtream2 and PH1-LD6b SoC pin control support.

   - Reverse-egineered the S/PDIF settings for the Allwinner sun4i
     driver.

   - Support for Qualcomm Technologies QDF2xxx ARM64 SoCs

   - A new Freescale i.mx6ul subdriver.

  Cleanup:

   - Remove platform data support in a number of SH PFC subdrivers"

* tag 'pinctrl-v4.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl: (95 commits)
  pinctrl: at91: fix null pointer dereference
  pinctrl: mediatek: Implement wake handler and suspend resume
  pinctrl: mediatek: Fix multiple registration issue.
  pinctrl: sh-pfc: r8a7794: add USB pin groups
  pinctrl: at91: Use generic irq_{request,release}_resources()
  pinctrl: cherryview: Use raw_spinlock for locking
  pinctrl: baytrail: Use raw_spinlock for locking
  pinctrl: imx6ul: Remove .owner field
  pinctrl: zynq: Fix typos in smc0_nand_grp and smc0_nor_grp
  pinctrl: sh-pfc: Implement pinconf power-source param for voltage switching
  clk: rockchip: add pclk_pd_pmu to the list of rk3288 critical clocks
  pinctrl: sun4i: add spdif to pin description.
  pinctrl: atlas7: clear ugly branch statements for pull and drivestrength
  pinctrl: baytrail: Serialize all register access
  pinctrl: baytrail: Drop FSF mailing address
  pinctrl: rockchip: only enable gpio clock when it setting
  pinctrl/mediatek: fix spelling mistake in dev_err error message
  pinctrl: cherryview: Serialize all register access
  pinctrl: UniPhier: PH1-Pro5: add I2C ch6 pin-mux setting
  pinctrl: nomadik: reflect current input value
  ...

9 years agoMerge tag 'gpio-v4.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux...
Linus Torvalds [Fri, 4 Sep 2015 17:07:45 +0000 (10:07 -0700)]
Merge tag 'gpio-v4.3-1' of git://git./linux/kernel/git/linusw/linux-gpio

Pull GPIO updates from Linus Walleij:
 "This is the bulk of GPIO changes for the v4.3 kernel cycle.

  There is quite a lot going on in the GPIO subsystem this merge window,
  so the main matter is decribed below.

  The hits in other subsystems when making the GPIO flags optional are
  all ACKed by their respective subsystem maintainers.

  Core changes:

   - Root out the wrapper devm_gpiod_get() and gpiod_get() etc versions
     of the descriptor calls that did not use the flags argument on the
     end.  This was around for too long and eventually Uwe Kleine-König
     took the time to clean it out and the last users are removed along
     with the macros in this tag.  In several cases the use of flags
     simplifies the code.  For this reason we have (ACKed) patches
     hitting in DRM, IIO, media, NFC, USB+PHY up until we hammer in the
     nail with removing the macros.

   - Add a fat document describing how much ready-made GPIO stuff we
     have i the kernel to discourage people from reinventing a square
     wheel in userspace, as so often happens.

   - Create a separate lockdep class for each instance of a GPIO IRQ
     chip instead of using one class for all chips, as the current code
     will not work with systems with several GPIO chips doing lockdep
     debugging.

   - Protect against driver unloading also when a GPIO line is only used
     as IRQ for the GPIOLIB_IRQCHIP helpers.

   - If the GPIO chip has no designated owner, assign the parent device
     driver owner as owner.

   - Consolidation of chained IRQ handler install/remove replacing all
     call sites where irq_set_handler_data() and
     irq_set_chained_handler() were done in succession with a combined
     call to irq_set_chained_handler_and_data().

     This series was created by Thomas Gleixner after the problem was
     observed by Russell King.

   - Tglx also made another series of patches switching
     __irq_set_handler_locked() for irq_set_handler_locked() which is
     way cleaner.

   - Tglx and Jiang Liu wrote a good bunch of patches to make use of
     irq_desc_get_xxx() accessors and avoid looking up irq_descs from
     IRQ numbers.  The goal is to get rid of the irq number from the
     handlers in the IRQ flow which is nice.

   - Rob Herring killed off the set_irq_flags() for all GPIO drivers.
     This was an ARM specific function that is replaced with the generic
     irq_modify_status() where special flags are actually needed.

   - When an OF node has a pin range for its GPIOs, return -EPROBE_DEFER
     if the pin controller isn't available.  Pretty logical, yet needed
     to be fixed.

   - If a driver using GPIOLIB_IRQCHIP has its own irq_*_resources call
     back, then call these instead of the defaults provided by the
     GPIOLIB.

   - Fix an undocumented ABI hole: named GPIOs were not properly
     documented.

  Driver improvements:

   - Add get_direction() support to the generic GPIO driver, it's
     strange that we didn't have that before.

   - Make it possible to have input-only GPIO chips using the generic
     GPIO driver.

   - Clean out platform data support from the Emma Mobile (EM) driver

   - Finegrained runtime PM support for the RCAR driver.

   - Support r8a7795 (R-car H3) in the RCAR driver.

   - Support interrupts on GPIOs 16 thru 31 in the DaVinci driver.

   - Some consolidation and new support in the MPC8xxx driver, we now
     support MPC5125.

   - Preempt-RT-friendly patches: the OMAP, MPC8xxx, drivers uses raw
     spinlocks making it work better with the realime patches.

   - Interrupt support for the EXTRAXFS GPIO driver.

   - Make the ETRAXFS GPIO driver support also ARTPEC-3.

   - Interrupt and wakeup support for the BRCMSTB driver, also for
     wakeup from S5 cold boot.

   - Mask MXC IRQs during suspend.

   - Improve OMAP2 GPIO set_debounce() to work according to spec.

   - The VF610 driver handles IRQs properly.

