* [PATCH v4 net-next 3/7] ipeh: Generic TLV parser
From: Tom Herbert @ 2019-08-23 19:13 UTC (permalink / raw)
To: davem, netdev; +Cc: Tom Herbert, Tom Herbert
In-Reply-To: <1566587643-16594-1-git-send-email-tom@herbertland.com>
From: Tom Herbert <tom@quantonium.net>
Create a generic TLV parser. This will be used with various
extension headers that carry options including Destination,
Hop-by-Hop, Segment Routing TLVs, and other cases of simple
stateless parsing.
Signed-off-by: Tom Herbert <tom@herbertland.com>
---
include/net/ipeh.h | 25 ++++++++
net/ipv6/exthdrs.c | 159 +++++++++++-----------------------------------
net/ipv6/exthdrs_common.c | 114 +++++++++++++++++++++++++++++++++
3 files changed, 177 insertions(+), 121 deletions(-)
diff --git a/include/net/ipeh.h b/include/net/ipeh.h
index 3b24831..c1aa7b6 100644
--- a/include/net/ipeh.h
+++ b/include/net/ipeh.h
@@ -31,4 +31,29 @@ struct ipv6_txoptions *ipeh_renew_options(struct sock *sk,
struct ipv6_txoptions *ipeh_fixup_options(struct ipv6_txoptions *opt_space,
struct ipv6_txoptions *opt);
+/* Generic extension header TLV parser */
+
+enum ipeh_parse_errors {
+ IPEH_PARSE_ERR_PAD1, /* Excessive PAD1 */
+ IPEH_PARSE_ERR_PADN, /* Excessive PADN */
+ IPEH_PARSE_ERR_PADNZ, /* Non-zero padding data */
+ IPEH_PARSE_ERR_EH_TOOBIG, /* Length of EH exceeds limit */
+ IPEH_PARSE_ERR_OPT_TOOBIG, /* Option size exceeds limit */
+ IPEH_PARSE_ERR_OPT_TOOMANY, /* Option count exceeds limit */
+ IPEH_PARSE_ERR_OPT_UNK_DISALW, /* Unknown option disallowed */
+ IPEH_PARSE_ERR_OPT_UNK, /* Unknown option */
+};
+
+/* The generic TLV parser assumes that the type value of PAD1 is 0, and PADN
+ * is 1. This is true for Destination, Hop-by-Hop and current definition
+ * of Segment Routing TLVs.
+ */
+#define IPEH_TLV_PAD1 0
+#define IPEH_TLV_PADN 1
+
+bool ipeh_parse_tlv(const struct tlvtype_proc *procs, struct sk_buff *skb,
+ int max_count, int off, int len,
+ bool (*parse_error)(struct sk_buff *skb,
+ int off, enum ipeh_parse_errors error));
+
#endif /* _NET_IPEH_H */
diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c
index e12d3a5..939d27c 100644
--- a/net/ipv6/exthdrs.c
+++ b/net/ipv6/exthdrs.c
@@ -54,135 +54,50 @@
Generic functions
*********************/
-/* An unknown option is detected, decide what to do */
-
-static bool ip6_tlvopt_unknown(struct sk_buff *skb, int optoff,
- bool disallow_unknowns)
+/* Handle parse errors from ipeh generic TLV parser */
+static bool ipv6_parse_error(struct sk_buff *skb, int off,
+ enum ipeh_parse_errors error)
{
- if (disallow_unknowns) {
- /* If unknown TLVs are disallowed by configuration
- * then always silently drop packet. Note this also
- * means no ICMP parameter problem is sent which
- * could be a good property to mitigate a reflection DOS
- * attack.
- */
-
- goto drop;
- }
-
- switch ((skb_network_header(skb)[optoff] & 0xC0) >> 6) {
- case 0: /* ignore */
- return true;
-
- case 1: /* drop packet */
- break;
-
- case 3: /* Send ICMP if not a multicast address and drop packet */
- /* Actually, it is redundant check. icmp_send
- will recheck in any case.
- */
- if (ipv6_addr_is_multicast(&ipv6_hdr(skb)->daddr))
+ switch (error) {
+ case IPEH_PARSE_ERR_OPT_UNK_DISALW:
+ /* Disallow unknown skip */
+ if (((skb_network_header(skb)[off] & 0xC0) >> 6) == 0) {
+ /* Silent drop */
break;
+ }
/* fall through */
- case 2: /* send ICMP PARM PROB regardless and drop packet */
- icmpv6_param_prob(skb, ICMPV6_UNK_OPTION, optoff);
- return false;
- }
-
-drop:
- kfree_skb(skb);
- return false;
-}
+ case IPEH_PARSE_ERR_OPT_UNK:
+ switch ((skb_network_header(skb)[off] & 0xC0) >> 6) {
+ case 0: /* ignore */
+ return true;
-/* Parse tlv encoded option header (hop-by-hop or destination) */
-
-static bool ip6_parse_tlv(const struct tlvtype_proc *procs,
- struct sk_buff *skb,
- int max_count)
-{
- int len = (skb_transport_header(skb)[1] + 1) << 3;
- const unsigned char *nh = skb_network_header(skb);
- int off = skb_network_header_len(skb);
- const struct tlvtype_proc *curr;
- bool disallow_unknowns = false;
- int tlv_count = 0;
- int padlen = 0;
-
- if (unlikely(max_count < 0)) {
- disallow_unknowns = true;
- max_count = -max_count;
- }
-
- if (skb_transport_offset(skb) + len > skb_headlen(skb))
- goto bad;
-
- off += 2;
- len -= 2;
-
- while (len > 0) {
- int optlen = nh[off + 1] + 2;
- int i;
-
- switch (nh[off]) {
- case IPV6_TLV_PAD1:
- optlen = 1;
- padlen++;
- if (padlen > 7)
- goto bad;
+ case 1: /* drop packet */
break;
- case IPV6_TLV_PADN:
- /* RFC 2460 states that the purpose of PadN is
- * to align the containing header to multiples
- * of 8. 7 is therefore the highest valid value.
- * See also RFC 4942, Section 2.1.9.5.
- */
- padlen += optlen;
- if (padlen > 7)
- goto bad;
- /* RFC 4942 recommends receiving hosts to
- * actively check PadN payload to contain
- * only zeroes.
+ case 3: /* Send ICMP if not a multicast address and drop packet
+ *
+ * Actually, it is redundant check. icmp_send
+ * will recheck in any case.
*/
- for (i = 2; i < optlen; i++) {
- if (nh[off + i] != 0)
- goto bad;
- }
- break;
+ if (ipv6_addr_is_multicast(&ipv6_hdr(skb)->daddr))
+ break;
- default: /* Other TLV code so scan list */
- if (optlen > len)
- goto bad;
-
- tlv_count++;
- if (tlv_count > max_count)
- goto bad;
-
- for (curr = procs; curr->type >= 0; curr++) {
- if (curr->type == nh[off]) {
- /* type specific length/alignment
- checks will be performed in the
- func(). */
- if (curr->func(skb, off) == false)
- return false;
- break;
- }
- }
- if (curr->type < 0 &&
- !ip6_tlvopt_unknown(skb, off, disallow_unknowns))
- return false;
-
- padlen = 0;
+ /* fall through */
+ case 2: /* send ICMP PARM PROB regardless and drop packet */
+ icmpv6_send(skb, ICMPV6_PARAMPROB,
+ ICMPV6_UNK_OPTION, off);
break;
}
- off += optlen;
- len -= optlen;
+ break;
+ default:
+ break;
}
- if (len == 0)
- return true;
-bad:
- kfree_skb(skb);
+ /* Will be dropping packet */
+
+ __IP6_INC_STATS(dev_net(skb->dev), __in6_dev_get(skb->dev),
+ IPSTATS_MIB_INHDRERRORS);
+
return false;
}
@@ -216,8 +131,9 @@ static int ipv6_destopt_rcv(struct sk_buff *skb)
dstbuf = opt->dst1;
#endif
- if (ip6_parse_tlv(tlvprocdestopt_lst, skb,
- init_net.ipv6.sysctl.max_dst_opts_cnt)) {
+ if (ipeh_parse_tlv(tlvprocdestopt_lst, skb,
+ init_net.ipv6.sysctl.max_dst_opts_cnt,
+ 2, extlen - 2, ipv6_parse_error)) {
skb->transport_header += extlen;
opt = IP6CB(skb);
#if IS_ENABLED(CONFIG_IPV6_MIP6)
@@ -639,8 +555,9 @@ int ipv6_parse_hopopts(struct sk_buff *skb)
goto fail_and_free;
opt->flags |= IP6SKB_HOPBYHOP;
- if (ip6_parse_tlv(tlvprochopopt_lst, skb,
- init_net.ipv6.sysctl.max_hbh_opts_cnt)) {
+ if (ipeh_parse_tlv(tlvprochopopt_lst, skb,
+ init_net.ipv6.sysctl.max_hbh_opts_cnt,
+ 2, extlen - 2, ipv6_parse_error)) {
skb->transport_header += extlen;
opt = IP6CB(skb);
opt->nhoff = sizeof(struct ipv6hdr);
diff --git a/net/ipv6/exthdrs_common.c b/net/ipv6/exthdrs_common.c
index 2c68184..99a0911 100644
--- a/net/ipv6/exthdrs_common.c
+++ b/net/ipv6/exthdrs_common.c
@@ -142,3 +142,117 @@ struct ipv6_txoptions *ipeh_fixup_options(struct ipv6_txoptions *opt_space,
return opt;
}
EXPORT_SYMBOL_GPL(ipeh_fixup_options);
+
+/* Generic extension header TLV parser
+ *
+ * Arguments:
+ * - skb_transport_header points to the extension header containing options
+ * - off is offset from skb_transport_header where first TLV is
+ * - len is length of TLV block
+ */
+bool ipeh_parse_tlv(const struct tlvtype_proc *procs, struct sk_buff *skb,
+ int max_count, int off, int len,
+ bool (*parse_error)(struct sk_buff *skb,
+ int off, enum ipeh_parse_errors error))
+{
+ const unsigned char *nh = skb_network_header(skb);
+ const struct tlvtype_proc *curr;
+ bool disallow_unknowns = false;
+ int tlv_count = 0;
+ int padlen = 0;
+
+ if (unlikely(max_count < 0)) {
+ disallow_unknowns = true;
+ max_count = -max_count;
+ }
+
+ if (skb_transport_offset(skb) + off + len > skb_headlen(skb)) {
+ if (!parse_error(skb, skb_transport_offset(skb),
+ IPEH_PARSE_ERR_EH_TOOBIG))
+ goto bad;
+
+ len = skb_headlen(skb) - skb_transport_offset(skb) - off;
+ }
+
+ /* ops function based offset on network header */
+ off += skb_network_header_len(skb);
+
+ while (len > 0) {
+ int optlen = nh[off + 1] + 2;
+ int i;
+
+ switch (nh[off]) {
+ case IPEH_TLV_PAD1:
+ optlen = 1;
+ padlen++;
+ if (padlen > 7 &&
+ !parse_error(skb, off, IPEH_PARSE_ERR_PAD1))
+ goto bad;
+
+ break;
+
+ case IPEH_TLV_PADN:
+ /* RFC 2460 states that the purpose of PadN is
+ * to align the containing header to multiples
+ * of 8. 7 is therefore the highest valid value.
+ * See also RFC 4942, Section 2.1.9.5.
+ */
+ padlen += optlen;
+ if (padlen > 7 &&
+ !parse_error(skb, off, IPEH_PARSE_ERR_PADN))
+ goto bad;
+
+ /* RFC 4942 recommends receiving hosts to
+ * actively check PadN payload to contain
+ * only zeroes.
+ */
+ for (i = 2; i < optlen; i++) {
+ if (nh[off + i] != 0 &&
+ !parse_error(skb, off + i,
+ IPEH_PARSE_ERR_PADNZ))
+ goto bad;
+ }
+ break;
+
+ default: /* Other TLV code so scan list */
+ if (optlen > len &&
+ !parse_error(skb, off, IPEH_PARSE_ERR_OPT_TOOBIG))
+ goto bad;
+
+ tlv_count++;
+ if (tlv_count > max_count &&
+ parse_error(skb, off, IPEH_PARSE_ERR_OPT_TOOMANY))
+ goto bad;
+
+ for (curr = procs; curr->type >= 0; curr++) {
+ if (curr->type == nh[off]) {
+ /* type specific length/alignment
+ * checks will be performed in the
+ * func().
+ */
+ if (curr->func(skb, off) == false)
+ return false;
+ break;
+ }
+ }
+ if (curr->type < 0 &&
+ !parse_error(skb, off,
+ disallow_unknowns ?
+ IPEH_PARSE_ERR_OPT_UNK_DISALW :
+ IPEH_PARSE_ERR_OPT_UNK))
+ goto bad;
+
+ padlen = 0;
+ break;
+ }
+ off += optlen;
+ len -= optlen;
+ }
+
+ if (len == 0)
+ return true;
+bad:
+ kfree_skb(skb);
+ return false;
+}
+EXPORT_SYMBOL(ipeh_parse_tlv);
--
2.7.4
^ permalink raw reply related
* [PATCH v4 net-next 2/7] ipeh: Move generic EH functions to exthdrs_common.c
From: Tom Herbert @ 2019-08-23 19:13 UTC (permalink / raw)
To: davem, netdev; +Cc: Tom Herbert, Tom Herbert
In-Reply-To: <1566587643-16594-1-git-send-email-tom@herbertland.com>
From: Tom Herbert <tom@quantonium.net>
Move generic functions in exthdrs.c to new exthdrs_common.c so that
exthdrs.c only contains functions that are specific to IPv6 processing,
and exthdrs_common.c contains functions that are generic. These
functions include those that will be used with IPv4 extension headers.
Generic extension header related functions are prefixed by ipeh_.
Signed-off-by: Tom Herbert <tom@herbertland.com>
---
include/net/ipeh.h | 12 ++++
include/net/ipv6.h | 9 ---
net/dccp/ipv6.c | 2 +-
net/ipv6/Kconfig | 4 ++
net/ipv6/Makefile | 1 +
net/ipv6/calipso.c | 6 +-
net/ipv6/exthdrs.c | 138 --------------------------------------------
net/ipv6/exthdrs_common.c | 144 ++++++++++++++++++++++++++++++++++++++++++++++
net/ipv6/ipv6_sockglue.c | 2 +-
net/ipv6/raw.c | 2 +-
net/ipv6/tcp_ipv6.c | 2 +-
net/ipv6/udp.c | 2 +-
net/l2tp/l2tp_ip6.c | 2 +-
net/sctp/ipv6.c | 2 +-
14 files changed, 171 insertions(+), 157 deletions(-)
create mode 100644 net/ipv6/exthdrs_common.c
diff --git a/include/net/ipeh.h b/include/net/ipeh.h
index ec2d186..3b24831 100644
--- a/include/net/ipeh.h
+++ b/include/net/ipeh.h
@@ -19,4 +19,16 @@ struct tlvtype_proc {
extern const struct tlvtype_proc tlvprocdestopt_lst[];
extern const struct tlvtype_proc tlvprochopopt_lst[];
+struct ipv6_txoptions;
+struct ipv6_opt_hdr;
+
+struct ipv6_txoptions *ipeh_dup_options(struct sock *sk,
+ struct ipv6_txoptions *opt);
+struct ipv6_txoptions *ipeh_renew_options(struct sock *sk,
+ struct ipv6_txoptions *opt,
+ int newtype,
+ struct ipv6_opt_hdr *newopt);
+struct ipv6_txoptions *ipeh_fixup_options(struct ipv6_txoptions *opt_space,
+ struct ipv6_txoptions *opt);
+
#endif /* _NET_IPEH_H */
diff --git a/include/net/ipv6.h b/include/net/ipv6.h
index ec10fca..1c6878b 100644
--- a/include/net/ipv6.h
+++ b/include/net/ipv6.h
@@ -429,15 +429,6 @@ int ip6_ra_control(struct sock *sk, int sel);
int ipv6_parse_hopopts(struct sk_buff *skb);
-struct ipv6_txoptions *ipv6_dup_options(struct sock *sk,
- struct ipv6_txoptions *opt);
-struct ipv6_txoptions *ipv6_renew_options(struct sock *sk,
- struct ipv6_txoptions *opt,
- int newtype,
- struct ipv6_opt_hdr *newopt);
-struct ipv6_txoptions *ipv6_fixup_options(struct ipv6_txoptions *opt_space,
- struct ipv6_txoptions *opt);
-
bool ipv6_opt_accepted(const struct sock *sk, const struct sk_buff *skb,
const struct inet6_skb_parm *opt);
struct ipv6_txoptions *ipv6_update_options(struct sock *sk,
diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c
index 1b7381f..0b83a04 100644
--- a/net/dccp/ipv6.c
+++ b/net/dccp/ipv6.c
@@ -514,7 +514,7 @@ static struct sock *dccp_v6_request_recv_sock(const struct sock *sk,
if (!opt)
opt = rcu_dereference(np->opt);
if (opt) {
- opt = ipv6_dup_options(newsk, opt);
+ opt = ipeh_dup_options(newsk, opt);
RCU_INIT_POINTER(newnp->opt, opt);
}
inet_csk(newsk)->icsk_ext_hdr_len = 0;
diff --git a/net/ipv6/Kconfig b/net/ipv6/Kconfig
index ae1344e..700fcea 100644
--- a/net/ipv6/Kconfig
+++ b/net/ipv6/Kconfig
@@ -3,9 +3,13 @@
# IPv6 configuration
#
+config EXTHDRS
+ bool
+
# IPv6 as module will cause a CRASH if you try to unload it
menuconfig IPV6
tristate "The IPv6 protocol"
+ select EXTHDRS
default y
---help---
Support for IP version 6 (IPv6).
diff --git a/net/ipv6/Makefile b/net/ipv6/Makefile
index df3919b..0bcab81 100644
--- a/net/ipv6/Makefile
+++ b/net/ipv6/Makefile
@@ -44,6 +44,7 @@ obj-$(CONFIG_IPV6_SIT) += sit.o
obj-$(CONFIG_IPV6_TUNNEL) += ip6_tunnel.o
obj-$(CONFIG_IPV6_GRE) += ip6_gre.o
obj-$(CONFIG_IPV6_FOU) += fou6.o
+obj-$(CONFIG_EXTHDRS) += exthdrs_common.o
obj-y += addrconf_core.o exthdrs_core.o ip6_checksum.o ip6_icmp.o
obj-$(CONFIG_INET) += output_core.o protocol.o $(ipv6-offload)
diff --git a/net/ipv6/calipso.c b/net/ipv6/calipso.c
index 221c81f..9c84848 100644
--- a/net/ipv6/calipso.c
+++ b/net/ipv6/calipso.c
@@ -785,7 +785,7 @@ static int calipso_opt_update(struct sock *sk, struct ipv6_opt_hdr *hop)
{
struct ipv6_txoptions *old = txopt_get(inet6_sk(sk)), *txopts;
- txopts = ipv6_renew_options(sk, old, IPV6_HOPOPTS, hop);
+ txopts = ipeh_renew_options(sk, old, IPV6_HOPOPTS, hop);
txopt_put(old);
if (IS_ERR(txopts))
return PTR_ERR(txopts);
@@ -1207,7 +1207,7 @@ static int calipso_req_setattr(struct request_sock *req,
if (IS_ERR(new))
return PTR_ERR(new);
- txopts = ipv6_renew_options(sk, req_inet->ipv6_opt, IPV6_HOPOPTS, new);
+ txopts = ipeh_renew_options(sk, req_inet->ipv6_opt, IPV6_HOPOPTS, new);
kfree(new);
@@ -1244,7 +1244,7 @@ static void calipso_req_delattr(struct request_sock *req)
if (calipso_opt_del(req_inet->ipv6_opt->hopopt, &new))
return; /* Nothing to do */
- txopts = ipv6_renew_options(sk, req_inet->ipv6_opt, IPV6_HOPOPTS, new);
+ txopts = ipeh_renew_options(sk, req_inet->ipv6_opt, IPV6_HOPOPTS, new);
if (!IS_ERR(txopts)) {
txopts = xchg(&req_inet->ipv6_opt, txopts);
diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c
index 664491e..e12d3a5 100644
--- a/net/ipv6/exthdrs.c
+++ b/net/ipv6/exthdrs.c
@@ -784,144 +784,6 @@ void ipv6_push_frag_opts(struct sk_buff *skb, struct ipv6_txoptions *opt, u8 *pr
}
EXPORT_SYMBOL(ipv6_push_frag_opts);
-struct ipv6_txoptions *
-ipv6_dup_options(struct sock *sk, struct ipv6_txoptions *opt)
-{
- struct ipv6_txoptions *opt2;
-
- opt2 = sock_kmalloc(sk, opt->tot_len, GFP_ATOMIC);
- if (opt2) {
- long dif = (char *)opt2 - (char *)opt;
- memcpy(opt2, opt, opt->tot_len);
- if (opt2->hopopt)
- *((char **)&opt2->hopopt) += dif;
- if (opt2->dst0opt)
- *((char **)&opt2->dst0opt) += dif;
- if (opt2->dst1opt)
- *((char **)&opt2->dst1opt) += dif;
- if (opt2->srcrt)
- *((char **)&opt2->srcrt) += dif;
- refcount_set(&opt2->refcnt, 1);
- }
- return opt2;
-}
-EXPORT_SYMBOL_GPL(ipv6_dup_options);
-
-static void ipv6_renew_option(int renewtype,
- struct ipv6_opt_hdr **dest,
- struct ipv6_opt_hdr *old,
- struct ipv6_opt_hdr *new,
- int newtype, char **p)
-{
- struct ipv6_opt_hdr *src;
-
- src = (renewtype == newtype ? new : old);
- if (!src)
- return;
-
- memcpy(*p, src, ipv6_optlen(src));
- *dest = (struct ipv6_opt_hdr *)*p;
- *p += CMSG_ALIGN(ipv6_optlen(*dest));
-}
-
-/**
- * ipv6_renew_options - replace a specific ext hdr with a new one.
