From 59435444a13ed52d3444c5df26b73d3086bcd57b Mon Sep 17 00:00:00 2001 From: Gerrit Renker Date: Sat, 26 Jul 2008 11:59:09 +0100 Subject: dccp: Allow to distinguish original and retransmitted packets This patch allows the sender to distinguish original and retransmitted packets, which is in particular needed for the retransmission of DCCP-Requests: * the first Request uses ISS (generated in net/dccp/ip*.c), and sets GSS = ISS; * all retransmitted Requests use GSS' = GSS + 1, so that the n-th retransmitted Request has sequence number ISS + n (mod 48). To add generic support, the patch reorganises existing code so that: * icsk_retransmits == 0 for the original packet and * icsk_retransmits = n > 0 for the n-th retransmitted packet at the time dccp_transmit_skb() is called, via dccp_retransmit_skb(). Thanks to Wei Yongjun for pointing this problem out. Further changes: ---------------- * removed the `skb' argument from dccp_retransmit_skb(), since sk_send_head is used for all retransmissions (the exception is client-Acks in PARTOPEN state, but these do not use sk_send_head); * since sk_send_head always contains the original skb (via dccp_entail()), skb_cloned() never evaluated to true and thus pskb_copy() was never used. Signed-off-by: Gerrit Renker --- net/dccp/dccp.h | 2 +- net/dccp/output.c | 20 ++++++++++++++++---- net/dccp/timer.c | 20 ++++---------------- 3 files changed, 21 insertions(+), 21 deletions(-) (limited to 'net') diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index 743d85fcd651..1c2e3ec2eb57 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -226,7 +226,7 @@ static inline void dccp_csum_outgoing(struct sk_buff *skb) extern void dccp_v4_send_check(struct sock *sk, int len, struct sk_buff *skb); -extern int dccp_retransmit_skb(struct sock *sk, struct sk_buff *skb); +extern int dccp_retransmit_skb(struct sock *sk); extern void dccp_send_ack(struct sock *sk); extern void dccp_reqsk_send_ack(struct sk_buff *sk, struct request_sock *rsk); diff --git a/net/dccp/output.c b/net/dccp/output.c index fe20068c5d8e..d19d48195013 100644 --- a/net/dccp/output.c +++ b/net/dccp/output.c @@ -284,14 +284,26 @@ void dccp_write_xmit(struct sock *sk, int block) } } -int dccp_retransmit_skb(struct sock *sk, struct sk_buff *skb) +/** + * dccp_retransmit_skb - Retransmit Request, Close, or CloseReq packets + * There are only four retransmittable packet types in DCCP: + * - Request in client-REQUEST state (sec. 8.1.1), + * - CloseReq in server-CLOSEREQ state (sec. 8.3), + * - Close in node-CLOSING state (sec. 8.3), + * - Acks in client-PARTOPEN state (sec. 8.1.5, handled by dccp_delack_timer()). + * This function expects sk->sk_send_head to contain the original skb. + */ +int dccp_retransmit_skb(struct sock *sk) { + WARN_ON(sk->sk_send_head == NULL); + if (inet_csk(sk)->icsk_af_ops->rebuild_header(sk) != 0) return -EHOSTUNREACH; /* Routing failure or similar. */ - return dccp_transmit_skb(sk, (skb_cloned(skb) ? - pskb_copy(skb, GFP_ATOMIC): - skb_clone(skb, GFP_ATOMIC))); + /* this count is used to distinguish original and retransmitted skb */ + inet_csk(sk)->icsk_retransmits++; + + return dccp_transmit_skb(sk, skb_clone(sk->sk_send_head, GFP_ATOMIC)); } struct sk_buff *dccp_make_response(struct sock *sk, struct dst_entry *dst, diff --git a/net/dccp/timer.c b/net/dccp/timer.c index 6a5b961b6f5c..54b3c7e9e016 100644 --- a/net/dccp/timer.c +++ b/net/dccp/timer.c @@ -98,22 +98,12 @@ static void dccp_retransmit_timer(struct sock *sk) goto backoff; } - /* - * sk->sk_send_head has to have one skb with - * DCCP_SKB_CB(skb)->dccpd_type set to one of the retransmittable DCCP - * packet types. The only packets eligible for retransmission are: - * -- Requests in client-REQUEST state (sec. 8.1.1) - * -- Acks in client-PARTOPEN state (sec. 8.1.5) - * -- CloseReq in server-CLOSEREQ state (sec. 8.3) - * -- Close in node-CLOSING state (sec. 8.3) */ - WARN_ON(sk->sk_send_head == NULL); - /* * More than than 4MSL (8 minutes) has passed, a RESET(aborted) was * sent, no need to retransmit, this sock is dead. */ if (dccp_write_timeout(sk)) - goto out; + return; /* * We want to know the number of packets retransmitted, not the @@ -122,30 +112,28 @@ static void dccp_retransmit_timer(struct sock *sk) if (icsk->icsk_retransmits == 0) DCCP_INC_STATS_BH(DCCP_MIB_TIMEOUTS); - if (dccp_retransmit_skb(sk, sk->sk_send_head) < 0) { + if (dccp_retransmit_skb(sk) != 0) { /* * Retransmission failed because of local congestion, * do not backoff. */ - if (icsk->icsk_retransmits == 0) + if (--icsk->icsk_retransmits == 0) icsk->icsk_retransmits = 1; inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, min(icsk->icsk_rto, TCP_RESOURCE_PROBE_INTERVAL), DCCP_RTO_MAX); - goto out; + return; } backoff: icsk->icsk_backoff++; - icsk->icsk_retransmits++; icsk->icsk_rto = min(icsk->icsk_rto << 1, DCCP_RTO_MAX); inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, icsk->icsk_rto, DCCP_RTO_MAX); if (icsk->icsk_retransmits > sysctl_dccp_retries1) __sk_dst_reset(sk); -out:; } static void dccp_write_timer(unsigned long data) -- cgit v1.2.1 From 73f18fdbca3f92b90aeaee16f5175fe30496e218 Mon Sep 17 00:00:00 2001 From: Gerrit Renker Date: Sat, 26 Jul 2008 11:59:10 +0100 Subject: dccp: Bug-Fix - AWL was never updated The AWL lower Ack validity window advances in proportion to GSS, the greatest sequence number sent. Updating AWL other than at connection setup (in the DCCP-Request sent by dccp_v{4,6}_connect()) was missing in the DCCP code. This bug lead to syslog messages such as "kernel: dccp_check_seqno: DCCP: Step 6 failed for DATAACK packet, [...] P.ackno exists or LAWL(82947089) <= P.ackno(82948208) <= S.AWH(82948728), sending SYNC..." The difference between AWL/AWH here is 1639 packets, while the expected value (the Sequence Window) would have been 100 (the default). A closer look showed that LAWL = AWL = 82947089 equalled the ISS on the Response. The patch now updates AWL with each increase of GSS. Further changes: ---------------- The patch also enforces more stringent checks on the ISS sequence number: * AWL is initialised to ISS at connection setup and remains at this value; * AWH is then always set to GSS (via dccp_update_gss()); * so on the first Request: AWL = AWH = ISS, and on the n-th Request: AWL = ISS, AWH = ISS + n. As a consequence, only Response packets that refer to Requests sent by this host will pass, all others are discarded. This is the intention and in effect implements the initial adjustments for AWL as specified in RFC 4340, 7.5.1. Signed-off-by: Gerrit Renker Acked-by: Ian McDonald --- net/dccp/output.c | 33 +++++++++++++++------------------ 1 file changed, 15 insertions(+), 18 deletions(-) (limited to 'net') diff --git a/net/dccp/output.c b/net/dccp/output.c index d19d48195013..d06945c7d3df 100644 --- a/net/dccp/output.c +++ b/net/dccp/output.c @@ -53,8 +53,11 @@ static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) dccp_packet_hdr_len(dcb->dccpd_type); int err, set_ack = 1; u64 ackno = dp->dccps_gsr; - - dccp_inc_seqno(&dp->dccps_gss); + /* + * Increment GSS here already in case the option code needs it. + * Update GSS for real only if option processing below succeeds. + */ + dcb->dccpd_seq = ADD48(dp->dccps_gss, 1); switch (dcb->dccpd_type) { case DCCP_PKT_DATA: @@ -66,6 +69,9 @@ static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) case DCCP_PKT_REQUEST: set_ack = 0; + /* Use ISS on the first (non-retransmitted) Request. */ + if (icsk->icsk_retransmits == 0) + dcb->dccpd_seq = dp->dccps_iss; /* fall through */ case DCCP_PKT_SYNC: @@ -84,8 +90,6 @@ static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) break; } - dcb->dccpd_seq = dp->dccps_gss; - if (dccp_insert_options(sk, skb)) { kfree_skb(skb); return -EPROTO; @@ -103,7 +107,7 @@ static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) /* XXX For now we're using only 48 bits sequence numbers */ dh->dccph_x = 1; - dp->dccps_awh = dp->dccps_gss; + dccp_update_gss(sk, dcb->dccpd_seq); dccp_hdr_set_seq(dh, dp->dccps_gss); if (set_ack) dccp_hdr_set_ack(dccp_hdr_ack_bits(skb), ackno); @@ -112,6 +116,11 @@ static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) case DCCP_PKT_REQUEST: dccp_hdr_request(skb)->dccph_req_service = dp->dccps_service; + /* + * Limit Ack window to ISS <= P.ackno <= GSS, so that + * only Responses to Requests we sent are considered. + */ + dp->dccps_awl = dp->dccps_iss; break; case DCCP_PKT_RESET: dccp_hdr_reset(skb)->dccph_reset_code = @@ -449,19 +458,7 @@ static inline void dccp_connect_init(struct sock *sk) dccp_sync_mss(sk, dst_mtu(dst)); - /* - * SWL and AWL are initially adjusted so that they are not less than - * the initial Sequence Numbers received and sent, respectively: - * SWL := max(GSR + 1 - floor(W/4), ISR), - * AWL := max(GSS - W' + 1, ISS). - * These adjustments MUST be applied only at the beginning of the - * connection. - */ - dccp_update_gss(sk, dp->dccps_iss); - dccp_set_seqno(&dp->dccps_awl, max48(dp->dccps_awl, dp->dccps_iss)); - - /* S.GAR - greatest valid acknowledgement number received on a non-Sync; - * initialized to S.ISS (sec. 8.5) */ + /* Initialise GAR as per 8.5; AWL/AWH are set in dccp_transmit_skb() */ dp->dccps_gar = dp->dccps_iss; icsk->icsk_retransmits = 0; -- cgit v1.2.1 From d68f0866f76e2bc4ddc07e88e2cb1bc8959a6d7e Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Sat, 26 Jul 2008 11:59:10 +0100 Subject: dccp: Fix sequence number check for ICMPv4 packets The payload of ICMP message is a part of the packet sent by ourself, so the sequence number check must use AWL and AWH, not SWL and SWH. For example: Endpoint A Endpoint B DATA-ACK --------> (SEQ=X) <-------- ICMP (Fragmentation Needed) (SEQ=X) Signed-off-by: Wei Yongjun Acked-by: Gerrit Renker --- net/dccp/ipv4.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index a835b88237cb..6a2f1879e183 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -238,7 +238,7 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info) dp = dccp_sk(sk); seq = dccp_hdr_seq(dh); if ((1 << sk->sk_state) & ~(DCCPF_REQUESTING | DCCPF_LISTEN) && - !between48(seq, dp->dccps_swl, dp->dccps_swh)) { + !between48(seq, dp->dccps_awl, dp->dccps_awh)) { NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); goto out; } -- cgit v1.2.1 From e0bcfb0c6a6ed9ebd68746b306298dc5797fd426 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Sat, 26 Jul 2008 11:59:10 +0100 Subject: dccp: Add check for sequence number in ICMPv6 message This adds a sequence number check for ICMPv6 DCCP error packets, in the same manner as it has been done for ICMPv4 in the previous patch. Signed-off-by: Wei Yongjun Acked-by: Gerrit Renker --- net/dccp/ipv6.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index da509127e00c..25826b1bf685 100644 --- a/net/dccp/ipv6.c +++ b/net/dccp/ipv6.c @@ -89,6 +89,7 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, { struct ipv6hdr *hdr = (struct ipv6hdr *)skb->data; const struct dccp_hdr *dh = (struct dccp_hdr *)(skb->data + offset); + struct dccp_sock *dp; struct ipv6_pinfo *np; struct sock *sk; int err; @@ -116,6 +117,14 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, if (sk->sk_state == DCCP_CLOSED) goto out; + dp = dccp_sk(sk); + seq = dccp_hdr_seq(dh); + if ((1 << sk->sk_state) & ~(DCCPF_REQUESTING | DCCPF_LISTEN) && + !between48(seq, dp->dccps_awl, dp->dccps_awh)) { + NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); + goto out; + } + np = inet6_sk(sk); if (type == ICMPV6_PKT_TOOBIG) { @@ -168,7 +177,6 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, icmpv6_err_convert(type, code, &err); - seq = dccp_hdr_seq(dh); /* Might be for an request_sock */ switch (sk->sk_state) { struct request_sock *req, **prev; -- cgit v1.2.1 From 18e1d836002ad970f42736bad09b7be9cfe99545 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Sat, 26 Jul 2008 11:59:10 +0100 Subject: dccp: Fix incorrect length check for ICMPv4 packets Unlike TCP, which only needs 8 octets of original packet data, DCCP requires minimally 12 or 16 bytes for ICMP-payload sequence number checks. This patch replaces the insufficient length constant of 8 with a two-stage test, making sure that 12 bytes are available, before computing the basic header length required for sequence number checks. Signed-off-by: Wei Yongjun Signed-off-by: Gerrit Renker --- net/dccp/ipv4.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 6a2f1879e183..882c5c4de69e 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -196,8 +196,8 @@ static inline void dccp_do_pmtu_discovery(struct sock *sk, static void dccp_v4_err(struct sk_buff *skb, u32 info) { const struct iphdr *iph = (struct iphdr *)skb->data; - const struct dccp_hdr *dh = (struct dccp_hdr *)(skb->data + - (iph->ihl << 2)); + const u8 offset = iph->ihl << 2; + const struct dccp_hdr *dh = (struct dccp_hdr *)(skb->data + offset); struct dccp_sock *dp; struct inet_sock *inet; const int type = icmp_hdr(skb)->type; @@ -207,7 +207,8 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info) int err; struct net *net = dev_net(skb->dev); - if (skb->len < (iph->ihl << 2) + 8) { + if (skb->len < offset + sizeof(*dh) || + skb->len < offset + __dccp_basic_hdr_len(dh)) { ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS); return; } -- cgit v1.2.1 From 860239c56bbc7c830bdbcec93b140f22a5a5219b Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Sat, 26 Jul 2008 11:59:11 +0100 Subject: dccp: Add check for truncated ICMPv6 DCCP error packets This patch adds a minimum-length check for ICMPv6 packets, as per the previous patch for ICMPv4 payloads. Signed-off-by: Wei Yongjun Signed-off-by: Gerrit Renker --- net/dccp/ipv6.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'net') diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index 25826b1bf685..5e1ee0da2c40 100644 --- a/net/dccp/ipv6.c +++ b/net/dccp/ipv6.c @@ -96,6 +96,12 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, __u64 seq; struct net *net = dev_net(skb->dev); + if (skb->len < offset + sizeof(*dh) || + skb->len < offset + __dccp_basic_hdr_len(dh)) { + ICMP6_INC_STATS_BH(__in6_dev_get(skb->dev), ICMP6_MIB_INERRORS); + return; + } + sk = inet6_lookup(net, &dccp_hashinfo, &hdr->daddr, dh->dccph_dport, &hdr->saddr, dh->dccph_sport, inet6_iif(skb)); -- cgit v1.2.1 From 0bc3cc03fa6e1c20aecb5a33356bcaae410640b9 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Thu, 24 Jul 2008 18:21:31 -0700 Subject: cpumask: change cpumask_of_cpu_ptr to use new cpumask_of_cpu * Replace previous instances of the cpumask_of_cpu_ptr* macros with a the new (lvalue capable) generic cpumask_of_cpu(). Signed-off-by: Mike Travis Cc: Andrew Morton Cc: Jack Steiner Cc: Rusty Russell Signed-off-by: Ingo Molnar --- net/sunrpc/svc.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'net') diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c index 835d27413083..5a32cb7c4bb4 100644 --- a/net/sunrpc/svc.c +++ b/net/sunrpc/svc.c @@ -310,8 +310,7 @@ svc_pool_map_set_cpumask(struct task_struct *task, unsigned int pidx) switch (m->mode) { case SVC_POOL_PERCPU: { - cpumask_of_cpu_ptr(cpumask, node); - set_cpus_allowed_ptr(task, cpumask); + set_cpus_allowed_ptr(task, &cpumask_of_cpu(node)); break; } case SVC_POOL_PERNODE: -- cgit v1.2.1 From 2c3abab7c95295f319dc8899b74cbd60140fcdfb Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sun, 27 Jul 2008 03:59:24 -0700 Subject: ipcomp: Fix warnings after ipcomp consolidation. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit net/ipv4/ipcomp.c: In function ‘ipcomp4_init_state’: net/ipv4/ipcomp.c:109: warning: unused variable ‘calg_desc’ net/ipv4/ipcomp.c:108: warning: unused variable ‘ipcd’ net/ipv4/ipcomp.c:107: warning: ‘err’ may be used uninitialized in this function net/ipv6/ipcomp6.c: In function ‘ipcomp6_init_state’: net/ipv6/ipcomp6.c:139: warning: unused variable ‘calg_desc’ net/ipv6/ipcomp6.c:138: warning: unused variable ‘ipcd’ net/ipv6/ipcomp6.c:137: warning: ‘err’ may be used uninitialized in this function Signed-off-by: David S. Miller --- net/ipv4/ipcomp.c | 4 +--- net/ipv6/ipcomp6.c | 4 +--- 2 files changed, 2 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipcomp.c b/net/ipv4/ipcomp.c index a42b64d040c4..38ccb6dfb02e 100644 --- a/net/ipv4/ipcomp.c +++ b/net/ipv4/ipcomp.c @@ -104,9 +104,7 @@ out: static int ipcomp4_init_state(struct xfrm_state *x) { - int err; - struct ipcomp_data *ipcd; - struct xfrm_algo_desc *calg_desc; + int err = -EINVAL; x->props.header_len = 0; switch (x->props.mode) { diff --git a/net/ipv6/ipcomp6.c b/net/ipv6/ipcomp6.c index 0cfcea42153a..4545e4306862 100644 --- a/net/ipv6/ipcomp6.c +++ b/net/ipv6/ipcomp6.c @@ -134,9 +134,7 @@ out: static int ipcomp6_init_state(struct xfrm_state *x) { - int err; - struct ipcomp_data *ipcd; - struct xfrm_algo_desc *calg_desc; + int err = -EINVAL; x->props.header_len = 0; switch (x->props.mode) { -- cgit v1.2.1 From 6f9f489a4eeaa3c8a8618e078a5270d2c4872b67 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 27 Jul 2008 04:40:51 -0700 Subject: net: missing bits of net-namespace / sysctl Piss-poor sysctl registration API strikes again, film at 11... What we really need is _pathname_ required to be present in already registered table, so that kernel could warn about bad order. That's the next target for sysctl stuff (and generally saner and more explicit order of initialization of ipv[46] internals wouldn't hurt either). For the time being, here are full fixups required by ..._rotable() stuff; we make per-net sysctl sets descendents of "ro" one and make sure that sufficient skeleton is there before we start registering per-net sysctls. Signed-off-by: Al Viro Signed-off-by: David S. Miller --- net/ipv4/route.c | 11 ++++++++++- net/ipv4/sysctl_net_ipv4.c | 14 -------------- net/ipv6/af_inet6.c | 12 ++++++++++++ net/ipv6/sysctl_net_ipv6.c | 16 ++++++++++++++++ net/sysctl_net.c | 4 +++- 5 files changed, 41 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index a507c5e27d0e..380d6474cf66 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -2914,7 +2914,7 @@ static int ipv4_sysctl_rtcache_flush_strategy(ctl_table *table, return 0; } -ctl_table ipv4_route_table[] = { +static ctl_table ipv4_route_table[] = { { .ctl_name = NET_IPV4_ROUTE_GC_THRESH, .procname = "gc_thresh", @@ -3216,6 +3216,15 @@ int __init ip_rt_init(void) return rc; } +/* + * We really need to sanitize the damn ipv4 init order, then all + * this nonsense will go away. + */ +void __init ip_static_sysctl_init(void) +{ + register_sysctl_paths(ipv4_route_path, ipv4_route_table); +} + EXPORT_SYMBOL(__ip_select_ident); EXPORT_SYMBOL(ip_route_input); EXPORT_SYMBOL(ip_route_output_key); diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c index d63e9388d92d..770d827f5ab8 100644 --- a/net/ipv4/sysctl_net_ipv4.c +++ b/net/ipv4/sysctl_net_ipv4.c @@ -401,13 +401,6 @@ static struct ctl_table ipv4_table[] = { .proc_handler = &ipv4_local_port_range, .strategy = &ipv4_sysctl_local_port_range, }, - { - .ctl_name = NET_IPV4_ROUTE, - .procname = "route", - .maxlen = 0, - .mode = 0555, - .child = ipv4_route_table - }, #ifdef CONFIG_IP_MULTICAST { .ctl_name = NET_IPV4_IGMP_MAX_MEMBERSHIPS, @@ -882,11 +875,4 @@ static __init int sysctl_ipv4_init(void) return 0; } -/* set enough of tree skeleton to get rid of ordering problems */ -void __init ip_static_sysctl_init(void) -{ - static ctl_table table[1]; - register_sysctl_paths(net_ipv4_ctl_path, table); -} - __initcall(sysctl_ipv4_init); diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index c708ca842298..95055f8c3f35 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -934,6 +934,11 @@ static int __init inet6_init(void) if (err) goto out_unregister_sock; +#ifdef CONFIG_SYSCTL + err = ipv6_static_sysctl_register(); + if (err) + goto static_sysctl_fail; +#endif /* * ipngwg API draft makes clear that the correct semantics * for TCP and UDP is to consider one TCP and UDP instance @@ -1058,6 +1063,10 @@ ipmr_fail: icmp_fail: unregister_pernet_subsys(&inet6_net_ops); register_pernet_fail: +#ifdef CONFIG_SYSCTL + ipv6_static_sysctl_unregister(); +static_sysctl_fail: +#endif cleanup_ipv6_mibs(); out_unregister_sock: sock_unregister(PF_INET6); @@ -1113,6 +1122,9 @@ static void __exit inet6_exit(void) rawv6_exit(); unregister_pernet_subsys(&inet6_net_ops); +#ifdef CONFIG_SYSCTL + ipv6_static_sysctl_unregister(); +#endif cleanup_ipv6_mibs(); proto_unregister(&rawv6_prot); proto_unregister(&udplitev6_prot); diff --git a/net/ipv6/sysctl_net_ipv6.c b/net/ipv6/sysctl_net_ipv6.c index 5c99274558bf..e6dfaeac6be3 100644 --- a/net/ipv6/sysctl_net_ipv6.c +++ b/net/ipv6/sysctl_net_ipv6.c @@ -150,3 +150,19 @@ void ipv6_sysctl_unregister(void) unregister_net_sysctl_table(ip6_header); unregister_pernet_subsys(&ipv6_sysctl_net_ops); } + +static struct ctl_table_header *ip6_base; + +int ipv6_static_sysctl_register(void) +{ + static struct ctl_table empty[1]; + ip6_base = register_net_sysctl_rotable(net_ipv6_ctl_path, empty); + if (ip6_base == NULL) + return -ENOMEM; + return 0; +} + +void ipv6_static_sysctl_unregister(void) +{ + unregister_net_sysctl_table(ip6_base); +} diff --git a/net/sysctl_net.c b/net/sysctl_net.c index cefbc367d8be..972201cd5fa7 100644 --- a/net/sysctl_net.c +++ b/net/sysctl_net.c @@ -73,7 +73,9 @@ static struct ctl_table_root net_sysctl_ro_root = { static int sysctl_net_init(struct net *net) { - setup_sysctl_set(&net->sysctls, NULL, is_seen); + setup_sysctl_set(&net->sysctls, + &net_sysctl_ro_root.default_set, + is_seen); return 0; } -- cgit v1.2.1 From 2fd9b2212e25e6411b6f309707f4e2683d164250 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 21 Jul 2008 21:18:17 -0300 Subject: rfkill: document rfkill_force_state as required (v2) While the rfkill class does work with just get_state(), it doesn't work well on devices that are subject to external events that cause rfkill state changes. Document that rfkill_force_state() is required in those cases. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 7a560b785097..022fe50ab0e0 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -252,7 +252,12 @@ EXPORT_SYMBOL_GPL(rfkill_epo); * a notification by the firmware/hardware of the current *real* * state of the radio rfkill switch. * - * It may not be called from an atomic context. + * Devices which are subject to external changes on their rfkill + * state (such as those caused by a hardware rfkill line) MUST + * have their driver arrange to call rfkill_force_state() as soon + * as possible after such a change. + * + * This function may not be called from an atomic context. */ int rfkill_force_state(struct rfkill *rfkill, enum rfkill_state state) { -- cgit v1.2.1 From 37f55e9d78d1b63047b1b7ae175cdce650547ba8 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 21 Jul 2008 21:18:18 -0300 Subject: rfkill: fix led-trigger unregister order in error unwind rfkill needs to unregister the led trigger AFTER a call to rfkill_remove_switch(), otherwise it will not update the LED state, possibly leaving it ON when it should be OFF. To make led-trigger unregistering safer, guard against unregistering a trigger twice, and also against issuing trigger events to a led trigger that was unregistered. This makes the error unwind paths more resilient. Refer to "rfkill: Register LED triggers before registering switch". Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Michael Buesch Cc: Dmitry Baryshkov Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 022fe50ab0e0..fc3a4fd88995 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -590,8 +590,10 @@ static void rfkill_led_trigger_register(struct rfkill *rfkill) static void rfkill_led_trigger_unregister(struct rfkill *rfkill) { #ifdef CONFIG_RFKILL_LEDS - if (rfkill->led_trigger.name) + if (rfkill->led_trigger.name) { led_trigger_unregister(&rfkill->led_trigger); + rfkill->led_trigger.name = NULL; + } #endif } @@ -627,8 +629,8 @@ int rfkill_register(struct rfkill *rfkill) error = device_add(dev); if (error) { - rfkill_led_trigger_unregister(rfkill); rfkill_remove_switch(rfkill); + rfkill_led_trigger_unregister(rfkill); return error; } -- cgit v1.2.1 From 064af1117b4aa64a0e52f6b741df7356ef055142 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 21 Jul 2008 21:18:20 -0300 Subject: rfkill: mutex fixes There are two mutexes in rfkill: rfkill->mutex, which protects some of the fields of a rfkill struct, and is also used for callback serialization. rfkill_mutex, which protects the global state, the list of registered rfkill structs and rfkill->claim. Make sure to use the correct mutex, and to not miss locking rfkill->mutex even when we already took rfkill_mutex. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 29 +++++++++++++++++++---------- 1 file changed, 19 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index fc3a4fd88995..ac205eceaf42 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -150,7 +150,7 @@ static void update_rfkill_state(struct rfkill *rfkill) * even if the radio is in RFKILL_STATE_HARD_BLOCKED state, so as to * give the driver a hint that it should double-BLOCK the transmitter. * - * Caller must have aquired rfkill_mutex. + * Caller must have acquired rfkill->mutex. */ static int rfkill_toggle_radio(struct rfkill *rfkill, enum rfkill_state state, @@ -216,8 +216,11 @@ void rfkill_switch_all(enum rfkill_type type, enum rfkill_state state) rfkill_states[type] = state; list_for_each_entry(rfkill, &rfkill_list, node) { - if ((!rfkill->user_claim) && (rfkill->type == type)) + if ((!rfkill->user_claim) && (rfkill->type == type)) { + mutex_lock(&rfkill->mutex); rfkill_toggle_radio(rfkill, state, 0); + mutex_unlock(&rfkill->mutex); + } } mutex_unlock(&rfkill_mutex); @@ -228,7 +231,7 @@ EXPORT_SYMBOL(rfkill_switch_all); * rfkill_epo - emergency power off all transmitters * * This kicks all rfkill devices to RFKILL_STATE_SOFT_BLOCKED, ignoring - * everything in its path but rfkill_mutex. + * everything in its path but rfkill_mutex and rfkill->mutex. */ void rfkill_epo(void) { @@ -236,7 +239,9 @@ void rfkill_epo(void) mutex_lock(&rfkill_mutex); list_for_each_entry(rfkill, &rfkill_list, node) { + mutex_lock(&rfkill->mutex); rfkill_toggle_radio(rfkill, RFKILL_STATE_SOFT_BLOCKED, 1); + mutex_unlock(&rfkill->mutex); } mutex_unlock(&rfkill_mutex); } @@ -372,6 +377,9 @@ static ssize_t rfkill_claim_store(struct device *dev, if (!capable(CAP_NET_ADMIN)) return -EPERM; + if (rfkill->user_claim_unsupported) + return -EOPNOTSUPP; + /* * Take the global lock to make sure the kernel is not in * the middle of rfkill_switch_all @@ -380,19 +388,17 @@ static ssize_t rfkill_claim_store(struct device *dev, if (error) return error; - if (rfkill->user_claim_unsupported) { - error = -EOPNOTSUPP; - goto out_unlock; - } if (rfkill->user_claim != claim) { - if (!claim) + if (!claim) { + mutex_lock(&rfkill->mutex); rfkill_toggle_radio(rfkill, rfkill_states[rfkill->type], 0); + mutex_unlock(&rfkill->mutex); + } rfkill->user_claim = claim; } -out_unlock: mutex_unlock(&rfkill_mutex); return error ? error : count; @@ -521,8 +527,11 @@ static void rfkill_remove_switch(struct rfkill *rfkill) { mutex_lock(&rfkill_mutex); list_del_init(&rfkill->node); - rfkill_toggle_radio(rfkill, RFKILL_STATE_SOFT_BLOCKED, 1); mutex_unlock(&rfkill_mutex); + + mutex_lock(&rfkill->mutex); + rfkill_toggle_radio(rfkill, RFKILL_STATE_SOFT_BLOCKED, 1); + mutex_unlock(&rfkill->mutex); } /** -- cgit v1.2.1 From 435307a365ceedc4f4e1813e405f583f434d98e4 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 21 Jul 2008 21:18:22 -0300 Subject: rfkill: yet more minor kernel-doc fixes For some stupid reason, I sent and old version of the patch minor kernel doc-fix patch, and it got merged before I noticed the problem. This is an incremental fix on top. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index ac205eceaf42..c6f2f388cb72 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -130,7 +130,6 @@ static void update_rfkill_state(struct rfkill *rfkill) /** * rfkill_toggle_radio - wrapper for toggle_radio hook - * * @rfkill: the rfkill struct to use * @force: calls toggle_radio even if cache says it is not needed, * and also makes sure notifications of the state will be @@ -141,8 +140,8 @@ static void update_rfkill_state(struct rfkill *rfkill) * calls and handling all the red tape such as issuing notifications * if the call is successful. * - * Note that @force cannot override a (possibly cached) state of - * RFKILL_STATE_HARD_BLOCKED. Any device making use of + * Note that the @force parameter cannot override a (possibly cached) + * state of RFKILL_STATE_HARD_BLOCKED. Any device making use of * RFKILL_STATE_HARD_BLOCKED implements either get_state() or * rfkill_force_state(), so the cache either is bypassed or valid. * @@ -200,12 +199,12 @@ static int rfkill_toggle_radio(struct rfkill *rfkill, /** * rfkill_switch_all - Toggle state of all switches of given type - * @type: type of interfaces to be affeceted + * @type: type of interfaces to be affected * @state: the new state * - * This function toggles state of all switches of given type unless - * a specific switch is claimed by userspace in which case it is - * left alone. + * This function toggles the state of all switches of given type, + * unless a specific switch is claimed by userspace (in which case, + * that switch is left alone). */ void rfkill_switch_all(enum rfkill_type type, enum rfkill_state state) { @@ -540,9 +539,10 @@ static void rfkill_remove_switch(struct rfkill *rfkill) * @type: type of the switch (RFKILL_TYPE_*) * * This function should be called by the network driver when it needs - * rfkill structure. Once the structure is allocated the driver shoud - * finish its initialization by setting name, private data, enable_radio + * rfkill structure. Once the structure is allocated the driver should + * finish its initialization by setting the name, private data, enable_radio * and disable_radio methods and then register it with rfkill_register(). + * * NOTE: If registration fails the structure shoudl be freed by calling * rfkill_free() otherwise rfkill_unregister() should be used. */ @@ -574,7 +574,7 @@ EXPORT_SYMBOL(rfkill_allocate); * rfkill_free - Mark rfkill structure for deletion * @rfkill: rfkill structure to be destroyed * - * Decrements reference count of rfkill structure so it is destroyed. + * Decrements reference count of the rfkill structure so it is destroyed. * Note that rfkill_free() should _not_ be called after rfkill_unregister(). */ void rfkill_free(struct rfkill *rfkill) -- cgit v1.2.1 From 1b0241656b658522a15e7aad570cb8ea6b255a2a Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Mon, 14 Jul 2008 12:43:23 +0200 Subject: mac80211: tx, use dev_kfree_skb_any for beacon_get Use dev_kfree_skb_any(); instead of dev_kfree_skb();, since ieee80211_beacon_get function might be called from atomic. (It's in a fail path.) Signed-off-by: Jiri Slaby Cc: Johannes Berg Cc: Michael Wu Signed-off-by: John W. Linville --- net/mac80211/tx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 0fbadd8b983c..7e99d1615192 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1931,7 +1931,7 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, "no rate found\n", wiphy_name(local->hw.wiphy)); } - dev_kfree_skb(skb); + dev_kfree_skb_any(skb); skb = NULL; goto out; } -- cgit v1.2.1 From 023a04bebe7030c1e6d5347bd3f27a3e49a1f222 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Mon, 14 Jul 2008 12:52:08 -0700 Subject: mac80211: return correct error return from ieee80211_wep_init Return the proper error code rather than a hard-coded ENOMEM from ieee80211_wep_init. Also, print the error code on failure. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: John W. Linville --- net/mac80211/main.c | 4 ++-- net/mac80211/wep.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index f1a83d450ea0..b5830f7055cf 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1731,8 +1731,8 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) result = ieee80211_wep_init(local); if (result < 0) { - printk(KERN_DEBUG "%s: Failed to initialize wep\n", - wiphy_name(local->hw.wiphy)); + printk(KERN_DEBUG "%s: Failed to initialize wep: %d\n", + wiphy_name(local->hw.wiphy), result); goto fail_wep; } diff --git a/net/mac80211/wep.c b/net/mac80211/wep.c index 872d2fcd1a5b..5c2bf0a3d4db 100644 --- a/net/mac80211/wep.c +++ b/net/mac80211/wep.c @@ -31,13 +31,13 @@ int ieee80211_wep_init(struct ieee80211_local *local) local->wep_tx_tfm = crypto_alloc_blkcipher("ecb(arc4)", 0, CRYPTO_ALG_ASYNC); if (IS_ERR(local->wep_tx_tfm)) - return -ENOMEM; + return PTR_ERR(local->wep_tx_tfm); local->wep_rx_tfm = crypto_alloc_blkcipher("ecb(arc4)", 0, CRYPTO_ALG_ASYNC); if (IS_ERR(local->wep_rx_tfm)) { crypto_free_blkcipher(local->wep_tx_tfm); - return -ENOMEM; + return PTR_ERR(local->wep_rx_tfm); } return 0; -- cgit v1.2.1 From 5422399518e8142198df888aab00acdac251f754 Mon Sep 17 00:00:00 2001 From: Rami Rosen Date: Thu, 24 Jul 2008 10:40:37 +0300 Subject: mac80211: append CONFIG_ to MAC80211_VERBOSE_PS_DEBUG in net/mac80211/tx.c. In net/mac80211/tx.c, there are some #ifdef which checks MAC80211_VERBOSE_PS_DEBUG (which in fact is never set) instead of CONFIG_MAC80211_VERBOSE_PS_DEBUG, as should be. This patch replaces MAC80211_VERBOSE_PS_DEBUG with CONFIG_MAC80211_VERBOSE_PS_DEBUG in these #ifdef commands in net/mac80211/tx.c. Signed-off-by: Rami Rosen Signed-off-by: John W. Linville --- net/mac80211/tx.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 7e99d1615192..c5f78059c6ca 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -305,7 +305,7 @@ static void purge_old_ps_buffers(struct ieee80211_local *local) rcu_read_unlock(); local->total_ps_buffered = total; -#ifdef MAC80211_VERBOSE_PS_DEBUG +#ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG printk(KERN_DEBUG "%s: PS buffers full - purged %d frames\n", wiphy_name(local->hw.wiphy), purged); #endif @@ -342,7 +342,7 @@ ieee80211_tx_h_multicast_ps_buf(struct ieee80211_tx_data *tx) purge_old_ps_buffers(tx->local); if (skb_queue_len(&tx->sdata->bss->ps_bc_buf) >= AP_MAX_BC_BUFFER) { -#ifdef MAC80211_VERBOSE_PS_DEBUG +#ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG if (net_ratelimit()) { printk(KERN_DEBUG "%s: BC TX buffer full - " "dropping the oldest frame\n", @@ -389,7 +389,7 @@ ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) purge_old_ps_buffers(tx->local); if (skb_queue_len(&sta->ps_tx_buf) >= STA_MAX_TX_BUFFER) { struct sk_buff *old = skb_dequeue(&sta->ps_tx_buf); -#ifdef MAC80211_VERBOSE_PS_DEBUG +#ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG if (net_ratelimit()) { printk(KERN_DEBUG "%s: STA %s TX " "buffer full - dropping oldest frame\n", -- cgit v1.2.1 From d0f09804144fd9471a13cf4d80e66842c7fa114f Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Tue, 29 Jul 2008 11:32:07 +0200 Subject: mac80211: partially fix skb->cb use This patch fixes mac80211 to not use the skb->cb over the queue step from virtual interfaces to the master. The patch also, for now, disables aggregation because that would still require requeuing, will fix that in a separate patch. There are two other places (software requeue and powersaving stations) where requeue can happen, but that is not currently used by any drivers/not possible to use respectively. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/core/skbuff.c | 3 +++ net/mac80211/main.c | 8 +------- net/mac80211/mlme.c | 8 ++------ net/mac80211/tx.c | 47 +++++++++++++++++++++-------------------------- net/mac80211/wme.c | 3 +++ 5 files changed, 30 insertions(+), 39 deletions(-) (limited to 'net') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 4e0c92274189..84640172d65d 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -485,6 +485,9 @@ static struct sk_buff *__skb_clone(struct sk_buff *n, struct sk_buff *skb) C(head); C(data); C(truesize); +#if defined(CONFIG_MAC80211) || defined(CONFIG_MAC80211_MODULE) + C(do_not_encrypt); +#endif atomic_set(&n->users, 1); atomic_inc(&(skb_shinfo(skb)->dataref)); diff --git a/net/mac80211/main.c b/net/mac80211/main.c index b5830f7055cf..a4c5b90de769 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1233,18 +1233,12 @@ static void ieee80211_tasklet_handler(unsigned long data) /* Remove added headers (e.g., QoS control), encryption header/MIC, etc. to * make a prepared TX frame (one that has been given to hw) to look like brand * new IEEE 802.11 frame that is ready to go through TX processing again. - * Also, tx_packet_data in cb is restored from tx_control. */ + */ static void ieee80211_remove_tx_extra(struct ieee80211_local *local, struct ieee80211_key *key, struct sk_buff *skb) { int hdrlen, iv_len, mic_len; - struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); - - info->flags &= IEEE80211_TX_CTL_REQ_TX_STATUS | - IEEE80211_TX_CTL_DO_NOT_ENCRYPT | - IEEE80211_TX_CTL_REQUEUE | - IEEE80211_TX_CTL_EAPOL_FRAME; hdrlen = ieee80211_get_hdrlen_from_skb(skb); diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index d7c371e36bf0..35eb767cbcbe 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -606,7 +606,6 @@ void ieee80211_sta_tx(struct net_device *dev, struct sk_buff *skb, int encrypt) { struct ieee80211_sub_if_data *sdata; - struct ieee80211_tx_info *info; sdata = IEEE80211_DEV_TO_SUB_IF(dev); skb->dev = sdata->local->mdev; @@ -614,11 +613,8 @@ void ieee80211_sta_tx(struct net_device *dev, struct sk_buff *skb, skb_set_network_header(skb, 0); skb_set_transport_header(skb, 0); - info = IEEE80211_SKB_CB(skb); - memset(info, 0, sizeof(struct ieee80211_tx_info)); - info->control.ifindex = sdata->dev->ifindex; - if (!encrypt) - info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; + skb->iif = sdata->dev->ifindex; + skb->do_not_encrypt = !encrypt; dev_queue_xmit(skb); } diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index c5f78059c6ca..69019e943873 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -439,14 +439,14 @@ ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx) struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); u16 fc = tx->fc; - if (unlikely(info->flags & IEEE80211_TX_CTL_DO_NOT_ENCRYPT)) + if (unlikely(tx->skb->do_not_encrypt)) tx->key = NULL; else if (tx->sta && (key = rcu_dereference(tx->sta->key))) tx->key = key; else if ((key = rcu_dereference(tx->sdata->default_key))) tx->key = key; else if (tx->sdata->drop_unencrypted && - !(info->flags & IEEE80211_TX_CTL_EAPOL_FRAME) && + (tx->skb->protocol != cpu_to_be16(ETH_P_PAE)) && !(info->flags & IEEE80211_TX_CTL_INJECTED)) { I802_DEBUG_INC(tx->local->tx_handlers_drop_unencrypted); return TX_DROP; @@ -476,7 +476,7 @@ ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx) } if (!tx->key || !(tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE)) - info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; + tx->skb->do_not_encrypt = 1; return TX_CONTINUE; } @@ -732,6 +732,7 @@ ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) memcpy(skb_put(frag, copylen), pos, copylen); memcpy(frag->cb, first->cb, sizeof(frag->cb)); skb_copy_queue_mapping(frag, first); + frag->do_not_encrypt = first->do_not_encrypt; pos += copylen; left -= copylen; @@ -852,7 +853,7 @@ __ieee80211_parse_tx_radiotap(struct ieee80211_tx_data *tx, sband = tx->local->hw.wiphy->bands[tx->channel->band]; - info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; + skb->do_not_encrypt = 1; info->flags |= IEEE80211_TX_CTL_INJECTED; tx->flags &= ~IEEE80211_TX_FRAGMENTED; @@ -925,8 +926,7 @@ __ieee80211_parse_tx_radiotap(struct ieee80211_tx_data *tx, skb_trim(skb, skb->len - FCS_LEN); } if (*iterator.this_arg & IEEE80211_RADIOTAP_F_WEP) - info->flags &= - ~IEEE80211_TX_CTL_DO_NOT_ENCRYPT; + tx->skb->do_not_encrypt = 0; if (*iterator.this_arg & IEEE80211_RADIOTAP_F_FRAG) tx->flags |= IEEE80211_TX_FRAGMENTED; break; @@ -1042,10 +1042,9 @@ static int ieee80211_tx_prepare(struct ieee80211_tx_data *tx, struct sk_buff *skb, struct net_device *mdev) { - struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct net_device *dev; - dev = dev_get_by_index(&init_net, info->control.ifindex); + dev = dev_get_by_index(&init_net, skb->iif); if (unlikely(dev && !is_ieee80211_device(dev, mdev))) { dev_put(dev); dev = NULL; @@ -1306,8 +1305,8 @@ int ieee80211_master_start_xmit(struct sk_buff *skb, bool may_encrypt; int ret; - if (info->control.ifindex) - odev = dev_get_by_index(&init_net, info->control.ifindex); + if (skb->iif) + odev = dev_get_by_index(&init_net, skb->iif); if (unlikely(odev && !is_ieee80211_device(odev, dev))) { dev_put(odev); odev = NULL; @@ -1321,9 +1320,13 @@ int ieee80211_master_start_xmit(struct sk_buff *skb, return 0; } + memset(info, 0, sizeof(*info)); + + info->flags |= IEEE80211_TX_CTL_REQ_TX_STATUS; + osdata = IEEE80211_DEV_TO_SUB_IF(odev); - may_encrypt = !(info->flags & IEEE80211_TX_CTL_DO_NOT_ENCRYPT); + may_encrypt = !skb->do_not_encrypt; headroom = osdata->local->tx_headroom; if (may_encrypt) @@ -1348,7 +1351,6 @@ int ieee80211_monitor_start_xmit(struct sk_buff *skb, struct net_device *dev) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct ieee80211_radiotap_header *prthdr = (struct ieee80211_radiotap_header *)skb->data; u16 len_rthdr; @@ -1371,11 +1373,11 @@ int ieee80211_monitor_start_xmit(struct sk_buff *skb, skb->dev = local->mdev; /* needed because we set skb device to master */ - info->control.ifindex = dev->ifindex; + skb->iif = dev->ifindex; - info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; - /* Interfaces should always request a status report */ - info->flags |= IEEE80211_TX_CTL_REQ_TX_STATUS; + /* sometimes we do encrypt injected frames, will be fixed + * up in radiotap parser if not wanted */ + skb->do_not_encrypt = 0; /* * fix up the pointers accounting for the radiotap @@ -1419,7 +1421,6 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, struct net_device *dev) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_tx_info *info; struct ieee80211_sub_if_data *sdata; int ret = 1, head_need; u16 ethertype, hdrlen, meshhdrlen = 0; @@ -1645,14 +1646,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, nh_pos += hdrlen; h_pos += hdrlen; - info = IEEE80211_SKB_CB(skb); - memset(info, 0, sizeof(*info)); - info->control.ifindex = dev->ifindex; - if (ethertype == ETH_P_PAE) - info->flags |= IEEE80211_TX_CTL_EAPOL_FRAME; - - /* Interfaces should always request a status report */ - info->flags |= IEEE80211_TX_CTL_REQ_TX_STATUS; + skb->iif = dev->ifindex; skb->dev = local->mdev; dev->stats.tx_packets++; @@ -1922,6 +1916,8 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, info = IEEE80211_SKB_CB(skb); + skb->do_not_encrypt = 1; + info->band = band; rate_control_get_rate(local->mdev, sband, skb, &rsel); @@ -1940,7 +1936,6 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, info->tx_rate_idx = rsel.rate_idx; info->flags |= IEEE80211_TX_CTL_NO_ACK; - info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; info->flags |= IEEE80211_TX_CTL_CLEAR_PS_FILT; info->flags |= IEEE80211_TX_CTL_ASSIGN_SEQ; if (sdata->bss_conf.use_short_preamble && diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 07edda0b8a5c..28437f0001db 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -188,6 +188,9 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, { int i; + /* XXX: currently broken due to cb/requeue use */ + return -EPERM; + /* prepare the filter and save it for the SW queue * matching the received HW queue */ -- cgit v1.2.1 From bba95fefb8e31f4799652666d05a4a9aad56e492 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Tue, 29 Jul 2008 13:22:51 +0200 Subject: nl80211: fix dump callbacks Julius Volz pointed out that the dump callbacks in nl80211 were broken and fixed one of them. This patch fixes the other three and also addresses the TODOs there. Signed-off-by: Johannes Berg Cc: Julius Volz Cc: Thomas Graf Signed-off-by: John W. Linville --- net/wireless/nl80211.c | 275 +++++++++++++++++++++++++++---------------------- 1 file changed, 153 insertions(+), 122 deletions(-) (limited to 'net') diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c index b7fefffd2d0d..59eb2cf42e5f 100644 --- a/net/wireless/nl80211.c +++ b/net/wireless/nl80211.c @@ -29,16 +29,16 @@ static struct genl_family nl80211_fam = { }; /* internal helper: get drv and dev */ -static int get_drv_dev_by_info_ifindex(struct genl_info *info, +static int get_drv_dev_by_info_ifindex(struct nlattr **attrs, struct cfg80211_registered_device **drv, struct net_device **dev) { int ifindex; - if (!info->attrs[NL80211_ATTR_IFINDEX]) + if (!attrs[NL80211_ATTR_IFINDEX]) return -EINVAL; - ifindex = nla_get_u32(info->attrs[NL80211_ATTR_IFINDEX]); + ifindex = nla_get_u32(attrs[NL80211_ATTR_IFINDEX]); *dev = dev_get_by_index(&init_net, ifindex); if (!