Merge branch 'devel-stable' into devel

19 files changed, 170 insertions, 93 deletions

diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c
index f96eff04e11a..a6395bdb26ae 100644
--- a/fs/binfmt_aout.c
+++ b/fs/binfmt_aout.c
@@ -134,10 +134,6 @@ static int aout_core_dump(struct coredump_params *cprm)
                 if (!dump_write(file, dump_start, dump_size))
                         goto end_coredump;
         }
-/* Finally dump the task struct.  Not be used by gdb, but could be useful */
-        set_fs(KERNEL_DS);
-        if (!dump_write(file, current, sizeof(*current)))
-                goto end_coredump;
 end_coredump:
         set_fs(fs);
         return has_dumped;
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 73c153092f72..5e9da996a151 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -2283,7 +2283,8 @@ static void handle_cap_grant(struct inode *inode, struct ceph_mds_caps *grant,
 {
         struct ceph_inode_info *ci = ceph_inode(inode);
         int mds = session->s_mds;
-        int seq = le32_to_cpu(grant->seq);
+        unsigned seq = le32_to_cpu(grant->seq);
+        unsigned issue_seq = le32_to_cpu(grant->issue_seq);
         int newcaps = le32_to_cpu(grant->caps);
         int issued, implemented, used, wanted, dirty;
         u64 size = le64_to_cpu(grant->size);
@@ -2295,8 +2296,8 @@ static void handle_cap_grant(struct inode *inode, struct ceph_mds_caps *grant,
         int revoked_rdcache = 0;
         int queue_invalidate = 0;
 
-        dout("handle_cap_grant inode %p cap %p mds%d seq %d %s\n",
-             inode, cap, mds, seq, ceph_cap_string(newcaps));
+        dout("handle_cap_grant inode %p cap %p mds%d seq %u/%u %s\n",
+             inode, cap, mds, seq, issue_seq, ceph_cap_string(newcaps));
         dout(" size %llu max_size %llu, i_size %llu\n", size, max_size,
                 inode->i_size);
 
@@ -2392,6 +2393,7 @@ static void handle_cap_grant(struct inode *inode, struct ceph_mds_caps *grant,
         }
 
         cap->seq = seq;
+        cap->issue_seq = issue_seq;
 
         /* file layout may have changed */
         ci->i_layout = grant->layout;
@@ -2774,15 +2776,7 @@ void ceph_handle_caps(struct ceph_mds_session *session,
                 if (op == CEPH_CAP_OP_IMPORT)
                         __queue_cap_release(session, vino.ino, cap_id,
                                             mseq, seq);
-
-                /*
-                 * send any full release message to try to move things
-                 * along for the mds (who clearly thinks we still have this
-                 * cap).
-                 */
-                ceph_add_cap_releases(mdsc, session);
-                ceph_send_cap_releases(mdsc, session);
-                goto done;
+                goto flush_cap_releases;
         }
 
         /* these will work even if we don't have a cap yet */
@@ -2810,7 +2804,7 @@ void ceph_handle_caps(struct ceph_mds_session *session,
                 dout(" no cap on %p ino %llx.%llx from mds%d\n",
                      inode, ceph_ino(inode), ceph_snap(inode), mds);
                 spin_unlock(&inode->i_lock);
-                goto done;
+                goto flush_cap_releases;
         }
 
         /* note that each of these drops i_lock for us */
@@ -2834,6 +2828,17 @@ void ceph_handle_caps(struct ceph_mds_session *session,
                        ceph_cap_op_name(op));
         }
 
+        goto done;
+
+flush_cap_releases:
+        /*
+         * send any full release message to try to move things
+         * along for the mds (who clearly thinks we still have this
+         * cap).
+         */
+        ceph_add_cap_releases(mdsc, session);
+        ceph_send_cap_releases(mdsc, session);
+
 done:
         mutex_unlock(&session->s_mutex);
 done_unlocked:
diff --git a/fs/ceph/export.c b/fs/ceph/export.c
index 4480cb1c63e7..e38423e82f2e 100644
--- a/fs/ceph/export.c
+++ b/fs/ceph/export.c
@@ -42,32 +42,37 @@ struct ceph_nfs_confh {
 static int ceph_encode_fh(struct dentry *dentry, u32 *rawfh, int *max_len,
                           int connectable)
 {
+        int type;
         struct ceph_nfs_fh *fh = (void *)rawfh;
         struct ceph_nfs_confh *cfh = (void *)rawfh;
         struct dentry *parent = dentry->d_parent;
         struct inode *inode = dentry->d_inode;
-        int type;
+        int connected_handle_length = sizeof(*cfh)/4;
+        int handle_length = sizeof(*fh)/4;
 
