Merge git://oss.sgi.com:8090/xfs/xfs-2.6

* git://oss.sgi.com:8090/xfs/xfs-2.6:
  [XFS] Remove KERNEL_VERSION macros from xfs_dmapi.h
  [XFS] Prevent a deadlock when xfslogd unpins inodes.
  [XFS] Clean up i_flags and i_flags_lock handling.
  [XFS] 956664: dm_read_invis() changes i_atime
  [XFS] rename uio_read() to xfs_uio_read()
  [XFS] Keep lockdep happy.
  [XFS] 956618: Linux crashes on boot with XFS-DMAPI filesystem when

fs/xfs/Makefile-linux-2.6

@@ -21,22 +21,7 @@ EXTRA_CFLAGS +=	 -Ifs/xfs -Ifs/xfs/linux-2.6 -funsigned-char
 XFS_LINUX := linux-2.6
 
 ifeq ($(CONFIG_XFS_DEBUG),y)
-	EXTRA_CFLAGS += -g -DSTATIC="" -DDEBUG
-	EXTRA_CFLAGS += -DXFS_BUF_LOCK_TRACKING
-endif
-ifeq ($(CONFIG_XFS_TRACE),y)
-	EXTRA_CFLAGS += -DXFS_ALLOC_TRACE
-	EXTRA_CFLAGS += -DXFS_ATTR_TRACE
-	EXTRA_CFLAGS += -DXFS_BLI_TRACE
-	EXTRA_CFLAGS += -DXFS_BMAP_TRACE
-	EXTRA_CFLAGS += -DXFS_BMBT_TRACE
-	EXTRA_CFLAGS += -DXFS_DIR2_TRACE
-	EXTRA_CFLAGS += -DXFS_DQUOT_TRACE
-	EXTRA_CFLAGS += -DXFS_ILOCK_TRACE
-	EXTRA_CFLAGS += -DXFS_LOG_TRACE
-	EXTRA_CFLAGS += -DXFS_RW_TRACE
-	EXTRA_CFLAGS += -DXFS_BUF_TRACE
-	EXTRA_CFLAGS += -DXFS_VNODE_TRACE
+	EXTRA_CFLAGS += -g
 endif
 
 obj-$(CONFIG_XFS_FS)		+= xfs.o

fs/xfs/linux-2.6/xfs_buf.c

@@ -15,6 +15,7 @@
  * along with this program; if not, write the Free Software Foundation,
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
+#include "xfs.h"
 #include <linux/stddef.h>
 #include <linux/errno.h>
 #include <linux/slab.h>
@@ -31,7 +32,6 @@
 #include <linux/kthread.h>
 #include <linux/migrate.h>
 #include <linux/backing-dev.h>
-#include "xfs_linux.h"
 
 STATIC kmem_zone_t *xfs_buf_zone;
 STATIC kmem_shaker_t xfs_buf_shake;
@@ -1406,7 +1406,7 @@ xfs_alloc_bufhash(
 	btp->bt_hashshift = external ? 3 : 8;	/* 8 or 256 buckets */
 	btp->bt_hashmask = (1 << btp->bt_hashshift) - 1;
 	btp->bt_hash = kmem_zalloc((1 << btp->bt_hashshift) *
-					sizeof(xfs_bufhash_t), KM_SLEEP);
+					sizeof(xfs_bufhash_t), KM_SLEEP | KM_LARGE);
 	for (i = 0; i < (1 << btp->bt_hashshift); i++) {
 		spin_lock_init(&btp->bt_hash[i].bh_lock);
 		INIT_LIST_HEAD(&btp->bt_hash[i].bh_list);

fs/xfs/linux-2.6/xfs_dmapi_priv.h

+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_DMAPI_PRIV_H__
+#define __XFS_DMAPI_PRIV_H__
+
+/*
+ *	Based on IO_ISDIRECT, decide which i_ flag is set.
+ */
+#define DM_SEM_FLAG_RD(ioflags) (((ioflags) & IO_ISDIRECT) ? \
+			      DM_FLAGS_IMUX : 0)
+#define DM_SEM_FLAG_WR	(DM_FLAGS_IALLOCSEM_WR | DM_FLAGS_IMUX)
+
+#endif /*__XFS_DMAPI_PRIV_H__*/

