xfs: add realtime reverse map inode to metadata directory

Add a metadir path to select the realtime rmap btree inode and load
it at mount time.  The rtrmapbt inode will have a unique extent format
code, which means that we also have to update the inode validation and
flush routines to look for it.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>

12 files changed, 90 insertions, 11 deletions

diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 6abaca457e82..b92035f68b76 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -862,7 +862,8 @@ enum xfs_dinode_fmt {
 	XFS_DINODE_FMT_LOCAL,		/* bulk data */
 	XFS_DINODE_FMT_EXTENTS,		/* struct xfs_bmbt_rec */
 	XFS_DINODE_FMT_BTREE,		/* struct xfs_bmdr_block */
-	XFS_DINODE_FMT_UUID		/* added long ago, but never used */
+	XFS_DINODE_FMT_UUID,		/* added long ago, but never used */
+	XFS_DINODE_FMT_RMAP,		/* reverse mapping btree */
 };
 
 #define XFS_INODE_FORMAT_STR \
@@ -870,7 +871,8 @@ enum xfs_dinode_fmt {
 	{ XFS_DINODE_FMT_LOCAL,		"local" }, \
 	{ XFS_DINODE_FMT_EXTENTS,	"extent" }, \
 	{ XFS_DINODE_FMT_BTREE,		"btree" }, \
-	{ XFS_DINODE_FMT_UUID,		"uuid" }
+	{ XFS_DINODE_FMT_UUID,		"uuid" }, \
+	{ XFS_DINODE_FMT_RMAP,		"rmap" }
 
 /*
  * Inode minimum and maximum sizes.
diff --git a/fs/xfs/libxfs/xfs_imeta.c b/fs/xfs/libxfs/xfs_imeta.c
index 8c0b8cbbf96d..a91bf747d12a 100644
--- a/fs/xfs/libxfs/xfs_imeta.c
+++ b/fs/xfs/libxfs/xfs_imeta.c
@@ -63,12 +63,14 @@
 /* Static metadata inode paths */
 static const char *rtbitmap_path[]	= {"realtime", "0.bitmap"};
 static const char *rtsummary_path[]	= {"realtime", "0.summary"};
+static const char *rtrmapbt_path[]	= {"realtime", "0.rmap"};
 static const char *usrquota_path[]	= {"quota", "user"};
 static const char *grpquota_path[]	= {"quota", "group"};
 static const char *prjquota_path[]	= {"quota", "project"};
 
 XFS_IMETA_DEFINE_PATH(XFS_IMETA_RTBITMAP,	rtbitmap_path);
 XFS_IMETA_DEFINE_PATH(XFS_IMETA_RTSUMMARY,	rtsummary_path);
+XFS_IMETA_DEFINE_PATH(XFS_IMETA_RTRMAPBT,	rtrmapbt_path);
 XFS_IMETA_DEFINE_PATH(XFS_IMETA_USRQUOTA,	usrquota_path);
 XFS_IMETA_DEFINE_PATH(XFS_IMETA_GRPQUOTA,	grpquota_path);
 XFS_IMETA_DEFINE_PATH(XFS_IMETA_PRJQUOTA,	prjquota_path);
diff --git a/fs/xfs/libxfs/xfs_imeta.h b/fs/xfs/libxfs/xfs_imeta.h
index 9633351b8839..3ee616678633 100644
--- a/fs/xfs/libxfs/xfs_imeta.h
+++ b/fs/xfs/libxfs/xfs_imeta.h
@@ -37,6 +37,7 @@ struct xfs_imeta_end {
 /* Lookup keys for static metadata inodes. */
 extern const struct xfs_imeta_path XFS_IMETA_RTBITMAP;
 extern const struct xfs_imeta_path XFS_IMETA_RTSUMMARY;
+extern const struct xfs_imeta_path XFS_IMETA_RTRMAPBT;
 extern const struct xfs_imeta_path XFS_IMETA_USRQUOTA;
 extern const struct xfs_imeta_path XFS_IMETA_GRPQUOTA;
 extern const struct xfs_imeta_path XFS_IMETA_PRJQUOTA;
diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 7892b1a01c3f..32a56e1a6a38 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -371,6 +371,12 @@ xfs_dinode_verify_fork(
 			return __this_address;
 		}
 		break;
+	case XFS_DINODE_FMT_RMAP:
+		if (!xfs_has_rtrmapbt(mp))
+			return __this_address;
+		if (!(dip->di_flags2 & cpu_to_be64(XFS_DIFLAG2_METADATA)))
+			return __this_address;
+		break;
 	default:
 		return __this_address;
 	}
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 4cb6d51aa31e..660f6fda122a 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -258,6 +258,11 @@ xfs_iformat_data_fork(
 			return xfs_iformat_extents(ip, dip, XFS_DATA_FORK);
 		case XFS_DINODE_FMT_BTREE:
 			return xfs_iformat_btree(ip, dip, XFS_DATA_FORK);
+		case XFS_DINODE_FMT_RMAP:
+			if (!xfs_has_rtrmapbt(ip->i_mount))
+				return -EFSCORRUPTED;
+			ASSERT(0); /* to be implemented later */
+			return -EFSCORRUPTED;
 		default:
 			xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__,
 					dip, sizeof(*dip), __this_address);
@@ -639,6 +644,10 @@ xfs_iflush_fork(
 		}
 		break;
 
