1 files changed, 126 insertions, 88 deletions

diff --git a/fs/xfs/libxfs/xfs_ag_resv.c b/fs/xfs/libxfs/xfs_ag_resv.c
index b008ff3250eb..8ac8230c3d3c 100644
--- a/fs/xfs/libxfs/xfs_ag_resv.c
+++ b/fs/xfs/libxfs/xfs_ag_resv.c
@@ -1,21 +1,7 @@
+// SPDX-License-Identifier: GPL-2.0+
 /*
  * Copyright (C) 2016 Oracle.  All Rights Reserved.
- *
  * Author: Darrick J. Wong <darrick.wong@oracle.com>
- *
- * 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; either version 2
- * of the License, or (at your option) any later version.
- *
- * 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.
  */
 #include "xfs.h"
 #include "xfs_fs.h"
@@ -23,23 +9,18 @@
 #include "xfs_format.h"
 #include "xfs_log_format.h"
 #include "xfs_trans_resv.h"
-#include "xfs_sb.h"
 #include "xfs_mount.h"
-#include "xfs_defer.h"
 #include "xfs_alloc.h"
+#include "xfs_errortag.h"
 #include "xfs_error.h"
 #include "xfs_trace.h"
-#include "xfs_cksum.h"
 #include "xfs_trans.h"
-#include "xfs_bit.h"
-#include "xfs_bmap.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_ag_resv.h"
-#include "xfs_trans_space.h"
 #include "xfs_rmap_btree.h"
 #include "xfs_btree.h"
 #include "xfs_refcount_btree.h"
 #include "xfs_ialloc_btree.h"
+#include "xfs_ag.h"
+#include "xfs_ag_resv.h"
 
 /*
  * Per-AG Block Reservations
@@ -89,18 +70,19 @@ xfs_ag_resv_critical(
 	struct xfs_perag		*pag,
 	enum xfs_ag_resv_type		type)
 {
+	struct xfs_mount		*mp = pag_mount(pag);
 	xfs_extlen_t			avail;
 	xfs_extlen_t			orig;
 
 	switch (type) {
 	case XFS_AG_RESV_METADATA:
-		avail = pag->pagf_freeblks - pag->pag_agfl_resv.ar_reserved;
+		avail = pag->pagf_freeblks - pag->pag_rmapbt_resv.ar_reserved;
 		orig = pag->pag_meta_resv.ar_asked;
 		break;
-	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_RMAPBT:
 		avail = pag->pagf_freeblks + pag->pagf_flcount -
 			pag->pag_meta_resv.ar_reserved;
-		orig = pag->pag_agfl_resv.ar_asked;
+		orig = pag->pag_rmapbt_resv.ar_asked;
 		break;
 	default:
 		ASSERT(0);
@@ -110,8 +92,8 @@ xfs_ag_resv_critical(
 	trace_xfs_ag_resv_critical(pag, type, avail);
 
 	/* Critically low if less than 10% or max btree height remains. */
-	return XFS_TEST_ERROR(avail < orig / 10 || avail < XFS_BTREE_MAXLEVELS,
-			pag->pag_mount, XFS_ERRTAG_AG_RESV_CRITICAL);
+	return avail < orig / 10 || avail < mp->m_agbtree_maxlevels ||
+		XFS_TEST_ERROR(mp, XFS_ERRTAG_AG_RESV_CRITICAL);
 }
 
 /*
@@ -125,12 +107,13 @@ xfs_ag_resv_needed(
 {
 	xfs_extlen_t			len;
 
-	len = pag->pag_meta_resv.ar_reserved + pag->pag_agfl_resv.ar_reserved;
+	len = pag->pag_meta_resv.ar_reserved + pag->pag_rmapbt_resv.ar_reserved;
 	switch (type) {
 	case XFS_AG_RESV_METADATA:
-	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_RMAPBT:
 		len -= xfs_perag_resv(pag, type)->ar_reserved;
 		break;
+	case XFS_AG_RESV_METAFILE:
 	case XFS_AG_RESV_NONE:
 		/* empty */
 		break;
@@ -144,50 +127,41 @@ xfs_ag_resv_needed(
 }
 