fs/xfs/linux-2.6/xfs_ioctl.c

@@ -341,8 +341,11 @@ xfs_open_by_handle(
 		put_unused_fd(new_fd);
 		return -XFS_ERROR(-PTR_ERR(filp));
 	}
-	if (inode->i_mode & S_IFREG)
+	if (inode->i_mode & S_IFREG) {
+		/* invisible operation should not change atime */
+		filp->f_flags |= O_NOATIME;
 		filp->f_op = &xfs_invis_file_operations;
+	}
 
 	fd_install(new_fd, filp);
 	return new_fd;

fs/xfs/linux-2.6/xfs_super.c

@@ -227,9 +227,7 @@ xfs_initialize_vnode(
 		xfs_revalidate_inode(XFS_BHVTOM(bdp), vp, ip);
 		xfs_set_inodeops(inode);
 
-		spin_lock(&ip->i_flags_lock);
-		ip->i_flags &= ~XFS_INEW;
-		spin_unlock(&ip->i_flags_lock);
+		xfs_iflags_clear(ip, XFS_INEW);
 		barrier();
 
 		unlock_new_inode(inode);

fs/xfs/support/debug.c

@@ -15,11 +15,9 @@
  * along with this program; if not, write the Free Software Foundation,
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
+#include <xfs.h>
 #include "debug.h"
 #include "spin.h"
-#include <asm/page.h>
-#include <linux/sched.h>
-#include <linux/kernel.h>
 
 static char		message[256];	/* keep it off the stack */
 static DEFINE_SPINLOCK(xfs_err_lock);

fs/xfs/support/move.c

@@ -22,7 +22,7 @@
  * as we go.
  */
 int
-uio_read(caddr_t src, size_t len, struct uio *uio)
+xfs_uio_read(caddr_t src, size_t len, struct uio *uio)
 {
 	size_t	count;
 

fs/xfs/support/move.h

@@ -65,6 +65,6 @@ struct uio {
 typedef struct uio uio_t;
 typedef struct iovec iovec_t;
 
-extern int	uio_read (caddr_t, size_t, uio_t *);
+extern int	xfs_uio_read (caddr_t, size_t, uio_t *);
 
 #endif  /* __XFS_SUPPORT_MOVE_H__ */

fs/xfs/xfs.h

@@ -17,5 +17,28 @@
  */
 #ifndef __XFS_H__
 #define __XFS_H__
+
+#ifdef CONFIG_XFS_DEBUG
+#define STATIC
+#define DEBUG 1
+#define XFS_BUF_LOCK_TRACKING 1
+/* #define QUOTADEBUG 1 */
+#endif
+
+#ifdef CONFIG_XFS_TRACE
+#define XFS_ALLOC_TRACE 1
+#define XFS_ATTR_TRACE 1
+#define XFS_BLI_TRACE 1
+#define XFS_BMAP_TRACE 1
+#define XFS_BMBT_TRACE 1
+#define XFS_DIR2_TRACE 1
+#define XFS_DQUOT_TRACE 1
+#define XFS_ILOCK_TRACE 1
+#define XFS_LOG_TRACE 1
+#define XFS_RW_TRACE 1
+#define XFS_BUF_TRACE 1
+#define XFS_VNODE_TRACE 1
+#endif
+
 #include <linux-2.6/xfs_linux.h>
 #endif	/* __XFS_H__ */

fs/xfs/xfs_dir2.c

@@ -678,7 +678,7 @@ xfs_dir2_put_dirent64_uio(
 	idbp->d_off = pa->cook;
 	idbp->d_name[namelen] = '\0';
 	memcpy(idbp->d_name, pa->name, namelen);
-	rval = uio_read((caddr_t)idbp, reclen, uio);
+	rval = xfs_uio_read((caddr_t)idbp, reclen, uio);
 	pa->done = (rval == 0);
 	return rval;
 }

fs/xfs/xfs_dmapi.h

@@ -157,27 +157,9 @@ typedef enum {
 #define DM_FLAGS_IALLOCSEM_WR	0x020	/* thread holds i_alloc_sem wr */
 
