block: fix DIO handling regressions in blkdev_read_iter()

Commit f32b0d14474c ("block: move direct_IO into our own read_iter
handler") introduced several regressions for bdev DIO:

1. read spanning EOF always returns 0 instead of the number of bytes
   read.  This is because "count" is assigned early and isn't updated
   when the iterator is truncated:

     $ lsblk -o name,size /dev/vdb
     NAME SIZE
     vdb    1G
     $ xfs_io -d -c 'pread -b 4M 1021M 4M' /dev/vdb
     read 0/4194304 bytes at offset 1070596096
     0.000000 bytes, 0 ops; 0.0007 sec (0.000000 bytes/sec and 0.0000 ops/sec)

     instead of

     $ xfs_io -d -c 'pread -b 4M 1021M 4M' /dev/vdb
     read 3145728/4194304 bytes at offset 1070596096
     3 MiB, 1 ops; 0.0007 sec (3.865 GiB/sec and 1319.2612 ops/sec)

2. truncated iterator isn't reexpanded
3. iterator isn't reverted on blkdev_direct_IO() error
4. zero size read no longer skips atime update

Fixes: f32b0d14474c ("block: move direct_IO into our own read_iter handler")
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20220201100420.25875-1-idryomov@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>

diff --git a/block/fops.c b/block/fops.c
index 26bf15c770d21a1c8fc062337691db037d59972d..4f59e0f5bf30966214064260c9a4119773b2bff7 100644 (file)
--- a/block/fops.c
+++ b/block/fops.c
@@ -566,34 +566,37 @@ static ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to)
 {
        struct block_device *bdev = iocb->ki_filp->private_data;
        loff_t size = bdev_nr_bytes(bdev);
-       size_t count = iov_iter_count(to);
        loff_t pos = iocb->ki_pos;
        size_t shorted = 0;
        ssize_t ret = 0;
+       size_t count;
 
-       if (unlikely(pos + count > size)) {
+       if (unlikely(pos + iov_iter_count(to) > size)) {
                if (pos >= size)
                        return 0;
                size -= pos;
-               if (count > size) {
-                       shorted = count - size;
-                       iov_iter_truncate(to, size);
-               }
+               shorted = iov_iter_count(to) - size;
+               iov_iter_truncate(to, size);
        }
 
+       count = iov_iter_count(to);
+       if (!count)
+               goto reexpand; /* skip atime */
+
        if (iocb->ki_flags & IOCB_DIRECT) {
                struct address_space *mapping = iocb->ki_filp->f_mapping;
 
                if (iocb->ki_flags & IOCB_NOWAIT) {
-                       if (filemap_range_needs_writeback(mapping, iocb->ki_pos,
-                                               iocb->ki_pos + count - 1))
-                               return -EAGAIN;
+                       if (filemap_range_needs_writeback(mapping, pos,
+                                                         pos + count - 1)) {
+                               ret = -EAGAIN;
+                               goto reexpand;
+                       }
                } else {
-                       ret = filemap_write_and_wait_range(mapping,
-                                               iocb->ki_pos,
-                                               iocb->ki_pos + count - 1);
+                       ret = filemap_write_and_wait_range(mapping, pos,
+                                                          pos + count - 1);
                        if (ret < 0)
-                               return ret;
+                               goto reexpand;
                }
 
                file_accessed(iocb->ki_filp);
@@ -603,12 +606,14 @@ static ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to)
                        iocb->ki_pos += ret;
                        count -= ret;
                }
+               iov_iter_revert(to, count - iov_iter_count(to));
                if (ret < 0 || !count)
-                       return ret;
+                       goto reexpand;
        }
 
        ret = filemap_read(iocb, to, ret);
 
+reexpand:
        if (unlikely(shorted))
                iov_iter_reexpand(to, iov_iter_count(to) + shorted);
        return ret;