1 files changed, 65 insertions, 20 deletions

diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index ef3ebd780aff..7b79a54a2789 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -188,7 +188,7 @@ static int ceph_releasepage(struct page *page, gfp_t g)
 /*
  * read a single page, without unlocking it.
  */
-static int readpage_nounlock(struct file *filp, struct page *page)
+static int ceph_do_readpage(struct file *filp, struct page *page)
 {
 	struct inode *inode = file_inode(filp);
 	struct ceph_inode_info *ci = ceph_inode(inode);
@@ -218,7 +218,7 @@ static int readpage_nounlock(struct file *filp, struct page *page)
 
 	err = ceph_readpage_from_fscache(inode, page);
 	if (err == 0)
-		goto out;
+		return -EINPROGRESS;
 
 	dout("readpage inode %p file %p page %p index %lu\n",
 	     inode, filp, page, page->index);
@@ -248,8 +248,11 @@ out:
 
 static int ceph_readpage(struct file *filp, struct page *page)
 {
-	int r = readpage_nounlock(filp, page);
-	unlock_page(page);
+	int r = ceph_do_readpage(filp, page);
+	if (r != -EINPROGRESS)
+		unlock_page(page);
+	else
+		r = 0;
 	return r;
 }
 
@@ -315,7 +318,32 @@ static int start_read(struct inode *inode, struct list_head *page_list, int max)
 	struct page **pages;
 	pgoff_t next_index;
 	int nr_pages = 0;
-	int ret;
+	int got = 0;
+	int ret = 0;
+
+	if (!current->journal_info) {
+		/* caller of readpages does not hold buffer and read caps
+		 * (fadvise, madvise and readahead cases) */
+		int want = CEPH_CAP_FILE_CACHE;
+		ret = ceph_try_get_caps(ci, CEPH_CAP_FILE_RD, want, &got);
+		if (ret < 0) {
+			dout("start_read %p, error getting cap\n", inode);
+		} else if (!(got & want)) {
+			dout("start_read %p, no cache cap\n", inode);
+			ret = 0;
+		}
+		if (ret <= 0) {
+			if (got)
+				ceph_put_cap_refs(ci, got);
+			while (!list_empty(page_list)) {
+				page = list_entry(page_list->prev,
+						  struct page, lru);
+				list_del(&page->lru);
+				put_page(page);
+			}
+			return ret;
+		}
+	}
 
 	off = (u64) page_offset(page);
 
@@ -338,15 +366,18 @@ static int start_read(struct inode *inode, struct list_head *page_list, int max)
 				    CEPH_OSD_FLAG_READ, NULL,
 				    ci->i_truncate_seq, ci->i_truncate_size,
 				    false);
-	if (IS_ERR(req))
-		return PTR_ERR(req);
+	if (IS_ERR(req)) {
+		ret = PTR_ERR(req);
+		goto out;
+	}
 
 	/* build page vector */
 	nr_pages = calc_pages_for(0, len);
 	pages = kmalloc(sizeof(*pages) * nr_pages, GFP_KERNEL);
-	ret = -ENOMEM;
-	if (!pages)
-		goto out;
+	if (!pages) {
+		ret = -ENOMEM;
+		goto out_put;
+	}
 	for (i = 0; i < nr_pages; ++i) {
 		page = list_entry(page_list->prev, struct page, lru);
 		BUG_ON(PageLocked(page));
@@ -363,6 +394,7 @@ static int start_read(struct inode *inode, struct list_head *page_list, int max)
 			nr_pages = i;
 			if (nr_pages > 0) {
 				len = nr_pages << PAGE_SHIFT;
+				osd_req_op_extent_update(req, 0, len);
 				break;
 			}
 			goto out_pages;
@@ -378,6 +410,12 @@ static int start_read(struct inode *inode, struct list_head *page_list, int max)
 	if (ret < 0)
 		goto out_pages;
 	ceph_osdc_put_request(req);
+
+	/* After adding locked pages to page cache, the inode holds cache cap.
+	 * So we can drop our cap refs. */
+	if (got)
+		ceph_put_cap_refs(ci, got);
+
 	return nr_pages;
 
 out_pages:
@@ -386,8 +424,11 @@ out_pages:
 		unlock_page(pages[i]);
 	}
 	ceph_put_page_vector(pages, nr_pages, false);
-out:
+out_put:
 	ceph_osdc_put_request(req);
+out:
+	if (got)
+		ceph_put_cap_refs(ci, got);
 	return ret;
 }
 
@@ -424,7 +465,6 @@ static int ceph_readpages(struct file *file, struct address_space *mapping,
 		rc = start_read(inode, page_list, max);
 		if (rc < 0)
 			goto out;
-		BUG_ON(rc == 0);
 	}
 out:
 	ceph_fscache_readpages_cancel(inode, page_list);
@@ -708,7 +748,7 @@ static int ceph_writepages_start(struct address_space *mapping,
 	struct pagevec pvec;
 	int done = 0;
 	int rc = 0;
-	unsigned wsize = 1 << inode->i_blkbits;
+	unsigned int wsize = i_blocksize(inode);
 	struct ceph_osd_request *req = NULL;
 	int do_sync = 0;
 	loff_t snap_size, i_size;
@@ -1198,7 +1238,7 @@ retry_locked:
 			goto retry_locked;
 		r = writepage_nounlock(page, NULL);
 		if (r < 0)
-			goto fail_nosnap;
+			goto fail_unlock;
 		goto retry_locked;
 	}
 
@@ -1226,11 +1266,14 @@ retry_locked:
 	}
 
 	/* we need to read it. */
-	r = readpage_nounlock(file, page);
-	if (r < 0)
-		goto fail_nosnap;
+	r = ceph_do_readpage(file, page);
+	if (r < 0) {
+		if (r == -EINPROGRESS)
+			return -EAGAIN;
+		goto fail_unlock;
+	}
 	goto retry_locked;
-fail_nosnap:
+fail_unlock:
 	unlock_page(page);
 	return r;
 }
@@ -1371,9 +1414,11 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	     inode, off, (size_t)PAGE_SIZE, ceph_cap_string(got));
 
 	if ((got & (CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO)) ||
-	    ci->i_inline_version == CEPH_INLINE_NONE)
+	    ci->i_inline_version == CEPH_INLINE_NONE) {
+		current->journal_info = vma->vm_file;
 		ret = filemap_fault(vma, vmf);
-	else
+		current->journal_info = NULL;
+	} else
 		ret = -EAGAIN;
 
 	dout("filemap_fault %p %llu~%zd dropping cap refs on %s ret %d\n",