         /* don't re-export snaps */
         if (ceph_snap(inode) != CEPH_NOSNAP)
                 return -EINVAL;
 
-        if (*max_len >= sizeof(*cfh)) {
+        if (*max_len >= connected_handle_length) {
                 dout("encode_fh %p connectable\n", dentry);
                 cfh->ino = ceph_ino(dentry->d_inode);
                 cfh->parent_ino = ceph_ino(parent->d_inode);
                 cfh->parent_name_hash = parent->d_name.hash;
-                *max_len = sizeof(*cfh);
+                *max_len = connected_handle_length;
                 type = 2;
-        } else if (*max_len > sizeof(*fh)) {
-                if (connectable)
-                        return -ENOSPC;
+        } else if (*max_len >= handle_length) {
+                if (connectable) {
+                        *max_len = connected_handle_length;
+                        return 255;
+                }
                 dout("encode_fh %p\n", dentry);
                 fh->ino = ceph_ino(dentry->d_inode);
-                *max_len = sizeof(*fh);
+                *max_len = handle_length;
                 type = 1;
         } else {
-                return -ENOSPC;
+                *max_len = handle_length;
+                return 255;
         }
         return type;
 }
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 8c044a4f0457..66e4da6dba22 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -697,7 +697,7 @@ more:
                          * start_request so that a tid has been assigned.
                          */
                         spin_lock(&ci->i_unsafe_lock);
-                        list_add(&ci->i_unsafe_writes, &req->r_unsafe_item);
+                        list_add(&req->r_unsafe_item, &ci->i_unsafe_writes);
                         spin_unlock(&ci->i_unsafe_lock);
                         ceph_get_cap_refs(ci, CEPH_CAP_FILE_WR);
                 }
diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c
index dfced1dacbcd..3b5571b8ce22 100644
--- a/fs/ceph/osd_client.c
+++ b/fs/ceph/osd_client.c
@@ -549,7 +549,7 @@ static void __unregister_request(struct ceph_osd_client *osdc,
  */
 static void __cancel_request(struct ceph_osd_request *req)
 {
-        if (req->r_sent) {
+        if (req->r_sent && req->r_osd) {
                 ceph_con_revoke(&req->r_osd->o_con, req->r_request);
                 req->r_sent = 0;
         }
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
index c65c3419dd37..7e83b356cc9e 100644
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -232,7 +232,7 @@ static int
 small_smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
                 void **request_buf)
 {
-        int rc = 0;
+        int rc;
 
         rc = cifs_reconnect_tcon(tcon, smb_command);
         if (rc)
@@ -250,7 +250,7 @@ small_smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
         if (tcon != NULL)
                 cifs_stats_inc(&tcon->num_smbs_sent);
 
-        return rc;
+        return 0;
 }
 
 int
@@ -281,16 +281,9 @@ small_smb_init_no_tc(const int smb_command, const int wct,
 
 /* If the return code is zero, this function must fill in request_buf pointer */
 static int
-smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
-         void **request_buf /* returned */ ,
-         void **response_buf /* returned */ )
+__smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
+                        void **request_buf, void **response_buf)
 {
-        int rc = 0;
-
-        rc = cifs_reconnect_tcon(tcon, smb_command);
-        if (rc)
-                return rc;
-
         *request_buf = cifs_buf_get();
         if (*request_buf == NULL) {
                 /* BB should we add a retry in here if not a writepage? */
@@ -309,7 +302,31 @@ smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
         if (tcon != NULL)
                 cifs_stats_inc(&tcon->num_smbs_sent);
 
-        return rc;
+        return 0;
+}
+
+/* If the return code is zero, this function must fill in request_buf pointer */
+static int
+smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
+         void **request_buf, void **response_buf)
+{
+        int rc;
+
+        rc = cifs_reconnect_tcon(tcon, smb_command);
+        if (rc)
+                return rc;
+
+        return __smb_init(smb_command, wct, tcon, request_buf, response_buf);
+}
+
+static int
+smb_init_no_reconnect(int smb_command, int wct, struct cifsTconInfo *tcon,
+                        void **request_buf, void **response_buf)
+{
+        if (tcon->ses->need_reconnect || tcon->need_reconnect)
+                return -EHOSTDOWN;
+
+        return __smb_init(smb_command, wct, tcon, request_buf, response_buf);
 }
 
 static int validate_t2(struct smb_t2_rsp *pSMB)
@@ -4534,8 +4551,8 @@ CIFSSMBQFSUnixInfo(const int xid, struct cifsTconInfo *tcon)
 
         cFYI(1, "In QFSUnixInfo");
 QFSUnixRetry:
-        rc = smb_init(SMB_COM_TRANSACTION2, 15, tcon, (void **) &pSMB,
-                      (void **) &pSMBr);
+        rc = smb_init_no_reconnect(SMB_COM_TRANSACTION2, 15, tcon,
+                                   (void **) &pSMB, (void **) &pSMBr);
         if (rc)
                 return rc;
 
@@ -4604,8 +4621,8 @@ CIFSSMBSetFSUnixInfo(const int xid, struct cifsTconInfo *tcon, __u64 cap)
         cFYI(1, "In SETFSUnixInfo");
 SETFSUnixRetry:
         /* BB switch to small buf init to save memory */
-        rc = smb_init(SMB_COM_TRANSACTION2, 15, tcon, (void **) &pSMB,
-                      (void **) &pSMBr);
+        rc = smb_init_no_reconnect(SMB_COM_TRANSACTION2, 15, tcon,
+                                        (void **) &pSMB, (void **) &pSMBr);
         if (rc)
                 return rc;
 
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 93f77d438d3c..53cce8cc2224 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -801,6 +801,8 @@ retry_iget5_locked:
                         inode->i_flags |= S_NOATIME | S_NOCMTIME;
                 if (inode->i_state & I_NEW) {
                         inode->i_ino = hash;
+                        if (S_ISREG(inode->i_mode))
+                                inode->i_data.backing_dev_info = sb->s_bdi;
 #ifdef CONFIG_CIFS_FSCACHE
                         /* initialize per-inode cache cookie pointer */
                         CIFS_I(inode)->fscache = NULL;
diff --git a/fs/exec.c b/fs/exec.c
index 828dd2461d6b..03278c984ba0 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -2014,3 +2014,41 @@ fail_creds:
 fail:
         return;
 }
+
+/*
+ * Core dumping helper functions.  These are the only things you should
+ * do on a core-file: use only these functions to write out all the
+ * necessary info.
+ */
+int dump_write(struct file *file, const void *addr, int nr)
+{
+        return access_ok(VERIFY_READ, addr, nr) && file->f_op->write(file, addr, nr, &file->f_pos) == nr;
+}
+
+int dump_seek(struct file *file, loff_t off)
+{
+        int ret = 1;
+
+        if (file->f_op->llseek && file->f_op->llseek != no_llseek) {
+                if (file->f_op->llseek(file, off, SEEK_CUR) < 0)
+                        return 0;
+        } else {
+                char *buf = (char *)get_zeroed_page(GFP_KERNEL);
+
+                if (!buf)
+                        return 0;
+                while (off > 0) {
+                        unsigned long n = off;
+
+                        if (n > PAGE_SIZE)
+                                n = PAGE_SIZE;
+                        if (!dump_write(file, buf, n)) {
+                                ret = 0;
+                                break;
+                        }
+                        off -= n;
+                }
+                free_page((unsigned long)buf);
+        }
+        return ret;
+}
diff --git a/fs/exofs/inode.c b/fs/exofs/inode.c
index eb7368ebd8cd..3eadd97324b1 100644
--- a/fs/exofs/inode.c
+++ b/fs/exofs/inode.c
@@ -54,6 +54,9 @@ struct page_collect {
         unsigned nr_pages;
         unsigned long length;
         loff_t pg_first; /* keep 64bit also in 32-arches */
+        bool read_4_write; /* This means two things: that the read is sync
+                            * And the pages should not be unlocked.
+                            */
 };
 
 static void _pcol_init(struct page_collect *pcol, unsigned expected_pages,
@@ -71,6 +74,7 @@ static void _pcol_init(struct page_collect *pcol, unsigned expected_pages,
         pcol->nr_pages = 0;
         pcol->length = 0;
         pcol->pg_first = -1;
+        pcol->read_4_write = false;
 }
 
 static void _pcol_reset(struct page_collect *pcol)
@@ -347,7 +351,8 @@ static int readpage_strip(void *data, struct page *page)
                 if (PageError(page))
                         ClearPageError(page);
 