- *
- * @sk: sock from which to allocate memory
- * @opt: original options
- * @newtype: option type to replace in @opt
- * @newopt: new option of type @newtype to replace (user-mem)
- * @newoptlen: length of @newopt
- *
- * Returns a new set of options which is a copy of @opt with the
- * option type @newtype replaced with @newopt.
- *
- * @opt may be NULL, in which case a new set of options is returned
- * containing just @newopt.
- *
- * @newopt may be NULL, in which case the specified option type is
- * not copied into the new set of options.
- *
- * The new set of options is allocated from the socket option memory
- * buffer of @sk.
- */
-struct ipv6_txoptions *
-ipv6_renew_options(struct sock *sk, struct ipv6_txoptions *opt,
- int newtype, struct ipv6_opt_hdr *newopt)
-{
- int tot_len = 0;
- char *p;
- struct ipv6_txoptions *opt2;
-
- if (opt) {
- if (newtype != IPV6_HOPOPTS && opt->hopopt)
- tot_len += CMSG_ALIGN(ipv6_optlen(opt->hopopt));
- if (newtype != IPV6_RTHDRDSTOPTS && opt->dst0opt)
- tot_len += CMSG_ALIGN(ipv6_optlen(opt->dst0opt));
- if (newtype != IPV6_RTHDR && opt->srcrt)
- tot_len += CMSG_ALIGN(ipv6_optlen(opt->srcrt));
- if (newtype != IPV6_DSTOPTS && opt->dst1opt)
- tot_len += CMSG_ALIGN(ipv6_optlen(opt->dst1opt));
- }
-
- if (newopt)
- tot_len += CMSG_ALIGN(ipv6_optlen(newopt));
-
- if (!tot_len)
- return NULL;
-
- tot_len += sizeof(*opt2);
- opt2 = sock_kmalloc(sk, tot_len, GFP_ATOMIC);
- if (!opt2)
- return ERR_PTR(-ENOBUFS);
-
- memset(opt2, 0, tot_len);
- refcount_set(&opt2->refcnt, 1);
- opt2->tot_len = tot_len;
- p = (char *)(opt2 + 1);
-
- ipv6_renew_option(IPV6_HOPOPTS, &opt2->hopopt,
- (opt ? opt->hopopt : NULL),
- newopt, newtype, &p);
- ipv6_renew_option(IPV6_RTHDRDSTOPTS, &opt2->dst0opt,
- (opt ? opt->dst0opt : NULL),
- newopt, newtype, &p);
- ipv6_renew_option(IPV6_RTHDR,
- (struct ipv6_opt_hdr **)&opt2->srcrt,
- (opt ? (struct ipv6_opt_hdr *)opt->srcrt : NULL),
- newopt, newtype, &p);
- ipv6_renew_option(IPV6_DSTOPTS, &opt2->dst1opt,
- (opt ? opt->dst1opt : NULL),
- newopt, newtype, &p);
-
- opt2->opt_nflen = (opt2->hopopt ? ipv6_optlen(opt2->hopopt) : 0) +
- (opt2->dst0opt ? ipv6_optlen(opt2->dst0opt) : 0) +
- (opt2->srcrt ? ipv6_optlen(opt2->srcrt) : 0);
- opt2->opt_flen = (opt2->dst1opt ? ipv6_optlen(opt2->dst1opt) : 0);
-
- return opt2;
-}
-
-struct ipv6_txoptions *ipv6_fixup_options(struct ipv6_txoptions *opt_space,
- struct ipv6_txoptions *opt)
-{
- /*
- * ignore the dest before srcrt unless srcrt is being included.
- * --yoshfuji
- */
- if (opt && opt->dst0opt && !opt->srcrt) {
- if (opt_space != opt) {
- memcpy(opt_space, opt, sizeof(*opt_space));
- opt = opt_space;
- }
- opt->opt_nflen -= ipv6_optlen(opt->dst0opt);
- opt->dst0opt = NULL;
- }
-
- return opt;
-}
-EXPORT_SYMBOL_GPL(ipv6_fixup_options);
-
/**
* fl6_update_dst - update flowi destination address with info given
* by srcrt option, if any.
diff --git a/net/ipv6/exthdrs_common.c b/net/ipv6/exthdrs_common.c
new file mode 100644
index 0000000..2c68184
--- /dev/null
+++ b/net/ipv6/exthdrs_common.c
@@ -0,0 +1,144 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/* Extension header and TLV library code that is not specific to IPv6. */
+#include <linux/export.h>
+#include <net/ipv6.h>
+
+struct ipv6_txoptions *
+ipeh_dup_options(struct sock *sk, struct ipv6_txoptions *opt)
+{
+ struct ipv6_txoptions *opt2;
+
+ opt2 = sock_kmalloc(sk, opt->tot_len, GFP_ATOMIC);
+ if (opt2) {
+ long dif = (char *)opt2 - (char *)opt;
+
+ memcpy(opt2, opt, opt->tot_len);
+ if (opt2->hopopt)
+ *((char **)&opt2->hopopt) += dif;
+ if (opt2->dst0opt)
+ *((char **)&opt2->dst0opt) += dif;
+ if (opt2->dst1opt)
+ *((char **)&opt2->dst1opt) += dif;
+ if (opt2->srcrt)
+ *((char **)&opt2->srcrt) += dif;
+ refcount_set(&opt2->refcnt, 1);
+ }
+ return opt2;
+}
+EXPORT_SYMBOL_GPL(ipeh_dup_options);
+
+static void ipeh_renew_option(int renewtype,
+ struct ipv6_opt_hdr **dest,
+ struct ipv6_opt_hdr *old,
+ struct ipv6_opt_hdr *new,
+ int newtype, char **p)
+{
+ struct ipv6_opt_hdr *src;
+
+ src = (renewtype == newtype ? new : old);
+ if (!src)
+ return;
+
+ memcpy(*p, src, ipv6_optlen(src));
+ *dest = (struct ipv6_opt_hdr *)*p;
+ *p += CMSG_ALIGN(ipv6_optlen(*dest));
+}
+
+/**
+ * ipeh_renew_options - replace a specific ext hdr with a new one.
+ *
+ * @sk: sock from which to allocate memory
+ * @opt: original options
+ * @newtype: option type to replace in @opt
+ * @newopt: new option of type @newtype to replace (user-mem)
+ * @newoptlen: length of @newopt
+ *
+ * Returns a new set of options which is a copy of @opt with the
+ * option type @newtype replaced with @newopt.
+ *
+ * @opt may be NULL, in which case a new set of options is returned
+ * containing just @newopt.
+ *
+ * @newopt may be NULL, in which case the specified option type is
+ * not copied into the new set of options.
+ *
+ * The new set of options is allocated from the socket option memory
+ * buffer of @sk.
+ */
+struct ipv6_txoptions *
+ipeh_renew_options(struct sock *sk, struct ipv6_txoptions *opt,
+ int newtype, struct ipv6_opt_hdr *newopt)
+{
+ int tot_len = 0;
+ char *p;
+ struct ipv6_txoptions *opt2;
+
+ if (opt) {
+ if (newtype != IPV6_HOPOPTS && opt->hopopt)
+ tot_len += CMSG_ALIGN(ipv6_optlen(opt->hopopt));
+ if (newtype != IPV6_RTHDRDSTOPTS && opt->dst0opt)
+ tot_len += CMSG_ALIGN(ipv6_optlen(opt->dst0opt));
+ if (newtype != IPV6_RTHDR && opt->srcrt)
+ tot_len += CMSG_ALIGN(ipv6_optlen(opt->srcrt));
+ if (newtype != IPV6_DSTOPTS && opt->dst1opt)
+ tot_len += CMSG_ALIGN(ipv6_optlen(opt->dst1opt));
+ }
+
+ if (newopt)
+ tot_len += CMSG_ALIGN(ipv6_optlen(newopt));
+
+ if (!tot_len)
+ return NULL;
+
+ tot_len += sizeof(*opt2);
+ opt2 = sock_kmalloc(sk, tot_len, GFP_ATOMIC);
+ if (!opt2)
+ return ERR_PTR(-ENOBUFS);
+
+ memset(opt2, 0, tot_len);
+ refcount_set(&opt2->refcnt, 1);
+ opt2->tot_len = tot_len;
+ p = (char *)(opt2 + 1);
+
+ ipeh_renew_option(IPV6_HOPOPTS, &opt2->hopopt,
+ (opt ? opt->hopopt : NULL),
+ newopt, newtype, &p);
+ ipeh_renew_option(IPV6_RTHDRDSTOPTS, &opt2->dst0opt,
+ (opt ? opt->dst0opt : NULL),
+ newopt, newtype, &p);
+ ipeh_renew_option(IPV6_RTHDR,
+ (struct ipv6_opt_hdr **)&opt2->srcrt,
+ (opt ? (struct ipv6_opt_hdr *)opt->srcrt : NULL),
+ newopt, newtype, &p);
+ ipeh_renew_option(IPV6_DSTOPTS, &opt2->dst1opt,
+ (opt ? opt->dst1opt : NULL),
+ newopt, newtype, &p);
+
+ opt2->opt_nflen = (opt2->hopopt ? ipv6_optlen(opt2->hopopt) : 0) +
+ (opt2->dst0opt ? ipv6_optlen(opt2->dst0opt) : 0) +
+ (opt2->srcrt ? ipv6_optlen(opt2->srcrt) : 0);
+ opt2->opt_flen = (opt2->dst1opt ? ipv6_optlen(opt2->dst1opt) : 0);
+
+ return opt2;
+}
+EXPORT_SYMBOL(ipeh_renew_options);
+
+struct ipv6_txoptions *ipeh_fixup_options(struct ipv6_txoptions *opt_space,
+ struct ipv6_txoptions *opt)
+{
+ /* ignore the dest before srcrt unless srcrt is being included.
+ * --yoshfuji
+ */
+ if (opt && opt->dst0opt && !opt->srcrt) {
+ if (opt_space != opt) {
+ memcpy(opt_space, opt, sizeof(*opt_space));
+ opt = opt_space;
+ }
+ opt->opt_nflen -= ipv6_optlen(opt->dst0opt);
+ opt->dst0opt = NULL;
+ }
+
+ return opt;
+}
+EXPORT_SYMBOL_GPL(ipeh_fixup_options);
diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c
index 264c292..8755ecc 100644
--- a/net/ipv6/ipv6_sockglue.c
+++ b/net/ipv6/ipv6_sockglue.c
@@ -427,7 +427,7 @@ static int do_ipv6_setsockopt(struct sock *sk, int level, int optname,
opt = rcu_dereference_protected(np->opt,
lockdep_sock_is_held(sk));
- opt = ipv6_renew_options(sk, opt, optname, new);
+ opt = ipeh_renew_options(sk, opt, optname, new);
kfree(new);
if (IS_ERR(opt)) {
retv = PTR_ERR(opt);
diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c
index 8a61319..5834ca4 100644
--- a/net/ipv6/raw.c
+++ b/net/ipv6/raw.c
@@ -888,7 +888,7 @@ static int rawv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
}
if (flowlabel)
opt = fl6_merge_options(&opt_space, flowlabel, opt);
- opt = ipv6_fixup_options(&opt_space, opt);
+ opt = ipeh_fixup_options(&opt_space, opt);
fl6.flowi6_proto = proto;
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 87f44d3..6bbf24f 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1243,7 +1243,7 @@ static struct sock *tcp_v6_syn_recv_sock(const struct sock *sk, struct sk_buff *
if (!opt)
opt = rcu_dereference(np->opt);
if (opt) {
- opt = ipv6_dup_options(newsk, opt);
+ opt = ipeh_dup_options(newsk, opt);
RCU_INIT_POINTER(newnp->opt, opt);
}
inet_csk(newsk)->icsk_ext_hdr_len = 0;
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index 827fe73..02ab20e 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -1384,7 +1384,7 @@ int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
}
if (flowlabel)
opt = fl6_merge_options(&opt_space, flowlabel, opt);
- opt = ipv6_fixup_options(&opt_space, opt);
+ opt = ipeh_fixup_options(&opt_space, opt);
ipc6.opt = opt;
fl6.flowi6_proto = sk->sk_protocol;
diff --git a/net/l2tp/l2tp_ip6.c b/net/l2tp/l2tp_ip6.c
index 687e23a..bc4925c 100644
--- a/net/l2tp/l2tp_ip6.c
+++ b/net/l2tp/l2tp_ip6.c
@@ -590,7 +590,7 @@ static int l2tp_ip6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
}
if (flowlabel)
opt = fl6_merge_options(&opt_space, flowlabel, opt);
- opt = ipv6_fixup_options(&opt_space, opt);
+ opt = ipeh_fixup_options(&opt_space, opt);
ipc6.opt = opt;
fl6.flowi6_proto = sk->sk_protocol;
diff --git a/net/sctp/ipv6.c b/net/sctp/ipv6.c
index e5f2fc7..b5b1349 100644
--- a/net/sctp/ipv6.c
+++ b/net/sctp/ipv6.c
@@ -441,7 +441,7 @@ static void sctp_v6_copy_ip_options(struct sock *sk, struct sock *newsk)
rcu_read_lock();
opt = rcu_dereference(np->opt);
if (opt) {
- opt = ipv6_dup_options(newsk, opt);
+ opt = ipeh_dup_options(newsk, opt);
if (!opt)
pr_err("%s: Failed to copy ip options\n", __func__);
}
--
2.7.4
^ permalink raw reply related
* [PATCH v4 net-next 1/7] ipeh: Create exthdrs_options.c and ipeh.h
From: Tom Herbert @ 2019-08-23 19:13 UTC (permalink / raw)
To: davem, netdev; +Cc: Tom Herbert, Tom Herbert
In-Reply-To: <1566587643-16594-1-git-send-email-tom@herbertland.com>
From: Tom Herbert <tom@quantonium.net>
Create exthdrs_options.c to hold code related to specific Hop-by-Hop
and Destination extension header options. Move related functions in
exthdrs.c to the new file.
Create include net/ipeh.h to contain common definitions for IP extension
headers.
Signed-off-by: Tom Herbert <tom@herbertland.com>
---
include/net/ipeh.h | 22 +++++
include/net/ipv6.h | 1 +
net/ipv6/Makefile | 2 +-
net/ipv6/exthdrs.c | 204 ---------------------------------------------
net/ipv6/exthdrs_options.c | 201 ++++++++++++++++++++++++++++++++++++++++++++
5 files changed, 225 insertions(+), 205 deletions(-)
create mode 100644 include/net/ipeh.h
create mode 100644 net/ipv6/exthdrs_options.c
diff --git a/include/net/ipeh.h b/include/net/ipeh.h
new file mode 100644
index 0000000..ec2d186
--- /dev/null
+++ b/include/net/ipeh.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _NET_IPEH_H
+#define _NET_IPEH_H
+
+#include <linux/skbuff.h>
+
+/*
+ * Parsing tlv encoded headers.
+ *
+ * Parsing function "func" returns true, if parsing succeed
+ * and false, if it failed.
+ * It MUST NOT touch skb->h.
+ */
+struct tlvtype_proc {
+ int type;
+ bool (*func)(struct sk_buff *skb, int offset);
+};
+
+extern const struct tlvtype_proc tlvprocdestopt_lst[];
+extern const struct tlvtype_proc tlvprochopopt_lst[];
+
+#endif /* _NET_IPEH_H */
diff --git a/include/net/ipv6.h b/include/net/ipv6.h
index 8dfc656..ec10fca 100644
--- a/include/net/ipv6.h
+++ b/include/net/ipv6.h
@@ -20,6 +20,7 @@
#include <net/flow_dissector.h>
#include <net/snmp.h>
#include <net/netns/hash.h>
+#include <net/ipeh.h>
#define SIN6_LEN_RFC2133 24
diff --git a/net/ipv6/Makefile b/net/ipv6/Makefile
index 8ccf355..df3919b 100644
--- a/net/ipv6/Makefile
+++ b/net/ipv6/Makefile
@@ -10,7 +10,7 @@ ipv6-objs := af_inet6.o anycast.o ip6_output.o ip6_input.o addrconf.o \
route.o ip6_fib.o ipv6_sockglue.o ndisc.o udp.o udplite.o \
raw.o icmp.o mcast.o reassembly.o tcp_ipv6.o ping.o \
exthdrs.o datagram.o ip6_flowlabel.o inet6_connection_sock.o \
- udp_offload.o seg6.o fib6_notifier.o
+ udp_offload.o seg6.o fib6_notifier.o exthdrs_options.o
ipv6-offload := ip6_offload.o tcpv6_offload.o exthdrs_offload.o
diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c
index ab5add0..664491e 100644
--- a/net/ipv6/exthdrs.c
+++ b/net/ipv6/exthdrs.c
@@ -39,7 +39,6 @@
#include <net/ndisc.h>
#include <net/ip6_route.h>
#include <net/addrconf.h>
-#include <net/calipso.h>
#if IS_ENABLED(CONFIG_IPV6_MIP6)
#include <net/xfrm.h>
#endif
@@ -51,19 +50,6 @@
#include <linux/uaccess.h>
-/*
- * Parsing tlv encoded headers.
- *
- * Parsing function "func" returns true, if parsing succeed
- * and false, if it failed.
- * It MUST NOT touch skb->h.
- */
-
-struct tlvtype_proc {
- int type;
- bool (*func)(struct sk_buff *skb, int offset);
-};
-
/*********************
Generic functions
*********************/
@@ -200,80 +186,6 @@ static bool ip6_parse_tlv(const struct tlvtype_proc *procs,
return false;
}
-/*****************************
- Destination options header.
- *****************************/
-
-#if IS_ENABLED(CONFIG_IPV6_MIP6)
-static bool ipv6_dest_hao(struct sk_buff *skb, int optoff)
-{
- struct ipv6_destopt_hao *hao;
- struct inet6_skb_parm *opt = IP6CB(skb);
- struct ipv6hdr *ipv6h = ipv6_hdr(skb);
- int ret;
-
- if (opt->dsthao) {
- net_dbg_ratelimited("hao duplicated\n");
- goto discard;
- }
- opt->dsthao = opt->dst1;
- opt->dst1 = 0;
-
- hao = (struct ipv6_destopt_hao *)(skb_network_header(skb) + optoff);
-
- if (hao->length != 16) {
- net_dbg_ratelimited("hao invalid option length = %d\n",
- hao->length);
- goto discard;
- }
-
- if (!(ipv6_addr_type(&hao->addr) & IPV6_ADDR_UNICAST)) {
- net_dbg_ratelimited("hao is not an unicast addr: %pI6\n",
- &hao->addr);
- goto discard;
- }
-
- ret = xfrm6_input_addr(skb, (xfrm_address_t *)&ipv6h->daddr,
- (xfrm_address_t *)&hao->addr, IPPROTO_DSTOPTS);
- if (unlikely(ret < 0))
- goto discard;
-
- if (skb_cloned(skb)) {
- if (pskb_expand_head(skb, 0, 0, GFP_ATOMIC))
- goto discard;
-
- /* update all variable using below by copied skbuff */
- hao = (struct ipv6_destopt_hao *)(skb_network_header(skb) +
- optoff);
- ipv6h = ipv6_hdr(skb);
- }
-
- if (skb->ip_summed == CHECKSUM_COMPLETE)
- skb->ip_summed = CHECKSUM_NONE;
-
- swap(ipv6h->saddr, hao->addr);
-
- if (skb->tstamp == 0)
- __net_timestamp(skb);
-
- return true;
-
- discard:
- kfree_skb(skb);
- return false;
-}
-#endif
-
-static const struct tlvtype_proc tlvprocdestopt_lst[] = {
-#if IS_ENABLED(CONFIG_IPV6_MIP6)
- {
- .type = IPV6_TLV_HAO,
- .func = ipv6_dest_hao,
- },
-#endif
- {-1, NULL}
-};
-
static int ipv6_destopt_rcv(struct sk_buff *skb)
{
struct inet6_dev *idev = __in6_dev_get(skb->dev);
@@ -702,122 +614,6 @@ void ipv6_exthdrs_exit(void)
inet6_del_protocol(&rthdr_protocol, IPPROTO_ROUTING);
}
-/**********************************
- Hop-by-hop options.