 /* Clean out a reservation */
-static int
+static void
 __xfs_ag_resv_free(
 	struct xfs_perag		*pag,
 	enum xfs_ag_resv_type		type)
 {
 	struct xfs_ag_resv		*resv;
 	xfs_extlen_t			oldresv;
-	int				error;
 
 	trace_xfs_ag_resv_free(pag, type, 0);
 
 	resv = xfs_perag_resv(pag, type);
-	pag->pag_mount->m_ag_max_usable += resv->ar_asked;
+	if (pag_agno(pag) == 0)
+		pag_mount(pag)->m_ag_max_usable += resv->ar_asked;
 	/*
-	 * AGFL blocks are always considered "free", so whatever
-	 * was reserved at mount time must be given back at umount.
+	 * RMAPBT blocks come from the AGFL and AGFL blocks are always
+	 * considered "free", so whatever was reserved at mount time must be
+	 * given back at umount.
 	 */
-	if (type == XFS_AG_RESV_AGFL)
+	if (type == XFS_AG_RESV_RMAPBT)
 		oldresv = resv->ar_orig_reserved;
 	else
 		oldresv = resv->ar_reserved;
-	error = xfs_mod_fdblocks(pag->pag_mount, oldresv, true);
+	xfs_add_fdblocks(pag_mount(pag), oldresv);
 	resv->ar_reserved = 0;
 	resv->ar_asked = 0;
-
-	if (error)
-		trace_xfs_ag_resv_free_error(pag->pag_mount, pag->pag_agno,
-				error, _RET_IP_);
-	return error;
+	resv->ar_orig_reserved = 0;
 }
 
 /* Free a per-AG reservation. */
-int
+void
 xfs_ag_resv_free(
 	struct xfs_perag		*pag)
 {
-	int				error;
-	int				err2;
-
-	error = __xfs_ag_resv_free(pag, XFS_AG_RESV_AGFL);
-	err2 = __xfs_ag_resv_free(pag, XFS_AG_RESV_METADATA);
-	if (err2 && !error)
-		error = err2;
-	return error;
+	__xfs_ag_resv_free(pag, XFS_AG_RESV_RMAPBT);
+	__xfs_ag_resv_free(pag, XFS_AG_RESV_METADATA);
 }
 
 static int
@@ -197,30 +171,62 @@ __xfs_ag_resv_init(
 	xfs_extlen_t			ask,
 	xfs_extlen_t			used)
 {
-	struct xfs_mount		*mp = pag->pag_mount;
+	struct xfs_mount		*mp = pag_mount(pag);
 	struct xfs_ag_resv		*resv;
 	int				error;
-	xfs_extlen_t			reserved;
+	xfs_extlen_t			hidden_space;
 
 	if (used > ask)
 		ask = used;
-	reserved = ask - used;
 
-	error = xfs_mod_fdblocks(mp, -(int64_t)reserved, true);
+	switch (type) {
+	case XFS_AG_RESV_RMAPBT:
+		/*
+		 * Space taken by the rmapbt is not subtracted from fdblocks
+		 * because the rmapbt lives in the free space.  Here we must
+		 * subtract the entire reservation from fdblocks so that we
+		 * always have blocks available for rmapbt expansion.
+		 */
+		hidden_space = ask;
+		break;
+	case XFS_AG_RESV_METADATA:
+		/*
+		 * Space taken by all other metadata btrees are accounted
+		 * on-disk as used space.  We therefore only hide the space
+		 * that is reserved but not used by the trees.
+		 */
+		hidden_space = ask - used;
+		break;
+	default:
+		ASSERT(0);
+		return -EINVAL;
+	}
+
+	if (XFS_TEST_ERROR(mp, XFS_ERRTAG_AG_RESV_FAIL))
+		error = -ENOSPC;
+	else
+		error = xfs_dec_fdblocks(mp, hidden_space, true);
 	if (error) {
-		trace_xfs_ag_resv_init_error(pag->pag_mount, pag->pag_agno,
-				error, _RET_IP_);
+		trace_xfs_ag_resv_init_error(pag, error, _RET_IP_);
 		xfs_warn(mp,
 "Per-AG reservation for AG %u failed.  Filesystem may run out of space.",
-				pag->pag_agno);
+				pag_agno(pag));
 		return error;
 	}
 