+	case XFS_DINODE_FMT_RMAP:
+		ASSERT(0); /* to be implemented later */
+		break;
+
 	default:
 		ASSERT(0);
 		break;
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 5c144a813cea..c4f95b8d98ab 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -423,6 +423,7 @@ xfs_rtrmapbt_commit_staged_btree(
 	int			flags = XFS_ILOG_CORE | XFS_ILOG_DBROOT;
 
 	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+	ASSERT(ifake->if_fork->if_format == XFS_DINODE_FMT_RMAP);
 
 	/*
 	 * Free any resources hanging off the real fork, then shallow-copy the
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index a13bd83573d0..ebca74b20499 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -2516,7 +2516,15 @@ xfs_iflush(
 			__func__, ip->i_ino, be16_to_cpu(dip->di_magic), dip);
 		goto flush_out;
 	}
-	if (S_ISREG(VFS_I(ip)->i_mode)) {
+	if (ip->i_df.if_format == XFS_DINODE_FMT_RMAP) {
+		if (!S_ISREG(VFS_I(ip)->i_mode) ||
+		    !(ip->i_diflags2 & XFS_DIFLAG2_METADATA)) {
+			xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
+				"%s: Bad rt rmapbt inode %Lu, ptr "PTR_FMT,
+				__func__, ip->i_ino, ip);
+			goto flush_out;
+		}
+	} else if (S_ISREG(VFS_I(ip)->i_mode)) {
 		if (XFS_TEST_ERROR(
 		    ip->i_df.if_format != XFS_DINODE_FMT_EXTENTS &&
 		    ip->i_df.if_format != XFS_DINODE_FMT_BTREE,
@@ -2556,6 +2564,15 @@ xfs_iflush(
 		goto flush_out;
 	}
 
+	if (ip->i_afp) {
+		if (ip->i_afp->if_format == XFS_DINODE_FMT_RMAP) {
+			xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
+				"%s: rt rmapbt in inode %Lu attr fork, ptr "PTR_FMT,
+				__func__, ip->i_ino, ip);
+			goto flush_out;
+		}
+	}
+
 	/*
 	 * Inode item log recovery for v2 inodes are dependent on the flushiter
 	 * count for correct sequencing.  We bump the flush iteration count so
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 90d8e591baf8..e7ac4adf1c55 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -61,6 +61,7 @@ xfs_inode_item_data_fork_size(
 		}
 		break;
 	case XFS_DINODE_FMT_BTREE:
+	case XFS_DINODE_FMT_RMAP:
 		if ((iip->ili_fields & XFS_ILOG_DBROOT) &&
 		    ip->i_df.if_broot_bytes > 0) {
 			*nbytes += ip->i_df.if_broot_bytes;
@@ -181,6 +182,7 @@ xfs_inode_item_format_data_fork(
 		}
 		break;
 	case XFS_DINODE_FMT_BTREE:
+	case XFS_DINODE_FMT_RMAP:
 		iip->ili_fields &=
 			~(XFS_ILOG_DDATA | XFS_ILOG_DEXT | XFS_ILOG_DEV);
 
diff --git a/fs/xfs/xfs_inode_item_recover.c b/fs/xfs/xfs_inode_item_recover.c
index 239dd2e3384e..5b6ebe5fd791 100644
--- a/fs/xfs/xfs_inode_item_recover.c
+++ b/fs/xfs/xfs_inode_item_recover.c
@@ -316,6 +316,7 @@ xlog_recover_inode_commit_pass2(
 
 	if (unlikely(S_ISREG(ldip->di_mode))) {
 		if ((ldip->di_format != XFS_DINODE_FMT_EXTENTS) &&
+		    (ldip->di_format != XFS_DINODE_FMT_RMAP) &&
 		    (ldip->di_format != XFS_DINODE_FMT_BTREE)) {
 			XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(3)",
 					 XFS_ERRLEVEL_LOW, mp, ldip,
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 16f133728f9d..934a9fb9e1f7 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -106,6 +106,7 @@ typedef struct xfs_mount {
 	struct xlog		*m_log;		/* log specific stuff */
 	struct xfs_inode	*m_rbmip;	/* pointer to bitmap inode */
 	struct xfs_inode	*m_rsumip;	/* pointer to summary inode */
+	struct xfs_inode	*m_rrmapip;	/* pointer to rmap inode */
 	struct xfs_inode	*m_rootip;	/* pointer to root directory */
 	struct xfs_inode	*m_metadirip;	/* ptr to metadata directory */
 	struct xfs_quotainfo	*m_quotainfo;	/* disk quota information */
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index bb7e874240fd..67774cf86998 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -24,6 +24,7 @@
 #include "xfs_da_format.h"
 #include "xfs_imeta.h"
 #include "xfs_quota.h"
+#include "xfs_error.h"
 