*dev) return -ENODEV; @@ -291,21 +291,31 @@ static int nl80211_dump_interface(struct sk_buff *skb, struct netlink_callback * mutex_lock(&cfg80211_drv_mutex); list_for_each_entry(dev, &cfg80211_drv_list, list) { - if (++wp_idx < wp_start) + if (wp_idx < wp_start) { + wp_idx++; continue; + } if_idx = 0; mutex_lock(&dev->devlist_mtx); list_for_each_entry(wdev, &dev->netdev_list, list) { - if (++if_idx < if_start) + if (if_idx < if_start) { + if_idx++; continue; + } if (nl80211_send_iface(skb, NETLINK_CB(cb->skb).pid, cb->nlh->nlmsg_seq, NLM_F_MULTI, - wdev->netdev) < 0) - break; + wdev->netdev) < 0) { + mutex_unlock(&dev->devlist_mtx); + goto out; + } + if_idx++; } mutex_unlock(&dev->devlist_mtx); + + wp_idx++; } + out: mutex_unlock(&cfg80211_drv_mutex); cb->args[0] = wp_idx; @@ -321,7 +331,7 @@ static int nl80211_get_interface(struct sk_buff *skb, struct genl_info *info) struct net_device *netdev; int err; - err = get_drv_dev_by_info_ifindex(info, &dev, &netdev); + err = get_drv_dev_by_info_ifindex(info->attrs, &dev, &netdev); if (err) return err; @@ -392,7 +402,7 @@ static int nl80211_set_interface(struct sk_buff *skb, struct genl_info *info) } else return -EINVAL; - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; ifindex = dev->ifindex; @@ -477,7 +487,7 @@ static int nl80211_del_interface(struct sk_buff *skb, struct genl_info *info) int ifindex, err; struct net_device *dev; - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; ifindex = dev->ifindex; @@ -545,7 +555,7 @@ static int nl80211_get_key(struct sk_buff *skb, struct genl_info *info) if (info->attrs[NL80211_ATTR_MAC]) mac_addr = nla_data(info->attrs[NL80211_ATTR_MAC]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -618,7 +628,7 @@ static int nl80211_set_key(struct sk_buff *skb, struct genl_info *info) if (!info->attrs[NL80211_ATTR_KEY_DEFAULT]) return -EINVAL; - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -699,7 +709,7 @@ static int nl80211_new_key(struct sk_buff *skb, struct genl_info *info) return -EINVAL; } - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -735,7 +745,7 @@ static int nl80211_del_key(struct sk_buff *skb, struct genl_info *info) if (info->attrs[NL80211_ATTR_MAC]) mac_addr = nla_data(info->attrs[NL80211_ATTR_MAC]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -764,7 +774,7 @@ static int nl80211_addset_beacon(struct sk_buff *skb, struct genl_info *info) struct beacon_parameters params; int haveinfo = 0; - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -843,7 +853,7 @@ static int nl80211_del_beacon(struct sk_buff *skb, struct genl_info *info) int err; struct net_device *dev; - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -937,67 +947,78 @@ static int nl80211_send_station(struct sk_buff *msg, u32 pid, u32 seq, } static int nl80211_dump_station(struct sk_buff *skb, - struct netlink_callback *cb) + struct netlink_callback *cb) { - int wp_idx = 0; - int if_idx = 0; - int sta_idx = cb->args[2]; - int wp_start = cb->args[0]; - int if_start = cb->args[1]; struct station_info sinfo; struct cfg80211_registered_device *dev; - struct wireless_dev *wdev; + struct net_device *netdev; u8 mac_addr[ETH_ALEN]; + int ifidx = cb->args[0]; + int sta_idx = cb->args[1]; int err; - int exit = 0; - /* TODO: filter by device */ - mutex_lock(&cfg80211_drv_mutex); - list_for_each_entry(dev, &cfg80211_drv_list, list) { - if (exit) + if (!ifidx) { + err = nlmsg_parse(cb->nlh, GENL_HDRLEN + nl80211_fam.hdrsize, + nl80211_fam.attrbuf, nl80211_fam.maxattr, + nl80211_policy); + if (err) + return err; + + if (!nl80211_fam.attrbuf[NL80211_ATTR_IFINDEX]) + return -EINVAL; + + ifidx = nla_get_u32(nl80211_fam.attrbuf[NL80211_ATTR_IFINDEX]); + if (!ifidx) + return -EINVAL; + } + + netdev = dev_get_by_index(&init_net, ifidx); + if (!netdev) + return -ENODEV; + + dev = cfg80211_get_dev_from_ifindex(ifidx); + if (IS_ERR(dev)) { + err = PTR_ERR(dev); + goto out_put_netdev; + } + + if (!dev->ops->dump_station) { + err = -ENOSYS; + goto out_err; + } + + rtnl_lock(); + + while (1) { + err = dev->ops->dump_station(&dev->wiphy, netdev, sta_idx, + mac_addr, &sinfo); + if (err == -ENOENT) break; - if (++wp_idx < wp_start) - continue; - if_idx = 0; + if (err) + goto out_err_rtnl; - mutex_lock(&dev->devlist_mtx); - list_for_each_entry(wdev, &dev->netdev_list, list) { - if (exit) - break; - if (++if_idx < if_start) - continue; - if (!dev->ops->dump_station) - continue; + if (nl80211_send_station(skb, + NETLINK_CB(cb->skb).pid, + cb->nlh->nlmsg_seq, NLM_F_MULTI, + netdev, mac_addr, + &sinfo) < 0) + goto out; - for (;; ++sta_idx) { - rtnl_lock(); - err = dev->ops->dump_station(&dev->wiphy, - wdev->netdev, sta_idx, mac_addr, - &sinfo); - rtnl_unlock(); - if (err) { - sta_idx = 0; - break; - } - if (nl80211_send_station(skb, - NETLINK_CB(cb->skb).pid, - cb->nlh->nlmsg_seq, NLM_F_MULTI, - wdev->netdev, mac_addr, - &sinfo) < 0) { - exit = 1; - break; - } - } - } - mutex_unlock(&dev->devlist_mtx); + sta_idx++; } - mutex_unlock(&cfg80211_drv_mutex); - cb->args[0] = wp_idx; - cb->args[1] = if_idx; - cb->args[2] = sta_idx; - return skb->len; + out: + cb->args[1] = sta_idx; + err = skb->len; + out_err_rtnl: + rtnl_unlock(); + out_err: + cfg80211_put_dev(dev); + out_put_netdev: + dev_put(netdev); + + return err; } static int nl80211_get_station(struct sk_buff *skb, struct genl_info *info) @@ -1016,7 +1037,7 @@ static int nl80211_get_station(struct sk_buff *skb, struct genl_info *info) mac_addr = nla_data(info->attrs[NL80211_ATTR_MAC]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -1112,7 +1133,7 @@ static int nl80211_set_station(struct sk_buff *skb, struct genl_info *info) params.plink_action = nla_get_u8(info->attrs[NL80211_ATTR_STA_PLINK_ACTION]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -1172,7 +1193,7 @@ static int nl80211_new_station(struct sk_buff *skb, struct genl_info *info) ¶ms.station_flags)) return -EINVAL; - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -1207,7 +1228,7 @@ static int nl80211_del_station(struct sk_buff *skb, struct genl_info *info) if (info->attrs[NL80211_ATTR_MAC]) mac_addr = nla_data(info->attrs[NL80211_ATTR_MAC]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -1277,68 +1298,78 @@ static int nl80211_send_mpath(struct sk_buff *msg, u32 pid, u32 seq, } static int nl80211_dump_mpath(struct sk_buff *skb, - struct netlink_callback *cb) + struct netlink_callback *cb) { - int wp_idx = 0; - int if_idx = 0; - int sta_idx = cb->args[2]; - int wp_start = cb->args[0]; - int if_start = cb->args[1]; struct mpath_info pinfo; struct cfg80211_registered_device *dev; - struct wireless_dev *wdev; + struct net_device *netdev; u8 dst[ETH_ALEN]; u8 next_hop[ETH_ALEN]; + int ifidx = cb->args[0]; + int path_idx = cb->args[1]; int err; - int exit = 0; - /* TODO: filter by device */ - mutex_lock(&cfg80211_drv_mutex); - list_for_each_entry(dev, &cfg80211_drv_list, list) { - if (exit) + if (!ifidx) { + err = nlmsg_parse(cb->nlh, GENL_HDRLEN + nl80211_fam.hdrsize, + nl80211_fam.attrbuf, nl80211_fam.maxattr, + nl80211_policy); + if (err) + return err; + + if (!nl80211_fam.attrbuf[NL80211_ATTR_IFINDEX]) + return -EINVAL; + + ifidx = nla_get_u32(nl80211_fam.attrbuf[NL80211_ATTR_IFINDEX]); + if (!ifidx) + return -EINVAL; + } + + netdev = dev_get_by_index(&init_net, ifidx); + if (!netdev) + return -ENODEV; + + dev = cfg80211_get_dev_from_ifindex(ifidx); + if (IS_ERR(dev)) { + err = PTR_ERR(dev); + goto out_put_netdev; + } + + if (!dev->ops->dump_mpath) { + err = -ENOSYS; + goto out_err; + } + + rtnl_lock(); + + while (1) { + err = dev->ops->dump_mpath(&dev->wiphy, netdev, path_idx, + dst, next_hop, &pinfo); + if (err == -ENOENT) break; - if (++wp_idx < wp_start) - continue; - if_idx = 0; + if (err) + goto out_err_rtnl; - mutex_lock(&dev->devlist_mtx); - list_for_each_entry(wdev, &dev->netdev_list, list) { - if (exit) - break; - if (++if_idx < if_start) - continue; - if (!dev->ops->dump_mpath) - continue; + if (nl80211_send_mpath(skb, NETLINK_CB(cb->skb).pid, + cb->nlh->nlmsg_seq, NLM_F_MULTI, + netdev, dst, next_hop, + &pinfo) < 0) + goto out; - for (;; ++sta_idx) { - rtnl_lock(); - err = dev->ops->dump_mpath(&dev->wiphy, - wdev->netdev, sta_idx, dst, - next_hop, &pinfo); - rtnl_unlock(); - if (err) { - sta_idx = 0; - break; - } - if (nl80211_send_mpath(skb, - NETLINK_CB(cb->skb).pid, - cb->nlh->nlmsg_seq, NLM_F_MULTI, - wdev->netdev, dst, next_hop, - &pinfo) < 0) { - exit = 1; - break; - } - } - } - mutex_unlock(&dev->devlist_mtx); + path_idx++; } - mutex_unlock(&cfg80211_drv_mutex); - cb->args[0] = wp_idx; - cb->args[1] = if_idx; - cb->args[2] = sta_idx; - return skb->len; + out: + cb->args[1] = path_idx; + err = skb->len; + out_err_rtnl: + rtnl_unlock(); + out_err: + cfg80211_put_dev(dev); + out_put_netdev: + dev_put(netdev); + + return err; } static int nl80211_get_mpath(struct sk_buff *skb, struct genl_info *info) @@ -1358,7 +1389,7 @@ static int nl80211_get_mpath(struct sk_buff *skb, struct genl_info *info) dst = nla_data(info->attrs[NL80211_ATTR_MAC]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -1411,7 +1442,7 @@ static int nl80211_set_mpath(struct sk_buff *skb, struct genl_info *info) dst = nla_data(info->attrs[NL80211_ATTR_MAC]); next_hop = nla_data(info->attrs[NL80211_ATTR_MPATH_NEXT_HOP]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -1446,7 +1477,7 @@ static int nl80211_new_mpath(struct sk_buff *skb, struct genl_info *info) dst = nla_data(info->attrs[NL80211_ATTR_MAC]); next_hop = nla_data(info->attrs[NL80211_ATTR_MPATH_NEXT_HOP]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; @@ -1475,7 +1506,7 @@ static int nl80211_del_mpath(struct sk_buff *skb, struct genl_info *info) if (info->attrs[NL80211_ATTR_MAC]) dst = nla_data(info->attrs[NL80211_ATTR_MAC]); - err = get_drv_dev_by_info_ifindex(info, &drv, &dev); + err = get_drv_dev_by_info_ifindex(info->attrs, &drv, &dev); if (err) return err; -- cgit v1.2.1 From 14db74bcc3f7a779cf395a47e26b06a28207571a Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Tue, 29 Jul 2008 13:22:52 +0200 Subject: mac80211: fix cfg80211 hooks for master interface The master interface is a virtual interface that is registered to mac80211, changing that does not seem like a good idea at the moment. However, since it has no sdata, we cannot accept any configuration for it. This patch makes the cfg80211 hooks reject any such attempt. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/cfg.c | 103 +++++++++++++++++++++++++++++++++++++++++++++-------- 1 file changed, 88 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c index 8e7ba0e62cf5..297c257864c7 100644 --- a/net/mac80211/cfg.c +++ b/net/mac80211/cfg.c @@ -81,6 +81,7 @@ static int ieee80211_change_iface(struct wiphy *wiphy, int ifindex, enum nl80211_iftype type, u32 *flags, struct vif_params *params) { + struct ieee80211_local *local = wiphy_priv(wiphy); struct net_device *dev; enum ieee80211_if_types itype; struct ieee80211_sub_if_data *sdata; @@ -95,6 +96,9 @@ static int ieee80211_change_iface(struct wiphy *wiphy, int ifindex, if (itype == IEEE80211_IF_TYPE_INVALID) return -EINVAL; + if (dev == local->mdev) + return -EOPNOTSUPP; + sdata = IEEE80211_DEV_TO_SUB_IF(dev); ret = ieee80211_if_change_type(sdata, itype); @@ -117,12 +121,16 @@ static int ieee80211_add_key(struct wiphy *wiphy, struct net_device *dev, u8 key_idx, u8 *mac_addr, struct key_params *params) { + struct ieee80211_local *local = wiphy_priv(wiphy); struct ieee80211_sub_if_data *sdata; struct sta_info *sta = NULL; enum ieee80211_key_alg alg; struct ieee80211_key *key; int err; + if (dev == local->mdev) + return -EOPNOTSUPP; + sdata = IEEE80211_DEV_TO_SUB_IF(dev); switch (params->cipher) { @@ -167,10 +175,14 @@ static int ieee80211_add_key(struct wiphy *wiphy, struct net_device *dev, static int ieee80211_del_key(struct wiphy *wiphy, struct net_device *dev, u8 key_idx, u8 *mac_addr) { + struct ieee80211_local *local = wiphy_priv(wiphy); struct ieee80211_sub_if_data *sdata; struct sta_info *sta; int ret; + if (dev == local->mdev) + return -EOPNOTSUPP; + sdata = IEEE80211_DEV_TO_SUB_IF(dev); rcu_read_lock(); @@ -211,7 +223,8 @@ static int ieee80211_get_key(struct wiphy *wiphy, struct net_device *dev, void (*callback)(void *cookie, struct key_params *params)) { - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct sta_info *sta = NULL; u8 seq[6] = {0}; struct key_params params; @@ -220,6 +233,11 @@ static int ieee80211_get_key(struct wiphy *wiphy, struct net_device *dev, u16 iv16; int err = -ENOENT; + if (dev == local->mdev) + return -EOPNOTSUPP; + + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + rcu_read_lock(); if (mac_addr) { @@ -293,8 +311,12 @@ static int ieee80211_config_default_key(struct wiphy *wiphy, struct net_device *dev, u8 key_idx) { + struct ieee80211_local *local = wiphy_priv(wiphy); struct ieee80211_sub_if_data *sdata; + if (dev == local->mdev) + return -EOPNOTSUPP; + rcu_read_lock(); sdata = IEEE80211_DEV_TO_SUB_IF(dev); @@ -475,9 +497,15 @@ static int ieee80211_config_beacon(struct ieee80211_sub_if_data *sdata, static int ieee80211_add_beacon(struct wiphy *wiphy, struct net_device *dev, struct beacon_parameters *params) { - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct beacon_data *old; + if (dev == local->mdev) + return -EOPNOTSUPP; + + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + if (sdata->vif.type != IEEE80211_IF_TYPE_AP) return -EINVAL; @@ -492,9 +520,15 @@ static int ieee80211_add_beacon(struct wiphy *wiphy, struct net_device *dev, static int ieee80211_set_beacon(struct wiphy *wiphy, struct net_device *dev, struct beacon_parameters *params) { - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct beacon_data *old; + if (dev == local->mdev) + return -EOPNOTSUPP; + + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + if (sdata->vif.type != IEEE80211_IF_TYPE_AP) return -EINVAL; @@ -508,9 +542,15 @@ static int ieee80211_set_beacon(struct wiphy *wiphy, struct net_device *dev, static int ieee80211_del_beacon(struct wiphy *wiphy, struct net_device *dev) { - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct beacon_data *old; + if (dev == local->mdev) + return -EOPNOTSUPP; + + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + if (sdata->vif.type != IEEE80211_IF_TYPE_AP) return -EINVAL; @@ -646,11 +686,14 @@ static void sta_apply_parameters(struct ieee80211_local *local, static int ieee80211_add_station(struct wiphy *wiphy, struct net_device *dev, u8 *mac, struct station_parameters *params) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct ieee80211_local *local = wiphy_priv(wiphy); struct sta_info *sta; struct ieee80211_sub_if_data *sdata; int err; + if (dev == local->mdev || params->vlan == local->mdev) + return -EOPNOTSUPP; + /* Prevent a race with changing the rate control algorithm */ if (!netif_running(dev)) return -ENETDOWN; @@ -701,10 +744,15 @@ static int ieee80211_add_station(struct wiphy *wiphy, struct net_device *dev, static int ieee80211_del_station(struct wiphy *wiphy, struct net_device *dev, u8 *mac) { - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); - struct ieee80211_local *local = sdata->local; + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct sta_info *sta; + if (dev == local->mdev) + return -EOPNOTSUPP; + + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + if (mac) { rcu_read_lock(); @@ -730,10 +778,13 @@ static int ieee80211_change_station(struct wiphy *wiphy, u8 *mac, struct station_parameters *params) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct ieee80211_local *local = wiphy_priv(wiphy); struct sta_info *sta; struct ieee80211_sub_if_data *vlansdata; + if (dev == local->mdev || params->vlan == local->mdev) + return -EOPNOTSUPP; + rcu_read_lock(); /* XXX: get sta belonging to dev */ @@ -752,7 +803,7 @@ static int ieee80211_change_station(struct wiphy *wiphy, return -EINVAL; } - sta->sdata = IEEE80211_DEV_TO_SUB_IF(params->vlan); + sta->sdata = vlansdata; ieee80211_send_layer2_update(sta); } @@ -767,15 +818,20 @@ static int ieee80211_change_station(struct wiphy *wiphy, static int ieee80211_add_mpath(struct wiphy *wiphy, struct net_device *dev, u8 *dst, u8 *next_hop) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct mesh_path *mpath; struct sta_info *sta; int err; + if (dev == local->mdev) + return -EOPNOTSUPP; + if (!netif_running(dev)) return -ENETDOWN; + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + if (sdata->vif.type != IEEE80211_IF_TYPE_MESH_POINT) return -ENOTSUPP; @@ -817,14 +873,19 @@ static int ieee80211_change_mpath(struct wiphy *wiphy, struct net_device *dev, u8 *dst, u8 *next_hop) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct mesh_path *mpath; struct sta_info *sta; + if (dev == local->mdev) + return -EOPNOTSUPP; + if (!netif_running(dev)) return -ENETDOWN; + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + if (sdata->vif.type != IEEE80211_IF_TYPE_MESH_POINT) return -ENOTSUPP; @@ -891,9 +952,15 @@ static int ieee80211_get_mpath(struct wiphy *wiphy, struct net_device *dev, u8 *dst, u8 *next_hop, struct mpath_info *pinfo) { - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct mesh_path *mpath; + if (dev == local->mdev) + return -EOPNOTSUPP; + + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + if (sdata->vif.type != IEEE80211_IF_TYPE_MESH_POINT) return -ENOTSUPP; @@ -913,9 +980,15 @@ static int ieee80211_dump_mpath(struct wiphy *wiphy, struct net_device *dev, int idx, u8 *dst, u8 *next_hop, struct mpath_info *pinfo) { - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wiphy_priv(wiphy); + struct ieee80211_sub_if_data *sdata; struct mesh_path *mpath; + if (dev == local->mdev) + return -EOPNOTSUPP; + + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + if (sdata->vif.type != IEEE80211_IF_TYPE_MESH_POINT) return -ENOTSUPP; -- cgit v1.2.1 From 56a6d13dfd49d90d72a1a962246206719dd9d143 Mon Sep 17 00:00:00 2001 From: Luis Carlos Cobo Date: Tue, 29 Jul 2008 19:59:31 +0200 Subject: mac80211: fix mesh beaconing This patch fixes mesh beaconing, which was broken by "mac80211: revamp beacon configuration". Signed-off-by: Luis Carlos Cobo Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 35eb767cbcbe..acb04133a95d 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -3299,6 +3299,7 @@ void ieee80211_start_mesh(struct net_device *dev) ifsta = &sdata->u.sta; ifsta->state = IEEE80211_MESH_UP; ieee80211_sta_timer((unsigned long)sdata); + ieee80211_if_config(sdata, IEEE80211_IFCC_BEACON); } #endif -- cgit v1.2.1 From 4a36702e016947a0ce6c0c024673bb5b16d3f618 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Tue, 29 Jul 2008 23:57:58 -0700 Subject: IPv6: datagram_send_ctl() should exit immediately when an error occured When an error occured, datagram_send_ctl() should exit immediately rather than continue to run the for loop. Otherwise, the variable err might be changed and the error might be hidden. Fix this bug by using "goto" instead of "break". Signed-off-by: Miao Xie Signed-off-by: David S. Miller --- net/ipv6/datagram.