public inbox for passt-dev@passt.top
 help / color / mirror / code / Atom feed
From: Stefano Brivio <sbrivio@redhat.com>
To: David Gibson <david@gibson.dropbear.id.au>
Cc: passt-dev@passt.top
Subject: Re: [PATCH v2 08/32] tcp: Add connection union type
Date: Fri, 18 Nov 2022 01:25:18 +0100	[thread overview]
Message-ID: <20221118012518.49a46b94@elisabeth> (raw)
In-Reply-To: <20221117055908.2782981-9-david@gibson.dropbear.id.au>

On Thu, 17 Nov 2022 16:58:44 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:

> Currently, the tables for spliced and non-spliced connections are entirely
> separate, with different types in different arrays.  We want to unify them.
> As a first step, create a union type which can represent either a spliced
> or non-spliced connection.  For them to be distinguishable, the individual
> types need to have a common header added, with a bit indicating which type
> this structure is.
> 
> This comes at the cost of increasing the size of tcp_tap_conn to over one
> (64 byte) cacheline.  This isn't ideal, but it makes things simpler for now
> and we'll re-optimize this later.
> 
> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> ---
>  tcp.c        |  4 ++++
>  tcp_conn.h   | 30 ++++++++++++++++++++++++++++++
>  tcp_splice.c |  2 ++
>  3 files changed, 36 insertions(+)
> 
> diff --git a/tcp.c b/tcp.c
> index 189041c..05eed85 100644
> --- a/tcp.c
> +++ b/tcp.c
> @@ -288,6 +288,7 @@
>  #include <sys/uio.h>
>  #include <unistd.h>
>  #include <time.h>
> +#include <assert.h>
>  
>  #include <linux/tcp.h> /* For struct tcp_info */
>  
> @@ -601,6 +602,7 @@ static inline struct tcp_tap_conn *conn_at_idx(int index)
>  {
>  	if ((index < 0) || (index >= TCP_MAX_CONNS))
>  		return NULL;
> +	assert(!(CONN(index)->c.spliced));
>  	return CONN(index);
>  }
>  
> @@ -2096,6 +2098,7 @@ static void tcp_conn_from_tap(struct ctx *c, int af, const void *addr,
>  	}
>  
>  	conn = CONN(c->tcp.conn_count++);
> +	conn->c.spliced = false;
>  	conn->sock = s;
>  	conn->timer = -1;
>  	conn_event(c, conn, TAP_SYN_RCVD);
> @@ -2764,6 +2767,7 @@ static void tcp_conn_from_sock(struct ctx *c, union epoll_ref ref,
>  		return;
>  
>  	conn = CONN(c->tcp.conn_count++);
> +	conn->c.spliced = false;
>  	conn->sock = s;
>  	conn->timer = -1;
>  	conn->ws_to_tap = conn->ws_from_tap = 0;
> diff --git a/tcp_conn.h b/tcp_conn.h
> index db4c2d9..39d104a 100644
> --- a/tcp_conn.h
> +++ b/tcp_conn.h
> @@ -11,8 +11,19 @@
>  
>  #define TCP_HASH_BUCKET_BITS		(TCP_CONN_INDEX_BITS + 1)
>  
> +/**
> + * struct tcp_conn_common - Common fields for spliced and non-spliced
> + * @spliced:		Is this a spliced connection?
> + */
> +struct tcp_conn_common {
> +	bool spliced	:1;
> +};
> +
> +extern const char *tcp_common_flag_str[];
> +
>  /**
>   * struct tcp_tap_conn - Descriptor for a TCP connection (not spliced)
> + * @c:			Fields common with tcp_splice_conn
>   * @next_index:		Connection index of next item in hash chain, -1 for none
>   * @tap_mss:		MSS advertised by tap/guest, rounded to 2 ^ TCP_MSS_BITS
>   * @sock:		Socket descriptor number
> @@ -40,6 +51,9 @@
>   * @seq_init_from_tap:	Initial sequence number from tap
>   */
>  struct tcp_tap_conn {
> +	/* Must be first element to match tcp_splice_conn */
> +	struct tcp_conn_common c;
> +
>  	int	 	next_index	:TCP_CONN_INDEX_BITS + 2;
>  
>  #define TCP_RETRANS_BITS		3
> @@ -122,6 +136,7 @@ struct tcp_tap_conn {
>  
>  /**
>   * struct tcp_splice_conn - Descriptor for a spliced TCP connection
> + * @c:			Fields common with tcp_tap_conn
>   * @a:			File descriptor number of socket for accepted connection
>   * @pipe_a_b:		Pipe ends for splice() from @a to @b
>   * @b:			File descriptor number of peer connected socket
> @@ -134,6 +149,9 @@ struct tcp_tap_conn {
>   * @b_written:		Bytes written to @b (not fully written from one @a read)
>  */
>  struct tcp_splice_conn {
> +	/* Must be first element to match tcp_tap_conn */
> +	struct tcp_conn_common c;
> +
>  	int a;
>  	int pipe_a_b[2];
>  	int b;
> @@ -165,4 +183,16 @@ struct tcp_splice_conn {
>  	uint32_t b_written;
>  };
>  
> +/**
> + * union tcp_conn - Descriptor for a TCP connection (spliced or non-spliced)
> + * @c:			Fields common between all variants
> + * @tap:		Fields specific to non-spliced connections
> + * @splice:		Fields specific to spliced connections
> +*/
> +union tcp_conn {
> +	struct tcp_conn_common c;
> +	struct tcp_tap_conn tap;
> +	struct tcp_splice_conn splice;
> +};