- **********************************/
-
-/*
- * Note: we cannot rely on skb_dst(skb) before we assign it in ip6_route_input().
- */
-static inline struct inet6_dev *ipv6_skb_idev(struct sk_buff *skb)
-{
- return skb_dst(skb) ? ip6_dst_idev(skb_dst(skb)) : __in6_dev_get(skb->dev);
-}
-
-static inline struct net *ipv6_skb_net(struct sk_buff *skb)
-{
- return skb_dst(skb) ? dev_net(skb_dst(skb)->dev) : dev_net(skb->dev);
-}
-
-/* Router Alert as of RFC 2711 */
-
-static bool ipv6_hop_ra(struct sk_buff *skb, int optoff)
-{
- const unsigned char *nh = skb_network_header(skb);
-
- if (nh[optoff + 1] == 2) {
- IP6CB(skb)->flags |= IP6SKB_ROUTERALERT;
- memcpy(&IP6CB(skb)->ra, nh + optoff + 2, sizeof(IP6CB(skb)->ra));
- return true;
- }
- net_dbg_ratelimited("ipv6_hop_ra: wrong RA length %d\n",
- nh[optoff + 1]);
- kfree_skb(skb);
- return false;
-}
-
-/* Jumbo payload */
-
-static bool ipv6_hop_jumbo(struct sk_buff *skb, int optoff)
-{
- const unsigned char *nh = skb_network_header(skb);
- struct inet6_dev *idev = __in6_dev_get_safely(skb->dev);
- struct net *net = ipv6_skb_net(skb);
- u32 pkt_len;
-
- if (nh[optoff + 1] != 4 || (optoff & 3) != 2) {
- net_dbg_ratelimited("ipv6_hop_jumbo: wrong jumbo opt length/alignment %d\n",
- nh[optoff+1]);
- __IP6_INC_STATS(net, idev, IPSTATS_MIB_INHDRERRORS);
- goto drop;
- }
-
- pkt_len = ntohl(*(__be32 *)(nh + optoff + 2));
- if (pkt_len <= IPV6_MAXPLEN) {
- __IP6_INC_STATS(net, idev, IPSTATS_MIB_INHDRERRORS);
- icmpv6_param_prob(skb, ICMPV6_HDR_FIELD, optoff+2);
- return false;
- }
- if (ipv6_hdr(skb)->payload_len) {
- __IP6_INC_STATS(net, idev, IPSTATS_MIB_INHDRERRORS);
- icmpv6_param_prob(skb, ICMPV6_HDR_FIELD, optoff);
- return false;
- }
-
- if (pkt_len > skb->len - sizeof(struct ipv6hdr)) {
- __IP6_INC_STATS(net, idev, IPSTATS_MIB_INTRUNCATEDPKTS);
- goto drop;
- }
-
- if (pskb_trim_rcsum(skb, pkt_len + sizeof(struct ipv6hdr)))
- goto drop;
-
- IP6CB(skb)->flags |= IP6SKB_JUMBOGRAM;
- return true;
-
-drop:
- kfree_skb(skb);
- return false;
-}
-
-/* CALIPSO RFC 5570 */
-
-static bool ipv6_hop_calipso(struct sk_buff *skb, int optoff)
-{
- const unsigned char *nh = skb_network_header(skb);
-
- if (nh[optoff + 1] < 8)
- goto drop;
-
- if (nh[optoff + 6] * 4 + 8 > nh[optoff + 1])
- goto drop;
-
- if (!calipso_validate(skb, nh + optoff))
- goto drop;
-
- return true;
-
-drop:
- kfree_skb(skb);
- return false;
-}
-
-static const struct tlvtype_proc tlvprochopopt_lst[] = {
- {
- .type = IPV6_TLV_ROUTERALERT,
- .func = ipv6_hop_ra,
- },
- {
- .type = IPV6_TLV_JUMBO,
- .func = ipv6_hop_jumbo,
- },
- {
- .type = IPV6_TLV_CALIPSO,
- .func = ipv6_hop_calipso,
- },
- { -1, }
-};
-
int ipv6_parse_hopopts(struct sk_buff *skb)
{
struct inet6_skb_parm *opt = IP6CB(skb);
diff --git a/net/ipv6/exthdrs_options.c b/net/ipv6/exthdrs_options.c
new file mode 100644
index 0000000..032e072
--- /dev/null
+++ b/net/ipv6/exthdrs_options.c
@@ -0,0 +1,201 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/errno.h>
+#include <linux/in6.h>
+#include <linux/net.h>
+#include <linux/netdevice.h>
+#include <linux/socket.h>
+#include <linux/types.h>
+#include <net/calipso.h>
+#include <net/ipv6.h>
+#include <net/ip6_route.h>
+#if IS_ENABLED(CONFIG_IPV6_MIP6)
+#include <net/xfrm.h>
+#endif
+
+/* Destination options header */
+
+#if IS_ENABLED(CONFIG_IPV6_MIP6)
+static bool ipv6_dest_hao(struct sk_buff *skb, int optoff)
+{
+ struct ipv6_destopt_hao *hao;
+ struct inet6_skb_parm *opt = IP6CB(skb);
+ struct ipv6hdr *ipv6h = ipv6_hdr(skb);
+ int ret;
+
+ if (opt->dsthao) {
+ net_dbg_ratelimited("hao duplicated\n");
+ goto discard;
+ }
+ opt->dsthao = opt->dst1;
+ opt->dst1 = 0;
+
+ hao = (struct ipv6_destopt_hao *)(skb_network_header(skb) + optoff);
+
+ if (hao->length != 16) {
+ net_dbg_ratelimited("hao invalid option length = %d\n",
+ hao->length);
+ goto discard;
+ }
+
+ if (!(ipv6_addr_type(&hao->addr) & IPV6_ADDR_UNICAST)) {
+ net_dbg_ratelimited("hao is not an unicast addr: %pI6\n",
+ &hao->addr);
+ goto discard;
+ }
+
+ ret = xfrm6_input_addr(skb, (xfrm_address_t *)&ipv6h->daddr,
+ (xfrm_address_t *)&hao->addr, IPPROTO_DSTOPTS);
+ if (unlikely(ret < 0))
+ goto discard;
+
+ if (skb_cloned(skb)) {
+ if (pskb_expand_head(skb, 0, 0, GFP_ATOMIC))
+ goto discard;
+
+ /* update all variable using below by copied skbuff */
+ hao = (struct ipv6_destopt_hao *)(skb_network_header(skb) +
+ optoff);
+ ipv6h = ipv6_hdr(skb);
+ }
+
+ if (skb->ip_summed == CHECKSUM_COMPLETE)
+ skb->ip_summed = CHECKSUM_NONE;
+
+ swap(ipv6h->saddr, hao->addr);
+
+ if (skb->tstamp == 0)
+ __net_timestamp(skb);
+
+ return true;
+
+ discard:
+ kfree_skb(skb);
+ return false;
+}
+#endif
+
+const struct tlvtype_proc tlvprocdestopt_lst[] = {
+#if IS_ENABLED(CONFIG_IPV6_MIP6)
+ {
+ .type = IPV6_TLV_HAO,
+ .func = ipv6_dest_hao,
+ },
+#endif
+ {-1, NULL}
+};
+
+/* Hop-by-hop options */
+
+/* Note: we cannot rely on skb_dst(skb) before we assign it in
+ * ip6_route_input().
+ */
+static inline struct inet6_dev *ipv6_skb_idev(struct sk_buff *skb)
+{
+ return skb_dst(skb) ? ip6_dst_idev(skb_dst(skb)) :
+ __in6_dev_get(skb->dev);
+}
+
+static inline struct net *ipv6_skb_net(struct sk_buff *skb)
+{
+ return skb_dst(skb) ? dev_net(skb_dst(skb)->dev) : dev_net(skb->dev);
+}
+
+/* Router Alert as of RFC 2711 */
+
+static bool ipv6_hop_ra(struct sk_buff *skb, int optoff)
+{
+ const unsigned char *nh = skb_network_header(skb);
+
+ if (nh[optoff + 1] == 2) {
+ IP6CB(skb)->flags |= IP6SKB_ROUTERALERT;
+ memcpy(&IP6CB(skb)->ra, nh + optoff + 2,
+ sizeof(IP6CB(skb)->ra));
+ return true;
+ }
+ net_dbg_ratelimited("%s: wrong RA length %d\n",
+ __func__, nh[optoff + 1]);
+ kfree_skb(skb);
+ return false;
+}
+
+/* Jumbo payload */
+
+static bool ipv6_hop_jumbo(struct sk_buff *skb, int optoff)
+{
+ const unsigned char *nh = skb_network_header(skb);
+ struct inet6_dev *idev = __in6_dev_get_safely(skb->dev);
+ struct net *net = ipv6_skb_net(skb);
+ u32 pkt_len;
+
+ if (nh[optoff + 1] != 4 || (optoff & 3) != 2) {
+ net_dbg_ratelimited("%s: wrong jumbo opt length/alignment %d\n",
+ __func__, nh[optoff + 1]);
+ __IP6_INC_STATS(net, idev, IPSTATS_MIB_INHDRERRORS);
+ goto drop;
+ }
+
+ pkt_len = ntohl(*(__be32 *)(nh + optoff + 2));
+ if (pkt_len <= IPV6_MAXPLEN) {
+ __IP6_INC_STATS(net, idev, IPSTATS_MIB_INHDRERRORS);
+ icmpv6_param_prob(skb, ICMPV6_HDR_FIELD, optoff + 2);
+ return false;
+ }
+ if (ipv6_hdr(skb)->payload_len) {
+ __IP6_INC_STATS(net, idev, IPSTATS_MIB_INHDRERRORS);
+ icmpv6_param_prob(skb, ICMPV6_HDR_FIELD, optoff);
+ return false;
+ }
+
+ if (pkt_len > skb->len - sizeof(struct ipv6hdr)) {
+ __IP6_INC_STATS(net, idev, IPSTATS_MIB_INTRUNCATEDPKTS);
+ goto drop;
+ }
+
+ if (pskb_trim_rcsum(skb, pkt_len + sizeof(struct ipv6hdr)))
+ goto drop;
+
+ IP6CB(skb)->flags |= IP6SKB_JUMBOGRAM;
+ return true;
+
+drop:
+ kfree_skb(skb);
+ return false;
+}
+
+/* CALIPSO RFC 5570 */
+
+static bool ipv6_hop_calipso(struct sk_buff *skb, int optoff)
+{
+ const unsigned char *nh = skb_network_header(skb);
+
+ if (nh[optoff + 1] < 8)
+ goto drop;
+
+ if (nh[optoff + 6] * 4 + 8 > nh[optoff + 1])
+ goto drop;
+
+ if (!calipso_validate(skb, nh + optoff))
+ goto drop;
+
+ return true;
+
+drop:
+ kfree_skb(skb);
+ return false;
+}
+
+const struct tlvtype_proc tlvprochopopt_lst[] = {
+ {
+ .type = IPV6_TLV_ROUTERALERT,
+ .func = ipv6_hop_ra,
+ },
+ {
+ .type = IPV6_TLV_JUMBO,
+ .func = ipv6_hop_jumbo,
+ },
+ {
+ .type = IPV6_TLV_CALIPSO,
+ .func = ipv6_hop_calipso,
+ },
+ { -1, }
+};
--
2.7.4
^ permalink raw reply related
* [PATCH v4 net-next 0/7] ipv6: Extension header infrastructure
From: Tom Herbert @ 2019-08-23 19:13 UTC (permalink / raw)
To: davem, netdev; +Cc: Tom Herbert
This patchset improves the IPv6 extension header infrastructure
to make extension headers more usable and scalable.
- Reorganize extension header files to separate out common
API components
- Create common TLV handler that will can be used in other use
cases (e.g. segment routing TLVs, UDP options)
- Allow registration of TLV handlers
- Elaborate on the TLV tables to include more characteristics
- Add a netlink interface to set TLV parameters (such as
alignment requirements, authorization to send, etc.)
- Enhance validation of TLVs being sent. Validation is strict
(unless overridden by admin) following that sending clause
of the robustness principle
- Allow non-privileged users to set Hop-by-Hop and Destination
Options if authorized by the admin
v2:
- Fix build errors from missing include file.
v3:
- Fix kbuild issue for ipv6_opt_hdr declared inside parameter list
in ipeh.h
v4:
- Resubmit
Tom Herbert (7):
ipeh: Create exthdrs_options.c and ipeh.h
ipeh: Move generic EH functions to exthdrs_common.c
ipeh: Generic TLV parser
ip6tlvs: Registration of TLV handlers and parameters
ip6tlvs: Add TX parameters
ip6tlvs: Add netlink interface
ip6tlvs: Validation of TX Destination and Hop-by-Hop options
include/net/ipeh.h | 208 ++++++++
include/net/ipv6.h | 12 +-
include/uapi/linux/in6.h | 6 +
include/uapi/linux/ipeh.h | 53 ++
net/dccp/ipv6.c | 2 +-
net/ipv6/Kconfig | 4 +
net/ipv6/Makefile | 3 +-
net/ipv6/calipso.c | 6 +-
net/ipv6/datagram.c | 51 +-
net/ipv6/exthdrs.c | 505 ++-----------------
net/ipv6/exthdrs_common.c | 1158 ++++++++++++++++++++++++++++++++++++++++++++
net/ipv6/exthdrs_options.c | 342 +++++++++++++
net/ipv6/ipv6_sockglue.c | 39 +-
net/ipv6/raw.c | 2 +-
net/ipv6/tcp_ipv6.c | 2 +-
net/ipv6/udp.c | 2 +-
net/l2tp/l2tp_ip6.c | 2 +-
net/sctp/ipv6.c | 2 +-
18 files changed, 1881 insertions(+), 518 deletions(-)
create mode 100644 include/net/ipeh.h
create mode 100644 include/uapi/linux/ipeh.h
create mode 100644 net/ipv6/exthdrs_common.c
create mode 100644 net/ipv6/exthdrs_options.c
--
2.7.4
^ permalink raw reply
* [PATCH] net: intel: Cleanup e1000 - add space between }}
From: Forrest Fleming @ 2019-08-23 19:14 UTC (permalink / raw)
To: Jeff Kirsher
Cc: Forrest Fleming, David S. Miller, intel-wired-lan, netdev,
linux-kernel
suggested by checkpatch
Signed-off-by: Forrest Fleming <ffleming@gmail.com>
---
.../net/ethernet/intel/e1000/e1000_param.c | 28 +++++++++----------
1 file changed, 14 insertions(+), 14 deletions(-)
diff --git a/drivers/net/ethernet/intel/e1000/e1000_param.c b/drivers/net/ethernet/intel/e1000/e1000_param.c
index d3f29ffe1e47..1a1f2f0237f9 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_param.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_param.c
@@ -266,7 +266,7 @@ void e1000_check_options(struct e1000_adapter *adapter)
.arg = { .r = {
.min = E1000_MIN_TXD,
.max = mac_type < e1000_82544 ? E1000_MAX_TXD : E1000_MAX_82544_TXD
- }}
+ } }
};
if (num_TxDescriptors > bd) {
@@ -295,7 +295,7 @@ void e1000_check_options(struct e1000_adapter *adapter)
.min = E1000_MIN_RXD,
.max = mac_type < e1000_82544 ? E1000_MAX_RXD :
E1000_MAX_82544_RXD
- }}
+ } }
};
if (num_RxDescriptors > bd) {
@@ -341,7 +341,7 @@ void e1000_check_options(struct e1000_adapter *adapter)
.err = "reading default settings from EEPROM",
.def = E1000_FC_DEFAULT,
.arg = { .l = { .nr = ARRAY_SIZE(fc_list),
- .p = fc_list }}
+ .p = fc_list } }
};
if (num_FlowControl > bd) {
@@ -359,7 +359,7 @@ void e1000_check_options(struct e1000_adapter *adapter)
.err = "using default of " __MODULE_STRING(DEFAULT_TIDV),
.def = DEFAULT_TIDV,
.arg = { .r = { .min = MIN_TXDELAY,
- .max = MAX_TXDELAY }}
+ .max = MAX_TXDELAY } }
};
if (num_TxIntDelay > bd) {
@@ -377,7 +377,7 @@ void e1000_check_options(struct e1000_adapter *adapter)
.err = "using default of " __MODULE_STRING(DEFAULT_TADV),
.def = DEFAULT_TADV,
.arg = { .r = { .min = MIN_TXABSDELAY,
- .max = MAX_TXABSDELAY }}
+ .max = MAX_TXABSDELAY } }
};
if (num_TxAbsIntDelay > bd) {
@@ -395,7 +395,7 @@ void e1000_check_options(struct e1000_adapter *adapter)
.err = "using default of " __MODULE_STRING(DEFAULT_RDTR),
.def = DEFAULT_RDTR,
.arg = { .r = { .min = MIN_RXDELAY,
- .max = MAX_RXDELAY }}
+ .max = MAX_RXDELAY } }
};
if (num_RxIntDelay > bd) {
@@ -413,7 +413,7 @@ void e1000_check_options(struct e1000_adapter *adapter)
.err = "using default of " __MODULE_STRING(DEFAULT_RADV),
.def = DEFAULT_RADV,
.arg = { .r = { .min = MIN_RXABSDELAY,
- .max = MAX_RXABSDELAY }}
+ .max = MAX_RXABSDELAY } }
};
if (num_RxAbsIntDelay > bd) {
@@ -431,7 +431,7 @@ void e1000_check_options(struct e1000_adapter *adapter)
.err = "using default of " __MODULE_STRING(DEFAULT_ITR),
.def = DEFAULT_ITR,
.arg = { .r = { .min = MIN_ITR,
- .max = MAX_ITR }}
+ .max = MAX_ITR } }
};
if (num_InterruptThrottleRate > bd) {
@@ -545,7 +545,7 @@ static void e1000_check_copper_options(struct e1000_adapter *adapter)
{ 0, "" },
{ SPEED_10, "" },
{ SPEED_100, "" },
- { SPEED_1000, "" }};
+ { SPEED_1000, "" } };
opt = (struct e1000_option) {
.type = list_option,
@@ -553,7 +553,7 @@ static void e1000_check_copper_options(struct e1000_adapter *adapter)
.err = "parameter ignored",
.def = 0,
.arg = { .l = { .nr = ARRAY_SIZE(speed_list),
- .p = speed_list }}
+ .p = speed_list } }
};
if (num_Speed > bd) {
@@ -567,7 +567,7 @@ static void e1000_check_copper_options(struct e1000_adapter *adapter)
static const struct e1000_opt_list dplx_list[] = {
{ 0, "" },
{ HALF_DUPLEX, "" },
- { FULL_DUPLEX, "" }};
+ { FULL_DUPLEX, "" } };
opt = (struct e1000_option) {
.type = list_option,
@@ -575,7 +575,7 @@ static void e1000_check_copper_options(struct e1000_adapter *adapter)
.err = "parameter ignored",
.def = 0,
.arg = { .l = { .nr = ARRAY_SIZE(dplx_list),
- .p = dplx_list }}
+ .p = dplx_list } }
};
if (num_Duplex > bd) {
@@ -623,7 +623,7 @@ static void e1000_check_copper_options(struct e1000_adapter *adapter)
{ 0x2c, AA "1000/FD, 100/FD, 100/HD" },
{ 0x2d, AA "1000/FD, 100/FD, 100/HD, 10/HD" },
{ 0x2e, AA "1000/FD, 100/FD, 100/HD, 10/FD" },
- { 0x2f, AA "1000/FD, 100/FD, 100/HD, 10/FD, 10/HD" }};
+ { 0x2f, AA "1000/FD, 100/FD, 100/HD, 10/FD, 10/HD" } };
opt = (struct e1000_option) {
.type = list_option,
@@ -631,7 +631,7 @@ static void e1000_check_copper_options(struct e1000_adapter *adapter)
.err = "parameter ignored",
.def = AUTONEG_ADV_DEFAULT,
.arg = { .l = { .nr = ARRAY_SIZE(an_list),
- .p = an_list }}
+ .p = an_list } }
};
if (num_AutoNeg > bd) {
--
2.17.1
^ permalink raw reply related
* RE: Is bug 200755 in anyone's queue??