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c index f7b535dec860..410046a8cc91 100644 --- a/net/ipv6/datagram.c +++ b/net/ipv6/datagram.c @@ -732,7 +732,7 @@ int datagram_send_ctl(struct net *net, LIMIT_NETDEBUG(KERN_DEBUG "invalid cmsg type: %d\n", cmsg->cmsg_type); err = -EINVAL; - break; + goto exit_f; } } -- cgit v1.2.1 From 8d50b53d66a8a6ae41bafbdcabe401467803f33a Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 30 Jul 2008 02:37:46 -0700 Subject: pkt_sched: Fix OOPS on ingress qdisc add. Bug report from Steven Jan Springl: Issuing the following command causes a kernel oops: tc qdisc add dev eth0 handle ffff: ingress The problem mostly stems from all of the special case handling of ingress qdiscs. So, to fix this, do the grafting operation the same way we do for TX qdiscs. Which means that dev_activate() and dev_deactivate() now do the "qdisc_sleeping <--> qdisc" transitions on dev->rx_queue too. Future simplifications are possible now, mainly because it is impossible for dev_queue->{qdisc,qdisc_sleeping} to be NULL. There are NULL checks all over to handle the ingress qdisc special case that used to exist before this commit. Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- net/sched/sch_api.c | 57 ++++++++++++++----------------------------------- net/sched/sch_generic.c | 8 ++++--- 3 files changed, 23 insertions(+), 46 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 8d13a9b9f1df..63d6bcddbf46 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2100,7 +2100,7 @@ static int ing_filter(struct sk_buff *skb) rxq = &dev->rx_queue; q = rxq->qdisc; - if (q) { + if (q != &noop_qdisc) { spin_lock(qdisc_lock(q)); result = qdisc_enqueue_root(skb, q); spin_unlock(qdisc_lock(q)); @@ -2113,7 +2113,7 @@ static inline struct sk_buff *handle_ing(struct sk_buff *skb, struct packet_type **pt_prev, int *ret, struct net_device *orig_dev) { - if (!skb->dev->rx_queue.qdisc) + if (skb->dev->rx_queue.qdisc == &noop_qdisc) goto out; if (*pt_prev) { diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index b0601642e227..4840aff47256 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -572,44 +572,21 @@ static u32 qdisc_alloc_handle(struct net_device *dev) static struct Qdisc *dev_graft_qdisc(struct netdev_queue *dev_queue, struct Qdisc *qdisc) { + struct Qdisc *oqdisc = dev_queue->qdisc_sleeping; spinlock_t *root_lock; - struct Qdisc *oqdisc; - int ingress; - - ingress = 0; - if (qdisc && qdisc->flags&TCQ_F_INGRESS) - ingress = 1; - - if (ingress) { - oqdisc = dev_queue->qdisc; - } else { - oqdisc = dev_queue->qdisc_sleeping; - } root_lock = qdisc_root_lock(oqdisc); spin_lock_bh(root_lock); - if (ingress) { - /* Prune old scheduler */ - if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) { - /* delete */ - qdisc_reset(oqdisc); - dev_queue->qdisc = NULL; - } else { /* new */ - dev_queue->qdisc = qdisc; - } + /* Prune old scheduler */ + if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) + qdisc_reset(oqdisc); - } else { - /* Prune old scheduler */ - if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) - qdisc_reset(oqdisc); - - /* ... and graft new one */ - if (qdisc == NULL) - qdisc = &noop_qdisc; - dev_queue->qdisc_sleeping = qdisc; - dev_queue->qdisc = &noop_qdisc; - } + /* ... and graft new one */ + if (qdisc == NULL) + qdisc = &noop_qdisc; + dev_queue->qdisc_sleeping = qdisc; + dev_queue->qdisc = &noop_qdisc; spin_unlock_bh(root_lock); @@ -678,7 +655,8 @@ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent, ingress = 0; num_q = dev->num_tx_queues; - if (q && q->flags & TCQ_F_INGRESS) { + if ((q && q->flags & TCQ_F_INGRESS) || + (new && new->flags & TCQ_F_INGRESS)) { num_q = 1; ingress = 1; } @@ -692,13 +670,10 @@ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent, if (!ingress) dev_queue = netdev_get_tx_queue(dev, i); - if (ingress) { - old = dev_graft_qdisc(dev_queue, q); - } else { - old = dev_graft_qdisc(dev_queue, new); - if (new && i > 0) - atomic_inc(&new->refcnt); - } + old = dev_graft_qdisc(dev_queue, new); + if (new && i > 0) + atomic_inc(&new->refcnt); + notify_and_destroy(skb, n, classid, old, new); } @@ -817,7 +792,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, goto err_out3; } } - if (parent) + if (parent && !(sch->flags & TCQ_F_INGRESS)) list_add_tail(&sch->list, &dev_queue->qdisc->list); return sch; diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index fd2a6cadb115..345838a2e369 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -596,7 +596,7 @@ static void transition_one_qdisc(struct net_device *dev, int *need_watchdog_p = _need_watchdog; rcu_assign_pointer(dev_queue->qdisc, new_qdisc); - if (new_qdisc != &noqueue_qdisc) + if (need_watchdog_p && new_qdisc != &noqueue_qdisc) *need_watchdog_p = 1; } @@ -619,6 +619,7 @@ void dev_activate(struct net_device *dev) need_watchdog = 0; netdev_for_each_tx_queue(dev, transition_one_qdisc, &need_watchdog); + transition_one_qdisc(dev, &dev->rx_queue, NULL); if (need_watchdog) { dev->trans_start = jiffies; @@ -677,6 +678,7 @@ void dev_deactivate(struct net_device *dev) bool running; netdev_for_each_tx_queue(dev, dev_deactivate_queue, &noop_qdisc); + dev_deactivate_queue(dev, &dev->rx_queue, &noop_qdisc); dev_watchdog_down(dev); @@ -718,7 +720,7 @@ static void dev_init_scheduler_queue(struct net_device *dev, void dev_init_scheduler(struct net_device *dev) { netdev_for_each_tx_queue(dev, dev_init_scheduler_queue, &noop_qdisc); - dev_init_scheduler_queue(dev, &dev->rx_queue, NULL); + dev_init_scheduler_queue(dev, &dev->rx_queue, &noop_qdisc); setup_timer(&dev->watchdog_timer, dev_watchdog, (unsigned long)dev); } @@ -745,6 +747,6 @@ static void shutdown_scheduler_queue(struct net_device *dev, void dev_shutdown(struct net_device *dev) { netdev_for_each_tx_queue(dev, shutdown_scheduler_queue, &noop_qdisc); - shutdown_scheduler_queue(dev, &dev->rx_queue, NULL); + shutdown_scheduler_queue(dev, &dev->rx_queue, &noop_qdisc); WARN_ON(timer_pending(&dev->watchdog_timer)); } -- cgit v1.2.1 From 785957d3e8c6fb37b18bf671923a76dbd8240025 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 30 Jul 2008 03:03:15 -0700 Subject: tcp: MD5: Use MIB counter instead of warning for MD5 mismatch. From a report by Matti Aarnio, and preliminary patch by Adam Langley. Signed-off-by: David S. Miller --- net/ipv4/proc.c | 2 ++ net/ipv4/tcp_ipv4.c | 10 ++-------- net/ipv6/tcp_ipv6.c | 27 ++++++++------------------- 3 files changed, 12 insertions(+), 27 deletions(-) (limited to 'net') diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 834356ea99df..8f5a403f6f6b 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -232,6 +232,8 @@ static const struct snmp_mib snmp4_net_list[] = { SNMP_MIB_ITEM("TCPDSACKIgnoredOld", LINUX_MIB_TCPDSACKIGNOREDOLD), SNMP_MIB_ITEM("TCPDSACKIgnoredNoUndo", LINUX_MIB_TCPDSACKIGNOREDNOUNDO), SNMP_MIB_ITEM("TCPSpuriousRTOs", LINUX_MIB_TCPSPURIOUSRTOS), + SNMP_MIB_ITEM("TCPMD5NotFound", LINUX_MIB_TCPMD5NOTFOUND), + SNMP_MIB_ITEM("TCPMD5Unexpected", LINUX_MIB_TCPMD5UNEXPECTED), SNMP_MIB_SENTINEL }; diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index a2b06d0cc26b..b3875c0d83c7 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -1116,18 +1116,12 @@ static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb) return 0; if (hash_expected && !hash_location) { - LIMIT_NETDEBUG(KERN_INFO "MD5 Hash expected but NOT found " - "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)\n", - NIPQUAD(iph->saddr), ntohs(th->source), - NIPQUAD(iph->daddr), ntohs(th->dest)); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5NOTFOUND); return 1; } if (!hash_expected && hash_location) { - LIMIT_NETDEBUG(KERN_INFO "MD5 Hash NOT expected but found " - "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)\n", - NIPQUAD(iph->saddr), ntohs(th->source), - NIPQUAD(iph->daddr), ntohs(th->dest)); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED); return 1; } diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index cff778b23a7f..1db45216b232 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -849,28 +849,17 @@ static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb) hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr); hash_location = tcp_parse_md5sig_option(th); - /* do we have a hash as expected? */ - if (!hash_expected) { - if (!hash_location) - return 0; - if (net_ratelimit()) { - printk(KERN_INFO "MD5 Hash NOT expected but found " - "(" NIP6_FMT ", %u)->" - "(" NIP6_FMT ", %u)\n", - NIP6(ip6h->saddr), ntohs(th->source), - NIP6(ip6h->daddr), ntohs(th->dest)); - } + /* We've parsed the options - do we have a hash? */ + if (!hash_expected && !hash_location) + return 0; + + if (hash_expected && !hash_location) { + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5NOTFOUND); return 1; } - if (!hash_location) { - if (net_ratelimit()) { - printk(KERN_INFO "MD5 Hash expected but NOT found " - "(" NIP6_FMT ", %u)->" - "(" NIP6_FMT ", %u)\n", - NIP6(ip6h->saddr), ntohs(th->source), - NIP6(ip6h->daddr), ntohs(th->dest)); - } + if (!hash_expected && hash_location) { + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED); return 1; } -- cgit v1.2.1 From 17ef51fce03758736e9051c4360eca237dd0aaeb Mon Sep 17 00:00:00 2001 From: Daniel Lezcano Date: Wed, 30 Jul 2008 03:12:31 -0700 Subject: ipv6: Fix useless proc net sockstat6 removal This call is no longer needed, sockstat6 is per namespace so it is removed at the namespace subsystem destruction. Signed-off-by: Daniel Lezcano Signed-off-by: David S. Miller --- net/ipv6/proc.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c index f82f6074cf85..0179b66864f1 100644 --- a/net/ipv6/proc.c +++ b/net/ipv6/proc.c @@ -286,7 +286,6 @@ proc_net_fail: void ipv6_misc_proc_exit(void) { - proc_net_remove(&init_net, "sockstat6"); proc_net_remove(&init_net, "dev_snmp6"); proc_net_remove(&init_net, "snmp6"); unregister_pernet_subsys(&ipv6_proc_ops); -- cgit v1.2.1 From 031cf19e6f63941506c9baf76ac7adac06edcf08 Mon Sep 17 00:00:00 2001 From: "Robert P. J. Day" Date: Wed, 30 Jul 2008 03:14:01 -0700 Subject: net: Make "networking" one-click deselectable. Use a menuconfig directive to make all of networking support one-click deselectable from the top-level menu. Signed-off-by: Robert P. J. Day Signed-off-by: David S. Miller --- net/Kconfig | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) (limited to 'net') diff --git a/net/Kconfig b/net/Kconfig index b98668751749..7612cc8c337c 100644 --- a/net/Kconfig +++ b/net/Kconfig @@ -2,9 +2,7 @@ # Network configuration # -menu "Networking" - -config NET +menuconfig NET bool "Networking support" ---help--- Unless you really know what you are doing, you should say Y here. @@ -22,7 +20,6 @@ config NET recommended to read the NET-HOWTO, available from . -# Make sure that all config symbols are dependent on NET if NET menu "Networking options" @@ -252,5 +249,3 @@ source "net/rfkill/Kconfig" source "net/9p/Kconfig" endif # if NET -endmenu # Networking - -- cgit v1.2.1 From 4adf0af6818f3ea52421dc0bae836cfaf20ef72a Mon Sep 17 00:00:00 2001 From: Simon Wunderlich Date: Wed, 30 Jul 2008 16:27:55 -0700 Subject: bridge: send correct MTU value in PMTU (revised) When bridging interfaces with different MTUs, the bridge correctly chooses the minimum of the MTUs of the physical devices as the bridges MTU. But when a frame is passed which fits through the incoming, but not through the outgoing interface, a "Fragmentation Needed" packet is generated. However, the propagated MTU is hardcoded to 1500, which is wrong in this situation. The sender will repeat the packet again with the same frame size, and the same problem will occur again. Instead of sending 1500, the (correct) MTU value of the bridge is now sent via PMTU. To achieve this, the corresponding rtable structure is stored in its net_bridge structure. Modified to get rid of fake_net_device as well. Signed-off-by: Simon Wunderlich Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/bridge/br_device.c | 9 ++++++- net/bridge/br_if.c | 3 +++ net/bridge/br_netfilter.c | 63 +++++++++++++++++++++++++---------------------- net/bridge/br_private.h | 6 +++++ 4 files changed, 51 insertions(+), 30 deletions(-) (limited to 'net') diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c index d9449df7cad5..9b58d70b0e7d 100644 --- a/net/bridge/br_device.c +++ b/net/bridge/br_device.c @@ -68,10 +68,17 @@ static int br_dev_stop(struct net_device *dev) static int br_change_mtu(struct net_device *dev, int new_mtu) { - if (new_mtu < 68 || new_mtu > br_min_mtu(netdev_priv(dev))) + struct net_bridge *br = netdev_priv(dev); + if (new_mtu < 68 || new_mtu > br_min_mtu(br)) return -EINVAL; dev->mtu = new_mtu; + +#ifdef CONFIG_BRIDGE_NETFILTER + /* remember the MTU in the rtable for PMTU */ + br->fake_rtable.u.dst.metrics[RTAX_MTU - 1] = new_mtu; +#endif + return 0; } diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index a072ea5ca6f5..63c18aacde8c 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -202,6 +202,9 @@ static struct net_device *new_bridge_dev(const char *name) br->topology_change = 0; br->topology_change_detected = 0; br->ageing_time = 300 * HZ; + + br_netfilter_rtable_init(br); + INIT_LIST_HEAD(&br->age_list); br_stp_timer_init(br); diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c index bb90cd7bace3..6e280a8a31ee 100644 --- a/net/bridge/br_netfilter.c +++ b/net/bridge/br_netfilter.c @@ -101,33 +101,30 @@ static inline __be16 pppoe_proto(const struct sk_buff *skb) pppoe_proto(skb) == htons(PPP_IPV6) && \ brnf_filter_pppoe_tagged) -/* We need these fake structures to make netfilter happy -- - * lots of places assume that skb->dst != NULL, which isn't - * all that unreasonable. - * +/* + * Initialize bogus route table used to keep netfilter happy. * Currently, we fill in the PMTU entry because netfilter * refragmentation needs it, and the rt_flags entry because * ipt_REJECT needs it. Future netfilter modules might - * require us to fill additional fields. */ -static struct net_device __fake_net_device = { - .hard_header_len = ETH_HLEN, -#ifdef CONFIG_NET_NS - .nd_net = &init_net, -#endif -}; + * require us to fill additional fields. + */ +void br_netfilter_rtable_init(struct net_bridge *br) +{ + struct rtable *rt = &br->fake_rtable; -static struct rtable __fake_rtable = { - .u = { - .dst = { - .__refcnt = ATOMIC_INIT(1), - .dev = &__fake_net_device, - .path = &__fake_rtable.u.dst, - .metrics = {[RTAX_MTU - 1] = 1500}, - .flags = DST_NOXFRM, - } - }, - .rt_flags = 0, -}; + atomic_set(&rt->u.dst.