Sorry, I could have noticed earlier: I understand that this is needed
to end up, at the end of the series, with a 64-byte tcp_conn, but it
doesn't really look like the most natural way of doing things. I would
have expected something like:

struct tcp_conn {
	struct tcp_conn_common c;
	union {
		struct tcp_tap_conn tap;
		struct tcp_splice_conn splice;
	} u;
};

but sure, if we do this, then we have 3 bytes between 'c' and 'u', and
struct tcp_conn becomes 68 bytes long.

It also confuses Coverity Scan, because in tcp_table_compact() we have:

	memset(hole, 0, sizeof(*hole));

and while the prototype is:

	void tcp_table_compact(struct ctx *c, union tcp_conn *hole)

it sees that we're passing, from tcp_splice_destroy(), something
smaller than that (48 bytes), but we're zeroing the whole thing.

Of course, it's not a real issue, that space is reserved for a
connection slot anyway, but given there are no other issues reported,
I'd try to keep Coverity happy if possible.

First try, failed: check hole->c.spliced and, if set, zero only
sizeof(struct tcp_splice_conn) bytes. This looks like a false
positive.

Another try, which should probably work (I just hit the daily build
submission quota, grr): explicitly pass the union tcp_conn containing
our struct tcp_splice_conn. This patch does it:

---
diff --git a/tcp.c b/tcp.c
index 8874789..d635a8e 100644
--- a/tcp.c
+++ b/tcp.c
@@ -591,7 +591,7 @@ static size_t tcp6_l2_flags_buf_bytes;
 union tcp_conn tc[TCP_MAX_CONNS];
 
 #define CONN(index)		(&tc[(index)].tap)
-#define CONN_IDX(conn)		((union tcp_conn *)(conn) - tc)
+#define CONN_IDX(conn)		(TCP_TAP_TO_COMMON(conn) - tc)
 
 /** conn_at_idx() - Find a connection by index, if present
  * @index:	Index of connection to lookup
@@ -1385,7 +1385,7 @@ static void tcp_conn_destroy(struct ctx *c, struct tcp_tap_conn *conn)
 		close(conn->timer);
 
 	tcp_hash_remove(c, conn);
-	tcp_table_compact(c, (union tcp_conn *)conn);
+	tcp_table_compact(c, TCP_TAP_TO_COMMON(conn));
 }
 
 static void tcp_rst_do(struct ctx *c, struct tcp_tap_conn *conn);
diff --git a/tcp_conn.h b/tcp_conn.h
index 4a8be29..fa407ad 100644
--- a/tcp_conn.h
+++ b/tcp_conn.h
@@ -176,6 +176,12 @@ union tcp_conn {
 	struct tcp_splice_conn splice;
 };
 
+#define TCP_TAP_TO_COMMON(x) \
+	((union tcp_conn *)((char *)(x) - offsetof(union tcp_conn, tap)))
+
+#define TCP_SPLICE_TO_COMMON(x) \
+	((union tcp_conn *)((char *)(x) - offsetof(union tcp_conn, splice)))
+
 /* TCP connections */
 extern union tcp_conn tc[];
 
diff --git a/tcp_splice.c b/tcp_splice.c
index e2f0ce1..7d3f17e 100644
--- a/tcp_splice.c
+++ b/tcp_splice.c
@@ -37,6 +37,7 @@
 #include <limits.h>
 #include <stdint.h>
 #include <stdbool.h>
+#include <stddef.h>
 #include <string.h>
 #include <time.h>
 #include <unistd.h>
@@ -74,7 +75,7 @@ static int splice_pipe_pool		[TCP_SPLICE_PIPE_POOL_SIZE][2][2];
 #define CONN_V4(x)			(!CONN_V6(x))
 #define CONN_HAS(conn, set)		((conn->events & (set)) == (set))
 #define CONN(index)			(&tc[(index)].splice)
-#define CONN_IDX(conn)			((union tcp_conn *)(conn) - tc)
+#define CONN_IDX(conn)			(TCP_SPLICE_TO_COMMON(conn) - tc)
 