From: Steve Zabele @ 2019-08-23 19:02 UTC (permalink / raw)
To: netdev
Cc: shum, vladimir116, saifi.khan, saifi.khan, daniel, on2k16nm,
'Stephen Hemminger'
In-Reply-To: <20190716070246.0745ee6f@hermes.lan>
Hi folks,
Is there a way to find out where the SO_REUSEPORT bug reported a year ago in
August (and apparently has been a bug with kernels later than 4.4) is being
addressed?
The bug characteristics, simple standalone test code demonstrating the bug,
and an assessment of the likely location/cause of the bug within the kernel
are all described here
https://bugzilla.kernel.org/show_bug.cgi?id=200755
I'm really hoping this gets fixed so we can move forward on updating our
kernels/Ubuntu release from our aging 4.4/16.04 release
Thanks!
Steve
-----Original Message-----
From: Stephen Hemminger [mailto:stephen@networkplumber.org]
Sent: Tuesday, July 16, 2019 10:03 AM
To: Steve Zabele
Cc: shum@canndrew.org; vladimir116@gmail.com; saifi.khan@DataSynergy.org;
saifi.khan@strikr.in; daniel@iogearbox.net; on2k16nm@gmail.com
Subject: Re: Is bug 200755 in anyone's queue??
On Tue, 16 Jul 2019 09:43:24 -0400
"Steve Zabele" <zabele@comcast.net> wrote:
> I came across bug report 200755 trying to figure out why some code I had
> provided to customers a while ago no longer works with the current Linux
> kernel. See
>
> https://bugzilla.kernel.org/show_bug.cgi?id=200755
>
> I've verified that, as reported, 'connect' no longer works for UDP.
> Moreover, it appears it has been broken since the 4.5 kernel has been
> released.
>
>
>
> It does also appear that the intended new feature of doing round robin
> assignments to different UDP sockets opened with SO_REUSEPORT also does
not
> work as described.
>
>
>
> Since the original bug report was made nearly a year ago for the 4.14
kernel
> (and the bug is also still present in the 4.15 kernel) I'm curious if
anyone
> is on the hook to get this fixed any time soon.
>
>
>
> I'd rather not have to do my own demultiplexing using a single socket in
> user space to work around what is clearly a (maybe not so recently
> introduced) kernel bug if at all possible. My code had worked just fine on
> 3.X kernels, and appears to work okay up through 4.4.
>
Kernel developers do not use bugzilla, I forward bug reports
to netdev@vger.kernel.org (after filtering).
^ permalink raw reply
* RE: Is bug 200755 in anyone's queue??
From: Steve Zabele @ 2019-08-23 19:04 UTC (permalink / raw)
To: 'Steve Zabele', netdev
Cc: shum, vladimir116, saifi.khan, daniel, on2k16nm,
'Stephen Hemminger'
In-Reply-To: <20190716070246.0745ee6f@hermes.lan>
Hi folks,
Is there a way to find out where the SO_REUSEPORT bug reported a year ago in
August (and apparently has been a bug with kernels later than 4.4) is being
addressed?
The bug characteristics, simple standalone test code demonstrating the bug,
and an assessment of the likely location/cause of the bug within the kernel
are all described here
https://bugzilla.kernel.org/show_bug.cgi?id=200755
I'm really hoping this gets fixed so we can move forward on updating our
kernels/Ubuntu release from our aging 4.4/16.04 release
Thanks!
Steve
-----Original Message-----
From: Stephen Hemminger [mailto:stephen@networkplumber.org]
Sent: Tuesday, July 16, 2019 10:03 AM
To: Steve Zabele
Cc: shum@canndrew.org; vladimir116@gmail.com; saifi.khan@DataSynergy.org;
saifi.khan@strikr.in; daniel@iogearbox.net; on2k16nm@gmail.com
Subject: Re: Is bug 200755 in anyone's queue??
On Tue, 16 Jul 2019 09:43:24 -0400
"Steve Zabele" <zabele@comcast.net> wrote:
> I came across bug report 200755 trying to figure out why some code I had
> provided to customers a while ago no longer works with the current Linux
> kernel. See
>
> https://bugzilla.kernel.org/show_bug.cgi?id=200755
>
> I've verified that, as reported, 'connect' no longer works for UDP.
> Moreover, it appears it has been broken since the 4.5 kernel has been
> released.
>
>
>
> It does also appear that the intended new feature of doing round robin
> assignments to different UDP sockets opened with SO_REUSEPORT also does
not
> work as described.
>
>
>
> Since the original bug report was made nearly a year ago for the 4.14
kernel
> (and the bug is also still present in the 4.15 kernel) I'm curious if
anyone
> is on the hook to get this fixed any time soon.
>
>
>
> I'd rather not have to do my own demultiplexing using a single socket in
> user space to work around what is clearly a (maybe not so recently
> introduced) kernel bug if at all possible. My code had worked just fine on
> 3.X kernels, and appears to work okay up through 4.4.
>
Kernel developers do not use bugzilla, I forward bug reports
to netdev@vger.kernel.org (after filtering).
^ permalink raw reply
* Re: New skb extension for use by LSMs (skb "security blob")?
From: Casey Schaufler @ 2019-08-23 18:56 UTC (permalink / raw)
To: David Miller; +Cc: fw, paul, netdev, linux-security-module, selinux, casey
In-Reply-To: <20190822.153642.10800077338364583.davem@davemloft.net>
On 8/22/2019 3:36 PM, David Miller wrote:
> From: Casey Schaufler <casey@schaufler-ca.com>
> Date: Thu, 22 Aug 2019 15:34:44 -0700
>
>> On 8/22/2019 3:28 PM, David Miller wrote:
>>> From: Casey Schaufler <casey@schaufler-ca.com>
>>> Date: Thu, 22 Aug 2019 14:59:37 -0700
>>>
>>>> Sure, you *can* do that, but it would be insane to do so.
>>> We look up the neighbour table entries on every single packet we
>>> transmit from the kernel in the same exact way.
>>>
>>> And it was exactly to get rid of a pointer in a data structure.
>> I very much expect that the lifecycle management issues would
>> be completely different, but I'll admit to having little understanding
>> of the details of the neighbour table.
> Neighbour table entries can live anywhere from essentially forever down
> to several microseconds.
>
> If your hash is good, and you use RCU locking on the read side, it's a
> single pointer dereference in cost.
The secmark is the data used by the netfilter system.
While it would be (Turing compatible, after all) possible,
we're talking multiple attributes with different lifecycles
being managed in a table (list, whatever) that may expand
explosively. Using a single ID to reference into a table that
could contain:
secmark from iptables for SELinux
secmark from iptables for AppArmor
SELinux secid/context for the packet
AppArmor secid/context for the packet
will be hairy. In the netfilter processing we may have to
allocate a new table entry. There's no way to identify that
the entry is no longer necessary, as there is no lifecycle
on a secmark. Is it possible to come up with something that
will limp along? Possibly. If there's a blob pointer, we know
how to do all this effectively.
^ permalink raw reply
* [PATCH net-next v2 06/10] net: sched: conditionally obtain rtnl lock in cls hw offloads API
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev; +Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
In order to remove dependency on rtnl lock from offloads code of
classifiers, take rtnl lock conditionally before executing driver
callbacks. Only obtain rtnl lock if block is bound to devices that require
it.
Block bind/unbind code is rtnl-locked and obtains block->cb_lock while
holding rtnl lock. Obtain locks in same order in tc_setup_cb_*() functions
to prevent deadlock.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
---
Changes from V1 to V2:
- Speculatively read block->lockeddevcnt in tc_setup_cb_*() to obtain
rtnl mutex without retry when block is bound to locked device.
net/sched/cls_api.c | 65 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 65 insertions(+)
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 87954f5370a4..f2dcecf34c6f 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -3076,11 +3076,28 @@ __tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
int tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
void *type_data, bool err_stop, bool rtnl_held)
{
+ bool take_rtnl = READ_ONCE(block->lockeddevcnt);
int ok_count;
+retry:
+ if (take_rtnl)
+ rtnl_lock();
down_read(&block->cb_lock);
+ /* Need to obtain rtnl lock if block is bound to devs that require it.
+ * In block bind code cb_lock is obtained while holding rtnl, so we must
+ * obtain the locks in same order here.
+ */
+ if (!rtnl_held && !take_rtnl && block->lockeddevcnt) {
+ up_read(&block->cb_lock);
+ take_rtnl = true;
+ goto retry;
+ }
+
ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
+
up_read(&block->cb_lock);
+ if (take_rtnl)
+ rtnl_unlock();
return ok_count;
}
EXPORT_SYMBOL(tc_setup_cb_call);
@@ -3095,9 +3112,23 @@ int tc_setup_cb_add(struct tcf_block *block, struct tcf_proto *tp,
enum tc_setup_type type, void *type_data, bool err_stop,
u32 *flags, unsigned int *in_hw_count, bool rtnl_held)
{
+ bool take_rtnl = READ_ONCE(block->lockeddevcnt);
int ok_count;
+retry:
+ if (take_rtnl)
+ rtnl_lock();
down_read(&block->cb_lock);
+ /* Need to obtain rtnl lock if block is bound to devs that require it.
+ * In block bind code cb_lock is obtained while holding rtnl, so we must
+ * obtain the locks in same order here.
+ */
+ if (!rtnl_held && !take_rtnl && block->lockeddevcnt) {
+ up_read(&block->cb_lock);
+ take_rtnl = true;
+ goto retry;
+ }
+
/* Make sure all netdevs sharing this block are offload-capable. */
if (block->nooffloaddevcnt && err_stop) {
ok_count = -EOPNOTSUPP;
@@ -3115,6 +3146,8 @@ int tc_setup_cb_add(struct tcf_block *block, struct tcf_proto *tp,
ok_count, true);
err_unlock:
up_read(&block->cb_lock);
+ if (take_rtnl)
+ rtnl_unlock();
return ok_count < 0 ? ok_count : 0;
}
EXPORT_SYMBOL(tc_setup_cb_add);
@@ -3131,9 +3164,23 @@ int tc_setup_cb_replace(struct tcf_block *block, struct tcf_proto *tp,
u32 *new_flags, unsigned int *new_in_hw_count,
bool rtnl_held)
{
+ bool take_rtnl = READ_ONCE(block->lockeddevcnt);
int ok_count;
+retry:
+ if (take_rtnl)
+ rtnl_lock();
down_read(&block->cb_lock);
+ /* Need to obtain rtnl lock if block is bound to devs that require it.
+ * In block bind code cb_lock is obtained while holding rtnl, so we must
+ * obtain the locks in same order here.
+ */
+ if (!rtnl_held && !take_rtnl && block->lockeddevcnt) {
+ up_read(&block->cb_lock);
+ take_rtnl = true;
+ goto retry;
+ }
+
/* Make sure all netdevs sharing this block are offload-capable. */
if (block->nooffloaddevcnt && err_stop) {
ok_count = -EOPNOTSUPP;
@@ -3155,6 +3202,8 @@ int tc_setup_cb_replace(struct tcf_block *block, struct tcf_proto *tp,
new_flags, ok_count, true);
err_unlock:
up_read(&block->cb_lock);
+ if (take_rtnl)
+ rtnl_unlock();
return ok_count < 0 ? ok_count : 0;
}
EXPORT_SYMBOL(tc_setup_cb_replace);
@@ -3167,9 +3216,23 @@ int tc_setup_cb_destroy(struct tcf_block *block, struct tcf_proto *tp,
enum tc_setup_type type, void *type_data, bool err_stop,
u32 *flags, unsigned int *in_hw_count, bool rtnl_held)
{
+ bool take_rtnl = READ_ONCE(block->lockeddevcnt);
int ok_count;
+retry:
+ if (take_rtnl)
+ rtnl_lock();
down_read(&block->cb_lock);
+ /* Need to obtain rtnl lock if block is bound to devs that require it.
+ * In block bind code cb_lock is obtained while holding rtnl, so we must
+ * obtain the locks in same order here.
+ */
+ if (!rtnl_held && !take_rtnl && block->lockeddevcnt) {
+ up_read(&block->cb_lock);
+ take_rtnl = true;
+ goto retry;
+ }
+
ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
tc_cls_offload_cnt_reset(block, tp, in_hw_count, flags);
@@ -3177,6 +3240,8 @@ int tc_setup_cb_destroy(struct tcf_block *block, struct tcf_proto *tp,
tp->ops->hw_del(tp, type_data);
up_read(&block->cb_lock);
+ if (take_rtnl)
+ rtnl_unlock();
return ok_count < 0 ? ok_count : 0;
}
EXPORT_SYMBOL(tc_setup_cb_destroy);
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 08/10] net: sched: take reference to action dev before calling offloads
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev
Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov, Jiri Pirko
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
In order to remove dependency on rtnl lock when calling hardware offload
API, take reference to action mirred dev when initializing flow_action
structure in tc_setup_flow_action(). Implement function
tc_cleanup_flow_action(), use it to release the device after hardware
offload API is done using it.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
include/net/pkt_cls.h | 2 ++
net/sched/cls_api.c | 32 ++++++++++++++++++++++++++++++++
net/sched/cls_flower.c | 2 ++
3 files changed, 36 insertions(+)
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index a48824bc1489..e553fc80eb23 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -505,6 +505,8 @@ tcf_match_indev(struct sk_buff *skb, int ifindex)
int tc_setup_flow_action(struct flow_action *flow_action,
const struct tcf_exts *exts, bool rtnl_held);
+void tc_cleanup_flow_action(struct flow_action *flow_action);
+
int tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
void *type_data, bool err_stop, bool rtnl_held);
int tc_setup_cb_add(struct tcf_block *block, struct tcf_proto *tp,
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index cb835d581b77..622146aafb06 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -3260,6 +3260,27 @@ int tc_setup_cb_reoffload(struct tcf_block *block, struct tcf_proto *tp,
}
EXPORT_SYMBOL(tc_setup_cb_reoffload);
+void tc_cleanup_flow_action(struct flow_action *flow_action)
+{
+ struct flow_action_entry *entry;
+ int i;
+
+ flow_action_for_each(i, entry, flow_action) {
+ switch (entry->id) {
+ case FLOW_ACTION_REDIRECT:
+ case FLOW_ACTION_MIRRED:
+ case FLOW_ACTION_REDIRECT_INGRESS:
+ case FLOW_ACTION_MIRRED_INGRESS:
+ if (entry->dev)
+ dev_put(entry->dev);
+ break;
+ default:
+ break;
+ }
+ }
+}
+EXPORT_SYMBOL(tc_cleanup_flow_action);
+
int tc_setup_flow_action(struct flow_action *flow_action,
const struct tcf_exts *exts, bool rtnl_held)
{
@@ -3289,15 +3310,23 @@ int tc_setup_flow_action(struct flow_action *flow_action,
} else if (is_tcf_mirred_egress_redirect(act)) {
entry->id = FLOW_ACTION_REDIRECT;
entry->dev = tcf_mirred_dev(act);
+ if (entry->dev)
+ dev_hold(entry->dev);
} else if (is_tcf_mirred_egress_mirror(act)) {
entry->id = FLOW_ACTION_MIRRED;
entry->dev = tcf_mirred_dev(act);
+ if (entry->dev)
+ dev_hold(entry->dev);
} else if (is_tcf_mirred_ingress_redirect(act)) {
entry->id = FLOW_ACTION_REDIRECT_INGRESS;
entry->dev = tcf_mirred_dev(act);
+ if (entry->dev)
+ dev_hold(entry->dev);
} else if (is_tcf_mirred_ingress_mirror(act)) {
entry->id = FLOW_ACTION_MIRRED_INGRESS;
entry->dev = tcf_mirred_dev(act);
+ if (entry->dev)
+ dev_hold(entry->dev);
} else if (is_tcf_vlan(act)) {
switch (tcf_vlan_action(act)) {
case TCA_VLAN_ACT_PUSH:
@@ -3405,6 +3434,9 @@ int tc_setup_flow_action(struct flow_action *flow_action,
if (!rtnl_held)
rtnl_unlock();
+ if (err)
+ tc_cleanup_flow_action(flow_action);
+
return err;
}
EXPORT_SYMBOL(tc_setup_flow_action);
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index d47d4e84d4e5..df141a67c17f 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -465,6 +465,7 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
err = tc_setup_cb_add(block, tp, TC_SETUP_CLSFLOWER, &cls_flower,
skip_sw, &f->flags, &f->in_hw_count, true);
+ tc_cleanup_flow_action(&cls_flower.rule->action);
kfree(cls_flower.rule);
if (err) {
@@ -1838,6 +1839,7 @@ static int fl_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
TC_SETUP_CLSFLOWER, &cls_flower,
cb_priv, &f->flags,
&f->in_hw_count);
+ tc_cleanup_flow_action(&cls_flower.rule->action);
kfree(cls_flower.rule);
if (err && add && tc_skip_sw(f->flags)) {
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 05/10] net: sched: add API for registering unlocked offload block callbacks
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev
Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov, Jiri Pirko
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
Extend struct flow_block_offload with "unlocked_driver_cb" flag to allow
registering and unregistering block hardware offload callbacks that do not
require caller to hold rtnl lock. Extend tcf_block with additional
lockeddevcnt counter that is incremented for each non-unlocked driver
callback attached to device. This counter is necessary to conditionally
obtain rtnl lock before calling hardware callbacks in following patches.
Register mlx5 tc block offload callbacks as "unlocked".
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 ++
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 3 +++
include/net/flow_offload.h | 1 +
include/net/sch_generic.h | 1 +
net/sched/cls_api.c | 6 ++++++
5 files changed, 13 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index fa4bf2d4bcd4..8592b98d0e70 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -3470,10 +3470,12 @@ static int mlx5e_setup_tc(struct net_device *dev, enum tc_setup_type type,
void *type_data)
{
struct mlx5e_priv *priv = netdev_priv(dev);
+ struct flow_block_offload *f = type_data;
switch (type) {
#ifdef CONFIG_MLX5_ESWITCH
case TC_SETUP_BLOCK:
+ f->unlocked_driver_cb = true;
return flow_block_cb_setup_simple(type_data,
&mlx5e_block_cb_list,
mlx5e_setup_tc_block_cb,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 3c0d36b2b91c..e7ac6233037d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -763,6 +763,7 @@ mlx5e_rep_indr_setup_tc_block(struct net_device *netdev,
if (f->binder_type != FLOW_BLOCK_BINDER_TYPE_CLSACT_INGRESS)
return -EOPNOTSUPP;
+ f->unlocked_driver_cb = true;
f->driver_block_list = &mlx5e_block_cb_list;
switch (f->command) {
@@ -1245,9 +1246,11 @@ static int mlx5e_rep_setup_tc(struct net_device *dev, enum tc_setup_type type,
void *type_data)
{
struct mlx5e_priv *priv = netdev_priv(dev);
+ struct flow_block_offload *f = type_data;
switch (type) {
case TC_SETUP_BLOCK:
+ f->unlocked_driver_cb = true;
return flow_block_cb_setup_simple(type_data,
&mlx5e_rep_block_cb_list,
mlx5e_rep_setup_tc_cb,
diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index 757fa84de654..fc881875f856 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h
@@ -284,6 +284,7 @@ struct flow_block_offload {
enum flow_block_command command;
enum flow_block_binder_type binder_type;
bool block_shared;
+ bool unlocked_driver_cb;
struct net *net;
struct flow_block *block;
struct list_head cb_list;
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index c4fbbaff30a2..43f5b7ed02bd 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -408,6 +408,7 @@ struct tcf_block {
bool keep_dst;
atomic_t offloadcnt; /* Number of oddloaded filters */
unsigned int nooffloaddevcnt; /* Number of devs unable to do offload */
+ unsigned int lockeddevcnt; /* Number of devs that require rtnl lock. */
struct {
struct tcf_chain *chain;
struct list_head filter_chain_list;
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index dd16cf171f51..87954f5370a4 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -1418,6 +1418,8 @@ static int tcf_block_bind(struct tcf_block *block,
bo->extack);
if (err)
goto err_unroll;
+ if (!bo->unlocked_driver_cb)
+ block->lockeddevcnt++;
i++;
}
@@ -1433,6 +1435,8 @@ static int tcf_block_bind(struct tcf_block *block,
block_cb->cb_priv, false,
tcf_block_offload_in_use(block),
NULL);
+ if (!bo->unlocked_driver_cb)
+ block->lockeddevcnt--;
}
flow_block_cb_free(block_cb);
}
@@ -1454,6 +1458,8 @@ static void tcf_block_unbind(struct tcf_block *block,
NULL);
list_del(&block_cb->list);
flow_block_cb_free(block_cb);
+ if (!bo->unlocked_driver_cb)
+ block->lockeddevcnt--;
}
}
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 09/10] net: sched: copy tunnel info when setting flow_action entry->tunnel
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev
Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov, Jiri Pirko
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
In order to remove dependency on rtnl lock, modify tc_setup_flow_action()
to copy tunnel info, instead of just saving pointer to tunnel_key action
tunnel info. This is necessary to prevent concurrent action overwrite from
releasing tunnel info while it is being used by rtnl-unlocked driver.