  New drivers:

   - ZTE ZX GPIO driver"

* tag 'gpio-v4.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio: (87 commits)
  Revert "gpio: extraxfs: fix returnvar.cocci warnings"
  gpio: tc3589x: use static container helper
  gpio: xlp: fix error return code
  gpio: vf610: handle level IRQ's properly
  gpio: max732x: Fix error handling in probe()
  gpio: omap: fix clk_prepare/unprepare usage
  gpio: omap: protect regs access in omap_gpio_irq_handler
  gpio: omap: fix omap2_set_gpio_debounce
  gpio: omap: switch to use platform_get_irq
  gpio: omap: remove wrong irq_domain_remove usage in probe
  gpiolib: add description for gpio irqchip fields in struct gpio_chip
  gpio: extraxfs: fix returnvar.cocci warnings
  gpiolib: irqchip: use different lockdep class for each gpio irqchip
  gpio/grgpio: fix deadlock in grgpio_irq_unmap()
  Documentation: gpio: consumer: describe active low property
  gpio: mxc: fix section mismatch warning
  gpio/mxc: mask gpio interrupts in suspend
  gpio: omap: Fix missing raw locks conversion
  gpio: brcmstb: support wakeup from S5 cold boot
  gpio: brcmstb: Add interrupt and wakeup source support
  ...

9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile
Linus Torvalds [Fri, 4 Sep 2015 15:59:53 +0000 (08:59 -0700)]
Merge git://git./linux/kernel/git/cmetcalf/linux-tile

Pull tile updates from Chris Metcalf:
 "This includes secure computing support as well as miscellaneous minor
  improvements"

* git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile:
  tile: correct some typos in opcode type names
  tile/vdso: emit a GNU hash as well
  tile: Remove finish_arch_switch
  tile: enable full SECCOMP support
  tile/time: Migrate to new 'set-state' interface

9 years agoMerge tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64...
Linus Torvalds [Fri, 4 Sep 2015 14:18:09 +0000 (07:18 -0700)]
Merge tag 'arm64-upstream' of git://git./linux/kernel/git/arm64/linux

Pull arm64 updates from Will Deacon:

 - Support for new architectural features introduced in ARMv8.1:
   * Privileged Access Never (PAN) to catch user pointer dereferences in
     the kernel
   * Large System Extension (LSE) for building scalable atomics and locks
     (depends on locking/arch-atomic from tip, which is included here)
   * Hardware Dirty Bit Management (DBM) for updating clean PTEs
     automatically

 - Move our PSCI implementation out into drivers/firmware/, where it can
   be shared with arch/arm/. RMK has also pulled this component branch
   and has additional patches moving arch/arm/ over. MAINTAINERS is
   updated accordingly.

 - Better BUG implementation based on the BRK instruction for trapping

 - Leaf TLB invalidation for unmapping user pages

 - Support for PROBE_ONLY PCI configurations

 - Various cleanups and non-critical fixes, including:
   * Always flush FP/SIMD state over exec()
   * Restrict memblock additions based on range of linear mapping
   * Ensure *(LIST_POISON) generates a fatal fault
   * Context-tracking syscall return no longer corrupts return value when
     not forced on.
   * Alternatives patching synchronisation/stability improvements
   * Signed sub-word cmpxchg compare fix (tickled by HAVE_CMPXCHG_LOCAL)
   * Force SMP=y
   * Hide direct DCC access from userspace
   * Fix EFI stub memory allocation when DRAM starts at 0x0

* tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: (92 commits)
  arm64: flush FP/SIMD state correctly after execve()
  arm64: makefile: fix perf_callchain.o kconfig dependency
  arm64: set MAX_MEMBLOCK_ADDR according to linear region size
  of/fdt: make memblock maximum physical address arch configurable
  arm64: Fix source code file path in comments
  arm64: entry: always restore x0 from the stack on syscall return
  arm64: mdscr_el1: avoid exposing DCC to userspace
  arm64: kconfig: Move LIST_POISON to a safe value
  arm64: Add __exception_irq_entry definition for function graph
  arm64: mm: ensure patched kernel text is fetched from PoU
  arm64: alternatives: ensure secondary CPUs execute ISB after patching
  arm64: make ll/sc __cmpxchg_case_##name asm consistent
  arm64: dma-mapping: Simplify pgprot handling
  arm64: restore cpu suspend/resume functionality
  ARM64: PCI: do not enable resources on PROBE_ONLY systems
  arm64: cmpxchg: truncate sub-word signed types before comparison
  arm64: alternative: put secondary CPUs into polling loop during patch
  arm64/Documentation: clarify wording regarding memory below the Image
  arm64: lse: fix lse cmpxchg code indentation
  arm64: remove redundant object file list
  ...

9 years agoMerge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus
Linus Torvalds [Thu, 3 Sep 2015 23:55:55 +0000 (16:55 -0700)]
Merge branch 'upstream' of git://git.linux-mips.org/ralf/upstream-linus

Pull MIPS updates from Ralf Baechle:
 "This is the main pull request for 4.3 for MIPS.  Here's the summary:

  Three fixes that didn't make 4.2-stable:

   - a -Os build might compile the kernel using the MIPS16 instruction
     set but the R2 optimized inline functions in <uapi/asm/swab.h> are
     implemented using 32-bit wide instructions which is invalid.

   - a build error in pgtable-bits.h for a particular kernel
     configuration.

   - accessing registers of the CM GCR might have been compiled to use
     64 bit accesses but these registers are onl 32 bit wide.

  And also a few new bits:

   - move the ATH79 GPIO driver to drivers/gpio

   - the definition of IRQCHIP_DECLARE has moved to linux/irqchip.h,
     change ATH79 accordingly.

   - fix definition of pgprot_writecombine

   - add an implementation of dma_map_ops.mmap

   - fix alignment of quiet build output for vmlinuz link

   - BCM47xx: Use kmemdup rather than duplicating its implementation

   - Netlogic: Fix 0x0x prefixes of constants.

   - merge Bjorn Helgaas' series to remove most of the weak keywords
     from function declarations.

   - CP0 and CP1 registers are best considered treated as unsigned
     values to avoid large values from becoming negative values.

   - improve support for the MIPS GIC timer.

   - enable common clock framework for Malta and SEAD3.

   - a number of improvments and fixes to dump_tlb().

   - document the MIPS TLB dump functionality in Magic SysRq.

   - Cavium Octeon CN68XX improvments.

   - NetLogic improvments.

   - irq: Use access helper irq_data_get_affinity_mask.

   - handle MSA unaligned accesses.

   - a number of R6-related math-emu fixes.

   - support for I6400.

   - improvments to MSA support.

   - add uprobes support.

   - move from deprecated __initcall to arch_initcall.

   - remove finish_arch_switch().

   - IRQ cleanups by Thomas Gleixner.

   - migrate to new 'set-state' interface.

