ceph: make page alignment explicit in osd interface

We used to infer alignment of IOs within a page based on the file offset,
which assumed they matched.  This broke with direct IO that was not aligned
to pages (e.g., 512-byte aligned IO).  We were also trusting the alignment
specified in the OSD reply, which could have been adjusted by the server.

Explicitly specify the page alignment when setting up OSD IO requests.

Signed-off-by: Sage Weil <sage@newdream.net>

3 files changed, 25 insertions, 9 deletions

diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 51bcc5ce3230..4aa857763037 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -204,7 +204,7 @@ static int readpage_nounlock(struct file *filp, struct page *page)
         err = ceph_osdc_readpages(osdc, ceph_vino(inode), &ci->i_layout,
                                   page->index << PAGE_CACHE_SHIFT, &len,
                                   ci->i_truncate_seq, ci->i_truncate_size,
-                                  &page, 1);
+                                  &page, 1, 0);
         if (err == -ENOENT)
                 err = 0;
         if (err < 0) {
@@ -287,7 +287,7 @@ static int ceph_readpages(struct file *file, struct address_space *mapping,
         rc = ceph_osdc_readpages(osdc, ceph_vino(inode), &ci->i_layout,
                                  offset, &len,
                                  ci->i_truncate_seq, ci->i_truncate_size,
-                                 pages, nr_pages);
+                                 pages, nr_pages, 0);
         if (rc == -ENOENT)
                 rc = 0;
         if (rc < 0)
@@ -782,7 +782,7 @@ get_more_pages:
                                             snapc, do_sync,
                                             ci->i_truncate_seq,
                                             ci->i_truncate_size,
-                                            &inode->i_mtime, true, 1);
+                                            &inode->i_mtime, true, 1, 0);
                                 max_pages = req->r_num_pages;
 
                                 alloc_page_vec(fsc, req);
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 603fd00af0a6..8d79b8912e31 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -282,11 +282,12 @@ int ceph_release(struct inode *inode, struct file *file)
 static int striped_read(struct inode *inode,
                         u64 off, u64 len,
                         struct page **pages, int num_pages,
-                        int *checkeof)
+                        int *checkeof, bool align_to_pages)
 {
         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
         struct ceph_inode_info *ci = ceph_inode(inode);
         u64 pos, this_len;
+        int io_align, page_align;
         int page_off = off & ~PAGE_CACHE_MASK; /* first byte's offset in page */
         int left, pages_left;
         int read;
@@ -302,14 +303,19 @@ static int striped_read(struct inode *inode,
         page_pos = pages;
         pages_left = num_pages;
         read = 0;
+        io_align = off & ~PAGE_MASK;
 
 more:
+        if (align_to_pages)
+                page_align = (pos - io_align) & ~PAGE_MASK;
+        else
+                page_align = pos & ~PAGE_MASK;
         this_len = left;
         ret = ceph_osdc_readpages(&fsc->client->osdc, ceph_vino(inode),
                                   &ci->i_layout, pos, &this_len,
                                   ci->i_truncate_seq,
                                   ci->i_truncate_size,
-                                  page_pos, pages_left);
+                                  page_pos, pages_left, page_align);
         hit_stripe = this_len < left;
         was_short = ret >= 0 && ret < this_len;
         if (ret == -ENOENT)
@@ -393,7 +399,8 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data,
         if (ret < 0)
                 goto done;
 
-        ret = striped_read(inode, off, len, pages, num_pages, checkeof);
+        ret = striped_read(inode, off, len, pages, num_pages, checkeof,
+                           file->f_flags & O_DIRECT);
 
         if (ret >= 0 && (file->f_flags & O_DIRECT) == 0)
                 ret = ceph_copy_page_vector_to_user(pages, data, off, ret);
@@ -448,6 +455,7 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data,
         int flags;
         int do_sync = 0;
         int check_caps = 0;
+        int page_align, io_align;
         int ret;
         struct timespec mtime = CURRENT_TIME;
 
@@ -462,6 +470,8 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data,
         else
                 pos = *offset;
 
+        io_align = pos & ~PAGE_MASK;
+
         ret = filemap_write_and_wait_range(inode->i_mapping, pos, pos + left);
         if (ret < 0)
                 return ret;
@@ -486,20 +496,26 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data,
          */
 more:
         len = left;
+        if (file->f_flags & O_DIRECT)
+                /* write from beginning of first page, regardless of
+                   io alignment */
+                page_align = (pos - io_align) & ~PAGE_MASK;
+        else
+                page_align = pos & ~PAGE_MASK;
         req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
                                     ceph_vino(inode), pos, &len,
                                     CEPH_OSD_OP_WRITE, flags,
                                     ci->i_snap_realm->cached_context,
                                     do_sync,
                                     ci->i_truncate_seq, ci->i_truncate_size,
-                                    &mtime, false, 2);
+                                    &mtime, false, 2, page_align);
         if (!req)
                 return -ENOMEM;
 
         num_pages = calc_pages_for(pos, len);
 
         if (file->f_flags & O_DIRECT) {
-                pages = ceph_get_direct_page_vector(data, num_pages, pos, len);
+                pages = ceph_get_direct_page_vector(data, num_pages);
                 if (IS_ERR(pages)) {
                         ret = PTR_ERR(pages);
                         goto out;
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 7bc0fbd26af2..8153ee5a8d74 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -1752,7 +1752,7 @@ int ceph_do_getattr(struct inode *inode, int mask)
                 return 0;
         }
 
-        dout("do_getattr inode %p mask %s\n", inode, ceph_cap_string(mask));
+        dout("do_getattr inode %p mask %s mode 0%o\n", inode, ceph_cap_string(mask), inode->i_mode);
         if (ceph_caps_issued_mask(ceph_inode(inode), mask, 1))
                 return 0;