 /* Display strings for connection events */
 static const char *tcp_splice_event_str[] __attribute((__unused__)) = {
@@ -283,7 +284,7 @@ void tcp_splice_destroy(struct ctx *c, struct tcp_splice_conn *conn)
 	debug("TCP (spliced): index %li, CLOSED", CONN_IDX(conn));
 
 	c->tcp.splice_conn_count--;
-	tcp_table_compact(c, (union tcp_conn *)conn);
+	tcp_table_compact(c, TCP_SPLICE_TO_COMMON(conn));
 }
 
 /**
---

I can add it on top if you agree, assuming it works.

I also tried to actually turn tcp_conn into a struct. It takes 68 bytes,
so I'm not pursuing this approach, but I'm including the diff just in
case you have some quick idea to fix it up:

---
diff --git a/tcp.c b/tcp.c
index 8874789..6ee5675 100644
--- a/tcp.c
+++ b/tcp.c
@@ -588,10 +588,10 @@ static unsigned int tcp6_l2_flags_buf_used;
 static size_t tcp6_l2_flags_buf_bytes;
 
 /* TCP connections */
-union tcp_conn tc[TCP_MAX_CONNS];
+struct tcp_conn tc[TCP_MAX_CONNS];
 
-#define CONN(index)		(&tc[(index)].tap)
-#define CONN_IDX(conn)		((union tcp_conn *)(conn) - tc)
+#define CONN(index)		(&tc[(index)].u.tap)
+#define CONN_IDX(conn)		(TO_TCP_CONN(conn) - tc)
 
 /** conn_at_idx() - Find a connection by index, if present
  * @index:	Index of connection to lookup
@@ -602,7 +602,7 @@ static inline struct tcp_tap_conn *conn_at_idx(int index)
 {
 	if ((index < 0) || (index >= TCP_MAX_CONNS))
 		return NULL;
-	assert(!(CONN(index)->c.spliced));
+	assert(!TO_TCP_CONN(CONN(index))->c.spliced);
 	return CONN(index);
 }
 
@@ -660,13 +660,13 @@ static void conn_flag_do(const struct ctx *c, struct tcp_tap_conn *conn,
  */
 static int tcp_epoll_ctl(const struct ctx *c, struct tcp_tap_conn *conn)
 {
-	int m = conn->c.in_epoll ? EPOLL_CTL_MOD : EPOLL_CTL_ADD;
+	int m = TO_TCP_CONN(conn)->c.in_epoll ? EPOLL_CTL_MOD : EPOLL_CTL_ADD;
 	union epoll_ref ref = { .r.proto = IPPROTO_TCP, .r.s = conn->sock,
 				.r.p.tcp.tcp.index = CONN_IDX(conn) };
 	struct epoll_event ev = { .data.u64 = ref.u64 };
 
 	if (conn->events == CLOSED) {
-		if (conn->c.in_epoll)
+		if (TO_TCP_CONN(conn)->c.in_epoll)
 			epoll_ctl(c->epollfd, EPOLL_CTL_DEL, conn->sock, &ev);
 		if (conn->timer != -1)
 			epoll_ctl(c->epollfd, EPOLL_CTL_DEL, conn->timer, &ev);
@@ -678,7 +678,7 @@ static int tcp_epoll_ctl(const struct ctx *c, struct tcp_tap_conn *conn)
 	if (epoll_ctl(c->epollfd, m, conn->sock, &ev))
 		return -errno;
 
