Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Pavel Begunkov <asml.silence@gmail.com>
To:     netdev@vger.kernel.org
Cc:     "David S . Miller" <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>, linux-kernel@vger.kernel.org,
        xen-devel@lists.xenproject.org, Wei Liu <wei.liu@kernel.org>,
        Paul Durrant <paul@xen.org>, kvm@vger.kernel.org,
        virtualization@lists.linux-foundation.org,
        "Michael S . Tsirkin" <mst@redhat.com>,
        Jason Wang <jasowang@redhat.com>,
        Pavel Begunkov <asml.silence@gmail.com>
Subject: [PATCH net-next 4/4] net: shrink struct ubuf_info
Date:   Fri, 23 Sep 2022 17:39:04 +0100
Message-Id: <fb899519db7573baea32309a158cfccbc849629b.1663892211.git.asml.silence@gmail.com>
In-Reply-To: <cover.1663892211.git.asml.silence@gmail.com>
References: <cover.1663892211.git.asml.silence@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

We can benefit from a smaller struct ubuf_info, so leave only mandatory
fields and let users to decide how they want to extend it. Convert
MSG_ZEROCOPY to struct ubuf_info_msgzc and remove duplicated fields.
This reduces the size from 48 bytes to just 16.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
 include/linux/skbuff.h | 22 ++++------------------
 net/core/skbuff.c      | 38 +++++++++++++++++++++-----------------
 net/ipv4/ip_output.c   |  2 +-
 net/ipv4/tcp.c         |  2 +-
 net/ipv6/ip6_output.c  |  2 +-
 5 files changed, 28 insertions(+), 38 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index fd7dcb977fdf..920eb6413fee 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -533,25 +533,8 @@ enum {
 struct ubuf_info {
 	void (*callback)(struct sk_buff *, struct ubuf_info *,
 			 bool zerocopy_success);
-	union {
-		struct {
-			unsigned long desc;
-			void *ctx;
-		};
-		struct {
-			u32 id;
-			u16 len;
-			u16 zerocopy:1;
-			u32 bytelen;
-		};
-	};
 	refcount_t refcnt;
 	u8 flags;
-
-	struct mmpin {
-		struct user_struct *user;
-		unsigned int num_pg;
-	} mmp;
 };
 
 struct ubuf_info_msgzc {
@@ -570,7 +553,10 @@ struct ubuf_info_msgzc {
 		};
 	};
 
-	struct mmpin mmp;
+	struct mmpin {
+		struct user_struct *user;
+		unsigned int num_pg;
+	} mmp;
 };
 
 #define skb_uarg(SKB)	((struct ubuf_info *)(skb_shinfo(SKB)->destructor_arg))
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index f1b8b20fc20b..bbcfb1c7f59e 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1188,7 +1188,7 @@ EXPORT_SYMBOL_GPL(mm_unaccount_pinned_pages);
 
 static struct ubuf_info *msg_zerocopy_alloc(struct sock *sk, size_t size)
 {
-	struct ubuf_info *uarg;
+	struct ubuf_info_msgzc *uarg;
 	struct sk_buff *skb;
 
 	WARN_ON_ONCE(!in_task());
@@ -1206,19 +1206,19 @@ static struct ubuf_info *msg_zerocopy_alloc(struct sock *sk, size_t size)
 		return NULL;
 	}
 
-	uarg->callback = msg_zerocopy_callback;
+	uarg->ubuf.callback = msg_zerocopy_callback;
 	uarg->id = ((u32)atomic_inc_return(&sk->sk_zckey)) - 1;
 	uarg->len = 1;
 	uarg->bytelen = size;
 	uarg->zerocopy = 1;
-	uarg->flags = SKBFL_ZEROCOPY_FRAG | SKBFL_DONT_ORPHAN;
-	refcount_set(&uarg->refcnt, 1);
+	uarg->ubuf.flags = SKBFL_ZEROCOPY_FRAG | SKBFL_DONT_ORPHAN;
+	refcount_set(&uarg->ubuf.refcnt, 1);
 	sock_hold(sk);
 
-	return uarg;
+	return &uarg->ubuf;
 }
 
-static inline struct sk_buff *skb_from_uarg(struct ubuf_info *uarg)
+static inline struct sk_buff *skb_from_uarg(struct ubuf_info_msgzc *uarg)
 {
 	return container_of((void *)uarg, struct sk_buff, cb);
 }
@@ -1227,6 +1227,7 @@ struct ubuf_info *msg_zerocopy_realloc(struct sock *sk, size_t size,
 				       struct ubuf_info *uarg)
 {
 	if (uarg) {
+		struct ubuf_info_msgzc *uarg_zc;
 		const u32 byte_limit = 1 << 19;		/* limit to a few TSO */
 		u32 bytelen, next;
 