__refcnt, 1); + rt->u.dst.dev = &br->dev; + rt->u.dst.path = &rt->u.dst; + rt->u.dst.metrics[RTAX_MTU - 1] = 1500; + rt->u.dst.flags = DST_NOXFRM; +} + +static inline struct rtable *bridge_parent_rtable(const struct net_device *dev) +{ + struct net_bridge_port *port = rcu_dereference(dev->br_port); + + return port ? &port->br->fake_rtable : NULL; +} static inline struct net_device *bridge_parent(const struct net_device *dev) { @@ -226,8 +223,12 @@ static int br_nf_pre_routing_finish_ipv6(struct sk_buff *skb) } nf_bridge->mask ^= BRNF_NF_BRIDGE_PREROUTING; - skb->rtable = &__fake_rtable; - dst_hold(&__fake_rtable.u.dst); + skb->rtable = bridge_parent_rtable(nf_bridge->physindev); + if (!skb->rtable) { + kfree_skb(skb); + return 0; + } + dst_hold(&skb->rtable->u.dst); skb->dev = nf_bridge->physindev; nf_bridge_push_encap_header(skb); @@ -391,8 +392,12 @@ bridged_dnat: skb->pkt_type = PACKET_HOST; } } else { - skb->rtable = &__fake_rtable; - dst_hold(&__fake_rtable.u.dst); + skb->rtable = bridge_parent_rtable(nf_bridge->physindev); + if (!skb->rtable) { + kfree_skb(skb); + return 0; + } + dst_hold(&skb->rtable->u.dst); } skb->dev = nf_bridge->physindev; @@ -611,8 +616,8 @@ static unsigned int br_nf_local_in(unsigned int hook, struct sk_buff *skb, const struct net_device *out, int (*okfn)(struct sk_buff *)) { - if (skb->rtable == &__fake_rtable) { - dst_release(&__fake_rtable.u.dst); + if (skb->rtable && skb->rtable == bridge_parent_rtable(in)) { + dst_release(&skb->rtable->u.dst); skb->rtable = NULL; } diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 815ed38925b2..c3dc18ddc043 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -15,6 +15,7 @@ #include #include +#include #define BR_HASH_BITS 8 #define BR_HASH_SIZE (1 << BR_HASH_BITS) @@ -92,6 +93,9 @@ struct net_bridge struct hlist_head hash[BR_HASH_SIZE]; struct list_head age_list; unsigned long feature_mask; +#ifdef CONFIG_BRIDGE_NETFILTER + struct rtable fake_rtable; +#endif unsigned long flags; #define BR_SET_MAC_ADDR 0x00000001 @@ -197,9 +201,11 @@ extern int br_ioctl_deviceless_stub(struct net *net, unsigned int cmd, void __us #ifdef CONFIG_BRIDGE_NETFILTER extern int br_netfilter_init(void); extern void br_netfilter_fini(void); +extern void br_netfilter_rtable_init(struct net_bridge *); #else #define br_netfilter_init() (0) #define br_netfilter_fini() do { } while(0) +#define br_netfilter_rtable_init(x) #endif /* br_stp.c */ -- cgit v1.2.1 From 6a8341b68b5269de71c32c6df91f4b0298da031d Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 30 Jul 2008 16:30:15 -0700 Subject: net: use the common ascii hex helpers Signed-off-by: Harvey Harrison Signed-off-by: Andrew Morton Signed-off-by: David S. Miller --- net/ipv4/netfilter/ipt_CLUSTERIP.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c index 1819ad7ab910..fafe8ebb4c55 100644 --- a/net/ipv4/netfilter/ipt_CLUSTERIP.c +++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c @@ -475,11 +475,10 @@ static void arp_print(struct arp_payload *payload) #define HBUFFERLEN 30 char hbuffer[HBUFFERLEN]; int j,k; - const char hexbuf[]= "0123456789abcdef"; for (k=0, j=0; k < HBUFFERLEN-3 && j < ETH_ALEN; j++) { - hbuffer[k++]=hexbuf[(payload->src_hw[j]>>4)&15]; - hbuffer[k++]=hexbuf[payload->src_hw[j]&15]; + hbuffer[k++] = hex_asc_hi(payload->src_hw[j]); + hbuffer[k++] = hex_asc_lo(payload->src_hw[j]); hbuffer[k++]=':'; } hbuffer[--k]='\0'; -- cgit v1.2.1 From cba5cbd1559f49bec76e54de6ed21b7df3742ada Mon Sep 17 00:00:00 2001 From: David Howells Date: Wed, 30 Jul 2008 16:31:46 -0700 Subject: atm: fix const assignment/discard warnings in the ATM networking driver Fix const assignment/discard warnings in the ATM networking driver. The lane2_assoc_ind() function needed its arguments changing to match changes in the lane2_ops struct (patch 61c33e012964ce358b42d2a1e9cd309af5dab02b "atm: use const where reasonable"). Signed-off-by: David Howells Acked-by: Chas Williams Signed-off-by: Andrew Morton Signed-off-by: David S. Miller --- net/atm/mpc.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/atm/mpc.c b/net/atm/mpc.c index 4fccaa1e07be..11b16d16661c 100644 --- a/net/atm/mpc.c +++ b/net/atm/mpc.c @@ -62,11 +62,13 @@ static void MPOA_cache_impos_rcvd(struct k_message *msg, struct mpoa_client *mpc static void set_mpc_ctrl_addr_rcvd(struct k_message *mesg, struct mpoa_client *mpc); static void set_mps_mac_addr_rcvd(struct k_message *mesg, struct mpoa_client *mpc); -static uint8_t *copy_macs(struct mpoa_client *mpc, uint8_t *router_mac, - uint8_t *tlvs, uint8_t mps_macs, uint8_t device_type); +static const uint8_t *copy_macs(struct mpoa_client *mpc, + const uint8_t *router_mac, + const uint8_t *tlvs, uint8_t mps_macs, + uint8_t device_type); static void purge_egress_shortcut(struct atm_vcc *vcc, eg_cache_entry *entry); -static void send_set_mps_ctrl_addr(char *addr, struct mpoa_client *mpc); +static void send_set_mps_ctrl_addr(const char *addr, struct mpoa_client *mpc); static void mpoad_close(struct atm_vcc *vcc); static int msg_from_mpoad(struct atm_vcc *vcc, struct sk_buff *skb); @@ -351,12 +353,12 @@ static const char *mpoa_device_type_string(char type) * lec sees a TLV it uses the pointer to call this function. * */ -static void lane2_assoc_ind(struct net_device *dev, uint8_t *mac_addr, - uint8_t *tlvs, uint32_t sizeoftlvs) +static void lane2_assoc_ind(struct net_device *dev, const u8 *mac_addr, + const u8 *tlvs, u32 sizeoftlvs) { uint32_t type; uint8_t length, mpoa_device_type, number_of_mps_macs; - uint8_t *end_of_tlvs; + const uint8_t *end_of_tlvs; struct mpoa_client *mpc; mpoa_device_type = number_of_mps_macs = 0; /* silence gcc */ @@ -430,8 +432,10 @@ static void lane2_assoc_ind(struct net_device *dev, uint8_t *mac_addr, * plus the possible MAC address(es) to mpc->mps_macs. * For a freshly allocated MPOA client mpc->mps_macs == 0. */ -static uint8_t *copy_macs(struct mpoa_client *mpc, uint8_t *router_mac, - uint8_t *tlvs, uint8_t mps_macs, uint8_t device_type) +static const uint8_t *copy_macs(struct mpoa_client *mpc, + const uint8_t *router_mac, + const uint8_t *tlvs, uint8_t mps_macs, + uint8_t device_type) { int num_macs; num_macs = (mps_macs > 1) ? mps_macs : 1; @@ -811,7 +815,7 @@ static int atm_mpoa_mpoad_attach (struct atm_vcc *vcc, int arg) return arg; } -static void send_set_mps_ctrl_addr(char *addr, struct mpoa_client *mpc) +static void send_set_mps_ctrl_addr(const char *addr, struct mpoa_client *mpc) { struct k_message mesg; -- cgit v1.2.1 From ae375044d31075a31de5a839e07ded7f67b660aa Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Thu, 31 Jul 2008 00:38:01 -0700 Subject: netfilter: nf_conntrack_tcp: decrease timeouts while data in unacknowledged In order to time out dead connections quicker, keep track of outstanding data and cap the timeout. Suggested by Herbert Xu. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/nf_conntrack_proto_tcp.c | 29 ++++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/netfilter/nf_conntrack_proto_tcp.c b/net/netfilter/nf_conntrack_proto_tcp.c index 420a10d8eb1e..6f61261888ef 100644 --- a/net/netfilter/nf_conntrack_proto_tcp.c +++ b/net/netfilter/nf_conntrack_proto_tcp.c @@ -67,7 +67,8 @@ static const char *const tcp_conntrack_names[] = { /* RFC1122 says the R2 limit should be at least 100 seconds. Linux uses 15 packets as limit, which corresponds to ~13-30min depending on RTO. */ -static unsigned int nf_ct_tcp_timeout_max_retrans __read_mostly = 5 MINS; +static unsigned int nf_ct_tcp_timeout_max_retrans __read_mostly = 5 MINS; +static unsigned int nf_ct_tcp_timeout_unacknowledged __read_mostly = 5 MINS; static unsigned int tcp_timeouts[TCP_CONNTRACK_MAX] __read_mostly = { [TCP_CONNTRACK_SYN_SENT] = 2 MINS, @@ -625,8 +626,10 @@ static bool tcp_in_window(const struct nf_conn *ct, swin = win + (sack - ack); if (sender->td_maxwin < swin) sender->td_maxwin = swin; - if (after(end, sender->td_end)) + if (after(end, sender->td_end)) { sender->td_end = end; + sender->flags |= IP_CT_TCP_FLAG_DATA_UNACKNOWLEDGED; + } /* * Update receiver data. */ @@ -637,6 +640,8 @@ static bool tcp_in_window(const struct nf_conn *ct, if (win == 0) receiver->td_maxend++; } + if (ack == receiver->td_end) + receiver->flags &= ~IP_CT_TCP_FLAG_DATA_UNACKNOWLEDGED; /* * Check retransmissions. @@ -951,9 +956,16 @@ static int tcp_packet(struct nf_conn *ct, if (old_state != new_state && new_state == TCP_CONNTRACK_FIN_WAIT) ct->proto.tcp.seen[dir].flags |= IP_CT_TCP_FLAG_CLOSE_INIT; - timeout = ct->proto.tcp.retrans >= nf_ct_tcp_max_retrans - && tcp_timeouts[new_state] > nf_ct_tcp_timeout_max_retrans - ? nf_ct_tcp_timeout_max_retrans : tcp_timeouts[new_state]; + + if (ct->proto.tcp.retrans >= nf_ct_tcp_max_retrans && + tcp_timeouts[new_state] > nf_ct_tcp_timeout_max_retrans) + timeout = nf_ct_tcp_timeout_max_retrans; + else if ((ct->proto.tcp.seen[0].flags | ct->proto.tcp.seen[1].flags) & + IP_CT_TCP_FLAG_DATA_UNACKNOWLEDGED && + tcp_timeouts[new_state] > nf_ct_tcp_timeout_unacknowledged) + timeout = nf_ct_tcp_timeout_unacknowledged; + else + timeout = tcp_timeouts[new_state]; write_unlock_bh(&tcp_lock); nf_conntrack_event_cache(IPCT_PROTOINFO_VOLATILE, skb); @@ -1235,6 +1247,13 @@ static struct ctl_table tcp_sysctl_table[] = { .mode = 0644, .proc_handler = &proc_dointvec_jiffies, }, + { + .procname = "nf_conntrack_tcp_timeout_unacknowledged", + .data = &nf_ct_tcp_timeout_unacknowledged, + .maxlen = sizeof(unsigned int), + .mode = 0644, + .proc_handler = &proc_dointvec_jiffies, + }, { .ctl_name = NET_NF_CONNTRACK_TCP_LOOSE, .procname = "nf_conntrack_tcp_loose", -- cgit v1.2.1 From a8ddc9163c6a16cd62531dba1ec5020484e33b02 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Thu, 31 Jul 2008 00:38:31 -0700 Subject: netfilter: ipt_recent: fix race between recent_mt_destroy and proc manipulations The thing is that recent_mt_destroy first flushes the entries from table with the recent_table_flush and only *after* this removes the proc file, corresponding to that table. Thus, if we manage to write to this file the '+XXX' command we will leak some entries. If we manage to write there a 'clean' command we'll race in two recent_table_flush flows, since the recent_mt_destroy calls this outside the recent_lock. The proper solution as I see it is to remove the proc file first and then go on with flushing the table. This flushing becomes safe w/o the lock, since the table is already inaccessible from the outside. Signed-off-by: Pavel Emelyanov Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv4/netfilter/ipt_recent.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/netfilter/ipt_recent.c b/net/ipv4/netfilter/ipt_recent.c index 21cb053f5d7d..3974d7cae5c0 100644 --- a/net/ipv4/netfilter/ipt_recent.c +++ b/net/ipv4/netfilter/ipt_recent.c @@ -305,10 +305,10 @@ static void recent_mt_destroy(const struct xt_match *match, void *matchinfo) spin_lock_bh(&recent_lock); list_del(&t->list); spin_unlock_bh(&recent_lock); - recent_table_flush(t); #ifdef CONFIG_PROC_FS remove_proc_entry(t->name, proc_dir); #endif + recent_table_flush(t); kfree(t); } mutex_unlock(&recent_mutex); -- cgit v1.2.1 From 967ab999a090b1a4e7d3c7febfd6d89b42fb4cf4 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Thu, 31 Jul 2008 00:38:52 -0700 Subject: netfilter: xt_hashlimit: fix race between htable_destroy and htable_gc Deleting a timer with del_timer doesn't guarantee, that the timer function is not running at the moment of deletion. Thus in the xt_hashlimit case we can get into a ticklish situation when the htable_gc rearms the timer back and we'll actually delete an entry with a pending timer. Fix it with using del_timer_sync(). AFAIK del_timer_sync checks for the timer to be pending by itself, so I remove the check. Signed-off-by: Pavel Emelyanov Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/xt_hashlimit.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'net') diff --git a/net/netfilter/xt_hashlimit.c b/net/netfilter/xt_hashlimit.c index 6809af542a2c..d9418a267812 100644 --- a/net/netfilter/xt_hashlimit.c +++ b/net/netfilter/xt_hashlimit.c @@ -367,9 +367,7 @@ static void htable_gc(unsigned long htlong) static void htable_destroy(struct xt_hashlimit_htable *hinfo) { - /* remove timer, if it is pending */ - if (timer_pending(&hinfo->timer)) - del_timer(&hinfo->timer); + del_timer_sync(&hinfo->timer); /* remove proc entry */ remove_proc_entry(hinfo->pde->name, -- cgit v1.2.1 From c3f26a269c2421f97f10cf8ed05d5099b573af4d Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 31 Jul 2008 16:58:50 -0700 Subject: netdev: Fix lockdep warnings in multiqueue configurations. When support for multiple TX queues were added, the netif_tx_lock() routines we converted to iterate over all TX queues and grab each queue's spinlock. This causes heartburn for lockdep and it's not a healthy thing to do with lots of TX queues anyways. So modify this to use a top-level lock and a "frozen" state for the individual TX queues. Signed-off-by: David S. Miller --- net/core/dev.c | 1 + net/core/netpoll.c | 1 + net/core/pktgen.c | 7 +++++-- net/sched/sch_generic.c | 6 ++++-- net/sched/sch_teql.c | 9 +++++---- 5 files changed, 16 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 63d6bcddbf46..69320a56a084 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4200,6 +4200,7 @@ static void netdev_init_queues(struct net_device *dev) { netdev_init_one_queue(dev, &dev->rx_queue, NULL); netdev_for_each_tx_queue(dev, netdev_init_one_queue, NULL); + spin_lock_init(&dev->tx_global_lock); } /** diff --git a/net/core/netpoll.c b/net/core/netpoll.c index c12720895ecf..6c7af390be0a 100644 --- a/net/core/netpoll.c +++ b/net/core/netpoll.c @@ -70,6 +70,7 @@ static void queue_process(struct work_struct *work) local_irq_save(flags); __netif_tx_lock(txq, smp_processor_id()); if (netif_tx_queue_stopped(txq) || + netif_tx_queue_frozen(txq) || dev->hard_start_xmit(skb, dev) != NETDEV_TX_OK) { skb_queue_head(&npinfo->txq, skb); __netif_tx_unlock(txq); diff --git a/net/core/pktgen.c b/net/core/pktgen.c index c7d484f7e1c4..3284605f2ec7 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -3305,6 +3305,7 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) txq = netdev_get_tx_queue(odev, queue_map); if (netif_tx_queue_stopped(txq) || + netif_tx_queue_frozen(txq) || need_resched()) { idle_start = getCurUs(); @@ -3320,7 +3321,8 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) pkt_dev->idle_acc += getCurUs() - idle_start; - if (netif_tx_queue_stopped(txq)) { + if (netif_tx_queue_stopped(txq) || + netif_tx_queue_frozen(txq)) { pkt_dev->next_tx_us = getCurUs(); /* TODO */ pkt_dev->next_tx_ns = 0; goto out; /* Try the next interface */ @@ -3352,7 +3354,8 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) txq = netdev_get_tx_queue(odev, queue_map); __netif_tx_lock_bh(txq); - if (!netif_tx_queue_stopped(txq)) { + if (!netif_tx_queue_stopped(txq) && + !netif_tx_queue_frozen(txq)) { atomic_inc(&(pkt_dev->skb->users)); retry_now: diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 345838a2e369..9c9cd4d94890 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -135,7 +135,8 @@ static inline int qdisc_restart(struct Qdisc *q) txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb)); HARD_TX_LOCK(dev, txq, smp_processor_id()); - if (!netif_subqueue_stopped(dev, skb)) + if (!netif_tx_queue_stopped(txq) && + !netif_tx_queue_frozen(txq)) ret = dev_hard_start_xmit(skb, dev, txq); HARD_TX_UNLOCK(dev, txq); @@ -162,7 +163,8 @@ static inline int qdisc_restart(struct Qdisc *q) break; } - if (ret && netif_tx_queue_stopped(txq)) + if (ret && (netif_tx_queue_stopped(txq) || + netif_tx_queue_frozen(txq))) ret = 0; return ret; diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 537223642b6e..2c35c678563b 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -305,10 +305,11 @@ restart: switch (teql_resolve(skb, skb_res, slave)) { case 0: - if (netif_tx_trylock(slave)) { - if (!