-	conn->c.in_epoll = true;
+	TO_TCP_CONN(conn)->c.in_epoll = true;
 
 	if (conn->timer != -1) {
 		union epoll_ref ref_t = { .r.proto = IPPROTO_TCP,
@@ -1347,9 +1347,9 @@ static struct tcp_tap_conn *tcp_hash_lookup(const struct ctx *c,
  * @c:		Execution context
  * @hole:	Pointer to recently closed connection
  */
-void tcp_table_compact(struct ctx *c, union tcp_conn *hole)
+void tcp_table_compact(struct ctx *c, struct tcp_conn *hole)
 {
-	union tcp_conn *from;
+	struct tcp_conn *from;
 
 	if (CONN_IDX(hole) == --c->tcp.conn_count) {
 		debug("TCP: table compaction: maximum index was %li (%p)",
@@ -1361,14 +1361,15 @@ void tcp_table_compact(struct ctx *c, union tcp_conn *hole)
 	from = tc + c->tcp.conn_count;
 	memcpy(hole, from, sizeof(*hole));
 
-	if (from->c.spliced)
-		tcp_splice_conn_update(c, &hole->splice);
+	if (TO_TCP_CONN(from)->c.spliced)
+		tcp_splice_conn_update(c, &hole->u.splice);
 	else
-		tcp_tap_conn_update(c, &from->tap, &hole->tap);
+		tcp_tap_conn_update(c, &from->u.tap, &hole->u.tap);
 
 	debug("TCP: table compaction (spliced=%d): old index %li, new index %li, "
 	      "from: %p, to: %p",
-	      from->c.spliced, CONN_IDX(from), CONN_IDX(hole), from, hole);
+	      TO_TCP_CONN(from)->c.spliced, CONN_IDX(from), CONN_IDX(hole),
+	      from, hole);
 
 	memset(from, 0, sizeof(*from));
 }
@@ -1385,7 +1386,7 @@ static void tcp_conn_destroy(struct ctx *c, struct tcp_tap_conn *conn)
 		close(conn->timer);
 
 	tcp_hash_remove(c, conn);
-	tcp_table_compact(c, (union tcp_conn *)conn);
+	tcp_table_compact(c, TO_TCP_CONN(conn));
 }
 
 static void tcp_rst_do(struct ctx *c, struct tcp_tap_conn *conn);
@@ -1523,7 +1524,7 @@ void tcp_defer_handler(struct ctx *c)
 {
 	int max_conns = c->tcp.conn_count / 100 * TCP_CONN_PRESSURE;
 	int max_files = c->nofile / 100 * TCP_FILE_PRESSURE;
-	union tcp_conn *conn;
+	struct tcp_conn *conn;
 
 	tcp_l2_flags_buf_flush(c);
 	tcp_l2_data_buf_flush(c);
@@ -1533,12 +1534,12 @@ void tcp_defer_handler(struct ctx *c)
 		return;
 
 	for (conn = tc + c->tcp.conn_count - 1; conn >= tc; conn--) {
-		if (conn->c.spliced) {
-			if (conn->splice.flags & CLOSING)
-				tcp_splice_destroy(c, &conn->splice);
+		if (TO_TCP_CONN(conn)->c.spliced) {
+			if (conn->u.splice.flags & CLOSING)
+				tcp_splice_destroy(c, &conn->u.splice);
 		} else {
-			if (conn->tap.events == CLOSED)
-				tcp_conn_destroy(c, &conn->tap);
+			if (conn->u.tap.events == CLOSED)
+				tcp_conn_destroy(c, &conn->u.tap);
 		}
 	}
 
@@ -2086,7 +2087,7 @@ static void tcp_conn_from_tap(struct ctx *c, int af, const void *addr,
 	}
 
 	conn = CONN(c->tcp.conn_count++);
-	conn->c.spliced = false;
+	TO_TCP_CONN(conn)->c.spliced = false;
 	conn->sock = s;
 	conn->timer = -1;
 	conn_event(c, conn, TAP_SYN_RCVD);
@@ -2770,7 +2771,7 @@ static void tcp_tap_conn_from_sock(struct ctx *c, union epoll_ref ref,
 				   struct sockaddr *sa,
 				   const struct timespec *now)
 {
-	conn->c.spliced = false;
+	TO_TCP_CONN(conn)->c.spliced = false;
 	conn->sock = s;
 	conn->timer = -1;
 	conn->ws_to_tap = conn->ws_from_tap = 0;
@@ -2804,7 +2805,7 @@ static void tcp_conn_from_sock(struct ctx *c, union epoll_ref ref,
 			       const struct timespec *now)
 {
 	struct sockaddr_storage sa;
-	union tcp_conn *conn;
+	struct tcp_conn *conn;
 	socklen_t sl;
 	int s;
 