   - random small cleanups"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus: (148 commits)
  MIPS: UAPI: Fix unrecognized opcode WSBH/DSBH/DSHD when using MIPS16.
  MIPS: Fix alignment of quiet build output for vmlinuz link
  MIPS: math-emu: Remove unused handle_dsemul function declaration
  MIPS: math-emu: Add support for the MIPS R6 MAX{, A} FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 MIN{, A} FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 CLASS FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 RINT FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 MSUBF FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 MADDF FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 SELNEZ FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 SELEQZ FPU instruction
  MIPS: math-emu: Add support for the CMP.condn.fmt R6 instruction
  MIPS: inst.h: Add new MIPS R6 FPU opcodes
  MIPS: Octeon: Fix management port MII address on Kontron S1901
  MIPS: BCM47xx: Use kmemdup rather than duplicating its implementation
  STAGING: Octeon: Use common helpers for determining interface and port
  MIPS: Octeon: Support interfaces 4 and 5
  MIPS: Octeon: Set up 1:1 mapping between CN68XX PKO queues and ports
  MIPS: Octeon: Initialize CN68XX PKO
  STAGING: Octeon: Support CN68XX style WQE
  ...

9 years agoMerge tag 'powerpc-4.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Thu, 3 Sep 2015 23:41:38 +0000 (16:41 -0700)]
Merge tag 'powerpc-4.3-1' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc updates from Michael Ellerman:

 - support "hybrid" iommu/direct DMA ops for coherent_mask < dma_mask
   from Benjamin Herrenschmidt

 - EEH fixes for SRIOV from Gavin

 - introduce rtas_get_sensor_fast() for IRQ handlers from Thomas Huth

 - use hardware RNG for arch_get_random_seed_* not arch_get_random_*
   from Paul Mackerras

 - seccomp filter support from Michael Ellerman

 - opal_cec_reboot2() handling for HMIs & machine checks from Mahesh
   Salgaonkar

 - add powerpc timebase as a trace clock source from Naveen N.  Rao

 - misc cleanups in the xmon, signal & SLB code from Anshuman Khandual

 - add an inline function to update POWER8 HID0 from Gautham R.  Shenoy

 - fix pte_pagesize_index() crash on 4K w/64K hash from Michael Ellerman

 - drop support for 64K local store on 4K kernels from Michael Ellerman

 - move dma_get_required_mask() from pnv_phb to pci_controller_ops from
   Andrew Donnellan

 - initialize distance lookup table from drconf path from Nikunj A
   Dadhania

 - enable RTC class support from Vaibhav Jain

 - disable automatically blocked PCI config from Gavin Shan

 - add LEDs driver for PowerNV platform from Vasant Hegde

 - fix endianness issues in the HVSI driver from Laurent Dufour

 - kexec endian fixes from Samuel Mendoza-Jonas

 - fix corrupted pdn list from Gavin Shan

 - fix fenced PHB caused by eeh_slot_error_detail() from Gavin Shan

 - Freescale updates from Scott: Highlights include 32-bit memcpy/memset
   optimizations, checksum optimizations, 85xx config fragments and
   updates, device tree updates, e6500 fixes for non-SMP, and misc
   cleanup and minor fixes.

 - a ton of cxl updates & fixes:
    - add explicit precision specifiers from Rasmus Villemoes
    - use more common format specifier from Rasmus Villemoes
    - destroy cxl_adapter_idr on module_exit from Johannes Thumshirn
    - destroy afu->contexts_idr on release of an afu from Johannes
      Thumshirn
    - compile with -Werror from Daniel Axtens
    - EEH support from Daniel Axtens
    - plug irq_bitmap getting leaked in cxl_context from Vaibhav Jain
    - add alternate MMIO error handling from Ian Munsie
    - allow release of contexts which have been OPENED but not STARTED
      from Andrew Donnellan
    - remove use of macro DEFINE_PCI_DEVICE_TABLE from Vaishali Thakkar
    - release irqs if memory allocation fails from Vaibhav Jain
    - remove racy attempt to force EEH invocation in reset from Daniel
      Axtens
    - fix + cleanup error paths in cxl_dev_context_init from Ian Munsie
    - fix force unmapping mmaps of contexts allocated through the kernel
      api from Ian Munsie
    - set up and enable PSL Timebase from Philippe Bergheaud

* tag 'powerpc-4.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (140 commits)
  cxl: Set up and enable PSL Timebase
  cxl: Fix force unmapping mmaps of contexts allocated through the kernel api
  cxl: Fix + cleanup error paths in cxl_dev_context_init
  powerpc/eeh: Fix fenced PHB caused by eeh_slot_error_detail()
  powerpc/pseries: Cleanup on pci_dn_reconfig_notifier()
  powerpc/pseries: Fix corrupted pdn list
  powerpc/powernv: Enable LEDS support
  powerpc/iommu: Set default DMA offset in dma_dev_setup
  cxl: Remove racy attempt to force EEH invocation in reset
  cxl: Release irqs if memory allocation fails
  cxl: Remove use of macro DEFINE_PCI_DEVICE_TABLE
  powerpc/powernv: Fix mis-merge of OPAL support for LEDS driver
  powerpc/powernv: Reset HILE before kexec_sequence()
  powerpc/kexec: Reset secondary cpu endianness before kexec
  powerpc/hvsi: Fix endianness issues in the HVSI driver
  leds/powernv: Add driver for PowerNV platform
  powerpc/powernv: Create LED platform device
  powerpc/powernv: Add OPAL interfaces for accessing and modifying system LED states
  powerpc/powernv: Fix the log message when disabling VF
  cxl: Allow release of contexts which have been OPENED but not STARTED
  ...

9 years agoMerge branch 'pcmcia' of git://ftp.arm.linux.org.uk/~rmk/linux-arm
Linus Torvalds [Thu, 3 Sep 2015 23:36:33 +0000 (16:36 -0700)]
Merge branch 'pcmcia' of git://ftp.arm.linux.org.uk/~rmk/linux-arm

Pull ARM pcmcia updates from Russell King:
 "A series of changes updating the PXA and SA11x0 PCMCIA code to use
  devm_* APIs, and resolve some resource leaks in doing so.  This
  results in a few small cleanups which are included in this set.

  FYI, the recommit of these today is to add Robert Jarzmik's
  reviewed-by tags, which I'd forgotten to add from mid-July"

* 'pcmcia' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  pcmcia: soc_common: remove skt_dev_info's clk pointer
  pcmcia: sa11xx_base.c: remove useless init/exit functions
  pcmcia: sa1111: simplify clk handing in sa1111_pcmcia_add()
  pcmcia: sa1111: update socket driver to use devm_clk_get() API
  pcmcia: pxa2xx: convert memory allocation to devm_* API
  pcmcia: pxa2xx: update socket driver to use devm_clk_get() API
  pcmcia: sa11x0: convert memory allocation to devm_* API
  pcmcia: sa11x0: fix missing clk_put() in sa11x0 socket drivers

9 years agoMerge branch 'for-linus' of git://ftp.arm.linux.org.uk/~rmk/linux-arm
Linus Torvalds [Thu, 3 Sep 2015 23:27:01 +0000 (16:27 -0700)]
Merge branch 'for-linus' of git://ftp.arm.linux.org.uk/~rmk/linux-arm

Pull ARM development updates from Russell King:
 "Included in this update:

   - moving PSCI code from ARM64/ARM to drivers/

   - removal of some architecture internals from global kernel view

   - addition of software based "privileged no access" support using the
     old domains register to turn off the ability for kernel
     loads/stores to access userspace.  Only the proper accessors will
     be usable.

   - addition of early fixup support for early console

   - re-addition (and reimplementation) of OMAP special interconnect
     barrier

   - removal of finish_arch_switch()

   - only expose cpuX/online in sysfs if hotpluggable

   - a number of code cleanups"

* 'for-linus' of git://ftp.arm.linux.org.uk/~rmk/linux-arm: (41 commits)
  ARM: software-based priviledged-no-access support
  ARM: entry: provide uaccess assembly macro hooks
  ARM: entry: get rid of multiple macro definitions
  ARM: 8421/1: smp: Collapse arch_cpu_idle_dead() into cpu_die()
  ARM: uaccess: provide uaccess_save_and_enable() and uaccess_restore()
  ARM: mm: improve do_ldrd_abort macro
  ARM: entry: ensure that IRQs are enabled when calling syscall_trace_exit()
  ARM: entry: efficiency cleanups
  ARM: entry: get rid of asm_trace_hardirqs_on_cond
  ARM: uaccess: simplify user access assembly
  ARM: domains: remove DOMAIN_TABLE
  ARM: domains: keep vectors in separate domain
  ARM: domains: get rid of manager mode for user domain
  ARM: domains: move initial domain setting value to asm/domains.h
  ARM: domains: provide domain_mask()
  ARM: domains: switch to keeping domain value in register
  ARM: 8419/1: dma-mapping: harmonize definition of DMA_ERROR_CODE
  ARM: 8417/1: refactor bitops functions with BIT_MASK() and BIT_WORD()
  ARM: 8416/1: Feroceon: use of_iomap() to map register base
  ARM: 8415/1: early fixmap support for earlycon
  ...

9 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 3 Sep 2015 23:15:41 +0000 (16:15 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar:
 "Tooling fixes plus a handful of late arriving tooling changes"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf tools: Fix link time error with sample_reg_masks on non x86
  perf build: Fix Intel PT instruction decoder dependency problem
  perf dwarf: Fix potential array out of bounds access
  perf record: Add ability to name registers to record
  perf/x86: Add list of register names
  perf script: Enable printing of interrupted machine state
  perf evlist: Open event on evsel cpus and threads
  bpf tools: New API to get name from a BPF object
  perf tools: Fix build on powerpc broken by pt/bts

9 years agoMerge branch 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 3 Sep 2015 22:46:07 +0000 (15:46 -0700)]
Merge branch 'locking-core-for-linus' of git://git./linux/kernel/git/tip/tip

Pull locking and atomic updates from Ingo Molnar:
 "Main changes in this cycle are:

   - Extend atomic primitives with coherent logic op primitives
     (atomic_{or,and,xor}()) and deprecate the old partial APIs
     (atomic_{set,clear}_mask())

     The old ops were incoherent with incompatible signatures across
     architectures and with incomplete support.  Now every architecture
     supports the primitives consistently (by Peter Zijlstra)

   - Generic support for 'relaxed atomics':

       - _acquire/release/relaxed() flavours of xchg(), cmpxchg() and {add,sub}_return()
       - atomic_read_acquire()
       - atomic_set_release()

     This came out of porting qwrlock code to arm64 (by Will Deacon)

   - Clean up the fragile static_key APIs that were causing repeat bugs,
     by introducing a new one:

       DEFINE_STATIC_KEY_TRUE(name);
       DEFINE_STATIC_KEY_FALSE(name);

     which define a key of different types with an initial true/false
     value.

     Then allow:

       static_branch_likely()
       static_branch_unlikely()

     to take a key of either type and emit the right instruction for the
     case.  To be able to know the 'type' of the static key we encode it
     in the jump entry (by Peter Zijlstra)

   - Static key self-tests (by Jason Baron)

   - qrwlock optimizations (by Waiman Long)

   - small futex enhancements (by Davidlohr Bueso)

   - ... and misc other changes"

* 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (63 commits)
  jump_label/x86: Work around asm build bug on older/backported GCCs
  locking, ARM, atomics: Define our SMP atomics in terms of _relaxed() operations
  locking, include/llist: Use linux/atomic.h instead of asm/cmpxchg.h
  locking/qrwlock: Make use of _{acquire|release|relaxed}() atomics
  locking/qrwlock: Implement queue_write_unlock() using smp_store_release()
  locking/lockref: Remove homebrew cmpxchg64_relaxed() macro definition
  locking, asm-generic: Add _{relaxed|acquire|release}() variants for 'atomic_long_t'
  locking, asm-generic: Rework atomic-long.h to avoid bulk code duplication
  locking/atomics: Add _{acquire|release|relaxed}() variants of some atomic operations
  locking, compiler.h: Cast away attributes in the WRITE_ONCE() magic
  locking/static_keys: Make verify_keys() static
  jump label, locking/static_keys: Update docs
  locking/static_keys: Provide a selftest
  jump_label: Provide a self-test
  s390/uaccess, locking/static_keys: employ static_branch_likely()
  x86, tsc, locking/static_keys: Employ static_branch_likely()
  locking/static_keys: Add selftest
  locking/static_keys: Add a new static_key interface
  locking/static_keys: Rework update logic
  locking/static_keys: Add static_key_{en,dis}able() helpers
  ...

