[Cluster-devel] [PATCH 00/16] GFS2: Pre-pull patch posting (merge window)

[Cluster-devel] [PATCH 00/16] GFS2: Pre-pull patch posting (merge window)

[Bob Peterson]

Hi,

Here is a list of patches we've accumulated for GFS2 for the current upstream
merge window. We have a good mixture this time. Here are some of the features:

1. Fix a problem with RO mounts writing to the journal.
2. Further improvements to quotas on GFS2.
3. Added support for rename2 and RENAME_EXCHANGE on GFS2.
4. Increase performance by making glock lru_list less of a bottleneck.
5. Increase performance by avoiding unnecessary buffer_head releases.
6. Increase performance by using average glock round trip time from all CPUs.
7. Fixes for some compiler warnings and minor white space issues.
8. Other misc. bug fixes

Regards,

Bob Peterson
Red Hat File Systems

----------------------------------------------------------------

Abhi Das (5):
  gfs2: handle NULL rgd in set_rgrp_preferences
  gfs2: fix quota updates on block boundaries
  gfs2: limit quota log messages
  gfs2: s64 cast for negative quota value
  gfs2: Don't support fallocate on jdata	files

Antonio Ospite (1):
  GFS2: inode.c: indent with TABs, not spaces

Benjamin Marzinski (3):
  GFS2: mark the journal idle to fix ro mounts
  GFS2: add support for rename2 and RENAME_EXCHANGE
  GFS2: make sure S_NOSEC flag isn't overwritten

Bob Peterson (4):
  GFS2: Use average srttb value in congestion calculations
  GFS2: Average in only non-zero round-trip times for congestion stats
  GFS2: Don't add all glocks to the lru
  GFS2: Don't brelse rgrp buffer_heads every allocation

Fabian Frederick (3):
  gfs2: convert simple_str to kstr
  gfs2: kerneldoc warning fixes
  gfs2: fix shadow warning in gfs2_rbm_find()

 fs/gfs2/aops.c       |  12 ++-
 fs/gfs2/file.c       |   4 +-
 fs/gfs2/glock.c      |   3 +-
 fs/gfs2/glops.c      |  20 +++--
 fs/gfs2/incore.h     |   2 +
 fs/gfs2/inode.c      | 211 +++++++++++++++++++++++++++++++++++++++++++++-----
 fs/gfs2/ops_fstype.c |   1 +
 fs/gfs2/quota.c      | 212 +++++++++++++++++++++++++++++++--------------------
 fs/gfs2/rgrp.c       |  48 ++++++++----
 fs/gfs2/rgrp.h       |   1 +
 fs/gfs2/sys.c        |  66 +++++++++++-----
 11 files changed, 435 insertions(+), 145 deletions(-)

-- 
2.1.0

[Cluster-devel] [PATCH 01/16] GFS2: Use average srttb value in congestion calculations

[Bob Peterson]

This patch changes function gfs2_rgrp_congested so that it uses an
average srttb (smoothed round trip time for blocking rgrp glocks)
rather than the CPU-specific value. If we use the CPU-specific value
it can incorrectly report no contention when there really is contention
due to the glock processing occurring on a different CPU.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Acked-by: Steven Whitehouse <swhiteho@redhat.com>
---
 fs/gfs2/rgrp.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index 6af2396..f39eedc 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -1850,14 +1850,19 @@ static bool gfs2_rgrp_congested(const struct gfs2_rgrpd *rgd, int loops)
 	const struct gfs2_sbd *sdp = gl->gl_sbd;
 	struct gfs2_lkstats *st;
 	s64 r_dcount, l_dcount;
-	s64 r_srttb, l_srttb;
+	s64 l_srttb, a_srttb = 0;
 	s64 srttb_diff;
 	s64 sqr_diff;
 	s64 var;
+	int cpu;
 
 	preempt_disable();
+	for_each_present_cpu(cpu) {
+		st = &per_cpu_ptr(sdp->sd_lkstats, cpu)->lkstats[LM_TYPE_RGRP];
+		a_srttb += st->stats[GFS2_LKS_SRTTB];
+	}
 	st = &this_cpu_ptr(sdp->sd_lkstats)->lkstats[LM_TYPE_RGRP];
-	r_srttb = st->stats[GFS2_LKS_SRTTB];
+	a_srttb /= num_present_cpus();
 	r_dcount = st->stats[GFS2_LKS_DCOUNT];
 	var = st->stats[GFS2_LKS_SRTTVARB] +
 	      gl->gl_stats.stats[GFS2_LKS_SRTTVARB];
@@ -1866,10 +1871,10 @@ static bool gfs2_rgrp_congested(const struct gfs2_rgrpd *rgd, int loops)
 	l_srttb = gl->gl_stats.stats[GFS2_LKS_SRTTB];
 	l_dcount = gl->gl_stats.stats[GFS2_LKS_DCOUNT];
 
-	if ((l_dcount < 1) || (r_dcount < 1) || (r_srttb == 0))
+	if ((l_dcount < 1) || (r_dcount < 1) || (a_srttb == 0))
 		return false;
 
-	srttb_diff = r_srttb - l_srttb;
+	srttb_diff = a_srttb - l_srttb;
 	sqr_diff = srttb_diff * srttb_diff;
 
 	var *= 2;
-- 
2.1.0

[Cluster-devel] [PATCH 02/16] GFS2: Average in only non-zero round-trip times for congestion stats

[Bob Peterson]

This patch changes function gfs2_rgrp_congested so that it only factors
in non-zero values into its average round trip time. If the round-trip
time is zero for a particular cpu, that cpu has obviously never dealt
with bouncing the resource group in question, so factoring in a zero
value will only skew the numbers. It also fixes a compile error on
some arches related to division.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Acked-by: Steven Whitehouse <swhiteho@redhat.com>
---
 fs/gfs2/rgrp.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index f39eedc..cb27065 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -1854,15 +1854,19 @@ static bool gfs2_rgrp_congested(const struct gfs2_rgrpd *rgd, int loops)
 	s64 srttb_diff;
 	s64 sqr_diff;
 	s64 var;
-	int cpu;
+	int cpu, nonzero = 0;
 
 	preempt_disable();
 	for_each_present_cpu(cpu) {
 		st = &per_cpu_ptr(sdp->sd_lkstats, cpu)->lkstats[LM_TYPE_RGRP];
-		a_srttb += st->stats[GFS2_LKS_SRTTB];
+		if (st->stats[GFS2_LKS_SRTTB]) {
+			a_srttb += st->stats[GFS2_LKS_SRTTB];
+			nonzero++;
+		}
 	}
 	st = &this_cpu_ptr(sdp->sd_lkstats)->lkstats[LM_TYPE_RGRP];
-	a_srttb /= num_present_cpus();
+	if (nonzero)
+		do_div(a_srttb, nonzero);
 	r_dcount = st->stats[GFS2_LKS_DCOUNT];
 	var = st->stats[GFS2_LKS_SRTTVARB] +
 	      gl->gl_stats.stats[GFS2_LKS_SRTTVARB];
-- 
2.1.0

[Cluster-devel] [PATCH 03/16] GFS2: mark the journal idle to fix ro mounts

[Bob Peterson]

From: Benjamin Marzinski <bmarzins@redhat.com>

When gfs2 was mounted read-only and then unmounted, it was writing a
header block to the journal in the syncing gfs2_log_flush() call from
kill_sb(). This is because the journal was not being marked as idle
until the first log header was written out, and on a read-only mount
there never was a log header written out. Since the journal was not
marked idle, gfs2_log_flush() was writing out a header lock to make
sure it was empty during the sync.  Not only did this cause IO to a
read-only filesystem, but the journalling isn't completely initialized
on read-only mounts, and so gfs2 was writing out the wrong sequence
number in the log header.

Now, the journal is marked idle on mount, and gfs2_log_flush() won't
write out anything until there starts being transactions to flush.

Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Acked-by: Steven Whitehouse <swhiteho@redhat.com>
---
 fs/gfs2/ops_fstype.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index efc8e25..e4fdecc 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -756,6 +756,7 @@ static int init_journal(struct gfs2_sbd *sdp, int undo)
 		}
 	}
 
+	sdp->sd_log_idle = 1;
 	set_bit(SDF_JOURNAL_CHECKED, &sdp->sd_flags);
 	gfs2_glock_dq_uninit(&ji_gh);
 	jindex = 0;
-- 
2.1.0

[Cluster-devel] [PATCH 04/16] GFS2: inode.c: indent with TABs, not spaces

[Bob Peterson]

From: Antonio Ospite <ao2@ao2.it>

Follow the same style used for the other functions in the same file.

Signed-off-by: Antonio Ospite <ao2@ao2.it>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/inode.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 08bc84d..2cd385d 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -1227,8 +1227,8 @@ static int gfs2_mknod(struct inode *dir, struct dentry *dentry, umode_t mode,
  */
 
 static int gfs2_atomic_open(struct inode *dir, struct dentry *dentry,
-                            struct file *file, unsigned flags,
-                            umode_t mode, int *opened)
+			    struct file *file, unsigned flags,
+			    umode_t mode, int *opened)
 {
 	struct dentry *d;
 	bool excl = !!(flags & O_EXCL);
-- 
2.1.0

[Cluster-devel] [PATCH 05/16] gfs2: handle NULL rgd in set_rgrp_preferences

[Bob Peterson]

From: Abhi Das <adas@redhat.com>

The function set_rgrp_preferences() does not handle the (rarely
returned) NULL value from gfs2_rgrpd_get_next() and this patch
fixes that.

The fs image in question is only 150MB in size which allows for
only 1 rgrp to be created. The in-memory rb tree has only 1 node
and when gfs2_rgrpd_get_next() is called on this sole rgrp, it
returns NULL. (Default behavior is to wrap around the rb tree and
return the first node to give the illusion of a circular linked
list. In the case of only 1 rgrp, we can't have
gfs2_rgrpd_get_next() return the same rgrp (first, last, next all
point to the same rgrp)... that would cause unintended consequences
and infinite loops.)

Signed-off-by: Abhi Das <adas@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/rgrp.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index cb27065..900e515 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -978,10 +978,10 @@ static void set_rgrp_preferences(struct gfs2_sbd *sdp)
 		rgd->rd_flags |= GFS2_RDF_PREFERRED;
 		for (i = 0; i < sdp->sd_journals; i++) {
 			rgd = gfs2_rgrpd_get_next(rgd);
-			if (rgd == first)
+			if (!rgd || rgd == first)
 				break;
 		}
-	} while (rgd != first);
+	} while (rgd && rgd != first);
 }
 
 /**
-- 
2.1.0

[Cluster-devel] [PATCH 06/16] GFS2: add support for rename2 and RENAME_EXCHANGE

[Bob Peterson]

From: Benjamin Marzinski <bmarzins@redhat.com>

gfs2 now uses the rename2 directory iop, and supports the
RENAME_EXCHANGE flag (as well as RENAME_NOREPLACE, which the vfs
takes care of).

Signed-off-by: Benjamin Marzinski <bmarzins redhat com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/inode.c | 205 +++++++++++++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 189 insertions(+), 16 deletions(-)

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 2cd385d..4d809eb 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -1307,6 +1307,35 @@ static int gfs2_ok_to_move(struct gfs2_inode *this, struct gfs2_inode *to)
 }
 
 /**
+ * update_moved_ino - Update an inode that's being moved
+ * @ip: The inode being moved
+ * @ndip: The parent directory of the new filename
+ * @dir_rename: True of ip is a directory
+ *
+ * Returns: errno
+ */
+
+static int update_moved_ino(struct gfs2_inode *ip, struct gfs2_inode *ndip,
+			    int dir_rename)
+{
+	int error;
+	struct buffer_head *dibh;
+
+	if (dir_rename)
+		return gfs2_dir_mvino(ip, &gfs2_qdotdot, ndip, DT_DIR);
+
+	error = gfs2_meta_inode_buffer(ip, &dibh);
+	if (error)
+		return error;
+	ip->i_inode.i_ctime = CURRENT_TIME;
+	gfs2_trans_add_meta(ip->i_gl, dibh);
+	gfs2_dinode_out(ip, dibh->b_data);
+	brelse(dibh);
+	return 0;
+}
+
+
+/**
  * gfs2_rename - Rename a file
  * @odir: Parent directory of old file name
  * @odentry: The old dentry of the file
@@ -1354,7 +1383,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry,
 
 		if (S_ISDIR(ip->i_inode.i_mode)) {
 			dir_rename = 1;
-			/* don't move a dirctory into it's subdir */
+			/* don't move a directory into its subdir */
 			error = gfs2_ok_to_move(ip, ndip);
 			if (error)
 				goto out_gunlock_r;
@@ -1494,20 +1523,9 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry,
 	if (nip)
 		error = gfs2_unlink_inode(ndip, ndentry);
 
-	if (dir_rename) {
-		error = gfs2_dir_mvino(ip, &gfs2_qdotdot, ndip, DT_DIR);
-		if (error)
-			goto out_end_trans;
-	} else {
-		struct buffer_head *dibh;
-		error = gfs2_meta_inode_buffer(ip, &dibh);
-		if (error)
-			goto out_end_trans;
-		ip->i_inode.i_ctime = CURRENT_TIME;
-		gfs2_trans_add_meta(ip->i_gl, dibh);
-		gfs2_dinode_out(ip, dibh->b_data);
-		brelse(dibh);
-	}
+	error = update_moved_ino(ip, ndip, dir_rename);
+	if (error)
+		goto out_end_trans;
 
 	error = gfs2_dir_del(odip, odentry);
 	if (error)
@@ -1539,6 +1557,161 @@ out:
 }
 
 /**
+ * gfs2_exchange - exchange two files
+ * @odir: Parent directory of old file name
+ * @odentry: The old dentry of the file
+ * @ndir: Parent directory of new file name
+ * @ndentry: The new dentry of the file
+ * @flags: The rename flags
+ *
+ * Returns: errno
+ */
+
+static int gfs2_exchange(struct inode *odir, struct dentry *odentry,
+			 struct inode *ndir, struct dentry *ndentry,
+			 unsigned int flags)
+{
+	struct gfs2_inode *odip = GFS2_I(odir);
+	struct gfs2_inode *ndip = GFS2_I(ndir);
+	struct gfs2_inode *oip = GFS2_I(odentry->d_inode);
+	struct gfs2_inode *nip = GFS2_I(ndentry->d_inode);
+	struct gfs2_sbd *sdp = GFS2_SB(odir);
+	struct gfs2_holder ghs[5], r_gh = { .gh_gl = NULL, };
+	unsigned int num_gh;
+	unsigned int x;
+	umode_t old_mode = oip->i_inode.i_mode;
+	umode_t new_mode = nip->i_inode.i_mode;
+	int error;
+
+	error = gfs2_rindex_update(sdp);
+	if (error)
+		return error;
+
+	if (odip != ndip) {
+		error = gfs2_glock_nq_init(sdp->sd_rename_gl, LM_ST_EXCLUSIVE,
+					   0, &r_gh);
+		if (error)
+			goto out;
+
+		if (S_ISDIR(old_mode)) {
+			/* don't move a directory into its subdir */
+			error = gfs2_ok_to_move(oip, ndip);
+			if (error)
+				goto out_gunlock_r;
+		}
+
+		if (S_ISDIR(new_mode)) {
+			/* don't move a directory into its subdir */
+			error = gfs2_ok_to_move(nip, odip);
+			if (error)
+				goto out_gunlock_r;
+		}
+	}
+
+	num_gh = 1;
+	gfs2_holder_init(odip->i_gl, LM_ST_EXCLUSIVE, 0, ghs);
+	if (odip != ndip) {
+		gfs2_holder_init(ndip->i_gl, LM_ST_EXCLUSIVE, 0, ghs + num_gh);
+		num_gh++;
+	}
+	gfs2_holder_init(oip->i_gl, LM_ST_EXCLUSIVE, 0, ghs + num_gh);
+	num_gh++;
+
+	gfs2_holder_init(nip->i_gl, LM_ST_EXCLUSIVE, 0, ghs + num_gh);
+	num_gh++;
+
+	for (x = 0; x < num_gh; x++) {
+		error = gfs2_glock_nq(ghs + x);
+		if (error)
+			goto out_gunlock;
+	}
+
+	error = -ENOENT;
+	if (oip->i_inode.i_nlink == 0 || nip->i_inode.i_nlink == 0)
+		goto out_gunlock;
+
+	error = gfs2_unlink_ok(odip, &odentry->d_name, oip);
+	if (error)
+		goto out_gunlock;
+	error = gfs2_unlink_ok(ndip, &ndentry->d_name, nip);
+	if (error)
+		goto out_gunlock;
+
+	if (S_ISDIR(old_mode)) {
+		error = gfs2_permission(odentry->d_inode, MAY_WRITE);
+		if (error)
+			goto out_gunlock;
+	}
+	if (S_ISDIR(new_mode)) {
+		error = gfs2_permission(ndentry->d_inode, MAY_WRITE);
+		if (error)
+			goto out_gunlock;
+	}
+	error = gfs2_trans_begin(sdp, 4 * RES_DINODE + 4 * RES_LEAF, 0);
+	if (error)
+		goto out_gunlock;
+
+	error = update_moved_ino(oip, ndip, S_ISDIR(old_mode));
+	if (error)
+		goto out_end_trans;
+
+	error = update_moved_ino(nip, odip, S_ISDIR(new_mode));
+	if (error)
+		goto out_end_trans;
+
+	error = gfs2_dir_mvino(ndip, &ndentry->d_name, oip,
+			       IF2DT(old_mode));
+	if (error)
+		goto out_end_trans;
+
+	error = gfs2_dir_mvino(odip, &odentry->d_name, nip,
+			       IF2DT(new_mode));
+	if (error)
+		goto out_end_trans;
+
+	if (odip != ndip) {
+		if (S_ISDIR(new_mode) && !S_ISDIR(old_mode)) {
+			inc_nlink(&odip->i_inode);
+			drop_nlink(&ndip->i_inode);
+		} else if (S_ISDIR(old_mode) && !S_ISDIR(new_mode)) {
+			inc_nlink(&ndip->i_inode);
+			drop_nlink(&odip->i_inode);
+		}
+	}
+	mark_inode_dirty(&ndip->i_inode);
+	if (odip != ndip)
+		mark_inode_dirty(&odip->i_inode);
+
+out_end_trans:
+	gfs2_trans_end(sdp);
+out_gunlock:
+	while (x--) {
+		gfs2_glock_dq(ghs + x);
+		gfs2_holder_uninit(ghs + x);
+	}
+out_gunlock_r:
+	if (r_gh.gh_gl)
+		gfs2_glock_dq_uninit(&r_gh);
+out:
+	return error;
+}
+
+static int gfs2_rename2(struct inode *odir, struct dentry *odentry,
+			struct inode *ndir, struct dentry *ndentry,
+			unsigned int flags)
+{
+	flags &= ~RENAME_NOREPLACE;
+
+	if (flags & ~RENAME_EXCHANGE)
+		return -EINVAL;
+
+	if (flags & RENAME_EXCHANGE)
+		return gfs2_exchange(odir, odentry, ndir, ndentry, flags);
+
+	return gfs2_rename(odir, odentry, ndir, ndentry);
+}
+
+/**
  * gfs2_follow_link - Follow a symbolic link
  * @dentry: The dentry of the link
  * @nd: Data that we pass to vfs_follow_link()
@@ -1943,7 +2116,7 @@ const struct inode_operations gfs2_dir_iops = {
 	.mkdir = gfs2_mkdir,
 	.rmdir = gfs2_unlink,
 	.mknod = gfs2_mknod,
-	.rename = gfs2_rename,
+	.rename2 = gfs2_rename2,
 	.permission = gfs2_permission,
 	.setattr = gfs2_setattr,
 	.getattr = gfs2_getattr,
-- 
2.1.0

[Cluster-devel] [PATCH 07/16] GFS2: make sure S_NOSEC flag isn't overwritten

[Bob Peterson]

From: Benjamin Marzinski <bmarzins@redhat.com>

At the end of gfs2_set_inode_flags inode->i_flags is set to flags, so
we should be modifying flags instead of inode->i_flags, so it isn't
overwritten.

Signed-off-by: Benjamin Marzinski <bmarzins redhat com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/file.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index 8ec43ab..c706c6d 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -180,7 +180,7 @@ void gfs2_set_inode_flags(struct inode *inode)
 
 	flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC|S_NOSEC);
 	if ((ip->i_eattr == 0) && !is_sxid(inode->i_mode))
-		inode->i_flags |= S_NOSEC;
+		flags |= S_NOSEC;
 	if (ip->i_diskflags & GFS2_DIF_IMMUTABLE)
 		flags |= S_IMMUTABLE;
 	if (ip->i_diskflags & GFS2_DIF_APPENDONLY)
-- 
2.1.0

[Cluster-devel] [PATCH 08/16] gfs2: convert simple_str to kstr

[Bob Peterson]

From: Fabian Frederick <fabf@skynet.be>

-Remove obsolete simple_str functions.
-Return error code when kstr failed.
-This patch also calls functions corresponding to destination type.

Thanks to Alexey Dobriyan for suggesting improvements in
block_store() and wdack_store()

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/sys.c | 66 +++++++++++++++++++++++++++++++++++++++++++----------------
 1 file changed, 48 insertions(+), 18 deletions(-)

diff --git a/fs/gfs2/sys.c b/fs/gfs2/sys.c
index ae8e881..c9ff1cf 100644
--- a/fs/gfs2/sys.c
+++ b/fs/gfs2/sys.c
@@ -101,8 +101,11 @@ static ssize_t freeze_show(struct gfs2_sbd *sdp, char *buf)
 
 static ssize_t freeze_store(struct gfs2_sbd *sdp, const char *buf, size_t len)
 {
-	int error;
-	int n = simple_strtol(buf, NULL, 0);
+	int error, n;
+
+	error = kstrtoint(buf, 0, &n);
+	if (error)
+		return error;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
@@ -134,10 +137,16 @@ static ssize_t withdraw_show(struct gfs2_sbd *sdp, char *buf)
 
 static ssize_t withdraw_store(struct gfs2_sbd *sdp, const char *buf, size_t len)
 {
+	int error, val;
+
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	if (simple_strtol(buf, NULL, 0) != 1)
+	error = kstrtoint(buf, 0, &val);
+	if (error)
+		return error;
+
+	if (val != 1)
 		return -EINVAL;
 
 	gfs2_lm_withdraw(sdp, "withdrawing from cluster at user's request\n");
@@ -148,10 +157,16 @@ static ssize_t withdraw_store(struct gfs2_sbd *sdp, const char *buf, size_t len)
 static ssize_t statfs_sync_store(struct gfs2_sbd *sdp, const char *buf,
 				 size_t len)
 {
+	int error, val;
+
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	if (simple_strtol(buf, NULL, 0) != 1)
+	error = kstrtoint(buf, 0, &val);
+	if (error)
+		return error;
+
+	if (val != 1)
 		return -EINVAL;
 
 	gfs2_statfs_sync(sdp->sd_vfs, 0);
@@ -161,10 +176,16 @@ static ssize_t statfs_sync_store(struct gfs2_sbd *sdp, const char *buf,
 static ssize_t quota_sync_store(struct gfs2_sbd *sdp, const char *buf,
 				size_t len)
 {
+	int error, val;
+
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	if (simple_strtol(buf, NULL, 0) != 1)
+	error = kstrtoint(buf, 0, &val);
+	if (error)
+		return error;
+
+	if (val != 1)
 		return -EINVAL;
 
 	gfs2_quota_sync(sdp->sd_vfs, 0);
@@ -181,7 +202,9 @@ static ssize_t quota_refresh_user_store(struct gfs2_sbd *sdp, const char *buf,
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	id = simple_strtoul(buf, NULL, 0);
+	error = kstrtou32(buf, 0, &id);
+	if (error)
+		return error;
 
 	qid = make_kqid(current_user_ns(), USRQUOTA, id);
 	if (!qid_valid(qid))
@@ -201,7 +224,9 @@ static ssize_t quota_refresh_group_store(struct gfs2_sbd *sdp, const char *buf,
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	id = simple_strtoul(buf, NULL, 0);
+	error = kstrtou32(buf, 0, &id);
+	if (error)
+		return error;
 
 	qid = make_kqid(current_user_ns(), GRPQUOTA, id);
 	if (!qid_valid(qid))
@@ -324,10 +349,11 @@ static ssize_t block_show(struct gfs2_sbd *sdp, char *buf)
 static ssize_t block_store(struct gfs2_sbd *sdp, const char *buf, size_t len)
 {
 	struct lm_lockstruct *ls = &sdp->sd_lockstruct;
-	ssize_t ret = len;
-	int val;
+	int ret, val;
 
-	val = simple_strtol(buf, NULL, 0);
+	ret = kstrtoint(buf, 0, &val);
+	if (ret)
+		return ret;
 
 	if (val == 1)
 		set_bit(DFL_BLOCK_LOCKS, &ls->ls_recover_flags);
@@ -336,9 +362,9 @@ static ssize_t block_store(struct gfs2_sbd *sdp, const char *buf, size_t len)
 		smp_mb__after_atomic();
 		gfs2_glock_thaw(sdp);
 	} else {
-		ret = -EINVAL;
+		return -EINVAL;
 	}
-	return ret;
+	return len;
 }
 
 static ssize_t wdack_show(struct gfs2_sbd *sdp, char *buf)
@@ -350,17 +376,18 @@ static ssize_t wdack_show(struct gfs2_sbd *sdp, char *buf)
 
 static ssize_t wdack_store(struct gfs2_sbd *sdp, const char *buf, size_t len)
 {
-	ssize_t ret = len;
-	int val;
+	int ret, val;
 
-	val = simple_strtol(buf, NULL, 0);
+	ret = kstrtoint(buf, 0, &val);
+	if (ret)
+		return ret;
 
 	if ((val == 1) &&
 	    !strcmp(sdp->sd_lockstruct.ls_ops->lm_proto_name, "lock_dlm"))
 		complete(&sdp->sd_wdack);
 	else
-		ret = -EINVAL;
-	return ret;
+		return -EINVAL;
+	return len;
 }
 
 static ssize_t lkfirst_show(struct gfs2_sbd *sdp, char *buf)
@@ -553,11 +580,14 @@ static ssize_t tune_set(struct gfs2_sbd *sdp, unsigned int *field,
 {
 	struct gfs2_tune *gt = &sdp->sd_tune;
 	unsigned int x;
+	int error;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	x = simple_strtoul(buf, NULL, 0);
+	error = kstrtouint(buf, 0, &x);
+	if (error)
+		return error;
 
 	if (check_zero && !x)
 		return -EINVAL;
-- 
2.1.0

[Cluster-devel] [PATCH 09/16] gfs2: kerneldoc warning fixes

[Bob Peterson]

From: Fabian Frederick <fabf@skynet.be>

Fixes the following kernel-doc warnings:
Warning(fs/gfs2/aops.c:180): No description found for parameter 'wbc'
Warning(fs/gfs2/aops.c:236): No description found for parameter 'end'
Warning(fs/gfs2/aops.c:236): No description found for parameter 'done_index'
Warning(fs/gfs2/aops.c:236): Excess function parameter 'writepage' description in 'gfs2_write_jdata_pagevec'
Warning(fs/gfs2/aops.c:346): Excess function parameter 'writepage' description in 'gfs2_write_cache_jdata'
Warning(fs/gfs2/aops.c:346): Excess function parameter 'data' description in 'gfs2_write_cache_jdata'
Warning(fs/gfs2/aops.c:605): No description found for parameter 'file'
Warning(fs/gfs2/aops.c:605): No description found for parameter 'mapping'
Warning(fs/gfs2/aops.c:605): No description found for parameter 'pages'
Warning(fs/gfs2/aops.c:605): No description found for parameter 'nr_pages'
Warning(fs/gfs2/aops.c:870): No description found for parameter 'copied'

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/aops.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index a6e6990..7c5a96c 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -171,6 +171,7 @@ static int __gfs2_jdata_writepage(struct page *page, struct writeback_control *w
 /**
  * gfs2_jdata_writepage - Write complete page
  * @page: Page to write
+ * @wbc: The writeback control
  *
  * Returns: errno
  *
@@ -221,9 +222,10 @@ static int gfs2_writepages(struct address_space *mapping,
  * gfs2_write_jdata_pagevec - Write back a pagevec's worth of pages
  * @mapping: The mapping
  * @wbc: The writeback control
- * @writepage: The writepage function to call for each page
  * @pvec: The vector of pages
  * @nr_pages: The number of pages to write
+ * @end: End position
+ * @done_index: Page index
  *
  * Returns: non-zero if loop should terminate, zero otherwise
  */
@@ -333,8 +335,6 @@ continue_unlock:
  * gfs2_write_cache_jdata - Like write_cache_pages but different
  * @mapping: The mapping to write
  * @wbc: The writeback control
- * @writepage: The writepage function to call
- * @data: The data to pass to writepage
  *
  * The reason that we use our own function here is that we need to
  * start transactions before we grab page locks. This allows us
@@ -588,6 +588,10 @@ int gfs2_internal_read(struct gfs2_inode *ip, char *buf, loff_t *pos,
 
 /**
  * gfs2_readpages - Read a bunch of pages at once
+ * @file: The file to read from
+ * @mapping: Address space info
+ * @pages: List of pages to read
+ * @nr_pages: Number of pages to read
  *
  * Some notes:
  * 1. This is only for readahead, so we can simply ignore any things
@@ -853,7 +857,7 @@ static int gfs2_stuffed_write_end(struct inode *inode, struct buffer_head *dibh,
  * @mapping: The address space to write to
  * @pos: The file position
  * @len: The length of the data
- * @copied:
+ * @copied: How much was actually copied by the VFS
  * @page: The page that has been written
  * @fsdata: The fsdata (unused in GFS2)
  *
-- 
2.1.0

[Cluster-devel] [PATCH 10/16] gfs2: fix shadow warning in gfs2_rbm_find()

[Bob Peterson]

From: Fabian Frederick <fabf@skynet.be>

bi was already declared and initialized globally in gfs2_rbm_find()

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/rgrp.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index 900e515..cd53d6e 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -1711,10 +1711,8 @@ static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 *minext,
 		return ret;
 
 bitmap_full:	/* Mark bitmap as full and fall through */
-		if ((state == GFS2_BLKST_FREE) && initial_offset == 0) {
-			struct gfs2_bitmap *bi = rbm_bi(rbm);
+		if ((state == GFS2_BLKST_FREE) && initial_offset == 0)
 			set_bit(GBF_FULL, &bi->bi_flags);
-		}
 
 next_bitmap:	/* Find next bitmap in the rgrp */
 		rbm->offset = 0;
-- 
2.1.0

[Cluster-devel] [PATCH 11/16] gfs2: fix quota updates on block boundaries

[Bob Peterson]

From: Abhi Das <adas@redhat.com>

For smaller block sizes (512B, 1K, 2K), some quotas straddle block
boundaries such that the usage value is on one block and the rest
of the quota is on the previous block. In such cases, the value
does not get updated correctly. This patch fixes that by addressing
the boundary conditions correctly.

This patch also adds a (s64) cast that was missing in a call to
gfs2_quota_change() in inode.c

Signed-off-by: Abhi Das <adas@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/inode.c |   2 +-
 fs/gfs2/quota.c | 197 +++++++++++++++++++++++++++++++++-----------------------
 2 files changed, 119 insertions(+), 80 deletions(-)

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 4d809eb..a088e54 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -1889,7 +1889,7 @@ static int setattr_chown(struct inode *inode, struct iattr *attr)
 
 	if (!uid_eq(ouid, NO_UID_QUOTA_CHANGE) ||
 	    !gid_eq(ogid, NO_GID_QUOTA_CHANGE)) {
-		gfs2_quota_change(ip, -ap.target, ouid, ogid);
+		gfs2_quota_change(ip, -(s64)ap.target, ouid, ogid);
 		gfs2_quota_change(ip, ap.target, nuid, ngid);
 	}
 
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index 5c27e48..7c20031 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -652,6 +652,112 @@ static void do_qc(struct gfs2_quota_data *qd, s64 change)
 	mutex_unlock(&sdp->sd_quota_mutex);
 }
 
+static int gfs2_write_buf_to_page(struct gfs2_inode *ip, unsigned long index,
+				  unsigned off, void *buf, unsigned bytes)
+{
+	struct inode *inode = &ip->i_inode;
+	struct gfs2_sbd *sdp = GFS2_SB(inode);
+	struct address_space *mapping = inode->i_mapping;
+	struct page *page;
+	struct buffer_head *bh;
+	void *kaddr;
+	u64 blk;
+	unsigned bsize = sdp->sd_sb.sb_bsize, bnum = 0, boff = 0;
+	unsigned to_write = bytes, pg_off = off;
+	int done = 0;
+
+	blk = index << (PAGE_CACHE_SHIFT - sdp->sd_sb.sb_bsize_shift);
+	boff = off % bsize;
+
+	page = find_or_create_page(mapping, index, GFP_NOFS);
+	if (!page)
+		return -ENOMEM;
+	if (!page_has_buffers(page))
+		create_empty_buffers(page, bsize, 0);
+
+	bh = page_buffers(page);
+	while (!done) {
+		/* Find the beginning block within the page */
+		if (pg_off >= ((bnum * bsize) + bsize)) {
+			bh = bh->b_this_page;
+			bnum++;
+			blk++;
+			continue;
+		}
+		if (!buffer_mapped(bh)) {
+			gfs2_block_map(inode, blk, bh, 1);
+			if (!buffer_mapped(bh))
+				goto unlock_out;
+			/* If it's a newly allocated disk block, zero it */
+			if (buffer_new(bh))
+				zero_user(page, bnum * bsize, bh->b_size);
+		}
+		if (PageUptodate(page))
+			set_buffer_uptodate(bh);
+		if (!buffer_uptodate(bh)) {
+			ll_rw_block(READ | REQ_META, 1, &bh);
+			wait_on_buffer(bh);
+			if (!buffer_uptodate(bh))
+				goto unlock_out;
+		}
+		gfs2_trans_add_data(ip->i_gl, bh);
+
+		/* If we need to write to the next block as well */
+		if (to_write > (bsize - boff)) {
+			pg_off += (bsize - boff);
+			to_write -= (bsize - boff);
+			boff = pg_off % bsize;
+			continue;
+		}
+		done = 1;
+	}
+
+	/* Write to the page, now that we have setup the buffer(s) */
+	kaddr = kmap_atomic(page);
+	memcpy(kaddr + off, buf, bytes);
+	flush_dcache_page(page);
+	kunmap_atomic(kaddr);
+	unlock_page(page);
+	page_cache_release(page);
+
+	return 0;
+
+unlock_out:
+	unlock_page(page);
+	page_cache_release(page);
+	return -EIO;
+}
+
+static int gfs2_write_disk_quota(struct gfs2_inode *ip, struct gfs2_quota *qp,
+				 loff_t loc)
+{
+	unsigned long pg_beg;
+	unsigned pg_off, nbytes, overflow = 0;
+	int pg_oflow = 0, error;
+	void *ptr;
+
+	nbytes = sizeof(struct gfs2_quota);
+
+	pg_beg = loc >> PAGE_CACHE_SHIFT;
+	pg_off = loc % PAGE_CACHE_SIZE;
+
+	/* If the quota straddles a page boundary, split the write in two */
+	if ((pg_off + nbytes) > PAGE_CACHE_SIZE) {
+		pg_oflow = 1;
+		overflow = (pg_off + nbytes) - PAGE_CACHE_SIZE;
+	}
+
+	ptr = qp;
+	error = gfs2_write_buf_to_page(ip, pg_beg, pg_off, ptr,
+				       nbytes - overflow);
+	/* If there's an overflow, write the remaining bytes to the next page */
+	if (!error && pg_oflow)
+		error = gfs2_write_buf_to_page(ip, pg_beg + 1, 0,
+					       ptr + nbytes - overflow,
+					       overflow);
+	return error;
+}
+
 /**
  * gfs2_adjust_quota - adjust record of current block usage
  * @ip: The quota inode
@@ -672,15 +778,8 @@ static int gfs2_adjust_quota(struct gfs2_inode *ip, loff_t loc,
 {
 	struct inode *inode = &ip->i_inode;
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
-	struct address_space *mapping = inode->i_mapping;
-	unsigned long index = loc >> PAGE_CACHE_SHIFT;
-	unsigned offset = loc & (PAGE_CACHE_SIZE - 1);
-	unsigned blocksize, iblock, pos;
-	struct buffer_head *bh;
-	struct page *page;
-	void *kaddr, *ptr;
 	struct gfs2_quota q;
-	int err, nbytes;
+	int err;
 	u64 size;
 
 	if (gfs2_is_stuffed(ip)) {
@@ -694,8 +793,11 @@ static int gfs2_adjust_quota(struct gfs2_inode *ip, loff_t loc,
 	if (err < 0)
 		return err;
 
+	loc -= sizeof(q); /* gfs2_internal_read would've advanced the loc ptr */
 	err = -EIO;
 	be64_add_cpu(&q.qu_value, change);
+	if (be64_to_cpu(q.qu_value) < 0)
+		q.qu_value = 0; /* Never go negative on quota usage */
 	qd->qd_qb.qb_value = q.qu_value;
 	if (fdq) {
 		if (fdq->d_fieldmask & QC_SPC_SOFT) {
@@ -712,79 +814,16 @@ static int gfs2_adjust_quota(struct gfs2_inode *ip, loff_t loc,
 		}
 	}
 
-	/* Write the quota into the quota file on disk */
-	ptr = &q;
-	nbytes = sizeof(struct gfs2_quota);
-get_a_page:
-	page = find_or_create_page(mapping, index, GFP_NOFS);
-	if (!page)
-		return -ENOMEM;
-
-	blocksize = inode->i_sb->s_blocksize;
-	iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits);
-
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, blocksize, 0);
-
-	bh = page_buffers(page);
-	pos = blocksize;
-	while (offset >= pos) {
-		bh = bh->b_this_page;
-		iblock++;
-		pos += blocksize;
-	}
-
-	if (!buffer_mapped(bh)) {
-		gfs2_block_map(inode, iblock, bh, 1);
-		if (!buffer_mapped(bh))
-			goto unlock_out;
-		/* If it's a newly allocated disk block for quota, zero it */
-		if (buffer_new(bh))
-			zero_user(page, pos - blocksize, bh->b_size);
-	}
-
-	if (PageUptodate(page))
-		set_buffer_uptodate(bh);
-
-	if (!buffer_uptodate(bh)) {
-		ll_rw_block(READ | REQ_META, 1, &bh);
-		wait_on_buffer(bh);
-		if (!buffer_uptodate(bh))
-			goto unlock_out;
-	}
-
-	gfs2_trans_add_data(ip->i_gl, bh);
-
-	kaddr = kmap_atomic(page);
-	if (offset + sizeof(struct gfs2_quota) > PAGE_CACHE_SIZE)
-		nbytes = PAGE_CACHE_SIZE - offset;
-	memcpy(kaddr + offset, ptr, nbytes);
-	flush_dcache_page(page);
-	kunmap_atomic(kaddr);
-	unlock_page(page);
-	page_cache_release(page);
-
-	/* If quota straddles page boundary, we need to update the rest of the
-	 * quota at the beginning of the next page */
-	if ((offset + sizeof(struct gfs2_quota)) > PAGE_CACHE_SIZE) {
-		ptr = ptr + nbytes;
-		nbytes = sizeof(struct gfs2_quota) - nbytes;
-		offset = 0;
-		index++;
-		goto get_a_page;
+	err = gfs2_write_disk_quota(ip, &q, loc);
+	if (!err) {
+		size = loc + sizeof(struct gfs2_quota);
+		if (size > inode->i_size)
+			i_size_write(inode, size);
+		inode->i_mtime = inode->i_atime = CURRENT_TIME;
+		mark_inode_dirty(inode);
+		set_bit(QDF_REFRESH, &qd->qd_flags);
 	}
 
-	size = loc + sizeof(struct gfs2_quota);
-	if (size > inode->i_size)
-		i_size_write(inode, size);
-	inode->i_mtime = inode->i_atime = CURRENT_TIME;
-	mark_inode_dirty(inode);
-	set_bit(QDF_REFRESH, &qd->qd_flags);
-	return 0;
-
-unlock_out:
-	unlock_page(page);
-	page_cache_release(page);
 	return err;
 }
 
-- 
2.1.0

[Cluster-devel] [PATCH 12/16] gfs2: limit quota log messages

[Bob Peterson]

From: Abhi Das <adas@redhat.com>

This patch makes the quota subsystem only report once that a
particular user/group has exceeded their allotted quota.

Previously, it was possible for a program to continuously try
exceeding quota (despite receiving EDQUOT) and in turn trigger
gfs2 to issue a kernel log message about quota exceed. In theory,
this could get out of hand and flood the log and the filesystem
hosting the log files.

Signed-off-by: Abhi Das <adas@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/incore.h |  1 +
 fs/gfs2/quota.c  | 15 +++++++++++----
 2 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index 58b75ab..304a223 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -432,6 +432,7 @@ enum {
 	QDF_CHANGE		= 1,
 	QDF_LOCKED		= 2,
 	QDF_REFRESH		= 3,
+	QDF_QMSG_QUIET          = 4,
 };
 
 struct gfs2_quota_data {
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index 7c20031..dcd598a 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -649,6 +649,8 @@ static void do_qc(struct gfs2_quota_data *qd, s64 change)
 		slot_hold(qd);
 	}
 
+	if (change < 0) /* Reset quiet flag if we freed some blocks */
+		clear_bit(QDF_QMSG_QUIET, &qd->qd_flags);
 	mutex_unlock(&sdp->sd_quota_mutex);
 }
 
@@ -1187,10 +1189,13 @@ int gfs2_quota_check(struct gfs2_inode *ip, kuid_t uid, kgid_t gid,
 			/* If no min_target specified or we don't meet
 			 * min_target, return -EDQUOT */
 			if (!ap->min_target || ap->min_target > ap->allowed) {
-				print_message(qd, "exceeded");
-				quota_send_warning(qd->qd_id,
-						   sdp->sd_vfs->s_dev,
-						   QUOTA_NL_BHARDWARN);
+				if (!test_and_set_bit(QDF_QMSG_QUIET,
+						      &qd->qd_flags)) {
+					print_message(qd, "exceeded");
+					quota_send_warning(qd->qd_id,
+							   sdp->sd_vfs->s_dev,
+							   QUOTA_NL_BHARDWARN);
+				}
 				error = -EDQUOT;
 				break;
 			}
@@ -1685,6 +1690,8 @@ static int gfs2_set_dqblk(struct super_block *sb, struct kqid qid,
 
 	/* Apply changes */
 	error = gfs2_adjust_quota(ip, offset, 0, qd, fdq);
+	if (!error)
+		clear_bit(QDF_QMSG_QUIET, &qd->qd_flags);
 
 	gfs2_trans_end(sdp);
 out_release:
-- 
2.1.0

[Cluster-devel] [PATCH 13/16] gfs2: s64 cast for negative quota value

[Bob Peterson]

From: Abhi Das <adas@redhat.com>

One-line fix to cast quota value to s64 before comparison.
By default the quantity is treated as u64.

Signed-off-by: Abhi Das <adas@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Acked-by: Steven Whitehouse <swhiteho@redhat.com>
---
 fs/gfs2/quota.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index dcd598a..c2607a2 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -798,7 +798,7 @@ static int gfs2_adjust_quota(struct gfs2_inode *ip, loff_t loc,
 	loc -= sizeof(q); /* gfs2_internal_read would've advanced the loc ptr */
 	err = -EIO;
 	be64_add_cpu(&q.qu_value, change);
-	if (be64_to_cpu(q.qu_value) < 0)
+	if (((s64)be64_to_cpu(q.qu_value)) < 0)
 		q.qu_value = 0; /* Never go negative on quota usage */
 	qd->qd_qb.qb_value = q.qu_value;
 	if (fdq) {
-- 
2.1.0

[Cluster-devel] [PATCH 14/16] gfs2: Don't support fallocate on jdata files

[Bob Peterson]

From: Abhi Das <adas@redhat.com>

We cannot provide an efficient implementation due to the headers
on the data blocks, so there doesn't seem much point in having it.

Signed-off-by: Abhi Das <adas@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/file.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index c706c6d..8252115 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -917,7 +917,7 @@ static long gfs2_fallocate(struct file *file, int mode, loff_t offset, loff_t le
 	struct gfs2_holder gh;
 	int ret;
 
-	if (mode & ~FALLOC_FL_KEEP_SIZE)
+	if ((mode & ~FALLOC_FL_KEEP_SIZE) || gfs2_is_jdata(ip))
 		return -EOPNOTSUPP;
 
 	mutex_lock(&inode->i_mutex);
-- 
2.1.0

[Cluster-devel] [PATCH 15/16] GFS2: Don't add all glocks to the lru

[Bob Peterson]

The glocks used for resource groups often come and go hundreds of
thousands of times per second. Adding them to the lru list just
adds unnecessary contention for the lru_lock spin_lock, especially
considering we're almost certainly going to re-use the glock and
take it back off the lru microseconds later. We never want the
glock shrinker to cull them anyway. This patch adds a new bit in
the glops that determines which glock types get put onto the lru
list and which ones don't.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Acked-by: Steven Whitehouse <swhiteho@redhat.com>
---
 fs/gfs2/glock.c  | 3 ++-
 fs/gfs2/glops.c  | 6 ++++--
 fs/gfs2/incore.h | 1 +
 3 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 0fa8062..a38e38f 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -1076,7 +1076,8 @@ void gfs2_glock_dq(struct gfs2_holder *gh)
 		    !test_bit(GLF_DEMOTE, &gl->gl_flags))
 			fast_path = 1;
 	}
-	if (!test_bit(GLF_LFLUSH, &gl->gl_flags) && demote_ok(gl))
+	if (!test_bit(GLF_LFLUSH, &gl->gl_flags) && demote_ok(gl) &&
+	    (glops->go_flags & GLOF_LRU))
 		gfs2_glock_add_to_lru(gl);
 
 	trace_gfs2_glock_queue(gh, 0);
diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index fe91951..1249b2b 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -561,7 +561,7 @@ const struct gfs2_glock_operations gfs2_inode_glops = {
 	.go_lock = inode_go_lock,
 	.go_dump = inode_go_dump,
 	.go_type = LM_TYPE_INODE,
-	.go_flags = GLOF_ASPACE,
+	.go_flags = GLOF_ASPACE | GLOF_LRU,
 };
 
 const struct gfs2_glock_operations gfs2_rgrp_glops = {
@@ -584,10 +584,12 @@ const struct gfs2_glock_operations gfs2_freeze_glops = {
 const struct gfs2_glock_operations gfs2_iopen_glops = {
 	.go_type = LM_TYPE_IOPEN,
 	.go_callback = iopen_go_callback,
+	.go_flags = GLOF_LRU,
 };
 
 const struct gfs2_glock_operations gfs2_flock_glops = {
 	.go_type = LM_TYPE_FLOCK,
+	.go_flags = GLOF_LRU,
 };
 
 const struct gfs2_glock_operations gfs2_nondisk_glops = {
@@ -596,7 +598,7 @@ const struct gfs2_glock_operations gfs2_nondisk_glops = {
 
 const struct gfs2_glock_operations gfs2_quota_glops = {
 	.go_type = LM_TYPE_QUOTA,
-	.go_flags = GLOF_LVB,
+	.go_flags = GLOF_LVB | GLOF_LRU,
 };
 
 const struct gfs2_glock_operations gfs2_journal_glops = {
diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index 304a223..a1ec7c2 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -225,6 +225,7 @@ struct gfs2_glock_operations {
 	const unsigned long go_flags;
 #define GLOF_ASPACE 1
 #define GLOF_LVB    2
+#define GLOF_LRU    4
 };
 
 enum {
-- 
2.1.0

[Cluster-devel] [PATCH 16/16] GFS2: Don't brelse rgrp buffer_heads every allocation

[Bob Peterson]

This patch allows the block allocation code to retain the buffers
for the resource groups so they don't need to be re-read from buffer
cache with every request. This is a performance improvement that's
especially noticeable when resource groups are very large. For
example, with 2GB resource groups and 4K blocks, there can be 33
blocks for every resource group. This patch allows those 33 buffers
to be kept around and not read in and thrown away with every
operation. The buffers are released when the resource group is
either synced or invalidated.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Reviewed-by: Steven Whitehouse <swhiteho@redhat.com>
Reviewed-by: Benjamin Marzinski <bmarzins@redhat.com>
---
 fs/gfs2/glops.c | 14 +++++++++++---
 fs/gfs2/rgrp.c  | 23 +++++++++++++++++++----
 fs/gfs2/rgrp.h  |  1 +
 3 files changed, 31 insertions(+), 7 deletions(-)

diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index 1249b2b..fa3fa5e 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -144,6 +144,12 @@ static void rgrp_go_sync(struct gfs2_glock *gl)
 	struct gfs2_rgrpd *rgd;
 	int error;
 
+	spin_lock(&gl->gl_spin);
+	rgd = gl->gl_object;
+	if (rgd)
+		gfs2_rgrp_brelse(rgd);
+	spin_unlock(&gl->gl_spin);
+
 	if (!test_and_clear_bit(GLF_DIRTY, &gl->gl_flags))
 		return;
 	GLOCK_BUG_ON(gl, gl->gl_state != LM_ST_EXCLUSIVE);
@@ -175,15 +181,17 @@ static void rgrp_go_inval(struct gfs2_glock *gl, int flags)
 {
 	struct gfs2_sbd *sdp = gl->gl_sbd;
 	struct address_space *mapping = &sdp->sd_aspace;
+	struct gfs2_rgrpd *rgd = gl->gl_object;
+
+	if (rgd)
+		gfs2_rgrp_brelse(rgd);
 
 	WARN_ON_ONCE(!(flags & DIO_METADATA));
 	gfs2_assert_withdraw(sdp, !atomic_read(&gl->gl_ail_count));
 	truncate_inode_pages_range(mapping, gl->gl_vm.start, gl->gl_vm.end);
 
-	if (gl->gl_object) {
-		struct gfs2_rgrpd *rgd = (struct gfs2_rgrpd *)gl->gl_object;
+	if (rgd)
 		rgd->rd_flags &= ~GFS2_RDF_UPTODATE;
-	}
 }
 
 /**
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index cd53d6e..c6c6232 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -1244,14 +1244,13 @@ int gfs2_rgrp_go_lock(struct gfs2_holder *gh)
 }
 
 /**
- * gfs2_rgrp_go_unlock - Release RG bitmaps read in with gfs2_rgrp_bh_get()
- * @gh: The glock holder for the resource group
+ * gfs2_rgrp_brelse - Release RG bitmaps read in with gfs2_rgrp_bh_get()
+ * @rgd: The resource group
  *
  */
 
-void gfs2_rgrp_go_unlock(struct gfs2_holder *gh)
+void gfs2_rgrp_brelse(struct gfs2_rgrpd *rgd)
 {
-	struct gfs2_rgrpd *rgd = gh->gh_gl->gl_object;
 	int x, length = rgd->rd_length;
 
 	for (x = 0; x < length; x++) {
@@ -1264,6 +1263,22 @@ void gfs2_rgrp_go_unlock(struct gfs2_holder *gh)
 
 }
 
+/**
+ * gfs2_rgrp_go_unlock - Unlock a rgrp glock
+ * @gh: The glock holder for the resource group
+ *
+ */
+
+void gfs2_rgrp_go_unlock(struct gfs2_holder *gh)
+{
+	struct gfs2_rgrpd *rgd = gh->gh_gl->gl_object;
+	int demote_requested = test_bit(GLF_DEMOTE, &gh->gh_gl->gl_flags) |
+		test_bit(GLF_PENDING_DEMOTE, &gh->gh_gl->gl_flags);
+
+	if (rgd && demote_requested)
+		gfs2_rgrp_brelse(rgd);
+}
+
 int gfs2_rgrp_send_discards(struct gfs2_sbd *sdp, u64 offset,
 			     struct buffer_head *bh,
 			     const struct gfs2_bitmap *bi, unsigned minlen, u64 *ptrimmed)
diff --git a/fs/gfs2/rgrp.h b/fs/gfs2/rgrp.h
index 68972ec..c0ab33f 100644
--- a/fs/gfs2/rgrp.h
+++ b/fs/gfs2/rgrp.h
@@ -36,6 +36,7 @@ extern void gfs2_clear_rgrpd(struct gfs2_sbd *sdp);
 extern int gfs2_rindex_update(struct gfs2_sbd *sdp);
 extern void gfs2_free_clones(struct gfs2_rgrpd *rgd);
 extern int gfs2_rgrp_go_lock(struct gfs2_holder *gh);
+extern void gfs2_rgrp_brelse(struct gfs2_rgrpd *rgd);
 extern void gfs2_rgrp_go_unlock(struct gfs2_holder *gh);
 
 extern struct gfs2_alloc *gfs2_alloc_get(struct gfs2_inode *ip);
-- 
2.1.0