@@ -2826,11 +2827,11 @@ static void tcp_conn_from_sock(struct ctx *c, union epoll_ref ref,
 	conn = tc + c->tcp.conn_count++;
 
 	if (c->mode == MODE_PASTA &&
-	    tcp_splice_conn_from_sock(c, ref, &conn->splice,
+	    tcp_splice_conn_from_sock(c, ref, &conn->u.splice,
 				      s, (struct sockaddr *)&sa))
 		return;
 
-	tcp_tap_conn_from_sock(c, ref, &conn->tap, s,
+	tcp_tap_conn_from_sock(c, ref, &conn->u.tap, s,
 			       (struct sockaddr *)&sa, now);
 }
 
@@ -2961,7 +2962,7 @@ static void tcp_tap_sock_handler(struct ctx *c, struct tcp_tap_conn *conn,
 void tcp_sock_handler(struct ctx *c, union epoll_ref ref, uint32_t events,
 		      const struct timespec *now)
 {
-	union tcp_conn *conn;
+	struct tcp_conn *conn;
 
 	if (ref.r.p.tcp.tcp.timer) {
 		tcp_timer_handler(c, ref);
@@ -2975,10 +2976,10 @@ void tcp_sock_handler(struct ctx *c, union epoll_ref ref, uint32_t events,
 
 	conn = tc + ref.r.p.tcp.tcp.index;
 
-	if (conn->c.spliced)
-		tcp_splice_sock_handler(c, &conn->splice, ref.r.s, events);
+	if (TO_TCP_CONN(conn)->c.spliced)
+		tcp_splice_sock_handler(c, &conn->u.splice, ref.r.s, events);
 	else
-		tcp_tap_sock_handler(c, &conn->tap, events);
+		tcp_tap_sock_handler(c, &conn->u.tap, events);
 }
 
 /**
@@ -3370,7 +3371,7 @@ static int tcp_port_rebind(void *arg)
 void tcp_timer(struct ctx *c, const struct timespec *ts)
 {
 	struct tcp_sock_refill_arg refill_arg = { c, 0 };
-	union tcp_conn *conn;
+	struct tcp_conn *conn;
 
 	(void)ts;
 
@@ -3394,11 +3395,11 @@ void tcp_timer(struct ctx *c, const struct timespec *ts)
 	}
 
 	for (conn = tc + c->tcp.conn_count - 1; conn >= tc; conn--) {
-		if (conn->c.spliced) {
-			tcp_splice_timer(c, &conn->splice);
+		if (TO_TCP_CONN(conn)->c.spliced) {
+			tcp_splice_timer(c, &conn->u.splice);
 		} else {
-			if (conn->tap.events == CLOSED)
-				tcp_conn_destroy(c, &conn->tap);
+			if (conn->u.tap.events == CLOSED)
+				tcp_conn_destroy(c, &conn->u.tap);
 		}
 	}
 
diff --git a/tcp_conn.h b/tcp_conn.h
index 4a8be29..3df7905 100644
--- a/tcp_conn.h
+++ b/tcp_conn.h
@@ -23,7 +23,6 @@ extern const char *tcp_common_flag_str[];
 
 /**
  * struct tcp_tap_conn - Descriptor for a TCP connection (not spliced)
- * @c:			Fields common with tcp_splice_conn
  * @next_index:		Connection index of next item in hash chain, -1 for none
  * @tap_mss:		MSS advertised by tap/guest, rounded to 2 ^ TCP_MSS_BITS
  * @sock:		Socket descriptor number
@@ -47,9 +46,6 @@ extern const char *tcp_common_flag_str[];
  * @seq_init_from_tap:	Initial sequence number from tap
  */
 struct tcp_tap_conn {
-	/* Must be first element to match tcp_splice_conn */
-	struct tcp_conn_common c;
-
 	int	 	next_index	:TCP_CONN_INDEX_BITS + 2;
 
 #define TCP_RETRANS_BITS		3
@@ -118,7 +114,6 @@ struct tcp_tap_conn {
 
 /**
  * struct tcp_splice_conn - Descriptor for a spliced TCP connection
- * @c:			Fields common with tcp_tap_conn
  * @a:			File descriptor number of socket for accepted connection
  * @pipe_a_b:		Pipe ends for splice() from @a to @b
  * @b:			File descriptor number of peer connected socket
@@ -131,9 +126,6 @@ struct tcp_tap_conn {
  * @b_written:		Bytes written to @b (not fully written from one @a read)
 */
 struct tcp_splice_conn {
-	/* Must be first element to match tcp_tap_conn */
-	struct tcp_conn_common c;
-
 	int a;
 	int pipe_a_b[2];
 	int b;
@@ -165,22 +157,27 @@ struct tcp_splice_conn {
 };
 