Implement helper tcf_tunnel_info_copy() that is used to copy tunnel info
with all its options to dynamically allocated memory block. Modify
tc_cleanup_flow_action() to free dynamically allocated tunnel info.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
include/net/tc_act/tc_tunnel_key.h | 17 +++++++++++++++++
net/sched/cls_api.c | 9 ++++++++-
2 files changed, 25 insertions(+), 1 deletion(-)
diff --git a/include/net/tc_act/tc_tunnel_key.h b/include/net/tc_act/tc_tunnel_key.h
index 7c3f777c168c..0689d9bcdf84 100644
--- a/include/net/tc_act/tc_tunnel_key.h
+++ b/include/net/tc_act/tc_tunnel_key.h
@@ -59,4 +59,21 @@ static inline struct ip_tunnel_info *tcf_tunnel_info(const struct tc_action *a)
return NULL;
#endif
}
+
+static inline struct ip_tunnel_info *
+tcf_tunnel_info_copy(const struct tc_action *a)
+{
+#ifdef CONFIG_NET_CLS_ACT
+ struct ip_tunnel_info *tun = tcf_tunnel_info(a);
+
+ if (tun) {
+ size_t tun_size = sizeof(*tun) + tun->options_len;
+ struct ip_tunnel_info *tun_copy = kmemdup(tun, tun_size,
+ GFP_KERNEL);
+
+ return tun_copy;
+ }
+#endif
+ return NULL;
+}
#endif /* __NET_TC_TUNNEL_KEY_H */
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 622146aafb06..c00c836cd155 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -3274,6 +3274,9 @@ void tc_cleanup_flow_action(struct flow_action *flow_action)
if (entry->dev)
dev_put(entry->dev);
break;
+ case FLOW_ACTION_TUNNEL_ENCAP:
+ kfree(entry->tunnel);
+ break;
default:
break;
}
@@ -3350,7 +3353,11 @@ int tc_setup_flow_action(struct flow_action *flow_action,
}
} else if (is_tcf_tunnel_set(act)) {
entry->id = FLOW_ACTION_TUNNEL_ENCAP;
- entry->tunnel = tcf_tunnel_info(act);
+ entry->tunnel = tcf_tunnel_info_copy(act);
+ if (!entry->tunnel) {
+ err = -ENOMEM;
+ goto err_out;
+ }
} else if (is_tcf_tunnel_release(act)) {
entry->id = FLOW_ACTION_TUNNEL_DECAP;
} else if (is_tcf_pedit(act)) {
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 10/10] net: sched: flower: don't take rtnl lock for cls hw offloads API
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev
Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov, Jiri Pirko
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
Don't manually take rtnl lock in flower classifier before calling cls
hardware offloads API. Instead, pass rtnl lock status via 'rtnl_held'
parameter.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
net/sched/cls_flower.c | 53 +++++++++++++-----------------------------
1 file changed, 16 insertions(+), 37 deletions(-)
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index df141a67c17f..fa1de7080c65 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -412,18 +412,13 @@ static void fl_hw_destroy_filter(struct tcf_proto *tp, struct cls_fl_filter *f,
struct tcf_block *block = tp->chain->block;
struct flow_cls_offload cls_flower = {};
- if (!rtnl_held)
- rtnl_lock();
-
tc_cls_common_offload_init(&cls_flower.common, tp, f->flags, extack);
cls_flower.command = FLOW_CLS_DESTROY;
cls_flower.cookie = (unsigned long) f;
tc_setup_cb_destroy(block, tp, TC_SETUP_CLSFLOWER, &cls_flower, false,
- &f->flags, &f->in_hw_count, true);
+ &f->flags, &f->in_hw_count, rtnl_held);
- if (!rtnl_held)
- rtnl_unlock();
}
static int fl_hw_replace_filter(struct tcf_proto *tp,
@@ -435,14 +430,9 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
bool skip_sw = tc_skip_sw(f->flags);
int err = 0;
- if (!rtnl_held)
- rtnl_lock();
-
cls_flower.rule = flow_rule_alloc(tcf_exts_num_actions(&f->exts));
- if (!cls_flower.rule) {
- err = -ENOMEM;
- goto errout;
- }
+ if (!cls_flower.rule)
+ return -ENOMEM;
tc_cls_common_offload_init(&cls_flower.common, tp, f->flags, extack);
cls_flower.command = FLOW_CLS_REPLACE;
@@ -453,36 +443,30 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
cls_flower.classid = f->res.classid;
err = tc_setup_flow_action(&cls_flower.rule->action, &f->exts,
- true);
+ rtnl_held);
if (err) {
kfree(cls_flower.rule);
- if (skip_sw)
+ if (skip_sw) {
NL_SET_ERR_MSG_MOD(extack, "Failed to setup flow action");
- else
- err = 0;
- goto errout;
+ return err;
+ }
+ return 0;
}
err = tc_setup_cb_add(block, tp, TC_SETUP_CLSFLOWER, &cls_flower,
- skip_sw, &f->flags, &f->in_hw_count, true);
+ skip_sw, &f->flags, &f->in_hw_count, rtnl_held);
tc_cleanup_flow_action(&cls_flower.rule->action);
kfree(cls_flower.rule);
if (err) {
- fl_hw_destroy_filter(tp, f, true, NULL);
- goto errout;
- }
-
- if (skip_sw && !(f->flags & TCA_CLS_FLAGS_IN_HW)) {
- err = -EINVAL;
- goto errout;
+ fl_hw_destroy_filter(tp, f, rtnl_held, NULL);
+ return err;
}
-errout:
- if (!rtnl_held)
- rtnl_unlock();
+ if (skip_sw && !(f->flags & TCA_CLS_FLAGS_IN_HW))
+ return -EINVAL;
- return err;
+ return 0;
}
static void fl_hw_update_stats(struct tcf_proto *tp, struct cls_fl_filter *f,
@@ -491,22 +475,17 @@ static void fl_hw_update_stats(struct tcf_proto *tp, struct cls_fl_filter *f,
struct tcf_block *block = tp->chain->block;
struct flow_cls_offload cls_flower = {};
- if (!rtnl_held)
- rtnl_lock();
-
tc_cls_common_offload_init(&cls_flower.common, tp, f->flags, NULL);
cls_flower.command = FLOW_CLS_STATS;
cls_flower.cookie = (unsigned long) f;
cls_flower.classid = f->res.classid;
- tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false, true);
+ tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false,
+ rtnl_held);
tcf_exts_stats_update(&f->exts, cls_flower.stats.bytes,
cls_flower.stats.pkts,
cls_flower.stats.lastused);
-
- if (!rtnl_held)
- rtnl_unlock();
}
static void __fl_put(struct cls_fl_filter *f)
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 02/10] net: sched: change tcf block offload counter type to atomic_t
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev
Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov, Jiri Pirko
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
As a preparation for running proto ops functions without rtnl lock, change
offload counter type to atomic. This is necessary to allow updating the
counter by multiple concurrent users when offloading filters to hardware
from unlocked classifiers.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
include/net/sch_generic.h | 7 ++++---
net/sched/cls_api.c | 2 +-
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index a3eaf5f9d28f..d778c502decd 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -14,6 +14,7 @@
#include <linux/workqueue.h>
#include <linux/mutex.h>
#include <linux/rwsem.h>
+#include <linux/atomic.h>
#include <net/gen_stats.h>
#include <net/rtnetlink.h>
#include <net/flow_offload.h>
@@ -401,7 +402,7 @@ struct tcf_block {
struct flow_block flow_block;
struct list_head owner_list;
bool keep_dst;
- unsigned int offloadcnt; /* Number of oddloaded filters */
+ atomic_t offloadcnt; /* Number of oddloaded filters */
unsigned int nooffloaddevcnt; /* Number of devs unable to do offload */
struct {
struct tcf_chain *chain;
@@ -443,7 +444,7 @@ static inline void tcf_block_offload_inc(struct tcf_block *block, u32 *flags)
if (*flags & TCA_CLS_FLAGS_IN_HW)
return;
*flags |= TCA_CLS_FLAGS_IN_HW;
- block->offloadcnt++;
+ atomic_inc(&block->offloadcnt);
}
static inline void tcf_block_offload_dec(struct tcf_block *block, u32 *flags)
@@ -451,7 +452,7 @@ static inline void tcf_block_offload_dec(struct tcf_block *block, u32 *flags)
if (!(*flags & TCA_CLS_FLAGS_IN_HW))
return;
*flags &= ~TCA_CLS_FLAGS_IN_HW;
- block->offloadcnt--;
+ atomic_dec(&block->offloadcnt);
}
static inline void
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 959b7ca1ca02..f2c2f8159e35 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -629,7 +629,7 @@ static void tc_indr_block_call(struct tcf_block *block,
static bool tcf_block_offload_in_use(struct tcf_block *block)
{
- return block->offloadcnt;
+ return atomic_read(&block->offloadcnt);
}
static int tcf_block_offload_cmd(struct tcf_block *block,
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 04/10] net: sched: notify classifier on successful offload add/delete
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev; +Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
To remove dependency on rtnl lock, extend classifier ops with new
ops->hw_add() and ops->hw_del() callbacks. Call them from cls API while
holding cb_lock every time filter if successfully added to or deleted from
hardware.
Implement the new API in flower classifier. Use it to manage hw_filters
list under cb_lock protection, instead of relying on rtnl lock to
synchronize with concurrent fl_reoffload() call.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
---
Changes from V1 to V2:
- Keep success flow unindented in tc_setup_cb_{add|replace}().
include/net/sch_generic.h | 4 ++++
net/sched/cls_api.c | 19 +++++++++++++++++--
net/sched/cls_flower.c | 33 ++++++++++++++++++++++++++-------
3 files changed, 47 insertions(+), 9 deletions(-)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index f90e3b2a3065..c4fbbaff30a2 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -312,6 +312,10 @@ struct tcf_proto_ops {
int (*reoffload)(struct tcf_proto *tp, bool add,
flow_setup_cb_t *cb, void *cb_priv,
struct netlink_ext_ack *extack);
+ void (*hw_add)(struct tcf_proto *tp,
+ void *type_data);
+ void (*hw_del)(struct tcf_proto *tp,
+ void *type_data);
void (*bind_class)(void *, u32, unsigned long);
void * (*tmplt_create)(struct net *net,
struct tcf_chain *chain,
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 157bf3599ce1..dd16cf171f51 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -3099,6 +3099,11 @@ int tc_setup_cb_add(struct tcf_block *block, struct tcf_proto *tp,
}
ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
+ if (ok_count < 0)
+ goto err_unlock;
+
+ if (tp->ops->hw_add)
+ tp->ops->hw_add(tp, type_data);
if (ok_count > 0)
tc_cls_offload_cnt_update(block, tp, in_hw_count, flags,
ok_count, true);
@@ -3130,11 +3135,18 @@ int tc_setup_cb_replace(struct tcf_block *block, struct tcf_proto *tp,
}
tc_cls_offload_cnt_reset(block, tp, old_in_hw_count, old_flags);
+ if (tp->ops->hw_del)
+ tp->ops->hw_del(tp, type_data);
ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
+ if (ok_count < 0)
+ goto err_unlock;
+
+ if (tp->ops->hw_add)
+ tp->ops->hw_add(tp, type_data);
if (ok_count > 0)
- tc_cls_offload_cnt_update(block, tp, new_in_hw_count, new_flags,
- ok_count, true);
+ tc_cls_offload_cnt_update(block, tp, new_in_hw_count,
+ new_flags, ok_count, true);
err_unlock:
up_read(&block->cb_lock);
return ok_count < 0 ? ok_count : 0;
@@ -3155,6 +3167,9 @@ int tc_setup_cb_destroy(struct tcf_block *block, struct tcf_proto *tp,
ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
tc_cls_offload_cnt_reset(block, tp, in_hw_count, flags);
+ if (tp->ops->hw_del)
+ tp->ops->hw_del(tp, type_data);
+
up_read(&block->cb_lock);
return ok_count < 0 ? ok_count : 0;
}
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 0978ae4a1428..9a7fd6bcd0a5 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -421,9 +421,6 @@ static void fl_hw_destroy_filter(struct tcf_proto *tp, struct cls_fl_filter *f,
tc_setup_cb_destroy(block, tp, TC_SETUP_CLSFLOWER, &cls_flower, false,
&f->flags, &f->in_hw_count, true);
- spin_lock(&tp->lock);
- list_del_init(&f->hw_list);
- spin_unlock(&tp->lock);
if (!rtnl_held)
rtnl_unlock();
@@ -433,7 +430,6 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
struct cls_fl_filter *f, bool rtnl_held,
struct netlink_ext_ack *extack)
{
- struct cls_fl_head *head = fl_head_dereference(tp);
struct tcf_block *block = tp->chain->block;
struct flow_cls_offload cls_flower = {};
bool skip_sw = tc_skip_sw(f->flags);
@@ -480,9 +476,6 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
goto errout;
}
- spin_lock(&tp->lock);
- list_add(&f->hw_list, &head->hw_filters);
- spin_unlock(&tp->lock);
errout:
if (!rtnl_held)
rtnl_unlock();
@@ -1856,6 +1849,30 @@ static int fl_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
return 0;
}
+static void fl_hw_add(struct tcf_proto *tp, void *type_data)
+{
+ struct flow_cls_offload *cls_flower = type_data;
+ struct cls_fl_filter *f =
+ (struct cls_fl_filter *) cls_flower->cookie;
+ struct cls_fl_head *head = fl_head_dereference(tp);
+
+ spin_lock(&tp->lock);
+ list_add(&f->hw_list, &head->hw_filters);
+ spin_unlock(&tp->lock);
+}
+
+static void fl_hw_del(struct tcf_proto *tp, void *type_data)
+{
+ struct flow_cls_offload *cls_flower = type_data;
+ struct cls_fl_filter *f =
+ (struct cls_fl_filter *) cls_flower->cookie;
+
+ spin_lock(&tp->lock);
+ if (!list_empty(&f->hw_list))
+ list_del_init(&f->hw_list);
+ spin_unlock(&tp->lock);
+}
+
static int fl_hw_create_tmplt(struct tcf_chain *chain,
struct fl_flow_tmplt *tmplt)
{
@@ -2516,6 +2533,8 @@ static struct tcf_proto_ops cls_fl_ops __read_mostly = {
.delete = fl_delete,
.walk = fl_walk,
.reoffload = fl_reoffload,
+ .hw_add = fl_hw_add,
+ .hw_del = fl_hw_del,
.dump = fl_dump,
.bind_class = fl_bind_class,
.tmplt_create = fl_tmplt_create,
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 03/10] net: sched: refactor block offloads counter usage
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev; +Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
Without rtnl lock protection filters can no longer safely manage block
offloads counter themselves. Refactor cls API to protect block offloadcnt
with tcf_block->cb_lock that is already used to protect driver callback
list and nooffloaddevcnt counter. The counter can be modified by concurrent
tasks by new functions that execute block callbacks (which is safe with
previous patch that changed its type to atomic_t), however, block
bind/unbind code that checks the counter value takes cb_lock in write mode
to exclude any concurrent modifications. This approach prevents race
conditions between bind/unbind and callback execution code but allows for
concurrency for tc rule update path.
Move block offload counter, filter in hardware counter and filter flags
management from classifiers into cls hardware offloads API. Make functions
tcf_block_offload_{inc|dec}() and tc_cls_offload_cnt_update() to be cls API
private. Implement following new cls API to be used instead:
tc_setup_cb_add() - non-destructive filter add. If filter that wasn't
already in hardware is successfully offloaded, increment block offloads
counter, set filter in hardware counter and flag. On failure, previously
offloaded filter is considered to be intact and offloads counter is not
decremented.
tc_setup_cb_replace() - destructive filter replace. Release existing
filter block offload counter and reset its in hardware counter and flag.
Set new filter in hardware counter and flag. On failure, previously
offloaded filter is considered to be destroyed and offload counter is
decremented.
tc_setup_cb_destroy() - filter destroy. Unconditionally decrement block
offloads counter.
tc_setup_cb_reoffload() - reoffload filter to single cb. Execute cb() and
call tc_cls_offload_cnt_update() if cb() didn't return an error.
Refactor all offload-capable classifiers to atomically offload filters to
hardware, change block offload counter, and set filter in hardware counter
and flag by means of the new cls API functions.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
---
Changes from V1 to V2:
- Remove redundant brackets around cnt.
- Rename 'errout' label to 'err_unlock'.
- Revert changing oldprog to cls_bpf.oldprog in conditional.
- Add new helper tc_setup_cb_reoffload().
- Make tc_cls_offload_cnt_update() static and remove its export.
- Change tc_setup_cb_*() helpers to only return zero or error code.
include/net/pkt_cls.h | 17 +++-
include/net/sch_generic.h | 31 -------
net/sched/cls_api.c | 171 ++++++++++++++++++++++++++++++++++----
net/sched/cls_bpf.c | 38 ++++-----
net/sched/cls_flower.c | 40 ++++-----
net/sched/cls_matchall.c | 27 +++---
net/sched/cls_u32.c | 29 +++----
7 files changed, 229 insertions(+), 124 deletions(-)
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index 64999ffcb486..612232492f67 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -506,7 +506,22 @@ tcf_match_indev(struct sk_buff *skb, int ifindex)
int tc_setup_flow_action(struct flow_action *flow_action,
const struct tcf_exts *exts);
int tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
- void *type_data, bool err_stop);
+ void *type_data, bool err_stop, bool rtnl_held);
+int tc_setup_cb_add(struct tcf_block *block, struct tcf_proto *tp,
+ enum tc_setup_type type, void *type_data, bool err_stop,
+ u32 *flags, unsigned int *in_hw_count, bool rtnl_held);
+int tc_setup_cb_replace(struct tcf_block *block, struct tcf_proto *tp,
+ enum tc_setup_type type, void *type_data, bool err_stop,
+ u32 *old_flags, unsigned int *old_in_hw_count,
+ u32 *new_flags, unsigned int *new_in_hw_count,
+ bool rtnl_held);
+int tc_setup_cb_destroy(struct tcf_block *block, struct tcf_proto *tp,
+ enum tc_setup_type type, void *type_data, bool err_stop,
+ u32 *flags, unsigned int *in_hw_count, bool rtnl_held);
+int tc_setup_cb_reoffload(struct tcf_block *block, struct tcf_proto *tp,
+ bool add, flow_setup_cb_t *cb,
+ enum tc_setup_type type, void *type_data,
+ void *cb_priv, u32 *flags, unsigned int *in_hw_count);
unsigned int tcf_exts_num_actions(struct tcf_exts *exts);
struct tc_cls_u32_knode {
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index d778c502decd..f90e3b2a3065 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -439,37 +439,6 @@ static inline bool lockdep_tcf_proto_is_locked(struct tcf_proto *tp)
#define tcf_proto_dereference(p, tp) \
rcu_dereference_protected(p, lockdep_tcf_proto_is_locked(tp))
-static inline void tcf_block_offload_inc(struct tcf_block *block, u32 *flags)
-{
- if (*flags & TCA_CLS_FLAGS_IN_HW)
- return;
- *flags |= TCA_CLS_FLAGS_IN_HW;
- atomic_inc(&block->offloadcnt);
-}
-
-static inline void tcf_block_offload_dec(struct tcf_block *block, u32 *flags)
-{
- if (!(*flags & TCA_CLS_FLAGS_IN_HW))
- return;
- *flags &= ~TCA_CLS_FLAGS_IN_HW;
- atomic_dec(&block->offloadcnt);
-}
-
-static inline void
-tc_cls_offload_cnt_update(struct tcf_block *block, u32 *cnt,
- u32 *flags, bool add)
-{
- if (add) {
- if (!*cnt)
- tcf_block_offload_inc(block, flags);
- (*cnt)++;
- } else {
- (*cnt)--;
- if (!*cnt)
- tcf_block_offload_dec(block, flags);
- }
-}
-
static inline void qdisc_cb_private_validate(const struct sk_buff *skb, int sz)
{
struct qdisc_skb_cb *qcb;
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index f2c2f8159e35..157bf3599ce1 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -3000,37 +3000,180 @@ int tcf_exts_dump_stats(struct sk_buff *skb, struct tcf_exts *exts)
}
EXPORT_SYMBOL(tcf_exts_dump_stats);
-int tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
- void *type_data, bool err_stop)
+static void tcf_block_offload_inc(struct tcf_block *block, u32 *flags)
+{
+ if (*flags & TCA_CLS_FLAGS_IN_HW)
+ return;
+ *flags |= TCA_CLS_FLAGS_IN_HW;
+ atomic_inc(&block->offloadcnt);
+}
+
+static void tcf_block_offload_dec(struct tcf_block *block, u32 *flags)
+{
+ if (!(*flags & TCA_CLS_FLAGS_IN_HW))
+ return;
+ *flags &= ~TCA_CLS_FLAGS_IN_HW;
+ atomic_dec(&block->offloadcnt);
+}
+
+static void tc_cls_offload_cnt_update(struct tcf_block *block,
+ struct tcf_proto *tp, u32 *cnt,
+ u32 *flags, u32 diff, bool add)
+{
+ lockdep_assert_held(&block->cb_lock);
+
+ spin_lock(&tp->lock);
+ if (add) {
+ if (!*cnt)
+ tcf_block_offload_inc(block, flags);
+ *cnt += diff;
+ } else {
+ *cnt -= diff;
+ if (!*cnt)
+ tcf_block_offload_dec(block, flags);
+ }
+ spin_unlock(&tp->lock);
+}
+
+static void
+tc_cls_offload_cnt_reset(struct tcf_block *block, struct tcf_proto *tp,
+ u32 *cnt, u32 *flags)
+{
+ lockdep_assert_held(&block->cb_lock);
+
+ spin_lock(&tp->lock);
+ tcf_block_offload_dec(block, flags);
+ *cnt = 0;
+ spin_unlock(&tp->lock);
+}
+
+static int
+__tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
+ void *type_data, bool err_stop)
{
struct flow_block_cb *block_cb;
int ok_count = 0;
int err;
- down_read(&block->cb_lock);
- /* Make sure all netdevs sharing this block are offload-capable. */
- if (block->nooffloaddevcnt && err_stop) {
- ok_count = -EOPNOTSUPP;
- goto err_unlock;
- }
-
list_for_each_entry(block_cb, &block->flow_block.cb_list, list) {
err = block_cb->cb(type, type_data, block_cb->cb_priv);
if (err) {
- if (err_stop) {
- ok_count = err;
- goto err_unlock;
- }
+ if (err_stop)
+ return err;
} else {
ok_count++;
}
}
-err_unlock:
+ return ok_count;
+}
+
+int tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
+ void *type_data, bool err_stop, bool rtnl_held)
+{
+ int ok_count;
+
+ down_read(&block->cb_lock);
+ ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
up_read(&block->cb_lock);
return ok_count;
}
EXPORT_SYMBOL(tc_setup_cb_call);
+/* Non-destructive filter add. If filter that wasn't already in hardware is
+ * successfully offloaded, increment block offloads counter. On failure,
+ * previously offloaded filter is considered to be intact and offloads counter
+ * is not decremented.