9 years agoMerge tag 'for-f2fs-4.3' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk...
Linus Torvalds [Thu, 3 Sep 2015 20:10:22 +0000 (13:10 -0700)]
Merge tag 'for-f2fs-4.3' of git://git./linux/kernel/git/jaegeuk/f2fs

Pull f2fs updates from Jaegeuk Kim:
 "The major work includes fixing and enhancing the existing extent_cache
  feature, which has been well settling down so far and now it becomes a
  default mount option accordingly.

  Also, this version newly registers a f2fs memory shrinker to reclaim
  several objects consumed by a couple of data structures in order to
  avoid memory pressures.

  Another new feature is to add ioctl(F2FS_GARBAGE_COLLECT) which
  triggers a cleaning job explicitly by users.

  Most of the other patches are to fix bugs occurred in the corner cases
  across the whole code area"

* tag 'for-f2fs-4.3' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (85 commits)
  f2fs: upset segment_info repair
  f2fs: avoid accessing NULL pointer in f2fs_drop_largest_extent
  f2fs: update extent tree in batches
  f2fs: fix to release inode correctly
  f2fs: handle f2fs_truncate error correctly
  f2fs: avoid unneeded initializing when converting inline dentry
  f2fs: atomically set inode->i_flags
  f2fs: fix wrong pointer access during try_to_free_nids
  f2fs: use __GFP_NOFAIL to avoid infinite loop
  f2fs: lookup neighbor extent nodes for merging later
  f2fs: split __insert_extent_tree_ret for readability
  f2fs: kill dead code in __insert_extent_tree
  f2fs: adjust showing of extent cache stat
  f2fs: add largest/cached stat in extent cache
  f2fs: fix incorrect mapping for bmap
  f2fs: add annotation for space utilization of regular/inline dentry
  f2fs: fix to update cached_en of extent tree properly
  f2fs: fix typo
  f2fs: check the node block address of newly allocated nid
  f2fs: go out for insert_inode_locked failure
  ...

9 years agoMerge tag 'dlm-4.3' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm
Linus Torvalds [Thu, 3 Sep 2015 19:57:48 +0000 (12:57 -0700)]
Merge tag 'dlm-4.3' of git://git./linux/kernel/git/teigland/linux-dlm

Pull dlm updates from David Teigland:
 "This set mainly includes a change to the way the dlm uses the SCTP API
  in the kernel, removing the direct dependency on the sctp module.
  Other odd SCTP-related fixes are also included.

  The other notable fix is for a long standing regression in the
  behavior of lock value blocks for user space locks"

* tag 'dlm-4.3' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm:
  dlm: print error from kernel_sendpage
  dlm: fix lvb copy for user locks
  dlm: sctp_accept_from_sock() can be static
  dlm: fix reconnecting but not sending data
  dlm: replace BUG_ON with a less severe handling
  dlm: use sctp 1-to-1 API
  dlm: fix not reconnecting on connecting error handling
  dlm: fix race while closing connections
  dlm: fix connection stealing if using SCTP

9 years agoMerge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso...
Linus Torvalds [Thu, 3 Sep 2015 19:52:19 +0000 (12:52 -0700)]
Merge tag 'ext4_for_linus' of git://git./linux/kernel/git/tytso/ext4

Pull ext4 updates from Ted Ts'o:
 "Pretty much all bug fixes and clean ups for 4.3, after a lot of
  features and other churn going into 4.2"

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  Revert "ext4: remove block_device_ejected"
  ext4: ratelimit the file system mounted message
  ext4: silence a format string false positive
  ext4: simplify some code in read_mmp_block()
  ext4: don't manipulate recovery flag when freezing no-journal fs
  jbd2: limit number of reserved credits
  ext4 crypto: remove duplicate header file
  ext4: update c/mtime on truncate up
  jbd2: avoid infinite loop when destroying aborted journal
  ext4, jbd2: add REQ_FUA flag when recording an error in the superblock
  ext4 crypto: fix spelling typo in comment
  ext4 crypto: exit cleanly if ext4_derive_key_aes() fails
  ext4: reject journal options for ext2 mounts
  ext4: implement cgroup writeback support
  ext4: replace ext4_io_submit->io_op with ->io_wbc
  ext4 crypto: check for too-short encrypted file names
  ext4 crypto: use a jbd2 transaction when adding a crypto policy
  jbd2: speedup jbd2_journal_dirty_metadata()

9 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Linus Torvalds [Thu, 3 Sep 2015 19:28:30 +0000 (12:28 -0700)]
Merge branch 'for_linus' of git://git./linux/kernel/git/jack/linux-fs

Pull ext3 removal, quota & udf fixes from Jan Kara:
 "The biggest change in the pull is the removal of ext3 filesystem
  driver (~28k lines removed).  Ext4 driver is a full featured
  replacement these days and both RH and SUSE use it for several years
  without issues.  Also there are some workarounds in VM & block layer
  mainly for ext3 which we could eventually get rid of.

  Other larger change is addition of proper error handling for
  dquot_initialize().  The rest is small fixes and cleanups"

[ I wasn't convinced about the ext3 removal and worried about things
  falling through the cracks for legacy users, but ext4 maintainers
  piped up and were all unanimously in favor of removal, and maintaining
  all legacy ext3 support inside ext4.   - Linus ]

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  udf: Don't modify filesystem for read-only mounts
  quota: remove an unneeded condition
  ext4: memory leak on error in ext4_symlink()
  mm/Kconfig: NEED_BOUNCE_POOL: clean-up condition
  ext4: Improve ext4 Kconfig test
  block: Remove forced page bouncing under IO
  fs: Remove ext3 filesystem driver
  doc: Update doc about journalling layer
  jfs: Handle error from dquot_initialize()
  reiserfs: Handle error from dquot_initialize()
  ocfs2: Handle error from dquot_initialize()
  ext4: Handle error from dquot_initialize()
  ext2: Handle error from dquot_initalize()
  quota: Propagate error from ->acquire_dquot()

9 years agoMerge branch 'hpfs' (patches from Mikulas)
Linus Torvalds [Thu, 3 Sep 2015 18:55:55 +0000 (11:55 -0700)]
Merge branch 'hpfs' (patches from Mikulas)

Merge hpfs upddate from Mikulas Patocka.