-                unlock_page(page);
+                if (!pcol->read_4_write)
+                        unlock_page(page);
                 EXOFS_DBGMSG("readpage_strip(0x%lx, 0x%lx) empty page,"
                              " splitting\n", inode->i_ino, page->index);
 
@@ -428,6 +433,7 @@ static int _readpage(struct page *page, bool is_sync)
         /* readpage_strip might call read_exec(,is_sync==false) at several
          * places but not if we have a single page.
          */
+        pcol.read_4_write = is_sync;
         ret = readpage_strip(&pcol, page);
         if (ret) {
                 EXOFS_ERR("_readpage => %d\n", ret);
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 5581122bd2c0..ab38fef1c9a1 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -72,22 +72,11 @@ int writeback_in_progress(struct backing_dev_info *bdi)
 static inline struct backing_dev_info *inode_to_bdi(struct inode *inode)
 {
         struct super_block *sb = inode->i_sb;
-        struct backing_dev_info *bdi = inode->i_mapping->backing_dev_info;
 
-        /*
-         * For inodes on standard filesystems, we use superblock's bdi. For
-         * inodes on virtual filesystems, we want to use inode mapping's bdi
-         * because they can possibly point to something useful (think about
-         * block_dev filesystem).
-         */
-        if (sb->s_bdi && sb->s_bdi != &noop_backing_dev_info) {
-                /* Some device inodes could play dirty tricks. Catch them... */
-                WARN(bdi != sb->s_bdi && bdi_cap_writeback_dirty(bdi),
-                        "Dirtiable inode bdi %s != sb bdi %s\n",
-                        bdi->name, sb->s_bdi->name);
-                return sb->s_bdi;
-        }
-        return bdi;
+        if (strcmp(sb->s_type->name, "bdev") == 0)
+                return inode->i_mapping->backing_dev_info;
+
+        return sb->s_bdi;
 }
 
 static void bdi_queue_work(struct backing_dev_info *bdi,
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index d367af1514ef..cde755cca564 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -1354,7 +1354,7 @@ static int fuse_retrieve(struct fuse_conn *fc, struct inode *inode,
         loff_t file_size;
         unsigned int num;
         unsigned int offset;
-        size_t total_len;
+        size_t total_len = 0;
 
         req = fuse_get_req(fc);
         if (IS_ERR(req))
diff --git a/fs/nfsd/nfsfh.h b/fs/nfsd/nfsfh.h
index cdfb8c6a4206..c16f8d8331b5 100644
--- a/fs/nfsd/nfsfh.h
+++ b/fs/nfsd/nfsfh.h
@@ -196,8 +196,6 @@ fh_lock(struct svc_fh *fhp)
 static inline void
 fh_unlock(struct svc_fh *fhp)
 {
-        BUG_ON(!fhp->fh_dentry);
-
         if (fhp->fh_locked) {
                 fill_post_wcc(fhp);
                 mutex_unlock(&fhp->fh_dentry->d_inode->i_mutex);
diff --git a/fs/notify/Kconfig b/fs/notify/Kconfig
index 22c629eedd82..b388443c3a09 100644
--- a/fs/notify/Kconfig
+++ b/fs/notify/Kconfig
@@ -3,4 +3,4 @@ config FSNOTIFY
 
 source "fs/notify/dnotify/Kconfig"
 source "fs/notify/inotify/Kconfig"
-source "fs/notify/fanotify/Kconfig"
+#source "fs/notify/fanotify/Kconfig"
diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c
index 32499d213fc4..9975457c981f 100644
--- a/fs/ocfs2/symlink.c
+++ b/fs/ocfs2/symlink.c
@@ -128,7 +128,7 @@ static void *ocfs2_fast_follow_link(struct dentry *dentry,
         }
 
         /* Fast symlinks can't be large */
-        len = strlen(target);
+        len = strnlen(target, ocfs2_fast_symlink_chars(inode->i_sb));
         link = kzalloc(len + 1, GFP_NOFS);
         if (!link) {
                 status = -ENOMEM;
diff --git a/fs/proc/base.c b/fs/proc/base.c
index a1c43e7c8a7b..8e4addaa5424 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -2675,7 +2675,7 @@ static const struct pid_entry tgid_base_stuff[] = {
         INF("auxv",       S_IRUSR, proc_pid_auxv),
         ONE("status",     S_IRUGO, proc_pid_status),
         ONE("personality", S_IRUSR, proc_pid_personality),
-        INF("limits",     S_IRUSR, proc_pid_limits),
+        INF("limits",     S_IRUGO, proc_pid_limits),
 #ifdef CONFIG_SCHED_DEBUG
         REG("sched",      S_IRUGO|S_IWUSR, proc_pid_sched_operations),
 #endif
@@ -3011,7 +3011,7 @@ static const struct pid_entry tid_base_stuff[] = {
         INF("auxv",      S_IRUSR, proc_pid_auxv),
         ONE("status",    S_IRUGO, proc_pid_status),
         ONE("personality", S_IRUSR, proc_pid_personality),
-        INF("limits",    S_IRUSR, proc_pid_limits),
+        INF("limits",    S_IRUGO, proc_pid_limits),
 #ifdef CONFIG_SCHED_DEBUG
         REG("sched",     S_IRUGO|S_IWUSR, proc_pid_sched_operations),
 #endif
diff --git a/fs/reiserfs/ioctl.c b/fs/reiserfs/ioctl.c
index f53505de0712..5cbb81e134ac 100644
--- a/fs/reiserfs/ioctl.c
+++ b/fs/reiserfs/ioctl.c
@@ -170,6 +170,7 @@ int reiserfs_prepare_write(struct file *f, struct page *page,
 int reiserfs_unpack(struct inode *inode, struct file *filp)
 {
         int retval = 0;
+        int depth;
         int index;
         struct page *page;
         struct address_space *mapping;
@@ -188,8 +189,8 @@ int reiserfs_unpack(struct inode *inode, struct file *filp)
         /* we need to make sure nobody is changing the file size beneath
          ** us
          */
-        mutex_lock(&inode->i_mutex);
-        reiserfs_write_lock(inode->i_sb);
+        reiserfs_mutex_lock_safe(&inode->i_mutex, inode->i_sb);
+        depth = reiserfs_write_lock_once(inode->i_sb);
 
         write_from = inode->i_size & (blocksize - 1);
         /* if we are on a block boundary, we are already unpacked.  */
@@ -224,6 +225,6 @@ int reiserfs_unpack(struct inode *inode, struct file *filp)
 
       out:
         mutex_unlock(&inode->i_mutex);
-        reiserfs_write_unlock(inode->i_sb);
+        reiserfs_write_unlock_once(inode->i_sb, depth);
         return retval;
 }
diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c
index d59c4a65d492..81976ffed7d6 100644
--- a/fs/xfs/linux-2.6/xfs_sync.c
+++ b/fs/xfs/linux-2.6/xfs_sync.c
@@ -668,14 +668,11 @@ xfs_inode_set_reclaim_tag(
         xfs_perag_put(pag);
 }
 
-void
-__xfs_inode_clear_reclaim_tag(
-        xfs_mount_t     *mp,
+STATIC void
+__xfs_inode_clear_reclaim(
         xfs_perag_t     *pag,
         xfs_inode_t     *ip)
 {
-        radix_tree_tag_clear(&pag->pag_ici_root,
-                        XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
         pag->pag_ici_reclaimable--;
         if (!pag->pag_ici_reclaimable) {
                 /* clear the reclaim tag from the perag radix tree */
@@ -689,6 +686,17 @@ __xfs_inode_clear_reclaim_tag(
         }
 }
 
+void
+__xfs_inode_clear_reclaim_tag(
+        xfs_mount_t     *mp,
+        xfs_perag_t     *pag,
+        xfs_inode_t     *ip)
+{
+        radix_tree_tag_clear(&pag->pag_ici_root,
+                        XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
+        __xfs_inode_clear_reclaim(pag, ip);
+}
+
 /*
  * Inodes in different states need to be treated differently, and the return
  * value of xfs_iflush is not sufficient to get this right. The following table
@@ -838,6 +846,7 @@ reclaim:
         if (!radix_tree_delete(&pag->pag_ici_root,
                                 XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino)))
                 ASSERT(0);
+        __xfs_inode_clear_reclaim(pag, ip);
         write_unlock(&pag->pag_ici_lock);
 
         /*
diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c
index ed575fb4b495..7e206fc1fa36 100644
--- a/fs/xfs/xfs_log_cil.c
+++ b/fs/xfs/xfs_log_cil.c
@@ -405,9 +405,15 @@ xlog_cil_push(
         new_ctx = kmem_zalloc(sizeof(*new_ctx), KM_SLEEP|KM_NOFS);
         new_ctx->ticket = xlog_cil_ticket_alloc(log);
 
-        /* lock out transaction commit, but don't block on background push */
+        /*
+         * Lock out transaction commit, but don't block for background pushes
+         * unless we are well over the CIL space limit. See the definition of
+         * XLOG_CIL_HARD_SPACE_LIMIT() for the full explanation of the logic
+         * used here.
+         */
         if (!down_write_trylock(&cil->xc_ctx_lock)) {
-                if (!push_seq)
+                if (!push_seq &&
+                    cil->xc_ctx->space_used < XLOG_CIL_HARD_SPACE_LIMIT(log))
                         goto out_free_ticket;
                 down_write(&cil->xc_ctx_lock);
         }
@@ -422,7 +428,7 @@ xlog_cil_push(
                 goto out_skip;
 
         /* check for a previously pushed seqeunce */
-        if (push_seq < cil->xc_ctx->sequence)
+        if (push_seq && push_seq < cil->xc_ctx->sequence)
                 goto out_skip;
 
         /*
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index ced52b98b322..edcdfe01617f 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -426,13 +426,13 @@ struct xfs_cil {
 };
 
 /*
- * The amount of log space we should the CIL to aggregate is difficult to size.
- * Whatever we chose we have to make we can get a reservation for the log space
- * effectively, that it is large enough to capture sufficient relogging to
- * reduce log buffer IO significantly, but it is not too large for the log or
- * induces too much latency when writing out through the iclogs. We track both
- * space consumed and the number of vectors in the checkpoint context, so we
- * need to decide which to use for limiting.
+ * The amount of log space we allow the CIL to aggregate is difficult to size.
+ * Whatever we choose, we have to make sure we can get a reservation for the
+ * log space effectively, that it is large enough to capture sufficient
+ * relogging to reduce log buffer IO significantly, but it is not too large for
+ * the log or induces too much latency when writing out through the iclogs. We
+ * track both space consumed and the number of vectors in the checkpoint
+ * context, so we need to decide which to use for limiting.
  *
  * Every log buffer we write out during a push needs a header reserved, which
  * is at least one sector and more for v2 logs. Hence we need a reservation of
@@ -459,16 +459,21 @@ struct xfs_cil {
  * checkpoint transaction ticket is specific to the checkpoint context, rather
  * than the CIL itself.
  *
- * With dynamic reservations, we can basically make up arbitrary limits for the
- * checkpoint size so long as they don't violate any other size rules.  Hence
- * the initial maximum size for the checkpoint transaction will be set to a
- * quarter of the log or 8MB, which ever is smaller. 8MB is an arbitrary limit
- * right now based on the latency of writing out a large amount of data through
- * the circular iclog buffers.
+ * With dynamic reservations, we can effectively make up arbitrary limits for
+ * the checkpoint size so long as they don't violate any other size rules.
+ * Recovery imposes a rule that no transaction exceed half the log, so we are
+ * limited by that.  Furthermore, the log transaction reservation subsystem
+ * tries to keep 25% of the log free, so we need to keep below that limit or we
+ * risk running out of free log space to start any new transactions.
+ *
+ * In order to keep background CIL push efficient, we will set a lower
+ * threshold at which background pushing is attempted without blocking current
+ * transaction commits.  A separate, higher bound defines when CIL pushes are
+ * enforced to ensure we stay within our maximum checkpoint size bounds.
+ * threshold, yet give us plenty of space for aggregation on large logs.
  */
-
-#define XLOG_CIL_SPACE_LIMIT(log)       \
-        (min((log->l_logsize >> 2), (8 * 1024 * 1024)))
+#define XLOG_CIL_SPACE_LIMIT(log)       (log->l_logsize >> 3)
+#define XLOG_CIL_HARD_SPACE_LIMIT(log)  (3 * (log->l_logsize >> 4))
 
 /*
  * The reservation head lsn is not made up of a cycle number and block number.