-	mp->m_ag_max_usable -= ask;
+	/*
+	 * Reduce the maximum per-AG allocation length by however much we're
+	 * trying to reserve for an AG.  Since this is a filesystem-wide
+	 * counter, we only make the adjustment for AG 0.  This assumes that
+	 * there aren't any AGs hungrier for per-AG reservation than AG 0.
+	 */
+	if (pag_agno(pag) == 0)
+		mp->m_ag_max_usable -= ask;
 
 	resv = xfs_perag_resv(pag, type);
 	resv->ar_asked = ask;
-	resv->ar_reserved = resv->ar_orig_reserved = reserved;
+	resv->ar_orig_reserved = hidden_space;
+	resv->ar_reserved = ask - used;
 
 	trace_xfs_ag_resv_init(pag, type, ask);
 	return 0;
@@ -229,23 +235,24 @@ __xfs_ag_resv_init(
 /* Create a per-AG block reservation. */
 int
 xfs_ag_resv_init(
-	struct xfs_perag		*pag)
+	struct xfs_perag		*pag,
+	struct xfs_trans		*tp)
 {
-	struct xfs_mount		*mp = pag->pag_mount;
-	xfs_agnumber_t			agno = pag->pag_agno;
+	struct xfs_mount		*mp = pag_mount(pag);
 	xfs_extlen_t			ask;
 	xfs_extlen_t			used;
-	int				error = 0;
+	int				error = 0, error2;
+	bool				has_resv = false;
 
 	/* Create the metadata reservation. */
 	if (pag->pag_meta_resv.ar_asked == 0) {
 		ask = used = 0;
 
-		error = xfs_refcountbt_calc_reserves(mp, agno, &ask, &used);
+		error = xfs_refcountbt_calc_reserves(mp, tp, pag, &ask, &used);
 		if (error)
 			goto out;
 
-		error = xfs_finobt_calc_reserves(mp, agno, &ask, &used);
+		error = xfs_finobt_calc_reserves(pag, tp, &ask, &used);
 		if (error)
 			goto out;
 
@@ -261,9 +268,9 @@ xfs_ag_resv_init(
 			 */
 			ask = used = 0;
 
-			mp->m_inotbt_nores = true;
+			mp->m_finobt_nores = true;
 
-			error = xfs_refcountbt_calc_reserves(mp, agno, &ask,
+			error = xfs_refcountbt_calc_reserves(mp, tp, pag, &ask,
 					&used);
 			if (error)
 				goto out;
@@ -273,32 +280,55 @@ xfs_ag_resv_init(
 			if (error)
 				goto out;
 		}
+		if (ask)
+			has_resv = true;
 	}
 
-	/* Create the AGFL metadata reservation */
-	if (pag->pag_agfl_resv.ar_asked == 0) {
+	/* Create the RMAPBT metadata reservation */
+	if (pag->pag_rmapbt_resv.ar_asked == 0) {
 		ask = used = 0;
 
-		error = xfs_rmapbt_calc_reserves(mp, agno, &ask, &used);
+		error = xfs_rmapbt_calc_reserves(mp, tp, pag, &ask, &used);
 		if (error)
 			goto out;
 
-		error = __xfs_ag_resv_init(pag, XFS_AG_RESV_AGFL, ask, used);
+		error = __xfs_ag_resv_init(pag, XFS_AG_RESV_RMAPBT, ask, used);
 		if (error)
 			goto out;
+		if (ask)
+			has_resv = true;
 	}
 
-#ifdef DEBUG
-	/* need to read in the AGF for the ASSERT below to work */
-	error = xfs_alloc_pagf_init(pag->pag_mount, NULL, pag->pag_agno, 0);
-	if (error)
-		return error;
-
-	ASSERT(xfs_perag_resv(pag, XFS_AG_RESV_METADATA)->ar_reserved +
-	       xfs_perag_resv(pag, XFS_AG_RESV_AGFL)->ar_reserved <=
-	       pag->pagf_freeblks + pag->pagf_flcount);
-#endif
 out:
+	/*
+	 * Initialize the pagf if we have at least one active reservation on the
+	 * AG. This may have occurred already via reservation calculation, but
+	 * fall back to an explicit init to ensure the in-core allocbt usage
+	 * counters are initialized as soon as possible. This is important
+	 * because filesystems with large perag reservations are susceptible to
+	 * free space reservation problems that the allocbt counter is used to
+	 * address.
+	 */
+	if (has_resv) {
+		error2 = xfs_alloc_read_agf(pag, tp, 0, NULL);
+		if (error2)
+			return error2;
+
+		/*
+		 * If there isn't enough space in the AG to satisfy the
+		 * reservation, let the caller know that there wasn't enough
+		 * space.  Callers are responsible for deciding what to do
+		 * next, since (in theory) we can stumble along with
+		 * insufficient reservation if data blocks are being freed to
+		 * replenish the AG's free space.
+		 */
+		if (!error &&
+		    xfs_perag_resv(pag, XFS_AG_RESV_METADATA)->ar_reserved +
+		    xfs_perag_resv(pag, XFS_AG_RESV_RMAPBT)->ar_reserved >
+		    pag->pagf_freeblks + pag->pagf_flcount)
+			error = -ENOSPC;
+	}
+
 	return error;
 }
 
@@ -316,13 +346,16 @@ xfs_ag_resv_alloc_extent(
 	trace_xfs_ag_resv_alloc_extent(pag, type, args->len);
 
 	switch (type) {
-	case XFS_AG_RESV_METADATA:
 	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_METAFILE:
+		return;
+	case XFS_AG_RESV_METADATA:
+	case XFS_AG_RESV_RMAPBT:
 		resv = xfs_perag_resv(pag, type);
 		break;
 	default:
 		ASSERT(0);
-		/* fall through */
+		fallthrough;
 	case XFS_AG_RESV_NONE:
 		field = args->wasdel ? XFS_TRANS_SB_RES_FDBLOCKS :
 				       XFS_TRANS_SB_FDBLOCKS;
@@ -332,7 +365,7 @@ xfs_ag_resv_alloc_extent(
 
 	len = min_t(xfs_extlen_t, args->len, resv->ar_reserved);
 	resv->ar_reserved -= len;
-	if (type == XFS_AG_RESV_AGFL)
+	if (type == XFS_AG_RESV_RMAPBT)
 		return;
 	/* Allocations of reserved blocks only need on-disk sb updates... */
 	xfs_trans_mod_sb(args->tp, XFS_TRANS_SB_RES_FDBLOCKS, -(int64_t)len);
@@ -356,21 +389,26 @@ xfs_ag_resv_free_extent(
 	trace_xfs_ag_resv_free_extent(pag, type, len);
 
 	switch (type) {
-	case XFS_AG_RESV_METADATA:
 	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_METAFILE:
+		return;
+	case XFS_AG_RESV_METADATA:
+	case XFS_AG_RESV_RMAPBT:
 		resv = xfs_perag_resv(pag, type);
 		break;
 	default:
 		ASSERT(0);
-		/* fall through */
+		fallthrough;
 	case XFS_AG_RESV_NONE:
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, (int64_t)len);
+		fallthrough;
+	case XFS_AG_RESV_IGNORE:
 		return;
 	}
 
 	leftover = min_t(xfs_extlen_t, len, resv->ar_asked - resv->ar_reserved);
 	resv->ar_reserved += leftover;
-	if (type == XFS_AG_RESV_AGFL)
+	if (type == XFS_AG_RESV_RMAPBT)
 		return;
 	/* Freeing into the reserved pool only requires on-disk update... */
 	xfs_trans_mod_sb(tp, XFS_TRANS_SB_RES_FDBLOCKS, len);