* emailed patches from Mikulas Patocka <mikulas@twibright.com>:
  hpfs: update ctime and mtime on directory modification
  hpfs: support hotfixes

9 years agohpfs: update ctime and mtime on directory modification
Mikulas Patocka [Wed, 2 Sep 2015 20:51:53 +0000 (22:51 +0200)]
hpfs: update ctime and mtime on directory modification

Update ctime and mtime when a directory is modified. (though OS/2 doesn't
update them anyway)

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@kernel.org # v3.3+
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agohpfs: support hotfixes
Mikulas Patocka [Wed, 2 Sep 2015 20:50:12 +0000 (22:50 +0200)]
hpfs: support hotfixes

When the OS/2 driver hits a disk write error, it writes the sector to
another location and adds the sector mapping to the hotfix map.

This patch makes the hpfs driver understand the hotfix map and remap
accesses accoring to it.

Signed-off-by: Mikulas Patocka <mikulas@twibright.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next
Linus Torvalds [Thu, 3 Sep 2015 15:08:17 +0000 (08:08 -0700)]
Merge git://git./linux/kernel/git/davem/net-next

Pull networking updates from David Miller:
 "Another merge window, another set of networking changes.  I've heard
  rumblings that the lightweight tunnels infrastructure has been voted
  networking change of the year.  But what do I know?

   1) Add conntrack support to openvswitch, from Joe Stringer.

   2) Initial support for VRF (Virtual Routing and Forwarding), which
      allows the segmentation of routing paths without using multiple
      devices.  There are some semantic kinks to work out still, but
      this is a reasonably strong foundation.  From David Ahern.

   3) Remove spinlock fro act_bpf fast path, from Alexei Starovoitov.

   4) Ignore route nexthops with a link down state in ipv6, just like
      ipv4.  From Andy Gospodarek.

   5) Remove spinlock from fast path of act_gact and act_mirred, from
      Eric Dumazet.

   6) Document the DSA layer, from Florian Fainelli.

   7) Add netconsole support to bcmgenet, systemport, and DSA.  Also
      from Florian Fainelli.

   8) Add Mellanox Switch Driver and core infrastructure, from Jiri
      Pirko.

   9) Add support for "light weight tunnels", which allow for
      encapsulation and decapsulation without bearing the overhead of a
      full blown netdevice.  From Thomas Graf, Jiri Benc, and a cast of
      others.

  10) Add Identifier Locator Addressing support for ipv6, from Tom
      Herbert.

  11) Support fragmented SKBs in iwlwifi, from Johannes Berg.

  12) Allow perf PMUs to be accessed from eBPF programs, from Kaixu Xia.

  13) Add BQL support to 3c59x driver, from Loganaden Velvindron.

  14) Stop using a zero TX queue length to mean that a device shouldn't
      have a qdisc attached, use an explicit flag instead.  From Phil
      Sutter.

  15) Use generic geneve netdevice infrastructure in openvswitch, from
      Pravin B Shelar.

  16) Add infrastructure to avoid re-forwarding a packet in software
      that was already forwarded by a hardware switch.  From Scott
      Feldman.

  17) Allow AF_PACKET fanout function to be implemented in a bpf
      program, from Willem de Bruijn"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1458 commits)
  netfilter: nf_conntrack: make nf_ct_zone_dflt built-in
  netfilter: nf_dup{4, 6}: fix build error when nf_conntrack disabled
  net: fec: clear receive interrupts before processing a packet
  ipv6: fix exthdrs offload registration in out_rt path
  xen-netback: add support for multicast control
  bgmac: Update fixed_phy_register()
  sock, diag: fix panic in sock_diag_put_filterinfo
  flow_dissector: Use 'const' where possible.
  flow_dissector: Fix function argument ordering dependency
  ixgbe: Resolve "initialized field overwritten" warnings
  ixgbe: Remove bimodal SR-IOV disabling
  ixgbe: Add support for reporting 2.5G link speed
  ixgbe: fix bounds checking in ixgbe_setup_tc for 82598
  ixgbe: support for ethtool set_rxfh
  ixgbe: Avoid needless PHY access on copper phys
  ixgbe: cleanup to use cached mask value
  ixgbe: Remove second instance of lan_id variable
  ixgbe: use kzalloc for allocating one thing
  flow: Move __get_hash_from_flowi{4,6} into flow_dissector.c
  ixgbe: Remove unused PCI bus types
  ...

9 years agopcmcia: soc_common: remove skt_dev_info's clk pointer
Russell King [Thu, 26 Mar 2015 11:01:04 +0000 (11:01 +0000)]
pcmcia: soc_common: remove skt_dev_info's clk pointer

We no longer need to store the clk pointer in struct skt_dev_info as we
no longer need to remember the clk pointer for the cleanup paths.

Reviewed-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
9 years agopcmcia: sa11xx_base.c: remove useless init/exit functions
Russell King [Thu, 26 Mar 2015 10:57:48 +0000 (10:57 +0000)]
pcmcia: sa11xx_base.c: remove useless init/exit functions

A library module is not required to have module init/exit functions.
Get rid of these unnecessary functions.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
9 years agopcmcia: sa1111: simplify clk handing in sa1111_pcmcia_add()
Russell King [Thu, 26 Mar 2015 10:46:35 +0000 (10:46 +0000)]
pcmcia: sa1111: simplify clk handing in sa1111_pcmcia_add()

clk_get(dev, NULL) will always refer to the same clock, so it's
pointless calling this multiple times for the same device.  As we no
longer have to worry about the cleanup (via use of devm_clk_get()) we
can simplify sa1111_pcmcia_add() too.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
9 years agopcmcia: sa1111: update socket driver to use devm_clk_get() API
Russell King [Thu, 26 Mar 2015 10:27:08 +0000 (10:27 +0000)]
pcmcia: sa1111: update socket driver to use devm_clk_get() API

Update the pxa2xx socket driver to use the devm_clk_get() API so that
the cleanup paths are simplified.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
9 years agopcmcia: pxa2xx: convert memory allocation to devm_* API
Russell King [Thu, 26 Mar 2015 10:33:16 +0000 (10:33 +0000)]
pcmcia: pxa2xx: convert memory allocation to devm_* API

Convert the pxa2xx socket driver memory allocation to use devm_kzalloc()
to simplify the cleanup path.