+ */
+
+int tc_setup_cb_add(struct tcf_block *block, struct tcf_proto *tp,
+ enum tc_setup_type type, void *type_data, bool err_stop,
+ u32 *flags, unsigned int *in_hw_count, bool rtnl_held)
+{
+ int ok_count;
+
+ down_read(&block->cb_lock);
+ /* Make sure all netdevs sharing this block are offload-capable. */
+ if (block->nooffloaddevcnt && err_stop) {
+ ok_count = -EOPNOTSUPP;
+ goto err_unlock;
+ }
+
+ ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
+ if (ok_count > 0)
+ tc_cls_offload_cnt_update(block, tp, in_hw_count, flags,
+ ok_count, true);
+err_unlock:
+ up_read(&block->cb_lock);
+ return ok_count < 0 ? ok_count : 0;
+}
+EXPORT_SYMBOL(tc_setup_cb_add);
+
+/* Destructive filter replace. If filter that wasn't already in hardware is
+ * successfully offloaded, increment block offload counter. On failure,
+ * previously offloaded filter is considered to be destroyed and offload counter
+ * is decremented.
+ */
+
+int tc_setup_cb_replace(struct tcf_block *block, struct tcf_proto *tp,
+ enum tc_setup_type type, void *type_data, bool err_stop,
+ u32 *old_flags, unsigned int *old_in_hw_count,
+ u32 *new_flags, unsigned int *new_in_hw_count,
+ bool rtnl_held)
+{
+ int ok_count;
+
+ down_read(&block->cb_lock);
+ /* Make sure all netdevs sharing this block are offload-capable. */
+ if (block->nooffloaddevcnt && err_stop) {
+ ok_count = -EOPNOTSUPP;
+ goto err_unlock;
+ }
+
+ tc_cls_offload_cnt_reset(block, tp, old_in_hw_count, old_flags);
+
+ ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
+ if (ok_count > 0)
+ tc_cls_offload_cnt_update(block, tp, new_in_hw_count, new_flags,
+ ok_count, true);
+err_unlock:
+ up_read(&block->cb_lock);
+ return ok_count < 0 ? ok_count : 0;
+}
+EXPORT_SYMBOL(tc_setup_cb_replace);
+
+/* Destroy filter and decrement block offload counter, if filter was previously
+ * offloaded.
+ */
+
+int tc_setup_cb_destroy(struct tcf_block *block, struct tcf_proto *tp,
+ enum tc_setup_type type, void *type_data, bool err_stop,
+ u32 *flags, unsigned int *in_hw_count, bool rtnl_held)
+{
+ int ok_count;
+
+ down_read(&block->cb_lock);
+ ok_count = __tc_setup_cb_call(block, type, type_data, err_stop);
+
+ tc_cls_offload_cnt_reset(block, tp, in_hw_count, flags);
+ up_read(&block->cb_lock);
+ return ok_count < 0 ? ok_count : 0;
+}
+EXPORT_SYMBOL(tc_setup_cb_destroy);
+
+int tc_setup_cb_reoffload(struct tcf_block *block, struct tcf_proto *tp,
+ bool add, flow_setup_cb_t *cb,
+ enum tc_setup_type type, void *type_data,
+ void *cb_priv, u32 *flags, unsigned int *in_hw_count)
+{
+ int err = cb(type, type_data, cb_priv);
+
+ if (!err)
+ tc_cls_offload_cnt_update(block, tp, in_hw_count, flags, 1,
+ add);
+ return err;
+}
+EXPORT_SYMBOL(tc_setup_cb_reoffload);
+
int tc_setup_flow_action(struct flow_action *flow_action,
const struct tcf_exts *exts)
{
diff --git a/net/sched/cls_bpf.c b/net/sched/cls_bpf.c
index 3f7a9c02b70c..9eef65a1b683 100644
--- a/net/sched/cls_bpf.c
+++ b/net/sched/cls_bpf.c
@@ -163,17 +163,19 @@ static int cls_bpf_offload_cmd(struct tcf_proto *tp, struct cls_bpf_prog *prog,
cls_bpf.exts_integrated = obj->exts_integrated;
if (oldprog)
- tcf_block_offload_dec(block, &oldprog->gen_flags);
+ err = tc_setup_cb_replace(block, tp, TC_SETUP_CLSBPF, &cls_bpf,
+ skip_sw, &oldprog->gen_flags,
+ &oldprog->in_hw_count,
+ &prog->gen_flags, &prog->in_hw_count,
+ true);
+ else
+ err = tc_setup_cb_add(block, tp, TC_SETUP_CLSBPF, &cls_bpf,
+ skip_sw, &prog->gen_flags,
+ &prog->in_hw_count, true);
- err = tc_setup_cb_call(block, TC_SETUP_CLSBPF, &cls_bpf, skip_sw);
- if (prog) {
- if (err < 0) {
- cls_bpf_offload_cmd(tp, oldprog, prog, extack);
- return err;
- } else if (err > 0) {
- prog->in_hw_count = err;
- tcf_block_offload_inc(block, &prog->gen_flags);
- }
+ if (prog && err) {
+ cls_bpf_offload_cmd(tp, oldprog, prog, extack);
+ return err;
}
if (prog && skip_sw && !(prog->gen_flags & TCA_CLS_FLAGS_IN_HW))
@@ -230,7 +232,7 @@ static void cls_bpf_offload_update_stats(struct tcf_proto *tp,
cls_bpf.name = prog->bpf_name;
cls_bpf.exts_integrated = prog->exts_integrated;
- tc_setup_cb_call(block, TC_SETUP_CLSBPF, &cls_bpf, false);
+ tc_setup_cb_call(block, TC_SETUP_CLSBPF, &cls_bpf, false, true);
}
static int cls_bpf_init(struct tcf_proto *tp)
@@ -673,15 +675,11 @@ static int cls_bpf_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb
cls_bpf.name = prog->bpf_name;
cls_bpf.exts_integrated = prog->exts_integrated;
- err = cb(TC_SETUP_CLSBPF, &cls_bpf, cb_priv);
- if (err) {
- if (add && tc_skip_sw(prog->gen_flags))
- return err;
- continue;
- }
-
- tc_cls_offload_cnt_update(block, &prog->in_hw_count,
- &prog->gen_flags, add);
+ err = tc_setup_cb_reoffload(block, tp, add, cb, TC_SETUP_CLSBPF,
+ &cls_bpf, cb_priv, &prog->gen_flags,
+ &prog->in_hw_count);
+ if (err && add && tc_skip_sw(prog->gen_flags))
+ return err;
}
return 0;
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 054123742e32..0978ae4a1428 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -419,10 +419,10 @@ static void fl_hw_destroy_filter(struct tcf_proto *tp, struct cls_fl_filter *f,
cls_flower.command = FLOW_CLS_DESTROY;
cls_flower.cookie = (unsigned long) f;
- tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false);
+ tc_setup_cb_destroy(block, tp, TC_SETUP_CLSFLOWER, &cls_flower, false,
+ &f->flags, &f->in_hw_count, true);
spin_lock(&tp->lock);
list_del_init(&f->hw_list);
- tcf_block_offload_dec(block, &f->flags);
spin_unlock(&tp->lock);
if (!rtnl_held)
@@ -466,18 +466,13 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
goto errout;
}
- err = tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, skip_sw);
+ err = tc_setup_cb_add(block, tp, TC_SETUP_CLSFLOWER, &cls_flower,
+ skip_sw, &f->flags, &f->in_hw_count, true);
kfree(cls_flower.rule);
- if (err < 0) {
+ if (err) {
fl_hw_destroy_filter(tp, f, true, NULL);
goto errout;
- } else if (err > 0) {
- f->in_hw_count = err;
- err = 0;
- spin_lock(&tp->lock);
- tcf_block_offload_inc(block, &f->flags);
- spin_unlock(&tp->lock);
}
if (skip_sw && !(f->flags & TCA_CLS_FLAGS_IN_HW)) {
@@ -509,7 +504,7 @@ static void fl_hw_update_stats(struct tcf_proto *tp, struct cls_fl_filter *f,
cls_flower.cookie = (unsigned long) f;
cls_flower.classid = f->res.classid;
- tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false);
+ tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false, true);
tcf_exts_stats_update(&f->exts, cls_flower.stats.bytes,
cls_flower.stats.pkts,
@@ -1844,21 +1839,16 @@ static int fl_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
cls_flower.classid = f->res.classid;
- err = cb(TC_SETUP_CLSFLOWER, &cls_flower, cb_priv);
+ err = tc_setup_cb_reoffload(block, tp, add, cb,
+ TC_SETUP_CLSFLOWER, &cls_flower,
+ cb_priv, &f->flags,
+ &f->in_hw_count);
kfree(cls_flower.rule);
- if (err) {
- if (add && tc_skip_sw(f->flags)) {
- __fl_put(f);
- return err;
- }
- goto next_flow;
+ if (err && add && tc_skip_sw(f->flags)) {
+ __fl_put(f);
+ return err;
}
-
- spin_lock(&tp->lock);
- tc_cls_offload_cnt_update(block, &f->in_hw_count, &f->flags,
- add);
- spin_unlock(&tp->lock);
next_flow:
__fl_put(f);
}
@@ -1886,7 +1876,7 @@ static int fl_hw_create_tmplt(struct tcf_chain *chain,
/* We don't care if driver (any of them) fails to handle this
* call. It serves just as a hint for it.
*/
- tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false);
+ tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false, true);
kfree(cls_flower.rule);
return 0;
@@ -1902,7 +1892,7 @@ static void fl_hw_destroy_tmplt(struct tcf_chain *chain,
cls_flower.command = FLOW_CLS_TMPLT_DESTROY;
cls_flower.cookie = (unsigned long) tmplt;
- tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false);
+ tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false, true);
}
static void *fl_tmplt_create(struct net *net, struct tcf_chain *chain,
diff --git a/net/sched/cls_matchall.c b/net/sched/cls_matchall.c
index 455ea2793f9b..7801851b55d5 100644
--- a/net/sched/cls_matchall.c
+++ b/net/sched/cls_matchall.c
@@ -75,8 +75,8 @@ static void mall_destroy_hw_filter(struct tcf_proto *tp,
cls_mall.command = TC_CLSMATCHALL_DESTROY;
cls_mall.cookie = cookie;
- tc_setup_cb_call(block, TC_SETUP_CLSMATCHALL, &cls_mall, false);
- tcf_block_offload_dec(block, &head->flags);
+ tc_setup_cb_destroy(block, tp, TC_SETUP_CLSMATCHALL, &cls_mall, false,
+ &head->flags, &head->in_hw_count, true);
}
static int mall_replace_hw_filter(struct tcf_proto *tp,
@@ -109,15 +109,13 @@ static int mall_replace_hw_filter(struct tcf_proto *tp,
return err;
}
- err = tc_setup_cb_call(block, TC_SETUP_CLSMATCHALL, &cls_mall, skip_sw);
+ err = tc_setup_cb_add(block, tp, TC_SETUP_CLSMATCHALL, &cls_mall,
+ skip_sw, &head->flags, &head->in_hw_count, true);
kfree(cls_mall.rule);
- if (err < 0) {
+ if (err) {
mall_destroy_hw_filter(tp, head, cookie, NULL);
return err;
- } else if (err > 0) {
- head->in_hw_count = err;
- tcf_block_offload_inc(block, &head->flags);
}
if (skip_sw && !(head->flags & TCA_CLS_FLAGS_IN_HW))
@@ -312,16 +310,13 @@ static int mall_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
return 0;
}
- err = cb(TC_SETUP_CLSMATCHALL, &cls_mall, cb_priv);
+ err = tc_setup_cb_reoffload(block, tp, add, cb, TC_SETUP_CLSMATCHALL,
+ &cls_mall, cb_priv, &head->flags,
+ &head->in_hw_count);
kfree(cls_mall.rule);
- if (err) {
- if (add && tc_skip_sw(head->flags))
- return err;
- return 0;
- }
-
- tc_cls_offload_cnt_update(block, &head->in_hw_count, &head->flags, add);
+ if (err && add && tc_skip_sw(head->flags))
+ return err;
return 0;
}
@@ -337,7 +332,7 @@ static void mall_stats_hw_filter(struct tcf_proto *tp,
cls_mall.command = TC_CLSMATCHALL_STATS;
cls_mall.cookie = cookie;
- tc_setup_cb_call(block, TC_SETUP_CLSMATCHALL, &cls_mall, false);
+ tc_setup_cb_call(block, TC_SETUP_CLSMATCHALL, &cls_mall, false, true);
tcf_exts_stats_update(&head->exts, cls_mall.stats.bytes,
cls_mall.stats.pkts, cls_mall.stats.lastused);
diff --git a/net/sched/cls_u32.c b/net/sched/cls_u32.c
index 8614088edd1b..a1672b180b19 100644
--- a/net/sched/cls_u32.c
+++ b/net/sched/cls_u32.c
@@ -480,7 +480,7 @@ static void u32_clear_hw_hnode(struct tcf_proto *tp, struct tc_u_hnode *h,
cls_u32.hnode.handle = h->handle;
cls_u32.hnode.prio = h->prio;
- tc_setup_cb_call(block, TC_SETUP_CLSU32, &cls_u32, false);
+ tc_setup_cb_call(block, TC_SETUP_CLSU32, &cls_u32, false, true);
}
static int u32_replace_hw_hnode(struct tcf_proto *tp, struct tc_u_hnode *h,
@@ -498,7 +498,7 @@ static int u32_replace_hw_hnode(struct tcf_proto *tp, struct tc_u_hnode *h,
cls_u32.hnode.handle = h->handle;
cls_u32.hnode.prio = h->prio;
- err = tc_setup_cb_call(block, TC_SETUP_CLSU32, &cls_u32, skip_sw);
+ err = tc_setup_cb_call(block, TC_SETUP_CLSU32, &cls_u32, skip_sw, true);
if (err < 0) {
u32_clear_hw_hnode(tp, h, NULL);
return err;
@@ -522,8 +522,8 @@ static void u32_remove_hw_knode(struct tcf_proto *tp, struct tc_u_knode *n,
cls_u32.command = TC_CLSU32_DELETE_KNODE;
cls_u32.knode.handle = n->handle;
- tc_setup_cb_call(block, TC_SETUP_CLSU32, &cls_u32, false);
- tcf_block_offload_dec(block, &n->flags);
+ tc_setup_cb_destroy(block, tp, TC_SETUP_CLSU32, &cls_u32, false,
+ &n->flags, &n->in_hw_count, true);
}
static int u32_replace_hw_knode(struct tcf_proto *tp, struct tc_u_knode *n,
@@ -552,13 +552,11 @@ static int u32_replace_hw_knode(struct tcf_proto *tp, struct tc_u_knode *n,
if (n->ht_down)
cls_u32.knode.link_handle = ht->handle;
- err = tc_setup_cb_call(block, TC_SETUP_CLSU32, &cls_u32, skip_sw);
- if (err < 0) {
+ err = tc_setup_cb_add(block, tp, TC_SETUP_CLSU32, &cls_u32, skip_sw,
+ &n->flags, &n->in_hw_count, true);
+ if (err) {
u32_remove_hw_knode(tp, n, NULL);
return err;
- } else if (err > 0) {
- n->in_hw_count = err;
- tcf_block_offload_inc(block, &n->flags);
}
if (skip_sw && !(n->flags & TCA_CLS_FLAGS_IN_HW))
@@ -1201,14 +1199,11 @@ static int u32_reoffload_knode(struct tcf_proto *tp, struct tc_u_knode *n,
cls_u32.knode.link_handle = ht->handle;
}
- err = cb(TC_SETUP_CLSU32, &cls_u32, cb_priv);
- if (err) {
- if (add && tc_skip_sw(n->flags))
- return err;
- return 0;
- }
-
- tc_cls_offload_cnt_update(block, &n->in_hw_count, &n->flags, add);
+ err = tc_setup_cb_reoffload(block, tp, add, cb, TC_SETUP_CLSU32,
+ &cls_u32, cb_priv, &n->flags,
+ &n->in_hw_count);
+ if (err && add && tc_skip_sw(n->flags))
+ return err;
return 0;
}
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 01/10] net: sched: protect block offload-related fields with rw_semaphore
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev; +Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
In order to remove dependency on rtnl lock, extend tcf_block with 'cb_lock'
rwsem and use it to protect flow_block->cb_list and related counters from
concurrent modification. The lock is taken in read mode for read-only
traversal of cb_list in tc_setup_cb_call() and write mode in all other
cases. This approach ensures that:
- cb_list is not changed concurrently while filters is being offloaded on
block.
