ceph: fix short sync reads from the OSD
If we get a short read from the OSD because the object is small, we need to
zero the remainder of the buffer. For O_DIRECT reads, the attempted range
is not trimmed to i_size by the VFS, so we were actually looping
indefinitely.
Fix by trimming by i_size, and the unconditionally zeroing the trailing
range.
Reported-by: Jeff Wu <cpwu@tnsoft.com.cn>
Signed-off-by: Sage Weil <sage@newdream.net>
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 8c5ac4e..b654f40 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -283,7 +283,7 @@
static int striped_read(struct inode *inode,
u64 off, u64 len,
struct page **pages, int num_pages,
- int *checkeof, bool align_to_pages,
+ int *checkeof, bool o_direct,
unsigned long buf_align)
{
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
@@ -308,7 +308,7 @@
io_align = off & ~PAGE_MASK;
more:
- if (align_to_pages)
+ if (o_direct)
page_align = (pos - io_align + buf_align) & ~PAGE_MASK;
else
page_align = pos & ~PAGE_MASK;
@@ -346,20 +346,22 @@
}
if (was_short) {
- /* was original extent fully inside i_size? */
- if (pos + left <= inode->i_size) {
- dout("zero tail\n");
- ceph_zero_page_vector_range(page_off + read, len - read,
- pages);
- read = len;
- goto out;
- }
+ /* did we bounce off eof? */
+ if (pos + left > inode->i_size)
+ *checkeof = 1;
- /* check i_size */
- *checkeof = 1;
+ /* zero trailing bytes (inside i_size) */
+ if (left > 0 && pos < inode->i_size) {
+ if (pos + left > inode->i_size)
+ left = inode->i_size - pos;
+
+ dout("zero tail %d\n", left);
+ ceph_zero_page_vector_range(page_off + read, left,
+ pages);
+ read += left;
+ }
}
-out:
if (ret >= 0)
ret = read;
dout("striped_read returns %d\n", ret);
@@ -659,7 +661,7 @@
/* hit EOF or hole? */
if (statret == 0 && *ppos < inode->i_size) {
- dout("aio_read sync_read hit hole, reading more\n");
+ dout("aio_read sync_read hit hole, ppos %lld < size %lld, reading more\n", *ppos, inode->i_size);
read += ret;
base += ret;
len -= ret;