 /*
- *	Based on IO_ISDIRECT, decide which i_ flag is set.
+ *	Pull in platform specific event flags defines
  */
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,0)
-#define DM_SEM_FLAG_RD(ioflags) (((ioflags) & IO_ISDIRECT) ? \
-			      DM_FLAGS_IMUX : 0)
-#define DM_SEM_FLAG_WR	(DM_FLAGS_IALLOCSEM_WR | DM_FLAGS_IMUX)
-#endif
-
-#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0)) && \
-    (LINUX_VERSION_CODE >= KERNEL_VERSION(2,4,22))
-#define DM_SEM_FLAG_RD(ioflags) (((ioflags) & IO_ISDIRECT) ? \
-			      DM_FLAGS_IALLOCSEM_RD : DM_FLAGS_IMUX)
-#define DM_SEM_FLAG_WR	(DM_FLAGS_IALLOCSEM_WR | DM_FLAGS_IMUX)
-#endif
-
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,4,21)
-#define DM_SEM_FLAG_RD(ioflags) (((ioflags) & IO_ISDIRECT) ? \
-			      0 : DM_FLAGS_IMUX)
-#define DM_SEM_FLAG_WR	(DM_FLAGS_IMUX)
-#endif
-
+#include "xfs_dmapi_priv.h"
 
 /*
  *	Macros to turn caller specified delay/block flags into

fs/xfs/xfs_iget.c

@@ -215,7 +215,7 @@ again:
 			 * If INEW is set this inode is being set up
 			 * we need to pause and try again.
 			 */
-			if (ip->i_flags & XFS_INEW) {
+			if (xfs_iflags_test(ip, XFS_INEW)) {
 				read_unlock(&ih->ih_lock);
 				delay(1);
 				XFS_STATS_INC(xs_ig_frecycle);
@@ -230,22 +230,50 @@ again:
 				 * on its way out of the system,
 				 * we need to pause and try again.
 				 */
-				if (ip->i_flags & XFS_IRECLAIM) {
+				if (xfs_iflags_test(ip, XFS_IRECLAIM)) {
 					read_unlock(&ih->ih_lock);
 					delay(1);
 					XFS_STATS_INC(xs_ig_frecycle);
 
 					goto again;
 				}
+				ASSERT(xfs_iflags_test(ip, XFS_IRECLAIMABLE));
+
+				/*
+				 * If lookup is racing with unlink, then we
+				 * should return an error immediately so we
+				 * don't remove it from the reclaim list and
+				 * potentially leak the inode.
+				 */
+				if ((ip->i_d.di_mode == 0) &&
+				    !(flags & XFS_IGET_CREATE)) {
+					read_unlock(&ih->ih_lock);
+					return ENOENT;
+				}
+
+				/*
+				 * There may be transactions sitting in the
+				 * incore log buffers or being flushed to disk
+				 * at this time.  We can't clear the
+				 * XFS_IRECLAIMABLE flag until these
+				 * transactions have hit the disk, otherwise we
+				 * will void the guarantee the flag provides
+				 * xfs_iunpin()
+				 */
+				if (xfs_ipincount(ip)) {
+					read_unlock(&ih->ih_lock);
+					xfs_log_force(mp, 0,
+						XFS_LOG_FORCE|XFS_LOG_SYNC);
+					XFS_STATS_INC(xs_ig_frecycle);
+					goto again;
+				}
 
 				vn_trace_exit(vp, "xfs_iget.alloc",
 					(inst_t *)__return_address);
 
 				XFS_STATS_INC(xs_ig_found);
 