@@ -1242,8 +1243,9 @@ struct ubuf_info *msg_zerocopy_realloc(struct sock *sk, size_t size,
 			return NULL;
 		}
 
-		bytelen = uarg->bytelen + size;
-		if (uarg->len == USHRT_MAX - 1 || bytelen > byte_limit) {
+		uarg_zc = uarg_to_msgzc(uarg);
+		bytelen = uarg_zc->bytelen + size;
+		if (uarg_zc->len == USHRT_MAX - 1 || bytelen > byte_limit) {
 			/* TCP can create new skb to attach new uarg */
 			if (sk->sk_type == SOCK_STREAM)
 				goto new_alloc;
@@ -1251,11 +1253,11 @@ struct ubuf_info *msg_zerocopy_realloc(struct sock *sk, size_t size,
 		}
 
 		next = (u32)atomic_read(&sk->sk_zckey);
-		if ((u32)(uarg->id + uarg->len) == next) {
-			if (mm_account_pinned_pages(&uarg->mmp, size))
+		if ((u32)(uarg_zc->id + uarg_zc->len) == next) {
+			if (mm_account_pinned_pages(&uarg_zc->mmp, size))
 				return NULL;
-			uarg->len++;
-			uarg->bytelen = bytelen;
+			uarg_zc->len++;
+			uarg_zc->bytelen = bytelen;
 			atomic_set(&sk->sk_zckey, ++next);
 
 			/* no extra ref when appending to datagram (MSG_MORE) */
@@ -1291,7 +1293,7 @@ static bool skb_zerocopy_notify_extend(struct sk_buff *skb, u32 lo, u16 len)
 	return true;
 }
 
-static void __msg_zerocopy_callback(struct ubuf_info *uarg)
+static void __msg_zerocopy_callback(struct ubuf_info_msgzc *uarg)
 {
 	struct sk_buff *tail, *skb = skb_from_uarg(uarg);
 	struct sock_exterr_skb *serr;
@@ -1344,19 +1346,21 @@ static void __msg_zerocopy_callback(struct ubuf_info *uarg)
 void msg_zerocopy_callback(struct sk_buff *skb, struct ubuf_info *uarg,
 			   bool success)
 {
-	uarg->zerocopy = uarg->zerocopy & success;
+	struct ubuf_info_msgzc *uarg_zc = uarg_to_msgzc(uarg);
+
+	uarg_zc->zerocopy = uarg_zc->zerocopy & success;
 
 	if (refcount_dec_and_test(&uarg->refcnt))
-		__msg_zerocopy_callback(uarg);
+		__msg_zerocopy_callback(uarg_zc);
 }
 EXPORT_SYMBOL_GPL(msg_zerocopy_callback);
 
 void msg_zerocopy_put_abort(struct ubuf_info *uarg, bool have_uref)
 {
-	struct sock *sk = skb_from_uarg(uarg)->sk;
+	struct sock *sk = skb_from_uarg(uarg_to_msgzc(uarg))->sk;
 
 	atomic_dec(&sk->sk_zckey);
-	uarg->len--;
+	uarg_to_msgzc(uarg)->len--;
 
 	if (have_uref)
 		msg_zerocopy_callback(NULL, uarg, true);
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 8201cd423ff9..1ae83ad629b2 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -1043,7 +1043,7 @@ static int __ip_append_data(struct sock *sk,
 				paged = true;
 				zc = true;
 			} else {
-				uarg->zerocopy = 0;
+				uarg_to_msgzc(uarg)->zerocopy = 0;
 				skb_zcopy_set(skb, uarg, &extra_uref);
 			}
 		}
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 829beee3fa32..34fd52c07534 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1239,7 +1239,7 @@ int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
 			}
 			zc = sk->sk_route_caps & NETIF_F_SG;
 			if (!zc)
-				uarg->zerocopy = 0;
+				uarg_to_msgzc(uarg)->zerocopy = 0;
 		}
 	}
 
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index e3dbb32b70dc..2f279cbd2648 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1569,7 +1569,7 @@ static int __ip6_append_data(struct sock *sk,
 				paged = true;
 				zc = true;
 			} else {
-				uarg->zerocopy = 0;
+				uarg_to_msgzc(uarg)->zerocopy = 0;
 				skb_zcopy_set(skb, uarg, &extra_uref);
 			}
 		}
-- 
2.37.2