wireguard-linux/net/qrtr, branch jd/unified-crypt-queue

net: qrtr: send msgs from local of same id as broadcast

2020-04-09T17:08:31Z

If the local node id(qrtr_local_nid) is not modified after its initialization, it equals to the broadcast node id(QRTR_NODE_BCAST). So the messages from local node should not be taken as broadcast and keep the process going to send them out anyway. The definitions are as follow: static unsigned int qrtr_local_nid = NUMA_NO_NODE; Fixes: fdf5fd397566 ("net: qrtr: Broadcast messages only from control port") Signed-off-by: Wang Wenhu Signed-off-by: David S. Miller

net: qrtr: Fix FIXME related to qrtr_ns_init()

2020-03-04T01:52:21Z

The 2 second delay before calling qrtr_ns_init() meant that the remote processors would register as endpoints in qrtr and the say_hello() call would therefor broadcast the outgoing HELLO to them. With the HELLO handshake corrected this delay is no longer needed. Reviewed-by: Manivannan Sadhasivam Tested-by: Manivannan Sadhasivam Signed-off-by: Bjorn Andersson Signed-off-by: David S. Miller

net: qrtr: Respond to HELLO message

2020-03-04T01:52:21Z

Lost in the translation from the user space implementation was the detail that HELLO mesages must be exchanged between each node pair. As such the incoming HELLO must be replied to. Similar to the previous implementation no effort is made to prevent two Linux boxes from continuously sending HELLO messages back and forth, this is left to a follow up patch. say_hello() is moved, to facilitate the new call site. Fixes: 0c2204a4ad71 ("net: qrtr: Migrate nameservice to kernel from userspace") Reviewed-by: Manivannan Sadhasivam Tested-by: Manivannan Sadhasivam Signed-off-by: Bjorn Andersson Signed-off-by: David S. Miller

net: qrtr: Fix error pointer vs NULL bugs

2020-02-27T04:52:31Z

The callers only expect NULL pointers, so returning an error pointer will lead to an Oops. Fixes: 0c2204a4ad71 ("net: qrtr: Migrate nameservice to kernel from userspace") Signed-off-by: Dan Carpenter Signed-off-by: David S. Miller

net: qrtr: fix spelling mistake "serivce" -> "service"

2020-02-24T23:41:14Z

There is a spelling mistake in a pr_err message. Fix it. Signed-off-by: Colin Ian King Signed-off-by: David S. Miller

net: qrtr: Fix the local node ID as 1

2020-02-21T19:59:12Z

In order to start the QRTR nameservice, the local node ID needs to be valid. Hence, fix it to 1. Previously, the node ID was configured through a userspace tool before starting the nameservice daemon. Since we have now integrated the nameservice handling to kernel, this change is necessary for making it functional. Signed-off-by: Manivannan Sadhasivam Signed-off-by: David S. Miller

net: qrtr: Migrate nameservice to kernel from userspace

2020-02-21T19:59:12Z

The QRTR nameservice has been maintained in userspace for some time. This commit migrates it to Linux kernel. This change is required in order to eliminate the need of starting a userspace daemon for making the WiFi functional for ath11k based devices. Since the QRTR NS is not usually packed in most of the distros, users need to clone, build and install it to get the WiFi working. It will become a hassle when the user doesn't have any other source of network connectivity. Signed-off-by: Manivannan Sadhasivam Signed-off-by: David S. Miller

net: qrtr: Remove receive worker

2020-01-15T02:36:42Z

Rather than enqueuing messages and scheduling a worker to deliver them to the individual sockets we can now, thanks to the previous work, move this directly into the endpoint callback. This saves us a context switch per incoming message and removes the possibility of an opportunistic suspend to happen between the message is coming from the endpoint until it ends up in the socket's receive buffer. Signed-off-by: Bjorn Andersson Signed-off-by: David S. Miller

net: qrtr: Make qrtr_port_lookup() use RCU

2020-01-15T02:36:42Z

The important part of qrtr_port_lookup() wrt synchronization is that the function returns a reference counted struct qrtr_sock, or fail. As such we need only to ensure that an decrement of the object's refcount happens inbetween the finding of the object in the idr and qrtr_port_lookup()'s own increment of the object. By using RCU and putting a synchronization point after we remove the mapping from the idr, but before it can be released we achieve this - with the benefit of not having to hold the mutex in qrtr_port_lookup(). Signed-off-by: Bjorn Andersson Signed-off-by: David S. Miller

net: qrtr: Migrate node lookup tree to spinlock

2020-01-15T02:36:42Z

Move operations on the qrtr_nodes radix tree under a separate spinlock and make the qrtr_nodes tree GFP_ATOMIC, to allow operation from atomic context in a subsequent patch. Signed-off-by: Bjorn Andersson Signed-off-by: David S. Miller