Merge tag 'ceph-for-4.18-rc1' of git://github.com/ceph/ceph-client

Pull ceph updates from Ilya Dryomov:
 "The main piece is a set of libceph changes that revamps how OSD
  requests are aborted, improving CephFS ENOSPC handling and making
  "umount -f" actually work (Zheng and myself).

  The rest is mostly mount option handling cleanups from Chengguang and
  assorted fixes from Zheng, Luis and Dongsheng.

* tag 'ceph-for-4.18-rc1' of git://github.com/ceph/ceph-client: (31 commits)
  rbd: flush rbd_dev->watch_dwork after watch is unregistered
  ceph: update description of some mount options
  ceph: show ino32 if the value is different with default
  ceph: strengthen rsize/wsize/readdir_max_bytes validation
  ceph: fix alignment of rasize
  ceph: fix use-after-free in ceph_statfs()
  ceph: prevent i_version from going back
  ceph: fix wrong check for the case of updating link count
  libceph: allocate the locator string with GFP_NOFAIL
  libceph: make abort_on_full a per-osdc setting
  libceph: don't abort reads in ceph_osdc_abort_on_full()
  libceph: avoid a use-after-free during map check
  libceph: don't warn if req->r_abort_on_full is set
  libceph: use for_each_request() in ceph_osdc_abort_on_full()
  libceph: defer __complete_request() to a workqueue
  libceph: move more code into __complete_request()
  libceph: no need to call flush_workqueue() before destruction
  ceph: flush pending works before shutdown super
  ceph: abort osd requests on force umount
  libceph: introduce ceph_osdc_abort_requests()
  ...

1 files changed, 43 insertions, 24 deletions

diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index ae056927080d..4fda7a9d4c9d 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -739,7 +739,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 	struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
 	struct ceph_mds_reply_inode *info = iinfo->in;
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	int issued = 0, implemented, new_issued;
+	int issued, new_issued, info_caps;
 	struct timespec mtime, atime, ctime;
 	struct ceph_buffer *xattr_blob = NULL;
 	struct ceph_string *pool_ns = NULL;
@@ -754,8 +754,10 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 	     inode, ceph_vinop(inode), le64_to_cpu(info->version),
 	     ci->i_version);
 
+	info_caps = le32_to_cpu(info->cap.caps);
+
 	/* prealloc new cap struct */
-	if (info->cap.caps && ceph_snap(inode) == CEPH_NOSNAP)
+	if (info_caps && ceph_snap(inode) == CEPH_NOSNAP)
 		new_cap = ceph_get_cap(mdsc, caps_reservation);
 
 	/*
@@ -792,9 +794,9 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 	     le64_to_cpu(info->version) > (ci->i_version & ~1)))
 		new_version = true;
 
-	issued = __ceph_caps_issued(ci, &implemented);
-	issued |= implemented | __ceph_caps_dirty(ci);
-	new_issued = ~issued & le32_to_cpu(info->cap.caps);
+	__ceph_caps_issued(ci, &issued);
+	issued |= __ceph_caps_dirty(ci);
+	new_issued = ~issued & info_caps;
 
 	/* update inode */
 	inode->i_rdev = le32_to_cpu(info->rdev);
@@ -826,6 +828,11 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 				&ctime, &mtime, &atime);
 	}
 
+	if (new_version || (info_caps & CEPH_CAP_FILE_SHARED)) {
+		ci->i_files = le64_to_cpu(info->files);
+		ci->i_subdirs = le64_to_cpu(info->subdirs);
+	}
+
 	if (new_version ||
 	    (new_issued & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR))) {
 		s64 old_pool = ci->i_layout.pool_id;
@@ -854,6 +861,18 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 		}
 	}
 
+	/* layout and rstat are not tracked by capability, update them if
+	 * the inode info is from auth mds */
+	if (new_version || (info->cap.flags & CEPH_CAP_FLAG_AUTH)) {
+		if (S_ISDIR(inode->i_mode)) {
+			ci->i_dir_layout = iinfo->dir_layout;
+			ci->i_rbytes = le64_to_cpu(info->rbytes);
+			ci->i_rfiles = le64_to_cpu(info->rfiles);
+			ci->i_rsubdirs = le64_to_cpu(info->rsubdirs);
+			ceph_decode_timespec(&ci->i_rctime, &info->rctime);
+		}
+	}
+
 	/* xattrs */
 	/* note that if i_xattrs.len <= 4, i_xattrs.data will still be NULL. */
 	if ((ci->i_xattrs.version == 0 || !(issued & CEPH_CAP_XATTR_EXCL))  &&
@@ -870,7 +889,8 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 	}
 
 	/* finally update i_version */
-	ci->i_version = le64_to_cpu(info->version);
+	if (le64_to_cpu(info->version) > ci->i_version)
+		ci->i_version = le64_to_cpu(info->version);
 
 	inode->i_mapping->a_ops = &ceph_aops;
 
@@ -918,15 +938,6 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 	case S_IFDIR:
 		inode->i_op = &ceph_dir_iops;
 		inode->i_fop = &ceph_dir_fops;
-
-		ci->i_dir_layout = iinfo->dir_layout;
-
-		ci->i_files = le64_to_cpu(info->files);
-		ci->i_subdirs = le64_to_cpu(info->subdirs);
-		ci->i_rbytes = le64_to_cpu(info->rbytes);
-		ci->i_rfiles = le64_to_cpu(info->rfiles);
-		ci->i_rsubdirs = le64_to_cpu(info->rsubdirs);
-		ceph_decode_timespec(&ci->i_rctime, &info->rctime);
 		break;
 	default:
 		pr_err("fill_inode %llx.%llx BAD mode 0%o\n",
@@ -934,12 +945,11 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 	}
 
 	/* were we issued a capability? */
-	if (info->cap.caps) {
+	if (info_caps) {
 		if (ceph_snap(inode) == CEPH_NOSNAP) {
-			unsigned caps = le32_to_cpu(info->cap.caps);
 			ceph_add_cap(inode, session,
 				     le64_to_cpu(info->cap.cap_id),
-				     cap_fmode, caps,
+				     cap_fmode, info_caps,
 				     le32_to_cpu(info->cap.wanted),
 				     le32_to_cpu(info->cap.seq),
 				     le32_to_cpu(info->cap.mseq),
@@ -949,7 +959,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 			/* set dir completion flag? */
 			if (S_ISDIR(inode->i_mode) &&
 			    ci->i_files == 0 && ci->i_subdirs == 0 &&
-			    (caps & CEPH_CAP_FILE_SHARED) &&
+			    (info_caps & CEPH_CAP_FILE_SHARED) &&
 			    (issued & CEPH_CAP_FILE_EXCL) == 0 &&
 			    !__ceph_dir_is_complete(ci)) {
 				dout(" marking %p complete (empty)\n", inode);
@@ -962,8 +972,8 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 			wake = true;
 		} else {
 			dout(" %p got snap_caps %s\n", inode,
-			     ceph_cap_string(le32_to_cpu(info->cap.caps)));
-			ci->i_snap_caps |= le32_to_cpu(info->cap.caps);
+			     ceph_cap_string(info_caps));
+			ci->i_snap_caps |= info_caps;
 			if (cap_fmode >= 0)
 				__ceph_get_fmode(ci, cap_fmode);
 		}
@@ -978,8 +988,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 		int cache_caps = CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO;
 		ci->i_inline_version = iinfo->inline_version;
 		if (ci->i_inline_version != CEPH_INLINE_NONE &&
-		    (locked_page ||
-		     (le32_to_cpu(info->cap.caps) & cache_caps)))
+		    (locked_page || (info_caps & cache_caps)))
 			fill_inline = true;
 	}
 
@@ -2178,6 +2187,7 @@ int __ceph_do_getattr(struct inode *inode, struct page *locked_page,
 	struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb);
 	struct ceph_mds_client *mdsc = fsc->mdsc;
 	struct ceph_mds_request *req;
+	int mode;
 	int err;
 
 	if (ceph_snap(inode) == CEPH_SNAPDIR) {
@@ -2190,7 +2200,8 @@ int __ceph_do_getattr(struct inode *inode, struct page *locked_page,
 	if (!force && ceph_caps_issued_mask(ceph_inode(inode), mask, 1))
 		return 0;
 
-	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_GETATTR, USE_ANY_MDS);
+	mode = (mask & CEPH_STAT_RSTAT) ? USE_AUTH_MDS : USE_ANY_MDS;
+	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_GETATTR, mode);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 	req->r_inode = inode;
@@ -2261,6 +2272,14 @@ int ceph_getattr(const struct path *path, struct kstat *stat,
 				stat->size = ci->i_files + ci->i_subdirs;
 			stat->blocks = 0;
 			stat->blksize = 65536;
+			/*
+			 * Some applications rely on the number of st_nlink
+			 * value on directories to be either 0 (if unlinked)
+			 * or 2 + number of subdirectories.
+			 */
+			if (stat->nlink == 1)
+				/* '.' + '..' + subdirs */
+				stat->nlink = 1 + 1 + ci->i_subdirs;
 		}
 	}
 	return err;