wireguard: peer: allocate in kmem_cache

commit a4e9f8e3287c9eb6bf70df982870980dd3341863 upstream.

With deployments having upwards of 600k peers now, this somewhat heavy
structure could benefit from more fine-grained allocations.
Specifically, instead of using a 2048-byte slab for a 1544-byte object,
we can now use 1544-byte objects directly, thus saving almost 25%
per-peer, or with 600k peers, that's a savings of 303 MiB. This also
makes wireguard's memory usage more transparent in tools like slabtop
and /proc/slabinfo.

Fixes: 8b5553ace83c ("wireguard: queueing: get rid of per-peer ring buffers")
Suggested-by: Arnd Bergmann <arnd@arndb.de>
Suggested-by: Matthew Wilcox <willy@infradead.org>
Cc: stable@vger.kernel.org
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/net/wireguard/main.c

@@ -28,6 +28,10 @@ static int __init mod_init(void)
 #endif
 	wg_noise_init();
 
+	ret = wg_peer_init();
+	if (ret < 0)
+		goto err_peer;
+
 	ret = wg_device_init();
 	if (ret < 0)
 		goto err_device;
@@ -44,6 +48,8 @@ static int __init mod_init(void)
 err_netlink:
 	wg_device_uninit();
 err_device:
+	wg_peer_uninit();
+err_peer:
 	return ret;
 }
 
@@ -51,6 +57,7 @@ static void __exit mod_exit(void)
 {
 	wg_genetlink_uninit();
 	wg_device_uninit();
+	wg_peer_uninit();
 }
 
 module_init(mod_init);