[PATCH for-3.19] vhost/net: fix up num_buffers endian-ness

In virtio 1.0 mode, when mergeable buffers are enabled on a big-endian
host, num_buffers wasn't byte-swapped correctly, so large incoming
packets got corrupted.

To fix, fill it in within hdr - this also makes sure it gets
the correct type.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---

It seems important to get this into 3.19, otherwise
VIRTIO_1 feature bit is unreliable.

Dave, do you plan another pull request for 3.19?
If yes pls merge this, if not pls let me know and I'll try to
send it to Linus directly.


 drivers/vhost/net.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
index d415d69..9484d56 100644
--- a/drivers/vhost/net.c
+++ b/drivers/vhost/net.c

@@ -650,8 +650,10 @@ static void handle_rx(struct vhost_net *net)
 			break;
 		}
 		/* TODO: Should check and handle checksum. */
+
+		hdr.num_buffers = cpu_to_vhost16(vq, headcount);
 		if (likely(mergeable) &&
-		    memcpy_toiovecend(nvq->hdr, (unsigned char *)&headcount,
+		    memcpy_toiovecend(nvq->hdr, (void *)&hdr.num_buffers,
 				      offsetof(typeof(hdr), num_buffers),
 				      sizeof hdr.num_buffers)) {
 			vq_err(vq, "Failed num_buffers write");

-- 
MST

From: "Michael S. Tsirkin" <mst@redhat.com>
Date: Tue, 3 Feb 2015 11:07:06 +0200

In virtio 1.0 mode, when mergeable buffers are enabled on a big-endian
host, num_buffers wasn't byte-swapped correctly, so large incoming
packets got corrupted.

To fix, fill it in within hdr - this also makes sure it gets
the correct type.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Applied.

On Wed, Feb 04, 2015 at 01:59:42PM -0800, David Miller wrote:

From: "Michael S. Tsirkin" <mst@redhat.com>
Date: Tue, 3 Feb 2015 11:07:06 +0200

quoted

In virtio 1.0 mode, when mergeable buffers are enabled on a big-endian
host, num_buffers wasn't byte-swapped correctly, so large incoming
packets got corrupted.

To fix, fill it in within hdr - this also makes sure it gets
the correct type.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Applied.

FWIW, conflict with iov_iter patches is trivial; once it shows up in
your kernel.org tree I can either rebase the series or just push
#merge-candidate - whichever you prefer.  Linus usually prefers the
second variant, but then he seriously dislikes rebasing of any sort;
I've no idea what your preferences are in that area...