All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jonathan Lemon <jonathan.lemon@gmail.com>
To: <netdev@vger.kernel.org>, <edumazet@google.com>,
	<willemdebruijn.kernel@gmail.com>
Cc: <kernel-team@fb.com>
Subject: [PATCH 6/9 v1 RFC] skbuff: Call sock_zerocopy_put_abort from skb_zcopy_put_abort
Date: Fri, 18 Dec 2020 12:16:30 -0800	[thread overview]
Message-ID: <20201218201633.2735367-7-jonathan.lemon@gmail.com> (raw)
In-Reply-To: <20201218201633.2735367-1-jonathan.lemon@gmail.com>

From: Jonathan Lemon <bsd@fb.com>

The sock_zerocopy_put_abort function contains logic which is
specific to the current zerocopy implementation.  Add a wrapper
which checks the callback and dispatches apppropriately.

Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
---
 include/linux/skbuff.h | 10 ++++++++++
 net/core/skbuff.c      | 12 +++++-------
 net/ipv4/ip_output.c   |  3 +--
 net/ipv4/tcp.c         |  2 +-
 net/ipv6/ip6_output.c  |  3 +--
 5 files changed, 18 insertions(+), 12 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 64ae6f3adcd5..a50d52b796a7 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -1480,6 +1480,16 @@ static inline void skb_zcopy_put(struct ubuf_info *uarg)
 		uarg->callback(NULL, uarg, true);
 }
 
+static inline void skb_zcopy_put_abort(struct ubuf_info *uarg, bool have_uref)
+{
+	if (uarg) {
+		if (uarg->callback == sock_zerocopy_callback)
+			sock_zerocopy_put_abort(uarg, have_uref);
+		else if (have_uref)
+			skb_zcopy_put(uarg);
+	}
+}
+
 /* Release a reference on a zerocopy structure */
 static inline void skb_zcopy_clear(struct sk_buff *skb, bool succsss)
 {
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 328385cd141e..8352da29f052 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1254,15 +1254,13 @@ EXPORT_SYMBOL_GPL(sock_zerocopy_callback);
 
 void sock_zerocopy_put_abort(struct ubuf_info *uarg, bool have_uref)
 {
-	if (uarg) {
-		struct sock *sk = skb_from_uarg(uarg)->sk;
+	struct sock *sk = skb_from_uarg(uarg)->sk;
 
-		atomic_dec(&sk->sk_zckey);
-		uarg->len--;
+	atomic_dec(&sk->sk_zckey);
+	uarg->len--;
 
-		if (have_uref)
-			skb_zcopy_put(uarg);
-	}
+	if (have_uref)
+		sock_zerocopy_callback(NULL, uarg, true);
 }
 EXPORT_SYMBOL_GPL(sock_zerocopy_put_abort);
 
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 879b76ae4435..65f2299fd682 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -1230,8 +1230,7 @@ static int __ip_append_data(struct sock *sk,
 error_efault:
 	err = -EFAULT;
 error:
-	if (uarg)
-		sock_zerocopy_put_abort(uarg, extra_uref);
+	skb_zcopy_put_abort(uarg, extra_uref);
 	cork->length -= length;
 	IP_INC_STATS(sock_net(sk), IPSTATS_MIB_OUTDISCARDS);
 	refcount_add(wmem_alloc_delta, &sk->sk_wmem_alloc);
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 5c38080df13f..900b6bb7b280 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1440,7 +1440,7 @@ int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
 	if (copied + copied_syn)
 		goto out;
 out_err:
-	sock_zerocopy_put_abort(uarg, true);
+	skb_zcopy_put_abort(uarg, true);
 	err = sk_stream_error(sk, flags, err);
 	/* make sure we wake any epoll edge trigger waiter */
 	if (unlikely(tcp_rtx_and_write_queues_empty(sk) && err == -EAGAIN)) {
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 749ad72386b2..c8c87891533a 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1715,8 +1715,7 @@ static int __ip6_append_data(struct sock *sk,
 error_efault:
 	err = -EFAULT;
 error:
-	if (uarg)
-		sock_zerocopy_put_abort(uarg, extra_uref);
+	skb_zcopy_put_abort(uarg, extra_uref);
 	cork->length -= length;
 	IP6_INC_STATS(sock_net(sk), rt->rt6i_idev, IPSTATS_MIB_OUTDISCARDS);
 	refcount_add(wmem_alloc_delta, &sk->sk_wmem_alloc);
-- 
2.24.1


  parent reply	other threads:[~2020-12-18 20:17 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-18 20:16 [PATCH 0/9 v1 RFC] Generic zcopy_* functions Jonathan Lemon
2020-12-18 20:16 ` [PATCH 1/9 v1 RFC] net: group skb_shinfo zerocopy related bits together Jonathan Lemon
2020-12-18 20:16 ` [PATCH 2/9 v1 RFC] skbuff: remove unused skb_zcopy_abort function Jonathan Lemon
2020-12-18 20:16 ` [PATCH 3/9 v1 RFC] skbuff: replace sock_zerocopy_put() with skb_zcopy_put() Jonathan Lemon
2020-12-19 18:46   ` Willem de Bruijn
2020-12-21 19:18     ` Jonathan Lemon
2020-12-21 22:49       ` Willem de Bruijn
2020-12-18 20:16 ` [PATCH 4/9 v1 RFC] skbuff: replace sock_zerocopy_get with skb_zcopy_get Jonathan Lemon
2020-12-18 20:16 ` [PATCH 5/9 v1 RFC] skbuff: Add skb parameter to the ubuf zerocopy callback Jonathan Lemon
2020-12-18 20:16 ` Jonathan Lemon [this message]
2020-12-18 20:16 ` [PATCH 7/9 v1 RFC] skbuff: add zc_flags to ubuf_info for ubuf setup Jonathan Lemon
2020-12-18 20:16 ` [PATCH 8/9 v1 RFC] tap/tun: use skb_zcopy_set() instead of open coded assignment Jonathan Lemon
2020-12-18 20:16 ` [PATCH 9/9 v1 RFC] skbuff: Call skb_zcopy_clear() before unref'ing fragments Jonathan Lemon
2020-12-18 20:49 ` [PATCH 0/9 v1 RFC] Generic zcopy_* functions Willem de Bruijn
2020-12-18 21:16   ` Jonathan Lemon
2020-12-19 19:00     ` Willem de Bruijn
2020-12-21 19:50       ` Jonathan Lemon
2020-12-21 22:52         ` Willem de Bruijn
2020-12-22  0:07           ` Jonathan Lemon

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=20201218201633.2735367-7-jonathan.lemon@gmail.com \
    --to=jonathan.lemon@gmail.com \
    --cc=edumazet@google.com \
    --cc=kernel-team@fb.com \
    --cc=netdev@vger.kernel.org \
    --cc=willemdebruijn.kernel@gmail.com \
    /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.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.