From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Michael S. Tsirkin" Subject: Re: [PATCH v6 3/3] skb_array: array based FIFO for skbs Date: Thu, 2 Jun 2016 12:27:06 +0300 Message-ID: <20160602122442-mutt-send-email-mst@redhat.com> References: <1464785601-3074-1-git-send-email-mst@redhat.com> <1464785601-3074-4-git-send-email-mst@redhat.com> <574FEB7E.8060902@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: linux-kernel@vger.kernel.org, Eric Dumazet , davem@davemloft.net, netdev@vger.kernel.org, Steven Rostedt , brouer@redhat.com, kvm@vger.kernel.org To: Jason Wang Return-path: Content-Disposition: inline In-Reply-To: <574FEB7E.8060902@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Thu, Jun 02, 2016 at 04:17:02PM +0800, Jason Wang wrote: >=20 >=20 > On 2016=E5=B9=B406=E6=9C=8801=E6=97=A5 20:54, Michael S. Tsirkin wrot= e: > >A simple array based FIFO of pointers. Intended for net stack so us= es > >skbs for type safety. Implemented as a set of wrappers around ptr_ar= ray. > > > >Signed-off-by: Michael S. Tsirkin > >--- > > include/linux/skb_array.h | 129 ++++++++++++++++++++++++++++++++++= ++++++++++++ > > 1 file changed, 129 insertions(+) > > create mode 100644 include/linux/skb_array.h > > > >diff --git a/include/linux/skb_array.h b/include/linux/skb_array.h > >new file mode 100644 > >index 0000000..cb67076 > >--- /dev/null > >+++ b/include/linux/skb_array.h > >@@ -0,0 +1,129 @@ > >+/* > >+ * Definitions for the 'struct skb_array' datastructure. > >+ * > >+ * Author: > >+ * Michael S. Tsirkin > >+ * > >+ * Copyright (C) 2016 Red Hat, Inc. > >+ * > >+ * This program is free software; you can redistribute it and/or mo= dify it > >+ * under the terms of the GNU General Public License as published b= y the > >+ * Free Software Foundation; either version 2 of the License, or (a= t your > >+ * option) any later version. > >+ * > >+ * Limited-size FIFO of skbs. Can be used more or less whenever > >+ * sk_buff_head can be used, except you need to know the queue size= in > >+ * advance. > >+ * Implemented as a type-safe wrapper around ptr_ring. > >+ */ > >+ > >+#ifndef _LINUX_SKB_ARRAY_H > >+#define _LINUX_SKB_ARRAY_H 1 > >+ > >+#ifdef __KERNEL__ > >+#include > >+#include > >+#endif > >+ > >+struct skb_array { > >+ struct ptr_ring ring; > >+}; > >+ > >+/* Might be slightly faster than skb_array_full below, but callers = invoking > >+ * this in a loop must use a compiler barrier, for example cpu_rela= x(). > >+ */ > >+static inline bool __skb_array_full(struct skb_array *a) > >+{ > >+ return __ptr_ring_full(&a->ring); > >+} > >+ > >+static inline bool skb_array_full(struct skb_array *a) > >+{ > >+ return ptr_ring_full(&a->ring); > >+} > >+ > >+static inline int skb_array_produce(struct skb_array *a, struct sk_= buff *skb) > >+{ > >+ return ptr_ring_produce(&a->ring, skb); > >+} > >+ > >+static inline int skb_array_produce_irq(struct skb_array *a, struct= sk_buff *skb) > >+{ > >+ return ptr_ring_produce_irq(&a->ring, skb); > >+} > >+ > >+static inline int skb_array_produce_bh(struct skb_array *a, struct = sk_buff *skb) > >+{ > >+ return ptr_ring_produce_bh(&a->ring, skb); > >+} > >+ > >+static inline int skb_array_produce_any(struct skb_array *a, struct= sk_buff *skb) > >+{ > >+ return ptr_ring_produce_any(&a->ring, skb); > >+} > >+ > >+/* Might be slightly faster than skb_array_empty below, but callers= invoking > >+ * this in a loop must take care to use a compiler barrier, for exa= mple > >+ * cpu_relax(). > >+ */ > >+static inline bool __skb_array_empty(struct skb_array *a) > >+{ > >+ return !__ptr_ring_peek(&a->ring); > >+} > >+ > >+static inline bool skb_array_empty(struct skb_array *a) > >+{ > >+ return ptr_ring_empty(&a->ring); > >+} > >+ > >+static inline struct sk_buff *skb_array_consume(struct skb_array *a= ) > >+{ > >+ return ptr_ring_consume(&a->ring); > >+} > >+ > >+static inline struct sk_buff *skb_array_consume_irq(struct skb_arra= y *a) > >+{ > >+ return ptr_ring_consume_irq(&a->ring); > >+} > >+ > >+static inline struct sk_buff *skb_array_consume_any(struct skb_arra= y *a) > >+{ > >+ return ptr_ring_consume_any(&a->ring); > >+} > >+ > >+static inline struct sk_buff *skb_array_consume_bh(struct skb_array= *a) > >+{ > >+ return ptr_ring_consume_bh(&a->ring); > >+} > >+ > >+static inline int skb_array_peek_len(struct skb_array *a) > >+{ > >+ return PTR_RING_PEEK_FIELD(&a->ring, struct sk_buff, len, 0); > >+} >=20 > An interesting case for tun is peek the length of a tagged packet. In= this > case, len should be increased accordingly. So this simple filed peeki= ng > won't work. >=20 > Do we need a callback here? I'd say let's just make it do the right thing wrt vlans. I'll post v7 with this change. > >+ > >+static inline int skb_array_peek_len_irq(struct skb_array *a) > >+{ > >+ return PTR_RING_PEEK_FIELD_IRQ(&a->ring, struct sk_buff, len, 0); > >+} > >+ > >+static inline int skb_array_peek_len_bh(struct skb_array *a) > >+{ > >+ return PTR_RING_PEEK_FIELD_BH(&a->ring, struct sk_buff, len, 0); > >+} > >+ > >+static inline int skb_array_peek_len_any(struct skb_array *a) > >+{ > >+ return PTR_RING_PEEK_FIELD_ANY(&a->ring, struct sk_buff, len, 0); > >+} > >+ > >+static inline int skb_array_init(struct skb_array *a, int size, gfp= _t gfp) > >+{ > >+ return ptr_ring_init(&a->ring, size, gfp); > >+} > >+ > >+static inline void skb_array_cleanup(struct skb_array *a) > >+{ > >+ ptr_ring_cleanup(&a->ring); > >+} > >+ > >+#endif /* _LINUX_SKB_ARRAY_H */