Re: [PATCH 3/3] netconsole: implement extended console support

Hello, Tetsuo.

On Sat, May 02, 2015 at 01:54:57PM +0900, Tetsuo Handa wrote:

Tejun Heo wrote:

quoted

+For example,
+
+ 6,416,1758426,-,ncfrag=0/33;the first chunk,
+ 6,416,1758426,-,ncfrag=16/33;the second chunk.
+

Wouldn't total-bytes > 1000 than 33 in this example?

It's trying to give an example on how the ncfrag header looks like but
yeah I'll tweak it a bit and add further explanation.

quoted

+/**
+ * send_ext_msg_udp - send extended log message to target
+ * @nt: target to send message to
+ * @msg: extended log message to send
+ * @msg_len: length of message
+ *
+ * Transfer extended log @msg to @nt.  If @msg is longer than
+ * MAX_PRINT_CHUNK, it'll be split and transmitted in multiple chunks with
+ * ncfrag header field added to identify them.
+ */
+static void send_ext_msg_udp(struct netconsole_target *nt, const char *msg,
+			     int msg_len)
+{
+	static char buf[MAX_PRINT_CHUNK];
+	const int max_extra_len = sizeof(",ncfrag=0000/0000");
+	const char *header, *body;
+	int header_len = msg_len, body_len = 0;
+	int chunk_len, nr_chunks, i;
+
+	if (msg_len <= MAX_PRINT_CHUNK) {
+		netpoll_send_udp(&nt->np, msg, msg_len);
+		return;
+	}
+
+	/* need to insert extra header fields, detect header and body */
+	header = msg;
+	body = memchr(msg, ';', msg_len);
+	if (body) {
+		header_len = body - header;
+		body_len = msg_len - header_len - 1;
+		body++;
+	}
+
+	chunk_len = MAX_PRINT_CHUNK - header_len - max_extra_len;
+	if (WARN_ON_ONCE(chunk_len <= 0))
+		return;

This path is executed only when msg_len > MAX_PRINT_CHUNK.
And since header_len == msg_len if body == NULL, chunk_len <= 0 is true.
We will hit this WARN_ON_ONCE() if memchr(msg, ';', msg_len) == NULL
which will fail to send the message. Is this what you want?

Yeah, extended consoles should be fed messages w/ proper headers.  If
not, we warn and bail.

quoted

+static void write_ext_msg(struct console *con, const char *msg,
+			  unsigned int len)
+{
+	struct netconsole_target *nt;
+	unsigned long flags;
+
+	if ((oops_only && !oops_in_progress) || list_empty(&target_list))
+		return;
+
+	spin_lock_irqsave(&target_list_lock, flags);
+	list_for_each_entry(nt, &target_list, list)

Don't you need to call netconsole_target_get() here

quoted

+		if (nt->extended && nt->enabled && netif_running(nt->np.dev))
+			send_ext_msg_udp(nt, msg, len);

and netconsole_target_put() here as with write_msg()?

Ah, I dropped the patch which removed it from write_msg() but the
following pattern doesn't do anything.  It's just pure confusion.

	lock;
	ref++ on an item on the index;
	do something w/o releasing lock;
	ref-- on the item;
	unlock;

netconsole_netdev_event() has to release lock to process the item so
refcnting is needed there.  It got duplicated in write_msg() for no
reason.  I'll add a patch to remove it.

Thanks.

-- 
tejun