 /**
- * union tcp_conn - Descriptor for a TCP connection (spliced or non-spliced)
+ * struct tcp_conn - Descriptor for a TCP connection (spliced or non-spliced)
  * @c:			Fields common between all variants
- * @tap:		Fields specific to non-spliced connections
- * @splice:		Fields specific to spliced connections
+ * @u.tap:		Fields specific to non-spliced connections
+ * @u.splice:		Fields specific to spliced connections
 */
-union tcp_conn {
+struct tcp_conn {
 	struct tcp_conn_common c;
-	struct tcp_tap_conn tap;
-	struct tcp_splice_conn splice;
+	union {
+		struct tcp_tap_conn tap;
+		struct tcp_splice_conn splice;
+	} u;
 };
 
+#define TO_TCP_CONN(x) \
+	((struct tcp_conn *)((char *)(x) - offsetof(struct tcp_conn, u)))
+
 /* TCP connections */
-extern union tcp_conn tc[];
+extern struct tcp_conn tc[];
 
 void tcp_splice_conn_update(struct ctx *c, struct tcp_splice_conn *new);
-void tcp_table_compact(struct ctx *c, union tcp_conn *hole);
+void tcp_table_compact(struct ctx *c, struct tcp_conn *hole);
 void tcp_splice_destroy(struct ctx *c, struct tcp_splice_conn *conn);
 void tcp_splice_timer(struct ctx *c, struct tcp_splice_conn *conn);
 void tcp_splice_pipe_refill(const struct ctx *c);
diff --git a/tcp_splice.c b/tcp_splice.c
index e2f0ce1..04fc513 100644
--- a/tcp_splice.c
+++ b/tcp_splice.c
@@ -37,6 +37,7 @@
 #include <limits.h>
 #include <stdint.h>
 #include <stdbool.h>
+#include <stddef.h>
 #include <string.h>
 #include <time.h>
 #include <unistd.h>
@@ -73,8 +74,8 @@ static int splice_pipe_pool		[TCP_SPLICE_PIPE_POOL_SIZE][2][2];
 #define CONN_V6(x)			(x->flags & SPLICE_V6)
 #define CONN_V4(x)			(!CONN_V6(x))
 #define CONN_HAS(conn, set)		((conn->events & (set)) == (set))
-#define CONN(index)			(&tc[(index)].splice)
-#define CONN_IDX(conn)			((union tcp_conn *)(conn) - tc)
+#define CONN(index)			(&tc[(index)].u.splice)
+#define CONN_IDX(conn)			(TO_TCP_CONN(conn) - tc)
 
 /* Display strings for connection events */
 static const char *tcp_splice_event_str[] __attribute((__unused__)) = {
@@ -165,7 +166,7 @@ static void conn_flag_do(const struct ctx *c, struct tcp_splice_conn *conn,
 static int tcp_splice_epoll_ctl(const struct ctx *c,
 				struct tcp_splice_conn *conn)
 {
-	int m = conn->c.in_epoll ? EPOLL_CTL_MOD : EPOLL_CTL_ADD;
+	int m = TO_TCP_CONN(conn)->c.in_epoll ? EPOLL_CTL_MOD : EPOLL_CTL_ADD;
 	union epoll_ref ref_a = { .r.proto = IPPROTO_TCP, .r.s = conn->a,
 				  .r.p.tcp.tcp.index = CONN_IDX(conn) };
 	union epoll_ref ref_b = { .r.proto = IPPROTO_TCP, .r.s = conn->b,
@@ -185,7 +186,7 @@ static int tcp_splice_epoll_ctl(const struct ctx *c,
 	    epoll_ctl(c->epollfd, m, conn->b, &ev_b))
 		goto delete;
 
-	conn->c.in_epoll = true;
+	TO_TCP_CONN(conn)->c.in_epoll = true;
 
 	return 0;
 
@@ -283,7 +284,7 @@ void tcp_splice_destroy(struct ctx *c, struct tcp_splice_conn *conn)
 	debug("TCP (spliced): index %li, CLOSED", CONN_IDX(conn));
 
 	c->tcp.splice_conn_count--;
-	tcp_table_compact(c, (union tcp_conn *)conn);
+	tcp_table_compact(c, TO_TCP_CONN(conn));
 }
 
 /**
@@ -535,7 +536,7 @@ bool tcp_splice_conn_from_sock(struct ctx *c, union epoll_ref ref,
 	if (setsockopt(s, SOL_TCP, TCP_QUICKACK, &((int){ 1 }), sizeof(int)))
 		trace("TCP (spliced): failed to set TCP_QUICKACK on %i", s);
 
-	conn->c.spliced = true;
+	TO_TCP_CONN(conn)->c.spliced = true;
 	c->tcp.splice_conn_count++;
 	conn->a = s;
 
---

I'm fine with this series in any case. If you don't have other ideas,
I would just try to get rid of that warning (Out-of-bounds access,
CWE-119) with the first diff here, or something similar.

