提交 139aacf7 编写于 作者: T Tomas Winkler 提交者: Greg Kroah-Hartman

mei: fix read after read scenario

mei read always has to be preceded by write but
'write write read read' scenario should work as well.
In this case the offset is not zero but new read should
be initiated
Signed-off-by: NTomas Winkler <tomas.winkler@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
上级 7131799b
......@@ -214,19 +214,21 @@ static ssize_t mei_read(struct file *file, char __user *ubuf,
goto out;
}
if (cl->read_cb && cl->read_cb->buf_idx > *offset) {
if (cl->read_cb) {
cb = cl->read_cb;
/* read what left */
if (cb->buf_idx > *offset)
goto copy_buffer;
} else if (cl->read_cb && cl->read_cb->buf_idx > 0 &&
cl->read_cb->buf_idx <= *offset) {
cb = cl->read_cb;
/* offset is beyond buf_idx we have no more data return 0 */
if (cb->buf_idx > 0 && cb->buf_idx <= *offset) {
rets = 0;
goto free;
} else if ((!cl->read_cb || !cl->read_cb->buf_idx) && *offset > 0) {
/*Offset needs to be cleaned for contiguous reads*/
}
/* Offset needs to be cleaned for contiguous reads*/
if (cb->buf_idx == 0 && *offset > 0)
*offset = 0;
} else if (*offset > 0) {
*offset = 0;
rets = 0;
goto out;
}
err = mei_cl_read_start(cl, length);
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册