Reviewed-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
9 years agopcmcia: pxa2xx: update socket driver to use devm_clk_get() API
Russell King [Thu, 26 Mar 2015 10:27:08 +0000 (10:27 +0000)]
pcmcia: pxa2xx: update socket driver to use devm_clk_get() API

Update the pxa2xx socket driver to use the devm_clk_get() API so that
the cleanup paths are simplified.

Reviewed-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
9 years agoMerge branch 'devel-stable' into for-linus
Russell King [Thu, 3 Sep 2015 14:28:50 +0000 (15:28 +0100)]
Merge branch 'devel-stable' into for-linus

Conflicts:
drivers/perf/arm_pmu.c

9 years agoMerge branches 'cleanup', 'fixes', 'misc', 'omap-barrier' and 'uaccess' into for...
Russell King [Thu, 3 Sep 2015 14:28:37 +0000 (15:28 +0100)]
Merge branches 'cleanup', 'fixes', 'misc', 'omap-barrier' and 'uaccess' into for-linus

9 years agoMerge branch '4.2-fixes' into mips-for-linux-next
Ralf Baechle [Thu, 3 Sep 2015 12:06:33 +0000 (14:06 +0200)]
Merge branch '4.2-fixes' into mips-for-linux-next

9 years agoMIPS: UAPI: Fix unrecognized opcode WSBH/DSBH/DSHD when using MIPS16.
Yousong Zhou [Thu, 3 Sep 2015 09:47:45 +0000 (17:47 +0800)]
MIPS: UAPI: Fix unrecognized opcode WSBH/DSBH/DSHD when using MIPS16.

The nomips16 has to be added both as function attribute and assembler
directive.

When only function attribute is specified, the compiler will inline the
function with -Os optimization.  The generated assembly code cannot be
correctly assembled because ISA mode switch has to be done through jump
instruction.

When only ".set nomips16" directive is used, the generated assembly code
will use MIPS32 code for the inline assembly template and MIPS16 for the
function return.  The compiled binary is invalid:

    00403100 <__arch_swab16>:
      403100:   7c0410a0    wsbh    v0,a0
      403104:   e820ea31    swc2    $0,-5583(at)

while correct code should be:

    00402650 <__arch_swab16>:
      402650:   7c0410a0    wsbh    v0,a0
      402654:   03e00008    jr  ra
      402658:   3042ffff    andi    v0,v0,0xffff

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
Cc: Chen Jie <chenj@lemote.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/11087/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Fix alignment of quiet build output for vmlinuz link
Alex Smith [Tue, 18 Aug 2015 14:03:10 +0000 (15:03 +0100)]
MIPS: Fix alignment of quiet build output for vmlinuz link

The "LD vmlinuz" line in the quiet build output is misaligned with the
rest of the output. Fix this.

Signed-off-by: Alex Smith <alex.smith@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/11019/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Remove unused handle_dsemul function declaration
Markos Chandras [Wed, 12 Aug 2015 09:14:35 +0000 (10:14 +0100)]
MIPS: math-emu: Remove unused handle_dsemul function declaration

handle_dsemul does not exist and it's not being used in the code at all
so remove its declaration. The deliberate DS emulation exception is
handled by the do_dsemulret C code.

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10950/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the MIPS R6 MAX{, A} FPU instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:36 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the MIPS R6 MAX{, A} FPU instruction

MIPS R6 introduced the following instruction:
Scalar Floating-Point Maximum and
Scalar Floating-Point argument with Maximum Absolute Value
MAX.fmt writes the maximum value of the inputs fs and ft to the
destination fd.
MAXA.fmt takes input arguments fs and ft and writes the argument with
the maximum absolute value to the destination fd.

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10961/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the MIPS R6 MIN{, A} FPU instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:35 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the MIPS R6 MIN{, A} FPU instruction

MIPS R6 introduced the following instruction:
Scalar Floating-Point Minimum and
Scalar Floating-Point argument with Minimum Absolute Value

MIN.fmt writes the minimum value of the inputs fs and ft to the
destination fd.
MINA.fmt takes input arguments fs and ft and writes the argument with
the minimum absolute value to the destination fd.

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10960/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the MIPS R6 CLASS FPU instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:34 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the MIPS R6 CLASS FPU instruction

MIPS R6 introduced the following instruction:
Stores in fd a bit mask reflecting the floating-point class of the
floating point scalar value fs.

CLASS.fmt: FPR[fd] = class(FPR[fs])

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10959/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the MIPS R6 RINT FPU instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:33 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the MIPS R6 RINT FPU instruction

MIPS R6 introduced the following instruction:
Floating-Point Round to Integral
Scalar floating-point round to integral floating point value.

RINT.fmt: FPR[fd] = round_int(FPR[fs])

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10958/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the MIPS R6 MSUBF FPU instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:32 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the MIPS R6 MSUBF FPU instruction

MIPS R6 introduced the following instruction:
Floating Point Fused Multiply Subtract:
MSUBF.fmt To perform a fused multiply-subtract of FP values.

MSUBF.fmt: FPR[fd] = FPR[fd] - (FPR[fs] x FPR[ft])

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10957/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the MIPS R6 MADDF FPU instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:31 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the MIPS R6 MADDF FPU instruction

MIPS R6 introduced the following instruction:
Floating Point Fused Multiply Add:
MADDF.fmt To perform a fused multiply-add of FP values.

MADDF.fmt: FPR[fd] = FPR[fd] + (FPR[fs] x FPR[ft])

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10956/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the MIPS R6 SELNEZ FPU instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:30 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the MIPS R6 SELNEZ FPU instruction

MIPS R6 introduced the following instruction:
SELNEZ.fmt: FPR[fd]  FPR[ft].bit0 ? FPR[fs] : 0

Add support for emulating the single and double precision
formats of the said instruction.

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10955/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the MIPS R6 SELEQZ FPU instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:29 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the MIPS R6 SELEQZ FPU instruction

MIPS R6 introduced the following instruction:
SELEQZ.fmt: FPR[fd]  FPR[ft].bit0 ? 0 : FPR[fs]

Add support for emulating the single and double precision formats
of the said instruction.

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10954/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: math-emu: Add support for the CMP.condn.fmt R6 instruction
Markos Chandras [Thu, 13 Aug 2015 07:56:28 +0000 (09:56 +0200)]
MIPS: math-emu: Add support for the CMP.condn.fmt R6 instruction

Add support for emulating the new CMP.condn.fmt R6 instructions and
return SIGILL for the old C.cond.fmt if R2 emulation is not enabled
since it's not supported by R6.

The functionality of the new CMP.condn.fmt is the following one:

If the comparison specified by the condn field of the instruction
is true for the operand values, the result is true; otherwise, the
result is false. If no exception is taken, the result is written into
FPR fd; true is all 1s and false is all 0s repeated the operand width
of fmt. All other bits beyond the operand width fmt are UNPREDICTABLE.

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10953/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: inst.h: Add new MIPS R6 FPU opcodes
Markos Chandras [Thu, 13 Aug 2015 07:56:27 +0000 (09:56 +0200)]
MIPS: inst.h: Add new MIPS R6 FPU opcodes

Add opcodes for the new MIPS R6 FPU instructions.

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10952/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Octeon: Fix management port MII address on Kontron S1901
Aaro Koskinen [Tue, 11 Aug 2015 07:56:28 +0000 (10:56 +0300)]
MIPS: Octeon: Fix management port MII address on Kontron S1901

Management port MII address is incorrect on Kontron S1901 resulting
in broken networking. Fix by providing definitions for the in-tree DT
pruning code.

Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10914/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: BCM47xx: Use kmemdup rather than duplicating its implementation
Andrzej Hajda [Fri, 7 Aug 2015 07:59:10 +0000 (09:59 +0200)]
MIPS: BCM47xx: Use kmemdup rather than duplicating its implementation

The patch was generated using fixed coccinelle semantic patch
scripts/coccinelle/api/memdup.cocci [1].

[1]: http://permalink.gmane.org/gmane.linux.kernel/2014320

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
Cc: Hauke Mehrtens <hauke@hauke-m.de>
Cc: Rafał Miłecki <zajec5@gmail.com>
Cc: Andrzej Hajda <a.hajda@samsung.com>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: linux-kernel@vger.kernel.org
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10898/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoSTAGING: Octeon: Use common helpers for determining interface and port
Janne Huttunen [Thu, 13 Aug 2015 13:21:46 +0000 (16:21 +0300)]
STAGING: Octeon: Use common helpers for determining interface and port

Currently the Octeon Ethernet driver hardcodes the mapping between
interface/port and IPD port number. Since we have generic helpers for
the very same purpose, we might as well use them instead. This prevents
having the same information in multiple places.

Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10975/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Octeon: Support interfaces 4 and 5
Janne Huttunen [Thu, 13 Aug 2015 13:21:45 +0000 (16:21 +0300)]
MIPS: Octeon: Support interfaces 4 and 5

Add the support for mapping between interface/port numbers and IPD port
numbers also for the additional interfaces some Octeon II models have.

Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10967/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Octeon: Set up 1:1 mapping between CN68XX PKO queues and ports
Janne Huttunen [Thu, 13 Aug 2015 13:21:44 +0000 (16:21 +0300)]
MIPS: Octeon: Set up 1:1 mapping between CN68XX PKO queues and ports

Use the internal port number also as the queue number on CN68XX.

Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10962/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Octeon: Initialize CN68XX PKO
Janne Huttunen [Thu, 13 Aug 2015 13:21:43 +0000 (16:21 +0300)]
MIPS: Octeon: Initialize CN68XX PKO

CN68XX requires a different PKO configuration. This patch provides
just enough setup to get the XAUI interfaces on CN6880 working with
default parameters.

Signed-off-by: Janne Huttunen <janne.huttunen@nsn.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10974/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoSTAGING: Octeon: Support CN68XX style WQE
Janne Huttunen [Thu, 13 Aug 2015 13:21:42 +0000 (16:21 +0300)]
STAGING: Octeon: Support CN68XX style WQE

CN68XX has a bit different WQE structure. This patch provides the new
definitions and converts the code to use the proper variant based on
the actual model.

Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10973/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoSTAGING: Octeon: Increase output command buffers
Janne Huttunen [Thu, 13 Aug 2015 13:21:41 +0000 (16:21 +0300)]
STAGING: Octeon: Increase output command buffers

The Octeon II models have more interfaces and thus require more output
command buffers. Increase the allocation to support these models.

Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10965/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoSTAGING: Octeon: Set SSO group mask properly on CN68XX
Aaro Koskinen [Thu, 13 Aug 2015 13:21:40 +0000 (16:21 +0300)]
STAGING: Octeon: Set SSO group mask properly on CN68XX

CN68XX uses SSO instead of POW.

Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10966/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoSTAGING: Octeon: Properly enable/disable SSO WQE interrupts
Aaro Koskinen [Thu, 13 Aug 2015 13:21:39 +0000 (16:21 +0300)]
STAGING: Octeon: Properly enable/disable SSO WQE interrupts

The Octeon models with SSO instead of POW need to use a different register
for configuring the WQE interrupt thresholds.

Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10964/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Octeon: Add definitions for setting up SSO
Janne Huttunen [Thu, 13 Aug 2015 13:21:38 +0000 (16:21 +0300)]
MIPS: Octeon: Add definitions for setting up SSO

Some Octeon II models have SSO instead of POW and use a different register
for setting the interrupt thresholds. Add the necessary definitions for
configuring the interrupts also on those models.

Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10972/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Octeon: Configure minimum PKO packet sizes on CN68XX
Janne Huttunen [Thu, 13 Aug 2015 13:21:37 +0000 (16:21 +0300)]
MIPS: Octeon: Configure minimum PKO packet sizes on CN68XX

CN68XX has common minimum packet size filters that need to be configured
for the traffic to work. Just set them to a default value.

Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10963/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Octeon: Configure XAUI pkinds
Janne Huttunen [Thu, 13 Aug 2015 13:21:36 +0000 (16:21 +0300)]
MIPS: Octeon: Configure XAUI pkinds

Configure the pkinds of XAUI interfaces on Octeon models that have
them. This simple configuration uses 1:1 mapping between the PIP input
port number and the selected pkind.

Signed-off-by: Janne Huttunen <janne.huttunen@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Acked-by: David Daney <david.daney@cavium.com>
Cc: David Daney <ddaney.cavm@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: Janne Huttunen <janne.huttunen@nokia.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: devel@driverdev.osuosl.org
Patchwork: https://patchwork.linux-mips.org/patch/10971/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>