kclient: check EOF on sync read to do short result when appropriate

We only want to return a short read if we read the end of file.
Otherwise, a short object read means we hit a hole and need to
zero-fill.

diff --git a/src/kernel/file.c b/src/kernel/file.c
index 156cb4735451cb47439fa1a520674c3ba959ff3c..89e6398e7993d162726a464f9d13a742ae668d22 100644 (file)
--- a/src/kernel/file.c
+++ b/src/kernel/file.c
@@ -440,13 +440,12 @@ static void zero_page_vector_range(int off, int len, struct page **pages)
  * Read a range of bytes striped over one or more objects.  Iterate over
  * objects we stripe over.  (That's not atomic, but good enough for now.)
  *
- * If @fill, zero any regions that are holes or past object EOF on
- * disk.
+ * If we get a short result from the OSD, check against i_size; we need to
+ * only return a short read to the caller if we hit EOF.
  */
 static int striped_read(struct inode *inode,
                        u64 off, u64 len,
-                       struct page **pages, int num_pages,
-                       bool fill)
+                       struct page **pages, int num_pages)
 {
        struct ceph_client *client = ceph_inode_to_client(inode);
        struct ceph_inode_info *ci = ceph_inode(inode);
@@ -456,7 +455,7 @@ static int striped_read(struct inode *inode,
        int read;
        struct page **page_pos;
        int ret;
-       bool was_short;
+       bool hit_stripe, was_short;
 
        /*
         * we may need to do multiple reads.  not atomic, unfortunately.
@@ -474,11 +473,12 @@ more:
                                  ci->i_truncate_seq,
                                  ci->i_truncate_size,
                                  page_pos, pages_left);
-       was_short = this_len < left;
+       hit_stripe = this_len < left;
+       was_short = ret >= 0 && ret < this_len;
        if (ret == -ENOENT)
                ret = 0;
-       dout("striped_read %llu~%u (read %u) got %d%s\n", pos, left, read, ret,
-            was_short ? " SHORT" : "");
+       dout("striped_read %llu~%u (read %u) got %d%s%s\n", pos, left, read,
+            ret, hit_stripe ? " HITSTRIPE" : "", was_short ? " SHORT" : "");
 
        if (ret > 0) {
                int didpages =
@@ -492,17 +492,36 @@ more:
                pos += ret;
                read = pos - off;
                left -= ret;
-               if (left && was_short) {
-                       page_pos += didpages;
-                       pages_left -= didpages;
+               page_pos += didpages;
+               pages_left -= didpages;
+
+               /* hit stripe? */
+               if (left && hit_stripe)
                        goto more;
-               }
-               if (fill) {
+       }
+
+       if (was_short) {
+               /* was original extent fully inside i_size? */
+               if (pos + left <= inode->i_size) {
                        dout("zero tail\n");
-                       zero_page_vector_range(page_off + read,
-                                              len - read, pages);
+                       zero_page_vector_range(page_off + read, len - read,
+                                              pages);
+                       goto out;
                }
+
+               /* check i_size */
+               ret = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE);
+               if (ret < 0)
+                       goto out;
+
+               /* hit EOF? */
+               if (pos >= inode->i_size)
+                       goto out;
+
+               goto more;
        }
+
+out:
        if (ret >= 0)
                ret = read;
        dout("striped_read returns %d\n", ret);
@@ -543,8 +562,7 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data,
        if (IS_ERR(pages))
                return PTR_ERR(pages);
 
-       ret = striped_read(inode, off, len, pages, num_pages,
-                          file->f_flags & O_DIRECT);
+       ret = striped_read(inode, off, len, pages, num_pages);
 
        if (ret >= 0 && (file->f_flags & O_DIRECT) == 0)
                ret = copy_page_vector_to_user(pages, data, off, ret);