- block->nooffloaddevcnt is checked while holding the lock in read mode,
but is only changed by bind/unbind code when holding the cb_lock in write
mode to prevent concurrent modification.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
---
Changes from V1 to V2:
- Rename 'errout' label to 'err_unlock'.
include/net/sch_generic.h | 2 ++
net/sched/cls_api.c | 45 +++++++++++++++++++++++++++++++--------
2 files changed, 38 insertions(+), 9 deletions(-)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index d9f359af0b93..a3eaf5f9d28f 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -13,6 +13,7 @@
#include <linux/refcount.h>
#include <linux/workqueue.h>
#include <linux/mutex.h>
+#include <linux/rwsem.h>
#include <net/gen_stats.h>
#include <net/rtnetlink.h>
#include <net/flow_offload.h>
@@ -396,6 +397,7 @@ struct tcf_block {
refcount_t refcnt;
struct net *net;
struct Qdisc *q;
+ struct rw_semaphore cb_lock; /* protects cb_list and offload counters */
struct flow_block flow_block;
struct list_head owner_list;
bool keep_dst;
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index e0d8b456e9f5..959b7ca1ca02 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -568,9 +568,11 @@ static void tc_indr_block_ing_cmd(struct net_device *dev,
bo.block = &block->flow_block;
+ down_write(&block->cb_lock);
cb(dev, cb_priv, TC_SETUP_BLOCK, &bo);
tcf_block_setup(block, &bo);
+ up_write(&block->cb_lock);
}
static struct tcf_block *tc_dev_ingress_block(struct net_device *dev)
@@ -661,6 +663,7 @@ static int tcf_block_offload_bind(struct tcf_block *block, struct Qdisc *q,
struct net_device *dev = q->dev_queue->dev;
int err;
+ down_write(&block->cb_lock);
if (!dev->netdev_ops->ndo_setup_tc)
goto no_offload_dev_inc;
@@ -669,24 +672,31 @@ static int tcf_block_offload_bind(struct tcf_block *block, struct Qdisc *q,
*/
if (!tc_can_offload(dev) && tcf_block_offload_in_use(block)) {
NL_SET_ERR_MSG(extack, "Bind to offloaded block failed as dev has offload disabled");
- return -EOPNOTSUPP;
+ err = -EOPNOTSUPP;
+ goto err_unlock;
}
err = tcf_block_offload_cmd(block, dev, ei, FLOW_BLOCK_BIND, extack);
if (err == -EOPNOTSUPP)
goto no_offload_dev_inc;
if (err)
- return err;
+ goto err_unlock;
tc_indr_block_call(block, dev, ei, FLOW_BLOCK_BIND, extack);
+ up_write(&block->cb_lock);
return 0;
no_offload_dev_inc:
- if (tcf_block_offload_in_use(block))
- return -EOPNOTSUPP;
+ if (tcf_block_offload_in_use(block)) {
+ err = -EOPNOTSUPP;
+ goto err_unlock;
+ }
+ err = 0;
block->nooffloaddevcnt++;
tc_indr_block_call(block, dev, ei, FLOW_BLOCK_BIND, extack);
- return 0;
+err_unlock:
+ up_write(&block->cb_lock);
+ return err;
}
static void tcf_block_offload_unbind(struct tcf_block *block, struct Qdisc *q,
@@ -695,6 +705,7 @@ static void tcf_block_offload_unbind(struct tcf_block *block, struct Qdisc *q,
struct net_device *dev = q->dev_queue->dev;
int err;
+ down_write(&block->cb_lock);
tc_indr_block_call(block, dev, ei, FLOW_BLOCK_UNBIND, NULL);
if (!dev->netdev_ops->ndo_setup_tc)
@@ -702,10 +713,12 @@ static void tcf_block_offload_unbind(struct tcf_block *block, struct Qdisc *q,
err = tcf_block_offload_cmd(block, dev, ei, FLOW_BLOCK_UNBIND, NULL);
if (err == -EOPNOTSUPP)
goto no_offload_dev_dec;
+ up_write(&block->cb_lock);
return;
no_offload_dev_dec:
WARN_ON(block->nooffloaddevcnt-- == 0);
+ up_write(&block->cb_lock);
}
static int
@@ -820,6 +833,7 @@ static struct tcf_block *tcf_block_create(struct net *net, struct Qdisc *q,
return ERR_PTR(-ENOMEM);
}
mutex_init(&block->lock);
+ init_rwsem(&block->cb_lock);
flow_block_init(&block->flow_block);
INIT_LIST_HEAD(&block->chain_list);
INIT_LIST_HEAD(&block->owner_list);
@@ -1355,6 +1369,8 @@ tcf_block_playback_offloads(struct tcf_block *block, flow_setup_cb_t *cb,
struct tcf_proto *tp, *tp_prev;
int err;
+ lockdep_assert_held(&block->cb_lock);
+
for (chain = __tcf_get_next_chain(block, NULL);
chain;
chain_prev = chain,
@@ -1393,6 +1409,8 @@ static int tcf_block_bind(struct tcf_block *block,
struct flow_block_cb *block_cb, *next;
int err, i = 0;
+ lockdep_assert_held(&block->cb_lock);
+
list_for_each_entry(block_cb, &bo->cb_list, list) {
err = tcf_block_playback_offloads(block, block_cb->cb,
block_cb->cb_priv, true,
@@ -1427,6 +1445,8 @@ static void tcf_block_unbind(struct tcf_block *block,
{
struct flow_block_cb *block_cb, *next;
+ lockdep_assert_held(&block->cb_lock);
+
list_for_each_entry_safe(block_cb, next, &bo->cb_list, list) {
tcf_block_playback_offloads(block, block_cb->cb,
block_cb->cb_priv, false,
@@ -2987,19 +3007,26 @@ int tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
int ok_count = 0;
int err;
+ down_read(&block->cb_lock);
/* Make sure all netdevs sharing this block are offload-capable. */
- if (block->nooffloaddevcnt && err_stop)
- return -EOPNOTSUPP;
+ if (block->nooffloaddevcnt && err_stop) {
+ ok_count = -EOPNOTSUPP;
+ goto err_unlock;
+ }
list_for_each_entry(block_cb, &block->flow_block.cb_list, list) {
err = block_cb->cb(type, type_data, block_cb->cb_priv);
if (err) {
- if (err_stop)
- return err;
+ if (err_stop) {
+ ok_count = err;
+ goto err_unlock;
+ }
} else {
ok_count++;
}
}
+err_unlock:
+ up_read(&block->cb_lock);
return ok_count;
}
EXPORT_SYMBOL(tc_setup_cb_call);
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 07/10] net: sched: take rtnl lock in tc_setup_flow_action()
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev
Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov, Jiri Pirko
In-Reply-To: <20190823185056.12536-1-vladbu@mellanox.com>
In order to allow using new flow_action infrastructure from unlocked
classifiers, modify tc_setup_flow_action() to accept new 'rtnl_held'
argument. Take rtnl lock before accessing tc_action data. This is necessary
to protect from concurrent action replace.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
---
include/net/pkt_cls.h | 2 +-
net/sched/cls_api.c | 17 +++++++++++++----
net/sched/cls_flower.c | 6 ++++--
net/sched/cls_matchall.c | 4 ++--
4 files changed, 20 insertions(+), 9 deletions(-)
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index 612232492f67..a48824bc1489 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -504,7 +504,7 @@ tcf_match_indev(struct sk_buff *skb, int ifindex)
}
int tc_setup_flow_action(struct flow_action *flow_action,
- const struct tcf_exts *exts);
+ const struct tcf_exts *exts, bool rtnl_held);
int tc_setup_cb_call(struct tcf_block *block, enum tc_setup_type type,
void *type_data, bool err_stop, bool rtnl_held);
int tc_setup_cb_add(struct tcf_block *block, struct tcf_proto *tp,
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index f2dcecf34c6f..cb835d581b77 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -3261,14 +3261,17 @@ int tc_setup_cb_reoffload(struct tcf_block *block, struct tcf_proto *tp,
EXPORT_SYMBOL(tc_setup_cb_reoffload);
int tc_setup_flow_action(struct flow_action *flow_action,
- const struct tcf_exts *exts)
+ const struct tcf_exts *exts, bool rtnl_held)
{
const struct tc_action *act;
- int i, j, k;
+ int i, j, k, err = 0;
if (!exts)
return 0;
+ if (!rtnl_held)
+ rtnl_lock();
+
j = 0;
tcf_exts_for_each_action(i, act, exts) {
struct flow_action_entry *entry;
@@ -3313,6 +3316,7 @@ int tc_setup_flow_action(struct flow_action *flow_action,
entry->vlan.prio = tcf_vlan_push_prio(act);
break;
default:
+ err = -EOPNOTSUPP;
goto err_out;
}
} else if (is_tcf_tunnel_set(act)) {
@@ -3330,6 +3334,7 @@ int tc_setup_flow_action(struct flow_action *flow_action,
entry->id = FLOW_ACTION_ADD;
break;
default:
+ err = -EOPNOTSUPP;
goto err_out;
}
entry->mangle.htype = tcf_pedit_htype(act, k);
@@ -3388,15 +3393,19 @@ int tc_setup_flow_action(struct flow_action *flow_action,
entry->id = FLOW_ACTION_PTYPE;
entry->ptype = tcf_skbedit_ptype(act);
} else {
+ err = -EOPNOTSUPP;
goto err_out;
}
if (!is_tcf_pedit(act))
j++;
}
- return 0;
+
err_out:
- return -EOPNOTSUPP;
+ if (!rtnl_held)
+ rtnl_unlock();
+
+ return err;
}
EXPORT_SYMBOL(tc_setup_flow_action);
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 9a7fd6bcd0a5..d47d4e84d4e5 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -452,7 +452,8 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
cls_flower.rule->match.key = &f->mkey;
cls_flower.classid = f->res.classid;
- err = tc_setup_flow_action(&cls_flower.rule->action, &f->exts);
+ err = tc_setup_flow_action(&cls_flower.rule->action, &f->exts,
+ true);
if (err) {
kfree(cls_flower.rule);
if (skip_sw)
@@ -1819,7 +1820,8 @@ static int fl_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
cls_flower.rule->match.mask = &f->mask->key;
cls_flower.rule->match.key = &f->mkey;
- err = tc_setup_flow_action(&cls_flower.rule->action, &f->exts);
+ err = tc_setup_flow_action(&cls_flower.rule->action, &f->exts,
+ true);
if (err) {
kfree(cls_flower.rule);
if (tc_skip_sw(f->flags)) {
diff --git a/net/sched/cls_matchall.c b/net/sched/cls_matchall.c
index 7801851b55d5..d53d81dfee03 100644
--- a/net/sched/cls_matchall.c
+++ b/net/sched/cls_matchall.c
@@ -97,7 +97,7 @@ static int mall_replace_hw_filter(struct tcf_proto *tp,
cls_mall.command = TC_CLSMATCHALL_REPLACE;
cls_mall.cookie = cookie;
- err = tc_setup_flow_action(&cls_mall.rule->action, &head->exts);
+ err = tc_setup_flow_action(&cls_mall.rule->action, &head->exts, true);
if (err) {
kfree(cls_mall.rule);
mall_destroy_hw_filter(tp, head, cookie, NULL);
@@ -300,7 +300,7 @@ static int mall_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
TC_CLSMATCHALL_REPLACE : TC_CLSMATCHALL_DESTROY;
cls_mall.cookie = (unsigned long)head;
- err = tc_setup_flow_action(&cls_mall.rule->action, &head->exts);
+ err = tc_setup_flow_action(&cls_mall.rule->action, &head->exts, true);
if (err) {
kfree(cls_mall.rule);
if (add && tc_skip_sw(head->flags)) {
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v2 00/10] Refactor cls hardware offload API to support rtnl-independent drivers
From: Vlad Buslov @ 2019-08-23 18:50 UTC (permalink / raw)
To: netdev; +Cc: jhs, xiyou.wangcong, jiri, davem, jakub.kicinski, pablo,
Vlad Buslov
Currently, all cls API hardware offloads driver callbacks require caller
to hold rtnl lock when calling them. This patch set introduces new API
that allows drivers to register callbacks that are not dependent on rtnl
lock and unlocked classifiers to offload filters without obtaining rtnl
lock first, which is intended to allow offloading tc rules in parallel.
Recently, new rtnl registration flag RTNL_FLAG_DOIT_UNLOCKED was added.
TC rule update handlers (RTM_NEWTFILTER, RTM_DELTFILTER, etc.) are
already registered with this flag and only take rtnl lock when qdisc or
classifier requires it. Classifiers can indicate that their ops
callbacks don't require caller to hold rtnl lock by setting the
TCF_PROTO_OPS_DOIT_UNLOCKED flag. Unlocked implementation of flower
classifier is now upstreamed. However, this implementation still obtains
rtnl lock before calling hardware offloads API.
Implement following cls API changes:
- Introduce new "unlocked_driver_cb" flag to struct flow_block_offload
to allow registering and unregistering block hardware offload
callbacks that do not require caller to hold rtnl lock. Drivers that
doesn't require users of its tc offload callbacks to hold rtnl lock
sets the flag to true on block bind/unbind. Internally tcf_block is
extended with additional lockeddevcnt counter that is used to count
number of devices that require rtnl lock that block is bound to. When
this counter is zero, tc_setup_cb_*() functions execute callbacks
without obtaining rtnl lock.
- Extend cls API single hardware rule update tc_setup_cb_call() function
with tc_setup_cb_add(), tc_setup_cb_replace(), tc_setup_cb_destroy()
and tc_setup_cb_reoffload() functions. These new APIs are needed to
move management of block offload counter, filter in hardware counter
and flag from classifier implementations to cls API, which is now
responsible for managing them in concurrency-safe manner. Access to
cb_list from callback execution code is synchronized by obtaining new
'cb_lock' rw_semaphore in read mode, which allows executing callbacks
in parallel, but excludes any modifications of data from
register/unregister code. tcf_block offloads counter type is changed
to atomic integer to allow updating the counter concurrently.
- Extend classifier ops with new ops->hw_add() and ops->hw_del()
callbacks which are used to notify unlocked classifiers when filter is
successfully added or deleted to hardware without releasing cb_lock.
This is necessary to update classifier state atomically with callback
list traversal and updating of all relevant counters and allows
unlocked classifiers to synchronize with concurrent reoffload without
requiring any changes to driver callback API implementations.
New tc flow_action infrastructure is also modified to allow its user to
execute without rtnl lock protection. Function tc_setup_flow_action() is
modified to conditionally obtain rtnl lock before accessing action
state. Action data that is accessed by reference is either copied or
reference counted to prevent concurrent action overwrite from
deallocating it. New function tc_cleanup_flow_action() is introduced to
cleanup/release all such data obtained by tc_setup_flow_action().
Flower classifier (only unlocked classifier at the moment) is modified
to use new cls hardware offloads API and no longer obtains rtnl lock
before calling it.
Vlad Buslov (10):
net: sched: protect block offload-related fields with rw_semaphore
net: sched: change tcf block offload counter type to atomic_t
net: sched: refactor block offloads counter usage
net: sched: notify classifier on successful offload add/delete
net: sched: add API for registering unlocked offload block callbacks
net: sched: conditionally obtain rtnl lock in cls hw offloads API
net: sched: take rtnl lock in tc_setup_flow_action()
net: sched: take reference to action dev before calling offloads
net: sched: copy tunnel info when setting flow_action entry->tunnel
net: sched: flower: don't take rtnl lock for cls hw offloads API
.../net/ethernet/mellanox/mlx5/core/en_main.c | 2 +
.../net/ethernet/mellanox/mlx5/core/en_rep.c | 3 +
include/net/flow_offload.h | 1 +
include/net/pkt_cls.h | 21 +-
include/net/sch_generic.h | 41 +--
include/net/tc_act/tc_tunnel_key.h | 17 +
net/sched/cls_api.c | 338 +++++++++++++++++-
net/sched/cls_bpf.c | 38 +-
net/sched/cls_flower.c | 126 +++----
net/sched/cls_matchall.c | 31 +-
net/sched/cls_u32.c | 29 +-
11 files changed, 474 insertions(+), 173 deletions(-)
--
2.21.0
^ permalink raw reply
* Re: [RFC 3/4] Move ACPI functionality out of r8152 driver
From: Bjørn Mork @ 2019-08-23 18:34 UTC (permalink / raw)
To: Charles.Hyde
Cc: linux-usb, linux-acpi, gregkh, Mario.Limonciello, oliver, netdev,
nic_swsd
In-Reply-To: <1566339738195.2913@Dellteam.com>
<Charles.Hyde@dellteam.com> writes:
> This change moves ACPI functionality out of the Realtek r8152 driver to
> its own source and header file, making it available to other drivers as
> needed now and into the future. At the time this ACPI snippet was
> introduced in 2016, only the Realtek driver made use of it in support of
> Dell's enterprise IT policy efforts. There comes now a need for this
> same support in a different driver, also in support of Dell's enterprise
> IT policy efforts.
Yes, and we told you so.
Bjørn
^ permalink raw reply
* Re: [PATCH net-next 03/10] net: sched: refactor block offloads counter usage
From: Jakub Kicinski @ 2019-08-23 18:26 UTC (permalink / raw)
To: Vlad Buslov
Cc: netdev@vger.kernel.org, jhs@mojatatu.com,
xiyou.wangcong@gmail.com, jiri@resnulli.us, davem@davemloft.net,
pablo@netfilter.org, Jiri Pirko, linux-doc
In-Reply-To: <vbfftls17yl.fsf@mellanox.com>
On Fri, 23 Aug 2019 10:39:50 +0000, Vlad Buslov wrote:
> >> +/* Destroy filter and decrement block offload counter, if filter was previously
> >> + * offloaded.
> >> + */
> >> +
> >
> > hm.. is this gap between comment and function it pertains to
> > intentional?
>
> Majority of function comments in cls_api.c have newline after them (not
> all of them though). I don't have any strong opinions regarding this.
> You suggest it is better not to have blank lines after function
> comments?
Ah, you're right. I think it's pretty strange to have a new line after
a comment which pertains only to the function which is immediately
following it. Often the new line is used as a separation, when the
comment describes whole section of the file..
I kind of wish kdoc allowed none of the parameters to be described.
Often you want to document the function but the parameters are kind
of obvious.
Anyway... feel free to leave this as is.
^ permalink raw reply
* Re: [PATCH net-next] net/rds: Whitelist rdma_cookie and rx_tstamp for usercopy
From: santosh.shilimkar @ 2019-08-23 18:17 UTC (permalink / raw)
To: Dag Moxnes, netdev, linux-rdma, rds-devel; +Cc: davem
In-Reply-To: <1566568998-26222-1-git-send-email-dag.moxnes@oracle.com>
On 8/23/19 7:03 AM, Dag Moxnes wrote:
> Add the RDMA cookie and RX timestamp to the usercopy whitelist.
>
> After the introduction of hardened usercopy whitelisting
> (https://lwn.net/Articles/727322/), a warning is displayed when the
> RDMA cookie or RX timestamp is copied to userspace:
>
> kernel: WARNING: CPU: 3 PID: 5750 at
> mm/usercopy.c:81 usercopy_warn+0x8e/0xa6
> [...]
> kernel: Call Trace:
> kernel: __check_heap_object+0xb8/0x11b
> kernel: __check_object_size+0xe3/0x1bc
> kernel: put_cmsg+0x95/0x115
> kernel: rds_recvmsg+0x43d/0x620 [rds]
> kernel: sock_recvmsg+0x43/0x4a
> kernel: ___sys_recvmsg+0xda/0x1e6
> kernel: ? __handle_mm_fault+0xcae/0xf79
> kernel: __sys_recvmsg+0x51/0x8a
> kernel: SyS_recvmsg+0x12/0x1c
> kernel: do_syscall_64+0x79/0x1ae
>
> When the whitelisting feature was introduced, the memory for the RDMA
> cookie and RX timestamp in RDS was not added to the whitelist, causing
> the warning above.
>
> Signed-off-by: Dag Moxnes <dag.moxnes@oracle.com>
> Tested-by: jenny.x.xu@oracle.com
> ---
Thanks Dag to get this out on list.
You might have to fix the Tested-by tag.
Tested-by: Jenny <jenny.x.xu@oracle.com
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
^ permalink raw reply
* Re: [net-next 4/8] net/mlx5e: Add device out of buffer counter
From: Jakub Kicinski @ 2019-08-23 18:16 UTC (permalink / raw)
To: Saeed Mahameed; +Cc: davem@davemloft.net, netdev@vger.kernel.org, Moshe Shemesh
In-Reply-To: <27f7cfa13d1b5e7717e2d75595ab453951b18a96.camel@mellanox.com>
On Fri, 23 Aug 2019 06:00:45 +0000, Saeed Mahameed wrote:
> On Thu, 2019-08-22 at 18:33 -0700, Jakub Kicinski wrote:
> > On Thu, 22 Aug 2019 23:35:52 +0000, Saeed Mahameed wrote:
> > > From: Moshe Shemesh <moshe@mellanox.com>
> > >
> > > Added the following packets drop counter:
> > > Device out of buffer - counts packets which were dropped due to
> > > full
> > > device internal receive queue.
> > > This counter will be shown on ethtool as a new counter called
> > > dev_out_of_buffer.
> > > The counter is read from FW by command QUERY_VNIC_ENV.
> > >
> > > Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
> > > Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
> >
> > Sounds like rx_fifo_errors, no? Doesn't rx_fifo_errors count RX
> > overruns?
>
> No, that is port buffer you are looking for and we got that fully
> covered in mlx5. this is different.
>
> This new counter is deep into the HW data path pipeline and it covers
> very rare and complex scenarios that got only recently introduced with
> swichdev mode and "some" lately added tunnels offloads that are routed
> between VFs/PFs.