-				spin_lock(&ip->i_flags_lock);
-				ip->i_flags &= ~XFS_IRECLAIMABLE;
-				spin_unlock(&ip->i_flags_lock);
+				xfs_iflags_clear(ip, XFS_IRECLAIMABLE);
 				version = ih->ih_version;
 				read_unlock(&ih->ih_lock);
 				xfs_ihash_promote(ih, ip, version);
@@ -299,10 +327,7 @@ finish_inode:
 			if (lock_flags != 0)
 				xfs_ilock(ip, lock_flags);
 
-			spin_lock(&ip->i_flags_lock);
-			ip->i_flags &= ~XFS_ISTALE;
-			spin_unlock(&ip->i_flags_lock);
-
+			xfs_iflags_clear(ip, XFS_ISTALE);
 			vn_trace_exit(vp, "xfs_iget.found",
 						(inst_t *)__return_address);
 			goto return_ip;
@@ -371,10 +396,7 @@ finish_inode:
 	ih->ih_next = ip;
 	ip->i_udquot = ip->i_gdquot = NULL;
 	ih->ih_version++;
-	spin_lock(&ip->i_flags_lock);
-	ip->i_flags |= XFS_INEW;
-	spin_unlock(&ip->i_flags_lock);
-
+	xfs_iflags_set(ip, XFS_INEW);
 	write_unlock(&ih->ih_lock);
 
 	/*
@@ -625,7 +647,7 @@ xfs_iput_new(xfs_inode_t	*ip,
 	vn_trace_entry(vp, "xfs_iput_new", (inst_t *)__return_address);
 
 	if ((ip->i_d.di_mode == 0)) {
-		ASSERT(!(ip->i_flags & XFS_IRECLAIMABLE));
+		ASSERT(!xfs_iflags_test(ip, XFS_IRECLAIMABLE));
 		vn_mark_bad(vp);
 	}
 	if (inode->i_state & I_NEW)
@@ -683,6 +705,7 @@ xfs_ireclaim(xfs_inode_t *ip)
 	/*
 	 * Free all memory associated with the inode.
 	 */
+	xfs_iunlock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
 	xfs_idestroy(ip);
 }
 

fs/xfs/xfs_inode.c

@@ -2193,7 +2193,7 @@ xfs_ifree_cluster(
 			/* Inode not in memory or we found it already,
 			 * nothing to do
 			 */
-			if (!ip || (ip->i_flags & XFS_ISTALE)) {
+			if (!ip || xfs_iflags_test(ip, XFS_ISTALE)) {
 				read_unlock(&ih->ih_lock);
 				continue;
 			}
@@ -2215,10 +2215,7 @@ xfs_ifree_cluster(
 
 			if (ip == free_ip) {
 				if (xfs_iflock_nowait(ip)) {
-					spin_lock(&ip->i_flags_lock);
-					ip->i_flags |= XFS_ISTALE;
-					spin_unlock(&ip->i_flags_lock);
-
+					xfs_iflags_set(ip, XFS_ISTALE);
 					if (xfs_inode_clean(ip)) {
 						xfs_ifunlock(ip);
 					} else {
@@ -2231,9 +2228,7 @@ xfs_ifree_cluster(
 
 			if (xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
 				if (xfs_iflock_nowait(ip)) {
-					spin_lock(&ip->i_flags_lock);
-					ip->i_flags |= XFS_ISTALE;
-					spin_unlock(&ip->i_flags_lock);
+					xfs_iflags_set(ip, XFS_ISTALE);
 
 					if (xfs_inode_clean(ip)) {
 						xfs_ifunlock(ip);
@@ -2263,9 +2258,7 @@ xfs_ifree_cluster(
 				AIL_LOCK(mp,s);
 				iip->ili_flush_lsn = iip->ili_item.li_lsn;
 				AIL_UNLOCK(mp, s);
-				spin_lock(&iip->ili_inode->i_flags_lock);
-				iip->ili_inode->i_flags |= XFS_ISTALE;
-				spin_unlock(&iip->ili_inode->i_flags_lock);
+				xfs_iflags_set(ip, XFS_ISTALE);
 				pre_flushed++;
 			}
 			lip = lip->li_bio_list;
@@ -2748,42 +2741,39 @@ xfs_iunpin(
 {
 	ASSERT(atomic_read(&ip->i_pincount) > 0);
 
-	if (atomic_dec_and_test(&ip->i_pincount)) {
+	if (atomic_dec_and_lock(&ip->i_pincount, &ip->i_flags_lock)) {
+
 		/*
-		 * If the inode is currently being reclaimed, the
-		 * linux inode _and_ the xfs vnode may have been
-		 * freed so we cannot reference either of them safely.
-		 * Hence we should not try to do anything to them
-		 * if the xfs inode is currently in the reclaim
-		 * path.
+		 * If the inode is currently being reclaimed, the link between
+		 * the bhv_vnode and the xfs_inode will be broken after the
+		 * XFS_IRECLAIM* flag is set. Hence, if these flags are not
+		 * set, then we can move forward and mark the linux inode dirty
+		 * knowing that it is still valid as it won't freed until after
+		 * the bhv_vnode<->xfs_inode link is broken in xfs_reclaim. The
+		 * i_flags_lock is used to synchronise the setting of the
+		 * XFS_IRECLAIM* flags and the breaking of the link, and so we
+		 * can execute atomically w.r.t to reclaim by holding this lock
+		 * here.
 		 *
-		 * However, we still need to issue the unpin wakeup
-		 * call as the inode reclaim may be blocked waiting for
-		 * the inode to become unpinned.
+		 * However, we still need to issue the unpin wakeup call as the
+		 * inode reclaim may be blocked waiting for the inode to become
+		 * unpinned.
 		 */
-		struct inode *inode = NULL;
 
-		spin_lock(&ip->i_flags_lock);
-		if (!(ip->i_flags & (XFS_IRECLAIM|XFS_IRECLAIMABLE))) {
+		if (!__xfs_iflags_test(ip, XFS_IRECLAIM|XFS_IRECLAIMABLE)) {
 			bhv_vnode_t	*vp = XFS_ITOV_NULL(ip);
+			struct inode *inode = NULL;
+
+			BUG_ON(vp == NULL);
+			inode = vn_to_inode(vp);
+			BUG_ON(inode->i_state & I_CLEAR);
 
 			/* make sync come back and flush this inode */
-			if (vp) {
-				inode = vn_to_inode(vp);
-
-				if (!(inode->i_state &
-						(I_NEW|I_FREEING|I_CLEAR))) {
-					inode = igrab(inode);
-					if (inode)
-						mark_inode_dirty_sync(inode);
-				} else
-					inode = NULL;
-			}
+			if (!(inode->i_state & (I_NEW|I_FREEING)))
+				mark_inode_dirty_sync(inode);
 		}
 		spin_unlock(&ip->i_flags_lock);
 		wake_up(&ip->i_ipin_wait);
-		if (inode)
-			iput(inode);
 	}
 }
 

fs/xfs/xfs_inode.h

@@ -305,6 +305,47 @@ typedef struct xfs_inode {
 #endif
 } xfs_inode_t;
 
+
+/*
+ * i_flags helper functions
+ */
+static inline void
+__xfs_iflags_set(xfs_inode_t *ip, unsigned short flags)
+{
+	ip->i_flags |= flags;
+}
+
+static inline void
+xfs_iflags_set(xfs_inode_t *ip, unsigned short flags)
+{
+	spin_lock(&ip->i_flags_lock);
+	__xfs_iflags_set(ip, flags);
+	spin_unlock(&ip->i_flags_lock);
+}
+
+static inline void
+xfs_iflags_clear(xfs_inode_t *ip, unsigned short flags)
+{
+	spin_lock(&ip->i_flags_lock);
+	ip->i_flags &= ~flags;
+	spin_unlock(&ip->i_flags_lock);
+}
+
+static inline int
+__xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
+{
+	return (ip->i_flags & flags);
+}
+
+static inline int
+xfs_iflags_test(xfs_inode_t *ip, unsigned short flags)
+{
+	int ret;
+	spin_lock(&ip->i_flags_lock);
+	ret = __xfs_iflags_test(ip, flags);
+	spin_unlock(&ip->i_flags_lock);
+	return ret;
+}
 #endif	/* __KERNEL__ */
 
 

fs/xfs/xfs_vnodeops.c

@@ -1013,7 +1013,7 @@ xfs_readlink(
 	pathlen = (int)ip->i_d.di_size;
 
 	if (ip->i_df.if_flags & XFS_IFINLINE) {
-		error = uio_read(ip->i_df.if_u1.if_data, pathlen, uiop);
+		error = xfs_uio_read(ip->i_df.if_u1.if_data, pathlen, uiop);
 	}
 	else {
 		/*
@@ -1044,7 +1044,7 @@ xfs_readlink(
 				byte_cnt = pathlen;
 			pathlen -= byte_cnt;
 
-			error = uio_read(XFS_BUF_PTR(bp), byte_cnt, uiop);
+			error = xfs_uio_read(XFS_BUF_PTR(bp), byte_cnt, uiop);
 			xfs_buf_relse (bp);
 		}
 
@@ -3827,11 +3827,16 @@ xfs_reclaim(
 	 */
 	xfs_synchronize_atime(ip);
 
-	/* If we have nothing to flush with this inode then complete the
-	 * teardown now, otherwise break the link between the xfs inode
-	 * and the linux inode and clean up the xfs inode later. This
-	 * avoids flushing the inode to disk during the delete operation
-	 * itself.
+	/*
+	 * If we have nothing to flush with this inode then complete the
+	 * teardown now, otherwise break the link between the xfs inode and the
+	 * linux inode and clean up the xfs inode later. This avoids flushing
+	 * the inode to disk during the delete operation itself.
+	 *
+	 * When breaking the link, we need to set the XFS_IRECLAIMABLE flag
+	 * first to ensure that xfs_iunpin() will never see an xfs inode
+	 * that has a linux inode being reclaimed. Synchronisation is provided
+	 * by the i_flags_lock.
 	 */
 	if (!ip->i_update_core && (ip->i_itemp == NULL)) {
 		xfs_ilock(ip, XFS_ILOCK_EXCL);
@@ -3840,13 +3845,13 @@ xfs_reclaim(
 	} else {
 		xfs_mount_t	*mp = ip->i_mount;
 
-		/* Protect sync from us */
+		/* Protect sync and unpin from us */
 		XFS_MOUNT_ILOCK(mp);
-		vn_bhv_remove(VN_BHV_HEAD(vp), XFS_ITOBHV(ip));
-		list_add_tail(&ip->i_reclaim, &mp->m_del_inodes);
 		spin_lock(&ip->i_flags_lock);
-		ip->i_flags |= XFS_IRECLAIMABLE;
+		__xfs_iflags_set(ip, XFS_IRECLAIMABLE);
+		vn_bhv_remove(VN_BHV_HEAD(vp), XFS_ITOBHV(ip));
 		spin_unlock(&ip->i_flags_lock);
+		list_add_tail(&ip->i_reclaim, &mp->m_del_inodes);
 		XFS_MOUNT_IUNLOCK(mp);
 	}
 	return 0;
@@ -3872,8 +3877,8 @@ xfs_finish_reclaim(
 	 */
 	write_lock(&ih->ih_lock);
 	spin_lock(&ip->i_flags_lock);
-	if ((ip->i_flags & XFS_IRECLAIM) ||
-	    (!(ip->i_flags & XFS_IRECLAIMABLE) && vp == NULL)) {
+	if (__xfs_iflags_test(ip, XFS_IRECLAIM) ||
+	    (!__xfs_iflags_test(ip, XFS_IRECLAIMABLE) && vp == NULL)) {
 		spin_unlock(&ip->i_flags_lock);
 		write_unlock(&ih->ih_lock);
 		if (locked) {
@@ -3882,7 +3887,7 @@ xfs_finish_reclaim(
 		}
 		return 1;
 	}
-	ip->i_flags |= XFS_IRECLAIM;
+	__xfs_iflags_set(ip, XFS_IRECLAIM);
 	spin_unlock(&ip->i_flags_lock);
 	write_unlock(&ih->ih_lock);