We're not supposed to kill all fids when a new attach request
arrives. This used to cause issues when the guest would send
multiple attach requests.
Signed-off-by: Sasha Levin <levinsasha...@gmail.com>
---
tools/kvm/virtio/9p.c | 22 ----------------------
1 file changed, 22 deletions(-)
diff --git a/tools/kvm/virtio/9p.c b/tools/kvm/virtio/9p.c
index 27ef57b..d668106 100644
--- a/tools/kvm/virtio/9p.c
+++ b/tools/kvm/virtio/9p.c
@@ -117,26 +117,6 @@ static void close_fid(struct p9_dev *p9dev, u32 fid)
free(pfid);
}
-static void clear_all_fids(struct p9_dev *p9dev)
-{
- struct rb_node *node = rb_first(&p9dev->fids);
-
- while (node) {
- struct p9_fid *fid = rb_entry(node, struct p9_fid, node);
-
- if (fid->fd > 0)
- close(fid->fd);
-
- if (fid->dir)
- closedir(fid->dir);
-
- rb_erase(&fid->node, &p9dev->fids);
- free(fid);
-
- node = rb_first(&p9dev->fids);
- }
-}
-
static void virtio_p9_set_reply_header(struct p9_pdu *pdu, u32 size)
{
u8 cmd;
@@ -443,8 +423,6 @@ static void virtio_p9_attach(struct p9_dev *p9dev,
free(uname);
free(aname);
- clear_all_fids(p9dev);
-
if (lstat(p9dev->root_dir, &st) < 0)
goto err_out;
--
1.7.12
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