>
> Normally the HW is lossless once the packet passes port buffers into
> the data plane pipeline, let's call that "fast lane", BUT for sriov
> configurations with switchdev mode enabled and some special hand
> crafted tc tunnel offloads that requires hairpin between VFs/PFs, the
> hw might decide to send some traffic to a "service lane" which is still
> fast path but unlike the "fast lane" it handles traffic through "HW
> internal" receive and send queues (just like we do with hairpin) that
> might drop packets. the whole thing is transparent to driver and it is
> HW implementation specific.
I see thanks for the explanation and sorry for the delayed response.
Would it perhaps make sense to indicate the hairpin in the name?
dev_out_of_buffer is quite a generic name, and there seems to be no
doc, nor does the commit message explains it as well as you have..
^ permalink raw reply
* Re: r8169: regression on MIPS/Loongson
From: Heiner Kallweit @ 2019-08-23 18:08 UTC (permalink / raw)
To: Aaro Koskinen; +Cc: netdev, linux-mips
In-Reply-To: <20190822233854.GG30291@darkstar.musicnaut.iki.fi>
On 23.08.2019 01:38, Aaro Koskinen wrote:
> Hi,
>
> On Fri, Aug 23, 2019 at 12:52:34AM +0200, Heiner Kallweit wrote:
>> On 23.08.2019 00:25, Aaro Koskinen wrote:
>>> After upgrading from v5.2 to v5.3-rc5 on MIPS/Loongson board, copying
>>> large files from network with scp started to fail with "Integrity error".
>>> Bisected to:
>>>
>>> f072218cca5b076dd99f3dfa3aaafedfd0023a51 is the first bad commit
>>> commit f072218cca5b076dd99f3dfa3aaafedfd0023a51
>>> Author: Heiner Kallweit <hkallweit1@gmail.com>
>>> Date: Thu Jun 27 23:19:09 2019 +0200
>>>
>>> r8169: remove not needed call to dma_sync_single_for_device
>>>
>>> Any idea what goes wrong? Should this change be reverted?
>>>
>> Typically the Realtek chips are used on Intel platforms and I haven't
>> seen any such report yet, so it seems to be platform-specific.
>
> Probably. On my AMD x86_64 box r8169 works fine.
>
>> Which board (DT config) is it, and can you provide a full dmesg?
>
> This board does not use DT (support files are under arch/mips/loongson64).
> dmesg is below:
>
> [ 0.000000] Linux version 5.3.0-rc4-lemote-los_1bf0c (aakoskin@amd-fx-6350) (gcc version 8.3.0 (GCC)) #1 Fri Aug 23 01:01:45 EEST 2019
> [ 0.000000] memsize=256, highmemsize=256
> [ 0.000000] CpuClock = 797800000
> [ 0.000000] printk: bootconsole [early0] enabled
> [ 0.000000] CPU0 revision is: 00006303 (ICT Loongson-2)
> [ 0.000000] FPU revision is: 00000501
> [ 0.000000] Checking for the multiply/shift bug... no.
> [ 0.000000] Checking for the daddiu bug... no.
> [ 0.000000] Determined physical RAM map:
> [ 0.000000] memory: 0000000010000000 @ 0000000000000000 (usable)
> [ 0.000000] memory: 0000000030000000 @ 0000000010000000 (reserved)
> [ 0.000000] memory: 0000000010000000 @ 0000000090000000 (usable)
> [ 0.000000] memory: 0000000010000000 @ 0000000080000000 (reserved)
> [ 0.000000] Initrd not found or empty - disabling initrd
> [ 0.000000] Primary instruction cache 64kB, VIPT, direct mapped, linesize 32 bytes.
> [ 0.000000] Primary data cache 64kB, 4-way, VIPT, no aliases, linesize 32 bytes
> [ 0.000000] Unified secondary cache 512kB 4-way, linesize 32 bytes.
> [ 0.000000] Zone ranges:
> [ 0.000000] Normal [mem 0x0000000000000000-0x000000009fffffff]
> [ 0.000000] Movable zone start for each node
> [ 0.000000] Early memory node ranges
> [ 0.000000] node 0: [mem 0x0000000000000000-0x000000003fffffff]
> [ 0.000000] node 0: [mem 0x0000000080000000-0x000000009fffffff]
> [ 0.000000] Initmem setup node 0 [mem 0x0000000000000000-0x000000009fffffff]
> [ 0.000000] On node 0 totalpages: 98304
> [ 0.000000] Normal zone: 336 pages used for memmap
> [ 0.000000] Normal zone: 0 pages reserved
> [ 0.000000] Normal zone: 98304 pages, LIFO batch:15
> [ 0.000000] pcpu-alloc: s0 r0 d32768 u32768 alloc=1*32768
> [ 0.000000] pcpu-alloc: [0] 0
> [ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 97968
> [ 0.000000] Kernel command line: console=tty console=ttyS0,115200
> [ 0.000000] Dentry cache hash table entries: 262144 (order: 7, 2097152 bytes, linear)
> [ 0.000000] Inode-cache hash table entries: 131072 (order: 6, 1048576 bytes, linear)
> [ 0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off
> [ 0.000000] Memory: 489664K/1572864K available (4863K kernel code, 467K rwdata, 876K rodata, 1968K init, 16616K bss, 1083200K reserved, 0K cma-reserved)
> [ 0.000000] SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
> [ 0.000000] NR_IRQS: 128
> [ 0.000000] random: get_random_bytes called from start_kernel+0x368/0x620 with crng_init=0
> [ 0.000000] Console: colour dummy device 80x25
> [ 0.000000] printk: console [tty0] enabled
> [ 0.000000] sched_clock: 64 bits at 250 Hz, resolution 4000000ns, wraps every 9007199254000000ns
> [ 0.004000] Calibrating delay loop... 528.38 BogoMIPS (lpj=1056768)
> [ 0.040000] pid_max: default: 32768 minimum: 301
> [ 0.044000] Mount-cache hash table entries: 4096 (order: 1, 32768 bytes, linear)
> [ 0.048000] Mountpoint-cache hash table entries: 4096 (order: 1, 32768 bytes, linear)
> [ 0.052000] *** VALIDATE proc ***
> [ 0.056000] Checking for the daddi bug... no.
> [ 0.064000] devtmpfs: initialized
> [ 0.068000] clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 7645041785100000 ns
> [ 0.072000] futex hash table entries: 256 (order: -2, 6144 bytes, linear)
> [ 0.076000] NET: Registered protocol family 16
> [ 0.080000] clocksource: mfgpt: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 133486551712 ns
> [ 0.124000] SCSI subsystem initialized
> [ 0.128000] usbcore: registered new interface driver usbfs
> [ 0.132000] usbcore: registered new interface driver hub
> [ 0.136000] usbcore: registered new device driver usb
> [ 0.140000] PCI host bridge to bus 0000:00
> [ 0.148000] pci_bus 0000:00: root bus resource [mem 0x40000000-0x7fffffff]
> [ 0.152000] pci_bus 0000:00: root bus resource [io 0x4000-0xffff]
> [ 0.156000] pci_bus 0000:00: root bus resource [??? 0x00000000 flags 0x0]
> [ 0.160000] pci_bus 0000:00: No busn resource found for root bus, will use [bus 00-ff]
> [ 0.164000] pci 0000:00:06.0: [10ec:8169] type 00 class 0x020000
> [ 0.168000] pci 0000:00:06.0: reg 0x10: [io 0xb100-0xb1ff]
> [ 0.172000] pci 0000:00:06.0: reg 0x14: [mem 0x04075000-0x040750ff]
> [ 0.176000] pci 0000:00:06.0: reg 0x30: [mem 0x04040000-0x0405ffff pref]
> [ 0.180000] pci 0000:00:06.0: supports D1 D2
> [ 0.184000] pci 0000:00:06.0: PME# supported from D1 D2 D3hot
> [ 0.188000] pci 0000:00:08.0: [1039:0325] type 00 class 0x030000
> [ 0.192000] pci 0000:00:08.0: reg 0x10: [mem 0x40000000-0x4fffffff pref]
> [ 0.196000] pci 0000:00:08.0: reg 0x14: [mem 0x04000000-0x0403ffff]
> [ 0.200000] pci 0000:00:08.0: reg 0x18: [io 0xb300-0xb37f]
> [ 0.204000] pci 0000:00:08.0: reg 0x30: [mem 0x04060000-0x0406ffff pref]
> [ 0.208000] pci 0000:00:08.0: supports D1 D2
> [ 0.212000] pci 0000:00:0e.0: [1022:2090] type 00 class 0x060100
> [ 0.216000] pci 0000:00:0e.0: reg 0x10: [io 0xb410-0xb417]
> [ 0.220000] pci 0000:00:0e.0: reg 0x14: [io 0xb000-0xb0ff]
> [ 0.224000] pci 0000:00:0e.0: reg 0x18: [io 0xb380-0xb3bf]
> [ 0.228000] pci 0000:00:0e.0: reg 0x20: [io 0xb280-0xb2ff]
> [ 0.232000] pci 0000:00:0e.0: reg 0x24: [io 0xb3c0-0xb3df]
> [ 0.240000] pci 0000:00:0e.2: [1022:209a] type 00 class 0x010180
> [ 0.244000] pci 0000:00:0e.2: reg 0x20: [io 0xb400-0xb40f]
> [ 0.252000] pci 0000:00:0e.2: legacy IDE quirk: reg 0x10: [io 0x01f0-0x01f7]
> [ 0.256000] pci 0000:00:0e.2: legacy IDE quirk: reg 0x14: [io 0x03f6]
> [ 0.260000] pci 0000:00:0e.2: legacy IDE quirk: reg 0x18: [io 0x0170-0x0177]
> [ 0.264000] pci 0000:00:0e.2: legacy IDE quirk: reg 0x1c: [io 0x0376]
> [ 0.268000] pci 0000:00:0e.3: [1022:2093] type 00 class 0x040100
> [ 0.272000] pci 0000:00:0e.3: reg 0x10: [io 0xb200-0xb27f]
> [ 0.276000] pci 0000:00:0e.4: [1022:2094] type 00 class 0x0c0310
> [ 0.280000] pci 0000:00:0e.4: reg 0x10: [mem 0x04074000-0x04074fff]
> [ 0.288000] pci 0000:00:0e.5: [1022:2095] type 00 class 0x0c0320
> [ 0.292000] pci 0000:00:0e.5: reg 0x10: [mem 0x04073000-0x04073fff]
> [ 0.296000] pci_bus 0000:00: busn_res: [bus 00-ff] end is updated to 00
> [ 0.300000] pci 0000:00:08.0: BAR 0: assigned [mem 0x40000000-0x4fffffff pref]
> [ 0.304000] pci 0000:00:08.0: BAR 1: assigned [mem 0x50000000-0x5003ffff]
> [ 0.308000] pci 0000:00:06.0: BAR 6: assigned [mem 0x50040000-0x5005ffff pref]
> [ 0.312000] pci 0000:00:08.0: BAR 6: assigned [mem 0x50060000-0x5006ffff pref]
> [ 0.316000] pci 0000:00:0e.4: BAR 0: assigned [mem 0x50070000-0x50070fff]
> [ 0.320000] pci 0000:00:0e.5: BAR 0: assigned [mem 0x50071000-0x50071fff]
> [ 0.324000] pci 0000:00:06.0: BAR 0: assigned [io 0x4000-0x40ff]
> [ 0.328000] pci 0000:00:06.0: BAR 1: assigned [mem 0x50072000-0x500720ff]
> [ 0.332000] pci 0000:00:0e.0: BAR 1: assigned [io 0x4400-0x44ff]
> [ 0.336000] pci 0000:00:08.0: BAR 2: assigned [io 0x4800-0x487f]
> [ 0.340000] pci 0000:00:0e.0: BAR 4: assigned [io 0x4880-0x48ff]
> [ 0.344000] pci 0000:00:0e.3: BAR 0: assigned [io 0x4c00-0x4c7f]
> [ 0.348000] pci 0000:00:0e.0: BAR 2: assigned [io 0x4c80-0x4cbf]
> [ 0.352000] pci 0000:00:0e.0: BAR 5: assigned [io 0x4cc0-0x4cdf]
> [ 0.356000] pci 0000:00:0e.2: BAR 4: assigned [io 0x4ce0-0x4cef]
> [ 0.360000] pci 0000:00:0e.0: BAR 0: assigned [io 0x4cf0-0x4cf7]
> [ 0.364000] clocksource: Switched to clocksource mfgpt
> [ 0.392000] NET: Registered protocol family 2
> [ 0.396000] tcp_listen_portaddr_hash hash table entries: 1024 (order: 0, 16384 bytes, linear)
> [ 0.400000] TCP established hash table entries: 16384 (order: 3, 131072 bytes, linear)
> [ 0.404000] TCP bind hash table entries: 16384 (order: 3, 131072 bytes, linear)
> [ 0.408000] TCP: Hash tables configured (established 16384 bind 16384)
> [ 0.412000] UDP hash table entries: 1024 (order: 1, 32768 bytes, linear)
> [ 0.416000] UDP-Lite hash table entries: 1024 (order: 1, 32768 bytes, linear)
> [ 0.420000] NET: Registered protocol family 1
> [ 0.424000] pci 0000:00:0e.4: enabling device (0000 -> 0002)
> [ 0.428000] PCI: CLS 32 bytes, default 32
> [ 0.888000] random: fast init done
> [ 1.804000] workingset: timestamp_bits=62 max_order=15 bucket_order=0
> [ 1.820000] NET: Registered protocol family 38
> [ 1.824000] io scheduler bfq registered
> [ 1.828000] slot: 8, pin: 1, irq: 38
> [ 1.832000] sisfb 0000:00:08.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0x3030
> [ 1.836000] sisfb: Video ROM not found
> [ 1.840000] sisfb: Video RAM at 0x40000000, mapped to 0x9000000040000000, size 32768k
> [ 1.844000] sisfb: MMIO at 0x50000000, mapped to 0x9000000050000000, size 256k
> [ 1.848000] sisfb: Memory heap starting at 32160K, size 32K
> [ 3.140000] sisfb: Detected SiS301C video bridge
> [ 3.220000] sisfb: Detected 1280x1024 flat panel
> [ 3.304000] sisfb: CRT2 DDC supported
> [ 3.304000] sisfb: CRT2 DDC level: 2
> [ 3.512000] sisfb: Monitor range H 30-81KHz, V 56-76Hz, Max. dotclock 140MHz
> [ 3.516000] sisfb: Default mode is 1280x1024x8 (60Hz)
> [ 3.520000] sisfb: Initial vbflags 0x10000022
> [ 4.008000] Console: switching to colour frame buffer device 160x64
> [ 4.068000] sisfb: 2D acceleration is enabled, y-panning enabled (auto-max)
> [ 4.072000] fb0: SiS 315PRO frame buffer device version 1.8.9
> [ 4.076000] sisfb: Copyright (C) 2001-2005 Thomas Winischhofer
> [ 4.156000] Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
> [ 4.308000] printk: console [ttyS0] disabled
> [ 4.316000] serial8250.0: ttyS0 at I/O 0x2f8 (irq = 3, base_baud = 115200) is a 16550A
> [ 4.320000] printk: console [ttyS0] enabled
> [ 4.324000] printk: bootconsole [early0] disabled
> [ 4.808000] brd: module loaded
> [ 5.036000] loop: module loaded
> [ 5.040000] Uniform Multi-Platform E-IDE driver
> [ 5.044000] amd74xx 0000:00:0e.2: UDMA100 controller
> [ 5.048000] amd74xx 0000:00:0e.2: IDE controller (0x1022:0x209a rev 0x01)
> [ 5.052000] amd74xx 0000:00:0e.2: IDE port disabled
> [ 5.056000] amd74xx 0000:00:0e.2: not 100% native mode: will probe irqs later
> [ 5.060000] legacy IDE will be removed in 2021, please switch to libata
> [ 5.060000] Report any missing HW support to linux-ide@vger.kernel.org
> [ 5.064000] ide0: BM-DMA at 0x4ce0-0x4ce7
> [ 5.068000] Probing IDE interface ide0...
> [ 5.428000] hda: WDC WD1600BEVS-00VAT0, ATA DISK drive
> [ 6.152000] hda: host max PIO5 wanted PIO255(auto-tune) selected PIO4
> [ 6.152000] hda: UDMA/100 mode selected
> [ 6.156000] ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
> [ 6.160000] ide-gd driver 1.18
> [ 6.164000] hda: max request size: 1024KiB
> [ 6.244000] hda: 312581808 sectors (160041 MB) w/8192KiB Cache, CHS=19457/255/63
> [ 6.248000] hda: cache flushes supported
> [ 6.268000] hda: hda1
> [ 6.272000] slot: 6, pin: 1, irq: 36
> [ 6.276000] libphy: r8169: probed
> [ 6.280000] r8169 0000:00:06.0 eth0: RTL8169sc/8110sc, 00:23:9e:00:0f:54, XID 980, IRQ 36
> [ 6.284000] r8169 0000:00:06.0 eth0: jumbo features [frames: 7152 bytes, tx checksumming: ok]
> [ 6.288000] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
> [ 6.292000] ehci-pci: EHCI PCI platform driver
> [ 6.296000] ehci-pci 0000:00:0e.5: EHCI Host Controller
> [ 6.300000] ehci-pci 0000:00:0e.5: new USB bus registered, assigned bus number 1
> [ 6.304000] ehci-pci 0000:00:0e.5: irq 11, io mem 0x50071000
> [ 6.440000] ehci-pci 0000:00:0e.5: USB 0.0 started, EHCI 1.00
> [ 6.448000] hub 1-0:1.0: USB hub found
> [ 6.460000] hub 1-0:1.0: 4 ports detected
> [ 6.468000] ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver
> [ 6.472000] ohci-pci: OHCI PCI platform driver
> [ 6.476000] ohci-pci 0000:00:0e.4: OHCI PCI host controller
> [ 6.480000] ohci-pci 0000:00:0e.4: new USB bus registered, assigned bus number 2
> [ 6.484000] ohci-pci 0000:00:0e.4: irq 11, io mem 0x50070000
> [ 6.580000] hub 2-0:1.0: USB hub found
> [ 6.592000] hub 2-0:1.0: 4 ports detected
> [ 6.600000] usbcore: registered new interface driver usb-storage
> [ 6.604000] loongson2_cpufreq: Loongson-2F CPU frequency driver
> [ 6.608000] usbcore: registered new interface driver usbhid
> [ 6.612000] usbhid: USB HID core driver
> [ 6.620000] NET: Registered protocol family 17
> [ 6.700000] Freeing unused kernel memory: 1968K
> [ 6.704000] This architecture does not have kernel memory protection.
> [ 6.708000] Run /init as init process
> [ 10.868000] EXT4-fs (hda1): mounting ext3 file system using the ext4 subsystem
> [ 12.756000] EXT4-fs (hda1): mounted filesystem with ordered data mode. Opts: (null)
> [ 15.800000] RTL8211B Gigabit Ethernet r8169-30:00: attached PHY driver [RTL8211B Gigabit Ethernet] (mii_bus:phy_addr=r8169-30:00, irq=IGNORE)
> [ 15.908000] r8169 0000:00:06.0 eth0: Link is Down
> [ 18.424000] r8169 0000:00:06.0 eth0: Link is Up - 1Gbps/Full - flow control rx/tx
>
> A.
>
Thanks for reporting and the log. I'll revert the patch.
Heiner
^ permalink raw reply
* [PATCH net-next] r8169: fix DMA issue on MIPS platform
From: Heiner Kallweit @ 2019-08-23 18:07 UTC (permalink / raw)
To: Realtek linux nic maintainers, David Miller
Cc: Aaro Koskinen, netdev@vger.kernel.org
As reported by Aaro this patch causes network problems on
MIPS Loongson platform. Therefore revert it.
Fixes: f072218cca5b ("r8169: remove not needed call to dma_sync_single_for_device")
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
drivers/net/ethernet/realtek/r8169_main.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
index 910944120..6182e7d33 100644
--- a/drivers/net/ethernet/realtek/r8169_main.c
+++ b/drivers/net/ethernet/realtek/r8169_main.c
@@ -5822,6 +5822,10 @@ static int rtl_rx(struct net_device *dev, struct rtl8169_private *tp, u32 budget
skb->tail += pkt_size;
skb->len = pkt_size;
+ dma_sync_single_for_device(tp_to_dev(tp),
+ le64_to_cpu(desc->addr),
+ pkt_size, DMA_FROM_DEVICE);
+
rtl8169_rx_csum(skb, status);
skb->protocol = eth_type_trans(skb, dev);
--
2.23.0
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox