linux.git/fs/gfs2, branch v3.17

GFS2: fix d_splice_alias() misuses

2014-09-12T19:58:55+00:00

Callers of d_splice_alias(dentry, inode) don't need iput(), neither
on success nor on failure.  Either the reference to inode is stored
in a previously negative dentry, or it's dropped.  In either case
inode reference the caller used to hold is consumed.

__gfs2_lookup() does iput() in case when d_splice_alias() has failed.
Double iput() if we ever hit that.  And gfs2_create_inode() ends up
not only with double iput(), but with link count dropped to zero - on
an inode it has just found in directory.

Cc: stable@vger.kernel.org # v3.14+
Signed-off-by: Al Viro 
Signed-off-by: Steven Whitehouse

GFS2: Don't use MAXQUOTAS value

2014-09-11T09:59:56+00:00

MAXQUOTAS value defines maximum number of quota types VFS supports.
This isn't necessarily the number of types gfs2 supports and with
addition of project quotas these two numbers stop matching. So make gfs2
use its private definition.

CC: cluster-devel@redhat.com
Signed-off-by: Jan Kara 
Signed-off-by: Steven Whitehouse

GFS2: Hash the negative dentry during inode lookup

2014-09-11T09:54:43+00:00

Fix a regression introduced by:
6d4ade986f9c8df31e68 GFS2: Add atomic_open support
where an early return misses d_splice_alias() which had been
adding the negative dentry.

Signed-off-by: Benjamin Coddington 
Signed-off-by: Bob Peterson 
Signed-off-by: Steven Whitehouse

GFS2: Request demote when a "try" flock fails

2014-08-21T09:22:52+00:00

This patch changes the flock code so that it uses the TRY_1CB flag
instead of the TRY flag on the first attempt. That forces any holding
nodes to issue a dlm callback, which requests a demote of the glock.
Then, if the "try" failed, it sleeps a small amount of time for the
demote to occur. Then it tries again, for an increasing amount of time.
Subsequent attempts to gain the "try" lock don't use "_1CB" so that
only one callback is issued.

Signed-off-by: Bob Peterson 
Signed-off-by: Steven Whitehouse

GFS2: Change maxlen variables to size_t

2014-08-21T09:22:23+00:00

This patch changes some variables (especially maxlen in function
gfs2_block_map) from unsigned int to size_t. We need 64-bit arithmetic
for very large files (e.g. 1PB) where the variables otherwise get
shifted to all 0's.

Signed-off-by: Bob Peterson 
Signed-off-by: Steven Whitehouse

GFS2: fs/gfs2/super.c: replace seq_printf by seq_puts

2014-08-21T09:22:05+00:00

fix checkpatch warnings:
"WARNING: Prefer seq_puts to seq_printf"

Cc: cluster-devel@redhat.com
Signed-off-by: Fabian Frederick 
Signed-off-by: Steven Whitehouse

Merge branch 'sched/urgent' into sched/core, to merge fixes before applying new changes

2014-07-28T08:03:00+00:00

Signed-off-by: Ingo Molnar

GFS2: fs/gfs2/rgrp.c: kernel-doc warning fixes

2014-07-18T10:15:14+00:00

Cc: cluster-devel@redhat.com
Signed-off-by: Fabian Frederick 
Signed-off-by: Steven Whitehouse

GFS2: memcontrol: Spelling s/invlidate/invalidate/

2014-07-18T10:14:31+00:00

Signed-off-by: Geert Uytterhoeven 
Cc: cluster-devel@redhat.com
Signed-off-by: Steven Whitehouse

GFS2: Allow caching of glocks for flock

2014-07-18T10:14:12+00:00

This patch removes the GLF_NOCACHE flag from the glocks associated with
flocks. There should be no good reason not to cache glocks for flocks:
they only force the glock to be demoted before they can be reacquired,
which can slow down performance and even cause glock hangs, especially
in cases where the flocks are held in Shared (SH) mode.

Signed-off-by: Bob Peterson 
Signed-off-by: Steven Whitehouse