-- 
@@ -37,6 +37,7 @@
 #include <limits.h>
 #include <stdint.h>
 #include <stdbool.h>
+#include <stddef.h>
 #include <string.h>
 #include <time.h>
 #include <unistd.h>
@@ -73,8 +74,8 @@ static int splice_pipe_pool		[TCP_SPLICE_PIPE_POOL_SIZE][2][2];
 #define CONN_V6(x)			(x->flags & SPLICE_V6)
 #define CONN_V4(x)			(!CONN_V6(x))
 #define CONN_HAS(conn, set)		((conn->events & (set)) == (set))
-#define CONN(index)			(&tc[(index)].splice)
-#define CONN_IDX(conn)			((union tcp_conn *)(conn) - tc)
+#define CONN(index)			(&tc[(index)].u.splice)
+#define CONN_IDX(conn)			(TO_TCP_CONN(conn) - tc)
 
 /* Display strings for connection events */
 static const char *tcp_splice_event_str[] __attribute((__unused__)) = {
@@ -165,7 +166,7 @@ static void conn_flag_do(const struct ctx *c, struct tcp_splice_conn *conn,
 static int tcp_splice_epoll_ctl(const struct ctx *c,
 				struct tcp_splice_conn *conn)
 {
-	int m = conn->c.in_epoll ? EPOLL_CTL_MOD : EPOLL_CTL_ADD;
+	int m = TO_TCP_CONN(conn)->c.in_epoll ? EPOLL_CTL_MOD : EPOLL_CTL_ADD;
 	union epoll_ref ref_a = { .r.proto = IPPROTO_TCP, .r.s = conn->a,
 				  .r.p.tcp.tcp.index = CONN_IDX(conn) };
 	union epoll_ref ref_b = { .r.proto = IPPROTO_TCP, .r.s = conn->b,
@@ -185,7 +186,7 @@ static int tcp_splice_epoll_ctl(const struct ctx *c,
 	    epoll_ctl(c->epollfd, m, conn->b, &ev_b))
 		goto delete;
 
-	conn->c.in_epoll = true;
+	TO_TCP_CONN(conn)->c.in_epoll = true;
 
 	return 0;
 
@@ -283,7 +284,7 @@ void tcp_splice_destroy(struct ctx *c, struct tcp_splice_conn *conn)
 	debug("TCP (spliced): index %li, CLOSED", CONN_IDX(conn));
 
 	c->tcp.splice_conn_count--;
-	tcp_table_compact(c, (union tcp_conn *)conn);
+	tcp_table_compact(c, TO_TCP_CONN(conn));
 }
 
 /**
@@ -535,7 +536,7 @@ bool tcp_splice_conn_from_sock(struct ctx *c, union epoll_ref ref,
 	if (setsockopt(s, SOL_TCP, TCP_QUICKACK, &((int){ 1 }), sizeof(int)))
 		trace("TCP (spliced): failed to set TCP_QUICKACK on %i", s);
 
-	conn->c.spliced = true;
+	TO_TCP_CONN(conn)->c.spliced = true;
 	c->tcp.splice_conn_count++;
 	conn->a = s;
 
---

I'm fine with this series in any case. If you don't have other ideas,
I would just try to get rid of that warning (Out-of-bounds access,
CWE-119) with the first diff here, or something similar.

-- 
Stefano


  reply	other threads:[~2022-11-18  0:25 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-17  5:58 [PATCH v2 00/32] Use dual stack sockets to listen for inbound TCP connections David Gibson
2022-11-17  5:58 ` [PATCH v2 01/32] clang-tidy: Suppress warning about assignments in if statements David Gibson
2022-11-17  5:58 ` [PATCH v2 02/32] style: Minor corrections to function comments David Gibson
2022-11-17  5:58 ` [PATCH v2 03/32] tcp_splice: #include tcp_splice.h in tcp_splice.c David Gibson
2022-11-17  5:58 ` [PATCH v2 04/32] tcp: Remove unused TCP_MAX_SOCKS constant David Gibson
2022-11-17  5:58 ` [PATCH v2 05/32] tcp: Better helpers for converting between connection pointer and index David Gibson
2022-11-17  5:58 ` [PATCH v2 06/32] tcp_splice: Helpers for converting from index to/from tcp_splice_conn David Gibson
2022-11-17  5:58 ` [PATCH v2 07/32] tcp: Move connection state structures into a shared header David Gibson
2022-11-17  5:58 ` [PATCH v2 08/32] tcp: Add connection union type David Gibson
2022-11-18  0:25   ` Stefano Brivio [this message]
2022-11-18  1:10     ` David Gibson
2022-11-19  8:39       ` Stefano Brivio
2022-11-17  5:58 ` [PATCH v2 09/32] tcp: Improved helpers to update connections after moving David Gibson
2022-11-17  5:58 ` [PATCH v2 10/32] tcp: Unify spliced and non-spliced connection tables David Gibson
2022-11-17  5:58 ` [PATCH v2 11/32] tcp: Unify tcp_defer_handler and tcp_splice_defer_handler() David Gibson
2022-11-17  5:58 ` [PATCH v2 12/32] tcp: Partially unify tcp_timer() and tcp_splice_timer() David Gibson
2022-11-17  5:58 ` [PATCH v2 13/32] tcp: Unify the IN_EPOLL flag David Gibson
2022-11-17  5:58 ` [PATCH v2 14/32] tcp: Separate helpers to create ns listening sockets David Gibson
2022-11-17  5:58 ` [PATCH v2 15/32] tcp: Unify part of spliced and non-spliced conn_from_sock path David Gibson
2022-11-17  5:58 ` [PATCH v2 16/32] tcp: Use the same sockets to listen for spliced and non-spliced connections David Gibson
2022-11-17  5:58 ` [PATCH v2 17/32] tcp: Remove splice from tcp_epoll_ref David Gibson
2022-11-17  5:58 ` [PATCH v2 18/32] tcp: Don't store hash bucket in connection structures David Gibson
2022-11-17  5:58 ` [PATCH v2 19/32] inany: Helper functions for handling addresses which could be IPv4 or IPv6 David Gibson
2022-11-17  5:58 ` [PATCH v2 20/32] tcp: Hash IPv4 and IPv4-mapped-IPv6 addresses the same David Gibson
2022-11-17  5:58 ` [PATCH v2 21/32] tcp: Take tcp_hash_insert() address from struct tcp_conn David Gibson
2022-11-17  5:58 ` [PATCH v2 22/32] tcp: Simplify tcp_hash_match() to take an inany_addr David Gibson
2022-11-17  5:58 ` [PATCH v2 23/32] tcp: Unify initial sequence number calculation for IPv4 and IPv6 David Gibson
2022-11-17  5:59 ` [PATCH v2 24/32] tcp: Have tcp_seq_init() take its parameters from struct tcp_conn David Gibson
2022-11-17  5:59 ` [PATCH v2 25/32] tcp: Fix small errors in tcp_seq_init() time handling David Gibson
2022-11-17  5:59 ` [PATCH v2 26/32] tcp: Remove v6 flag from tcp_epoll_ref David Gibson
2022-11-17  5:59 ` [PATCH v2 27/32] tcp: NAT IPv4-mapped IPv6 addresses like IPv4 addresses David Gibson
2022-11-17  5:59 ` [PATCH v2 28/32] tcp_splice: Allow splicing of connections from IPv4-mapped loopback David Gibson
2022-11-17  5:59 ` [PATCH v2 29/32] tcp: Consolidate tcp_sock_init[46] David Gibson
2022-11-17  5:59 ` [PATCH v2 30/32] util: Allow sock_l4() to open dual stack sockets David Gibson
2022-11-17  5:59 ` [PATCH v2 31/32] util: Always return -1 on error in sock_l4() David Gibson
2022-11-17  5:59 ` [PATCH v2 32/32] tcp: Use dual stack sockets for port forwarding when possible David Gibson
2022-11-25  9:22 ` [PATCH v2 00/32] Use dual stack sockets to listen for inbound TCP connections Stefano Brivio

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20221118012518.49a46b94@elisabeth \
    --to=sbrivio@redhat.com \
    --cc=david@gibson.dropbear.id.au \
    --cc=passt-dev@passt.top \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://passt.top/passt

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for IMAP folder(s).