Skip to contentSkip to conversation

mld: avoid unnecessary high order page allocation in mld_newpack()

PATCH Applied Taehee Yoo 2 participants 2 messages Jun 13, 2021 – Jun 14, 2021

Subsystems (3)
lore
Share & tools
lore

Taehee Yoo

author
#0 lore raw

PATCH 1/1 mld: avoid unnecessary high order page allocation in mld_newpack()

If link mtu is too big, mld_newpack() allocates high-order page. But most mld packets don't need high-order page.
So, it might waste unnecessary pages.
To avoid this, it makes mld_newpack() try to allocate order-0 page.

---
 net/ipv6/mcast.c | 25 ++++++++++++++-----------
 1 file changed, 14 insertions(+), 11 deletions(-)
net/ipv6/mcast.c +14 -11
index d36ef9d25e73..54ec163fbafa 100644
@@ -1729,22 +1729,25 @@ static void ip6_mc_hdr(struct sock *sk, struct sk_buff *skb,
 
 static struct sk_buff *mld_newpack(struct inet6_dev *idev, unsigned int mtu)
 {
+	u8 ra[8] = { IPPROTO_ICMPV6, 0, IPV6_TLV_ROUTERALERT,
+		     2, 0, 0, IPV6_TLV_PADN, 0 };
 	struct net_device *dev = idev->dev;
-	struct net *net = dev_net(dev);
-	struct sock *sk = net->ipv6.igmp_sk;
-	struct sk_buff *skb;
-	struct mld2_report *pmr;
-	struct in6_addr addr_buf;
-	const struct in6_addr *saddr;
 	int hlen = LL_RESERVED_SPACE(dev);
 	int tlen = dev->needed_tailroom;
-	unsigned int size = mtu + hlen + tlen;
+	struct net *net = dev_net(dev);
+	const struct in6_addr *saddr;
+	struct in6_addr addr_buf;
+	struct mld2_report *pmr;
+	struct sk_buff *skb;
+	unsigned int size;
+	struct sock *sk;
 	int err;
-	u8 ra[8] = { IPPROTO_ICMPV6, 0,
-		     IPV6_TLV_ROUTERALERT, 2, 0, 0,
-		     IPV6_TLV_PADN, 0 };
 
-	/* we assume size > sizeof(ra) here */
+	sk = net->ipv6.igmp_sk;
+	/* we assume size > sizeof(ra) here
+	 * Also try to not allocate high-order pages for big MTU
+	 */
+	size = min_t(int, mtu, PAGE_SIZE / 2) + hlen + tlen;
 	skb = sock_alloc_send_skb(sk, size, 1, &err);
 	if (!skb)
 		return NULL;
Signature · 1 line
2.17.1
#1 lore raw
This patch was applied to netdev/net-next.git (refs/heads/master):

Hello:

This patch was applied to netdev/net-next.git (refs/heads/master):

Quoted 9 lines · On Sun, 13 Jun 2021 14:43:44 +0000 you wrote:

If link mtu is too big, mld_newpack() allocates high-order page. But most mld packets don't need high-order page.
So, it might waste unnecessary pages.
To avoid this, it makes mld_newpack() try to allocate order-0 page.

[...]

Here is the summary with links:
- [net-next] mld: avoid unnecessary high order page allocation in mld_newpack()

    https://git.kernel.org/netdev/net-next/c/ffa85b73c3c4

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html