 /*
  * Read and return the summary information for a given extent size,
@@ -1316,6 +1317,7 @@ xfs_rt_resv_init(
  */
 static struct lock_class_key xfs_rbmip_key;
 static struct lock_class_key xfs_rsumip_key;
+static struct lock_class_key xfs_rrmapip_key;
 
 static inline int
 xfs_rt_iget(
@@ -1338,12 +1340,13 @@ xfs_rt_iget(
  * Get the bitmap and summary inodes and the summary cache into the mount
  * structure at mount time.
  */
-int					/* error */
+int
 xfs_rtmount_inodes(
-	xfs_mount_t	*mp)		/* file system mount structure */
+	struct xfs_mount	*mp)
 {
-	int		error;		/* error return value */
-	xfs_sb_t	*sbp;
+	struct xfs_sb		*sbp;
+	xfs_ino_t		ino;
+	int			error;
 
 	sbp = &mp->m_sb;
 	error = xfs_rt_iget(mp, mp->m_sb.sb_rbmino, &xfs_rbmip_key,
@@ -1358,13 +1361,38 @@ xfs_rtmount_inodes(
 			&mp->m_rsumip);
 	if (xfs_metadata_is_sick(error))
 		xfs_rt_mark_sick(mp, XFS_SICK_RT_SUMMARY);
-	if (error) {
-		xfs_imeta_irele(mp->m_rbmip);
-		return error;
-	}
+	if (error)
+		goto out_rbm;
 	ASSERT(mp->m_rsumip != NULL);
+
+	/* If we have rmap and a realtime device, look for the rtrmapbt. */
+	if (xfs_has_rtrmapbt(mp)) {
+		error = xfs_imeta_lookup(mp, &XFS_IMETA_RTRMAPBT, &ino);
+		if (error)
+			goto out_rsum;
+
+		error = xfs_rt_iget(mp, ino, &xfs_rrmapip_key,
+				&mp->m_rrmapip);
+		if (error)
+			goto out_rsum;
+
+		if (XFS_IS_CORRUPT(mp,
+				   mp->m_rrmapip->i_df.if_format !=
+				   XFS_DINODE_FMT_RMAP)) {
+			error = -EFSCORRUPTED;
+			goto out_rrmap;
+		}
+	}
+
 	xfs_alloc_rsum_cache(mp, sbp->sb_rbmblocks);
 	return 0;
+out_rrmap:
+	xfs_imeta_irele(mp->m_rrmapip);
+out_rsum:
+	xfs_imeta_irele(mp->m_rsumip);
+out_rbm:
+	xfs_imeta_irele(mp->m_rbmip);
+	return error;
 }
 
 /* Attach dquots for realtime metadata files. */
@@ -1382,6 +1410,12 @@ xfs_rtmount_dqattach(
 	if (error)
 		return error;
 
+	if (xfs_has_rtrmapbt(mp)) {
+		error = xfs_qm_dqattach(mp->m_rrmapip);
+		if (error)
+			return error;
+	}
+
 	return 0;
 }
 
@@ -1390,6 +1424,8 @@ xfs_rtunmount_inodes(
 	struct xfs_mount	*mp)
 {
 	kmem_free(mp->m_rsum_cache);
+	if (mp->m_rrmapip)
+		xfs_imeta_irele(mp->m_rrmapip);
 	if (mp->m_rbmip)
 		xfs_imeta_irele(mp->m_rbmip);
 	if (mp->m_rsumip)
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 7da298008ea7..3988ab860564 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -2201,6 +2201,7 @@ TRACE_DEFINE_ENUM(XFS_DINODE_FMT_LOCAL);
 TRACE_DEFINE_ENUM(XFS_DINODE_FMT_EXTENTS);
 TRACE_DEFINE_ENUM(XFS_DINODE_FMT_BTREE);
 TRACE_DEFINE_ENUM(XFS_DINODE_FMT_UUID);
+TRACE_DEFINE_ENUM(XFS_DINODE_FMT_RMAP);
 
 DECLARE_EVENT_CLASS(xfs_swap_extent_class,
 	TP_PROTO(struct xfs_inode *ip, int which),