__netif_subqueue_stopped(slave, subq) && + if (__netif_tx_trylock(slave_txq)) { + if (!netif_tx_queue_stopped(slave_txq) && + !netif_tx_queue_frozen(slave_txq) && slave->hard_start_xmit(skb, slave) == 0) { - netif_tx_unlock(slave); + __netif_tx_unlock(slave_txq); master->slaves = NEXT_SLAVE(q); netif_wake_queue(dev); master->stats.tx_packets++; @@ -316,7 +317,7 @@ restart: qdisc_pkt_len(skb); return 0; } - netif_tx_unlock(slave); + __netif_tx_unlock(slave_txq); } if (netif_queue_stopped(dev)) busy = 1; -- cgit v1.2.1 From 77e2f14f71d68d05945f1d30ca55b5194d6ab1ce Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Thu, 31 Jul 2008 20:46:47 -0700 Subject: ipv6: Fix ip6_xmit to send fragments if ipfragok is true SCTP used ip6_xmit() to send fragments after received ICMP packet too big message. But while send packet used ip6_xmit, the skb->local_df is not initialized. So when skb if enter ip6_fragment(), the following code will discard the skb. ip6_fragment(...) { if (!skb->local_df) { ... return -EMSGSIZE; } ... } SCTP do the following step: 1. send packet ip6_xmit(skb, ipfragok=0) 2. received ICMP packet too big message 3. if PMTUD_ENABLE: ip6_xmit(skb, ipfragok=1) This patch fixed the problem by set local_df if ipfragok is true. Signed-off-by: Wei Yongjun Acked-by: Herbert Xu Signed-off-by: David S. Miller --- net/ipv6/ip6_output.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'net') diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 6811901e6b1e..a027003d69a4 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -236,6 +236,10 @@ int ip6_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl, skb_reset_network_header(skb); hdr = ipv6_hdr(skb); + /* Allow local fragmentation. */ + if (ipfragok) + skb->local_df = 1; + /* * Fill in the IPv6 header */ -- cgit v1.2.1 From 90b7e1120bb43ffaabb88d28f80a0c2e13167b15 Mon Sep 17 00:00:00 2001 From: Adam Langley Date: Thu, 31 Jul 2008 20:49:48 -0700 Subject: tcp: MD5: Fix MD5 signatures on certain ACK packets I noticed, looking at tcpdumps, that timewait ACKs were getting sent with an incorrect MD5 signature when signatures were enabled. I broke this in 49a72dfb8814c2d65bd9f8c9c6daf6395a1ec58d ("tcp: Fix MD5 signatures for non-linear skbs"). I didn't take into account that the skb passed to tcp_*_send_ack was the inbound packet, thus the source and dest addresses need to be swapped when calculating the MD5 pseudoheader. Signed-off-by: Adam Langley Signed-off-by: David S. Miller --- net/ipv4/tcp_ipv4.c | 4 ++-- net/ipv6/tcp_ipv6.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index b3875c0d83c7..91a8cfddf1c4 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -655,8 +655,8 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, rep.th.doff = arg.iov[0].iov_len/4; tcp_v4_md5_hash_hdr((__u8 *) &rep.opt[offset], - key, ip_hdr(skb)->daddr, - ip_hdr(skb)->saddr, &rep.th); + key, ip_hdr(skb)->saddr, + ip_hdr(skb)->daddr, &rep.th); } #endif arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr, diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 1db45216b232..1bcdcbc71d68 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1094,8 +1094,8 @@ static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 *topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | (TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG); tcp_v6_md5_hash_hdr((__u8 *)topt, key, - &ipv6_hdr(skb)->daddr, - &ipv6_hdr(skb)->saddr, t1); + &ipv6_hdr(skb)->saddr, + &ipv6_hdr(skb)->daddr, t1); } #endif -- cgit v1.2.1 From 8a9204db665365354b349ed5b0bc054f0433a2a4 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 31 Jul 2008 20:51:22 -0700 Subject: net/ipv4/route.c: fix build error fix: net/ipv4/route.c: In function 'ip_static_sysctl_init': net/ipv4/route.c:3225: error: 'ipv4_route_path' undeclared (first use in this function) net/ipv4/route.c:3225: error: (Each undeclared identifier is reported only once net/ipv4/route.c:3225: error: for each function it appears in.) net/ipv4/route.c:3225: error: 'ipv4_route_table' undeclared (first use in this function) Signed-off-by: Ingo Molnar Signed-off-by: David S. Miller --- net/ipv4/route.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 380d6474cf66..d7bf0564b5f4 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -3222,7 +3222,9 @@ int __init ip_rt_init(void) */ void __init ip_static_sysctl_init(void) { +#ifdef CONFIG_SYSCTL register_sysctl_paths(ipv4_route_path, ipv4_route_table); +#endif } EXPORT_SYMBOL(__ip_select_ident); -- cgit v1.2.1 From 00b1304c4ca81dd893973cc620b87a5c3ff3f660 Mon Sep 17 00:00:00 2001 From: Adam Langley Date: Thu, 31 Jul 2008 21:36:07 -0700 Subject: tcp: MD5: Fix IPv6 signatures Reported by Stefanos Harhalakis; although 2.6.27-rc1 talks to itself using IPv6 TCP MD5 packets just fine, Stefanos noted that tcpdump claimed that the signatures were invalid. I broke this in 49a72dfb8814c2d65bd9f8c9c6daf6395a1ec58d ("tcp: Fix MD5 signatures for non-linear skbs"), it was just a typo. Note that tcpdump will still sometimes claim that the signatures are incorrect. A patch to tcpdump has been submitted for this[1]. [1] http://tinyurl.com/6a4fl2 Signed-off-by: Adam Langley Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 1bcdcbc71d68..78185a409212 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -748,7 +748,7 @@ static int tcp_v6_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp, ipv6_addr_copy(&bp->saddr, saddr); ipv6_addr_copy(&bp->daddr, daddr); bp->protocol = cpu_to_be32(IPPROTO_TCP); - bp->len = cpu_to_be16(nbytes); + bp->len = cpu_to_be32(nbytes); sg_init_one(&sg, bp, sizeof(*bp)); return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp)); -- cgit v1.2.1 From a1bc6eb4b499ae67ada9a01660010580b6569403 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 30 Jul 2008 06:32:52 -0400 Subject: [PATCH] ipv4_static_sysctl_init() should be under CONFIG_SYSCTL Signed-off-by: Al Viro --- net/ipv4/route.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 380d6474cf66..a72a5ad46ec5 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -3216,6 +3216,7 @@ int __init ip_rt_init(void) return rc; } +#ifdef CONFIG_SYSCTL /* * We really need to sanitize the damn ipv4 init order, then all * this nonsense will go away. @@ -3224,6 +3225,7 @@ void __init ip_static_sysctl_init(void) { register_sysctl_paths(ipv4_route_path, ipv4_route_table); } +#endif EXPORT_SYMBOL(__ip_select_ident); EXPORT_SYMBOL(ip_route_input); -- cgit v1.2.1 From 6e28fbef0f330d7c1cade345eeae003d4e5d6070 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Thu, 31 Jul 2008 10:53:57 -0300 Subject: rfkill: query EV_SW states when rfkill-input (re)?connects to a input device Every time a new input device that is capable of one of the rfkill EV_SW events (currently only SW_RFKILL_ALL) is connected to rfkill-input, we must check the states of the input EV_SW switches and take action. Otherwise, we will ignore the initial switch state. We also need to re-check the states of the EV_SW switches after a device that was under an exclusive grab is released back to us, since we got no input events from that device while it was grabbed. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Dmitry Torokhov Signed-off-by: John W. Linville --- net/rfkill/rfkill-input.c | 54 ++++++++++++++++++++++++++++++++++------------- 1 file changed, 39 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill-input.c b/net/rfkill/rfkill-input.c index 8aa822730145..e5b69556bb5b 100644 --- a/net/rfkill/rfkill-input.c +++ b/net/rfkill/rfkill-input.c @@ -109,6 +109,25 @@ static DEFINE_RFKILL_TASK(rfkill_uwb, RFKILL_TYPE_UWB); static DEFINE_RFKILL_TASK(rfkill_wimax, RFKILL_TYPE_WIMAX); static DEFINE_RFKILL_TASK(rfkill_wwan, RFKILL_TYPE_WWAN); +static void rfkill_schedule_evsw_rfkillall(int state) +{ + /* EVERY radio type. state != 0 means radios ON */ + /* handle EPO (emergency power off) through shortcut */ + if (state) { + rfkill_schedule_set(&rfkill_wwan, + RFKILL_STATE_UNBLOCKED); + rfkill_schedule_set(&rfkill_wimax, + RFKILL_STATE_UNBLOCKED); + rfkill_schedule_set(&rfkill_uwb, + RFKILL_STATE_UNBLOCKED); + rfkill_schedule_set(&rfkill_bt, + RFKILL_STATE_UNBLOCKED); + rfkill_schedule_set(&rfkill_wlan, + RFKILL_STATE_UNBLOCKED); + } else + rfkill_schedule_epo(); +} + static void rfkill_event(struct input_handle *handle, unsigned int type, unsigned int code, int data) { @@ -132,21 +151,7 @@ static void rfkill_event(struct input_handle *handle, unsigned int type, } else if (type == EV_SW) { switch (code) { case SW_RFKILL_ALL: - /* EVERY radio type. data != 0 means radios ON */ - /* handle EPO (emergency power off) through shortcut */ - if (data) { - rfkill_schedule_set(&rfkill_wwan, - RFKILL_STATE_UNBLOCKED); - rfkill_schedule_set(&rfkill_wimax, - RFKILL_STATE_UNBLOCKED); - rfkill_schedule_set(&rfkill_uwb, - RFKILL_STATE_UNBLOCKED); - rfkill_schedule_set(&rfkill_bt, - RFKILL_STATE_UNBLOCKED); - rfkill_schedule_set(&rfkill_wlan, - RFKILL_STATE_UNBLOCKED); - } else - rfkill_schedule_epo(); + rfkill_schedule_evsw_rfkillall(data); break; default: break; @@ -168,6 +173,7 @@ static int rfkill_connect(struct input_handler *handler, struct input_dev *dev, handle->handler = handler; handle->name = "rfkill"; + /* causes rfkill_start() to be called */ error = input_register_handle(handle); if (error) goto err_free_handle; @@ -185,6 +191,23 @@ static int rfkill_connect(struct input_handler *handler, struct input_dev *dev, return error; } +static void rfkill_start(struct input_handle *handle) +{ + /* Take event_lock to guard against configuration changes, we + * should be able to deal with concurrency with rfkill_event() + * just fine (which event_lock will also avoid). */ + spin_lock_irq(&handle->dev->event_lock); + + if (test_bit(EV_SW, handle->dev->evbit)) { + if (test_bit(SW_RFKILL_ALL, handle->dev->swbit)) + rfkill_schedule_evsw_rfkillall(test_bit(SW_RFKILL_ALL, + handle->dev->sw)); + /* add resync for further EV_SW events here */ + } + + spin_unlock_irq(&handle->dev->event_lock); +} + static void rfkill_disconnect(struct input_handle *handle) { input_close_device(handle); @@ -225,6 +248,7 @@ static struct input_handler rfkill_handler = { .event = rfkill_event, .connect = rfkill_connect, .disconnect = rfkill_disconnect, + .start = rfkill_start, .name = "rfkill", .id_table = rfkill_ids, }; -- cgit v1.2.1 From 7c4f4578fc85d42d149f86b47f76c28626a20d92 Mon Sep 17 00:00:00 2001 From: Dmitry Baryshkov Date: Tue, 22 Jul 2008 14:17:37 +0400 Subject: RFKILL: allow one to specify led trigger name Allow the rfkill driver to specify led trigger name. By default it still defaults to the name of rfkill switch. Signed-off-by: Dmitry Baryshkov Acked-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index c6f2f388cb72..5c24f3647189 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -589,7 +589,8 @@ static void rfkill_led_trigger_register(struct rfkill *rfkill) #ifdef CONFIG_RFKILL_LEDS int error; - rfkill->led_trigger.name = rfkill->dev.bus_id; + if (!rfkill->led_trigger.name) + rfkill->led_trigger.name = rfkill->dev.bus_id; error = led_trigger_register(&rfkill->led_trigger); if (error) rfkill->led_trigger.name = NULL; -- cgit v1.2.1 From 96185664f10e79d038c084305d3cacff9b52204f Mon Sep 17 00:00:00 2001 From: Dmitry Baryshkov Date: Tue, 22 Jul 2008 14:21:59 +0400 Subject: RFKILL: set the status of the leds on activation. Provide default activate function to set the state of the led when the led becomes bound to the trigger Signed-off-by: Dmitry Baryshkov Acked-by: Ivo van Doorn Acked-by: Henrique de Moraes Holschuh Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 5c24f3647189..d2d45655cd1a 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -105,6 +105,16 @@ static void rfkill_led_trigger(struct rfkill *rfkill, #endif /* CONFIG_RFKILL_LEDS */ } +#ifdef CONFIG_RFKILL_LEDS +static void rfkill_led_trigger_activate(struct led_classdev *led) +{ + struct rfkill *rfkill = container_of(led->trigger, + struct rfkill, led_trigger); + + rfkill_led_trigger(rfkill, rfkill->state); +} +#endif /* CONFIG_RFKILL_LEDS */ + static void notify_rfkill_state_change(struct rfkill *rfkill) { blocking_notifier_call_chain(&rfkill_notifier_list, @@ -591,6 +601,8 @@ static void rfkill_led_trigger_register(struct rfkill *rfkill) if (!rfkill->led_trigger.name) rfkill->led_trigger.name = rfkill->dev.bus_id; + if (!rfkill->led_trigger.activate) + rfkill->led_trigger.activate = rfkill_led_trigger_activate; error = led_trigger_register(&rfkill->led_trigger); if (error) rfkill->led_trigger.name = NULL; -- cgit v1.2.1 From f8e79ddd31c3615ddca26b9a469c44a7adbd4e13 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Thu, 24 Jul 2008 18:46:44 +0300 Subject: mac80211: fix fragmentation kludge This patch make mac80211 transmit correctly fragmented packet after queue was stopped Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 1 + net/mac80211/tx.c | 17 ++++++++++++----- net/mac80211/util.c | 1 + 3 files changed, 14 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index a4f9a832722a..a2e200f9811e 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -586,6 +586,7 @@ struct ieee80211_local { struct timer_list sta_cleanup; unsigned long queues_pending[BITS_TO_LONGS(IEEE80211_MAX_QUEUES)]; + unsigned long queues_pending_run[BITS_TO_LONGS(IEEE80211_MAX_QUEUES)]; struct ieee80211_tx_stored_packet pending_packet[IEEE80211_MAX_QUEUES]; struct tasklet_struct tx_pending_tasklet; diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 69019e943873..771ec68b848d 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1060,13 +1060,14 @@ static int ieee80211_tx_prepare(struct ieee80211_tx_data *tx, static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, struct ieee80211_tx_data *tx) { - struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); + struct ieee80211_tx_info *info; int ret, i; - if (netif_subqueue_stopped(local->mdev, skb)) - return IEEE80211_TX_AGAIN; - if (skb) { + if (netif_subqueue_stopped(local->mdev, skb)) + return IEEE80211_TX_AGAIN; + info = IEEE80211_SKB_CB(skb); + ieee80211_dump_frame(wiphy_name(local->hw.wiphy), "TX to low-level driver", skb); ret = local->ops->tx(local_to_hw(local), skb); @@ -1215,6 +1216,7 @@ retry: if (ret == IEEE80211_TX_FRAG_AGAIN) skb = NULL; + set_bit(queue, local->queues_pending); smp_mb(); /* @@ -1708,14 +1710,19 @@ void ieee80211_tx_pending(unsigned long data) netif_tx_lock_bh(dev); for (i = 0; i < ieee80211_num_regular_queues(&local->hw); i++) { /* Check that this queue is ok */ - if (__netif_subqueue_stopped(local->mdev, i)) + if (__netif_subqueue_stopped(local->mdev, i) && + !test_bit(i, local->queues_pending_run)) continue; if (!test_bit(i, local->queues_pending)) { + clear_bit(i, local->queues_pending_run); ieee80211_wake_queue(&local->hw, i); continue; } + clear_bit(i, local->queues_pending_run); + netif_start_subqueue(local->mdev, i); + store = &local->pending_packet[i]; tx.extra_frag = store->extra_frag; tx.num_extra_frag = store->num_extra_frag; diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 19f85e1b3695..0d463c80c404 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -361,6 +361,7 @@ void ieee80211_wake_queue(struct ieee80211_hw *hw, int queue) struct ieee80211_local *local = hw_to_local(hw); if (test_bit(queue, local->queues_pending)) { + set_bit(queue, local->queues_pending_run); tasklet_schedule(&local->tx_pending_tasklet); } else { netif_wake_subqueue(local->mdev, queue); -- cgit v1.2.1 From 5fb662297b8a4bdadd60371c34b760efca948ebc Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sat, 2 Aug 2008 20:02:43 -0700 Subject: pkt_sched: Use qdisc_lock() on already sampled root qdisc. Based upon a bug report by Jeff Kirsher. Don't use qdisc_root_lock() in these cases as the root qdisc could have been changed, and we'd thus lock the wrong object. Tested by Emil S Tantilov who confirms that this seems to fix the problem. Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- net/sched/sch_generic.c | 12 ++++++------ 2 files changed, 8 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 69320a56a084..da7acacf02b5 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1796,7 +1796,7 @@ gso: skb->tc_verd = SET_TC_AT(skb->tc_verd,AT_EGRESS); #endif if (q->enqueue) { - spinlock_t *root_lock = qdisc_root_lock(q); + spinlock_t *root_lock = qdisc_lock(q); spin_lock(root_lock); @@ -1995,7 +1995,7 @@ static void net_tx_action(struct softirq_action *h) smp_mb__before_clear_bit(); clear_bit(__QDISC_STATE_SCHED, &q->state); - root_lock = qdisc_root_lock(q); + root_lock = qdisc_lock(q); if (spin_trylock(root_lock)) { qdisc_run(q); spin_unlock(root_lock); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 9c9cd4d94890..7cf83b37459d 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -29,7 +29,7 @@ /* Main transmission queue. */ /* Modifications to data participating in scheduling must be protected with - * qdisc_root_lock(qdisc) spinlock. + * qdisc_lock(qdisc) spinlock. * * The idea is the following: * - enqueue, dequeue are serialized via qdisc root lock @@ -126,7 +126,7 @@ static inline int qdisc_restart(struct Qdisc *q) if (unlikely((skb = dequeue_skb(q)) == NULL)) return 0; - root_lock = qdisc_root_lock(q); + root_lock = qdisc_lock(q); /* And release qdisc */ spin_unlock(root_lock); @@ -507,7 +507,7 @@ errout: } EXPORT_SYMBOL(qdisc_create_dflt); -/* Under qdisc_root_lock(qdisc) and BH! */ +/* Under qdisc_lock(qdisc) and BH! */ void qdisc_reset(struct Qdisc *qdisc) { @@ -543,7 +543,7 @@ static void __qdisc_destroy(struct rcu_head *head) kfree((char *) qdisc - qdisc->padded); } -/* Under qdisc_root_lock(qdisc) and BH! */ +/* Under qdisc_lock(qdisc) and BH! */ void qdisc_destroy(struct Qdisc *qdisc) { @@ -659,7 +659,7 @@ static bool some_qdisc_is_running(struct net_device *dev, int lock) dev_queue = netdev_get_tx_queue(dev, i); q = dev_queue->qdisc; - root_lock = qdisc_root_lock(q); + root_lock = qdisc_lock(q); if (lock) spin_lock_bh(root_lock); @@ -735,7 +735,7 @@ static void shutdown_scheduler_queue(struct net_device *dev, struct Qdisc *qdisc_default = _qdisc_default; if (qdisc) { - spinlock_t *root_lock = qdisc_root_lock(qdisc); + spinlock_t *root_lock = qdisc_lock(qdisc); dev_queue->qdisc = qdisc_default; dev_queue->qdisc_sleeping = qdisc_default; -- cgit v1.2.1 From 35ed4e75989c4e84a44b25569bbf09b98f923880 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sat, 2 Aug 2008 23:25:50 -0700 Subject: mac80211: Use queue_lock() in ieee80211_ht_agg_queue_remove(). qdisc_root_lock() is only %100 safe to use when the RTNL semaphore is held. Signed-off-by: David S. Miller --- net/mac80211/wme.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 28437f0001db..4310e2f65661 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -241,12 +241,14 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, } else { struct netdev_queue *txq; spinlock_t *root_lock; + struct Qdisc *q; txq = netdev_get_tx_queue(local->mdev, agg_queue); - root_lock = qdisc_root_lock(txq->qdisc); + q = rcu_dereference(txq->qdisc); + root_lock = qdisc_lock(q); spin_lock_bh(root_lock); - qdisc_reset(txq->qdisc); + qdisc_reset(q); spin_unlock_bh(root_lock); } } -- cgit v1.2.1 From bff69732c9947f821a64a8477f7dcaa9c30e6a69 Mon Sep 17 00:00:00 2001 From: Chris Larson Date: Sun, 3 Aug 2008 01:02:41 -0700 Subject: net: in the first call to neigh_seq_next, call neigh_get_first, not neigh_get_idx. neigh_seq_next won't be called both with *pos > 0 && v == SEQ_START_TOKEN, so there's no point calling neigh_get_idx when we're on the start token, just call neigh_get_first directly. Signed-off-by: Chris Larson Signed-off-by: David S. Miller --- net/core/neighbour.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/core/neighbour.c b/net/core/neighbour.c index f62c8af85d38..a57de755c8cc 100644 --- a/net/core/neighbour.c +++ b/net/core/neighbour.c @@ -2385,7 +2385,7 @@ void *neigh_seq_next(struct seq_file *seq, void *v, loff_t *pos) void *rc; if (v == SEQ_START_TOKEN) { - rc = neigh_get_idx(seq, pos); + rc = neigh_get_first(seq); goto out; } -- cgit v1.2.1 From 745e203164a9057e0de769ff4649e6e455daf753 Mon Sep 17 00:00:00 2001 From: Chris Larson Date: Sun, 3 Aug 2008 01:10:55 -0700 Subject: net: fix missing pneigh entries in the neighbor seq_file code When pneigh entries exist, but the user's read buffer isn't sufficient to hold them all, one of the pneigh entries will be missing from the results. In neigh_get_idx_any, the number of elements which neigh_get_idx encountered is not correctly subtracted from the position number before the call to pneigh_get_idx. neigh_get_idx reduces the position by 1 for each call to neigh_get_next, but it does not reduce it by one for the first element (neigh_get_first). The patch alters the neigh_get_idx and pneigh_get_idx functions to subtract one from pos, for the first element, when pos is non-zero. Signed-off-by: Chris Larson Signed-off-by: David S. Miller --- net/core/neighbour.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/core/neighbour.c b/net/core/neighbour.c index a57de755c8cc..9d92e41826e7 100644 --- a/net/core/neighbour.c +++ b/net/core/neighbour.c @@ -2281,6 +2281,7 @@ static struct neighbour *neigh_get_idx(struct seq_file *seq, loff_t *pos) struct neighbour *n = neigh_get_first(seq); if (n) { + --(*pos); while (*pos) { n = neigh_get_next(seq, n, pos); if (!n) @@ -2341,6 +2342,7 @@ static struct pneigh_entry *pneigh_get_idx(struct seq_file *seq, loff_t *pos) struct pneigh_entry *pn = pneigh_get_first(seq); if (pn) { + --(*pos); while (*pos) { pn = pneigh_get_next(seq, pn, pos); if (!pn) @@ -2354,10 +2356,11 @@ static void *neigh_get_idx_any(struct seq_file *seq, loff_t *pos) { struct neigh_seq_state *state = seq->private; void *rc; + loff_t idxpos = *pos; - rc = neigh_get_idx(seq, pos); + rc = neigh_get_idx(seq, &idxpos); if (!rc && !(state->flags & NEIGH_SEQ_NEIGH_ONLY)) - rc = pneigh_get_idx(seq, pos); + rc = pneigh_get_idx(seq, &idxpos); return rc; } @@ -2366,7 +2369,6 @@ void *neigh_seq_start(struct seq_file *seq, loff_t *pos, struct neigh_table *tbl __acquires(tbl->lock) { struct neigh_seq_state *state = seq->private; - loff_t pos_minus_one; state->tbl = tbl; state->bucket = 0; @@ -2374,8 +2376,7 @@ void *neigh_seq_start(struct seq_file *seq, loff_t *pos, struct neigh_table *tbl read_lock_bh(&tbl->lock); - pos_minus_one = *pos - 1; - return *pos ? neigh_get_idx_any(seq, &pos_minus_one) : SEQ_START_TOKEN; + return *pos ? neigh_get_idx_any(seq, pos) : SEQ_START_TOKEN; } EXPORT_SYMBOL(neigh_seq_start); -- cgit v1.2.1 From e5a4a72d4f88f4389e9340d383ca67031d1b8536 Mon Sep 17 00:00:00 2001 From: Lennert Buytenhek Date: Sun, 3 Aug 2008 01:23:10 -0700 Subject: net: use software GSO for SG+CSUM capable netdevices If a netdevice does not support hardware GSO, allowing the stack to use GSO anyway and then splitting the GSO skb into MSS-sized pieces as it is handed to the netdevice for transmitting is likely still a win as far as throughput and/or CPU usage are concerned, since it reduces the number of trips through the output path. This patch enables the use of GSO on any netdevice that supports SG. If a GSO skb is then sent to a netdevice that supports SG but does not support hardware GSO, net/core/dev.c:dev_hard_start_xmit() will take care of doing the necessary GSO segmentation in software. Signed-off-by: Lennert Buytenhek Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index da7acacf02b5..cbf80098980c 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3988,6 +3988,10 @@ int register_netdevice(struct net_device *dev) } } + /* Enable software GSO if SG is supported. */ + if (dev->features & NETIF_F_SG) + dev->features |= NETIF_F_GSO; + netdev_initialize_kobject(dev); ret = netdev_register_kobject(dev); if (ret) -- cgit v1.2.1 From adf044c8778de98dae29c5ce9973b7e43964674f Mon Sep 17 00:00:00 2001 From: Sven Wegener Date: Sun, 3 Aug 2008 14:06:44 -0700 Subject: net: Add missing extra2 parameter for ip_default_ttl sysctl Commit 76e6ebfb40a2455c18234dcb0f9df37533215461 ("netns: add namespace parameter to rt_cache_flush") acceses the extra2 parameter of the ip_default_ttl ctl_table, but it is never set to a meaningful value. When e84f84f276473dcc673f360e8ff3203148bdf0e2 ("netns: place rt_genid into struct net") is applied, we'll oops in rt_cache_invalidate(). Set extra2 to init_net, to avoid that. Reported-by: Marcin Slusarz Signed-off-by: Sven Wegener Tested-by: Marcin Slusarz Acked-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/sysctl_net_ipv4.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c index 770d827f5ab8..e0689fd7b798 100644 --- a/net/ipv4/sysctl_net_ipv4.c +++ b/net/ipv4/sysctl_net_ipv4.c @@ -232,6 +232,7 @@ static struct ctl_table ipv4_table[] = { .mode = 0644, .proc_handler = &ipv4_doint_and_flush, .strategy = &ipv4_doint_and_flush_strategy, + .extra2 = &init_net, }, { .ctl_name = NET_IPV4_NO_PMTU_DISC, -- cgit v1.2.1 From 1730554f253deb65fe5112c54b2f898d5318a328 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Sun, 3 Aug 2008 18:13:44 -0700 Subject: ipv6: syncookies: free reqsk on xfrm_lookup error cookie_v6_check() did not call reqsk_free() if xfrm_lookup() fails, leaking the request sock. Signed-off-by: Florian Westphal Signed-off-by: David S. Miller --- net/ipv6/syncookies.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) (limited to 'net') diff --git a/net/ipv6/syncookies.c b/net/ipv6/syncookies.c index a46badd1082d..ec394cf5a19b 100644 --- a/net/ipv6/syncookies.c +++ b/net/ipv6/syncookies.c @@ -199,10 +199,8 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb) ireq6 = inet6_rsk(req); treq = tcp_rsk(req); - if (security_inet_conn_request(sk, skb, req)) { - reqsk_free(req); - goto out; - } + if (security_inet_conn_request(sk, skb, req)) + goto out_free; req->mss = mss; ireq->rmt_port = th->source; @@ -255,14 +253,13 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb) fl.fl_ip_dport = inet_rsk(req)->rmt_port; fl.fl_ip_sport = inet_sk(sk)->sport; security_req_classify_flow(req, &fl); - if (ip6_dst_lookup(sk, &dst, &fl)) { - reqsk_free(req); - goto out; - } + if (ip6_dst_lookup(sk, &dst, &fl)) + goto out_free; + if (final_p) ipv6_addr_copy(&fl.fl6_dst, final_p); if ((xfrm_lookup(&dst, &fl, sk, 0)) < 0) - goto out; + goto out_free; } req->window_clamp = tp->window_clamp ? :dst_metric(dst, RTAX_WINDOW); @@ -273,7 +270,10 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb) ireq->rcv_wscale = rcv_wscale; ret = get_cookie_sock(sk, skb, req, dst); - -out: return ret; +out: + return ret; +out_free: + reqsk_free(req); + return NULL; } -- cgit v1.2.1 From cfb266c0ee0ea0b7bfa8189e3a3a80344dec6112 Mon Sep 17 00:00:00 2001 From: Yang Hongyang Date: Sun, 3 Aug 2008 18:16:15 -0700 Subject: ipv6: Fix the return value of Set Hop-by-Hop options header with NULL data pointer When Set Hop-by-Hop options header with NULL data pointer and optlen is not zero use setsockopt(), the kernel successfully return 0 instead of return error EINVAL or EFAULT. This patch fix the problem. Signed-off-by: Yang Hongyang Signed-off-by: David S. Miller --- net/ipv6/ipv6_sockglue.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net') diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c index ea33b26512c2..741cfcd96f88 100644 --- a/net/ipv6/ipv6_sockglue.c +++ b/net/ipv6/ipv6_sockglue.c @@ -346,6 +346,8 @@ static int do_ipv6_setsockopt(struct sock *sk, int level, int optname, */ if (optlen == 0) optval = NULL; + else if (optval == NULL) + goto e_inval; else if (optlen < sizeof(struct ipv6_opt_hdr) || optlen & 0x7 || optlen > 8 * 255) goto e_inval; -- cgit v1.2.1 From f880374c2fe37aad3fa62253a4bc125d7a933aad Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Sun, 3 Aug 2008 21:15:08 -0700 Subject: sctp: Drop ipfargok in sctp_xmit function The ipfragok flag controls whether the packet may be fragmented either on the local host on beyond. The latter is only valid on IPv4. In fact, we never want to do the latter even on IPv4 when PMTU is enabled. This is because even though we can't fragment packets within SCTP due to the prtocol's inherent faults, we can still fragment it at IP layer. By setting the DF bit we will improve the PMTU process. RFC 2960 only says that we SHOULD clear the DF bit in this case, so we're compliant even if we set the DF bit. In fact RFC 4960 no longer has this statement. Once we make this change, we only need to control the local fragmentation. There is already a bit in the skb which controls that, local_df. So this patch sets that instead of using the ipfragok argument. The only complication is that there isn't a struct sock object per transport, so for IPv4 we have to resort to changing the pmtudisc field for every packet. This should be safe though as the protocol is single-threaded. Note that after this patch we can remove ipfragok from the rest of the stack too. Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- net/sctp/ipv6.c | 8 +++++--- net/sctp/output.c | 6 ++---- net/sctp/protocol.c | 9 +++++++-- 3 files changed, 14 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/sctp/ipv6.c b/net/sctp/ipv6.c index a238d6834b33..483a01d0740a 100644 --- a/net/sctp/ipv6.c +++ b/net/sctp/ipv6.c @@ -195,8 +195,7 @@ out: } /* Based on tcp_v6_xmit() in tcp_ipv6.c. */ -static int sctp_v6_xmit(struct sk_buff *skb, struct sctp_transport *transport, - int ipfragok) +static int sctp_v6_xmit(struct sk_buff *skb, struct sctp_transport *transport) { struct sock *sk = skb->sk; struct ipv6_pinfo *np = inet6_sk(sk); @@ -231,7 +230,10 @@ static int sctp_v6_xmit(struct sk_buff *skb, struct sctp_transport *transport, SCTP_INC_STATS(SCTP_MIB_OUTSCTPPACKS); - return ip6_xmit(sk, skb, &fl, np->opt, ipfragok); + if (!(transport->param_flags & SPP_PMTUD_ENABLE)) + skb->local_df = 1; + + return ip6_xmit(sk, skb, &fl, np->opt, 0); } /* Returns the dst cache entry for the given source and destination ip diff --git a/net/sctp/output.c b/net/sctp/output.c index 45684646b1db..0dc4a7dfb234 100644 --- a/net/sctp/output.c +++ b/net/sctp/output.c @@ -586,10 +586,8 @@ int sctp_packet_transmit(struct sctp_packet *packet) SCTP_DEBUG_PRINTK("***sctp_transmit_packet*** skb len %d\n", nskb->len); - if (tp->param_flags & SPP_PMTUD_ENABLE) - (*tp->af_specific->sctp_xmit)(nskb, tp, packet->ipfragok); - else - (*tp->af_specific->sctp_xmit)(nskb, tp, 1); + nskb->local_df = packet->ipfragok; + (*tp->af_specific->sctp_xmit)(nskb, tp); out: packet->size = packet->overhead; diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c index a6e0818bcff5..0b65354aaf64 100644 --- a/net/sctp/protocol.c +++ b/net/sctp/protocol.c @@ -862,16 +862,21 @@ static int sctp_inet_supported_addrs(const struct sctp_sock *opt, /* Wrapper routine that calls the ip transmit routine. */ static inline int sctp_v4_xmit(struct sk_buff *skb, - struct sctp_transport *transport, int ipfragok) + struct sctp_transport *transport) { + struct inet_sock *inet = inet_sk(skb->sk); + SCTP_DEBUG_PRINTK("%s: skb:%p, len:%d, " "src:%u.%u.%u.%u, dst:%u.%u.%u.%u\n", __func__, skb, skb->len, NIPQUAD(skb->rtable->rt_src), NIPQUAD(skb->rtable->rt_dst)); + inet->pmtudisc = transport->param_flags & SPP_PMTUD_ENABLE ? + IP_PMTUDISC_DO : IP_PMTUDISC_DONT; + SCTP_INC_STATS(SCTP_MIB_OUTSCTPPACKS); - return ip_queue_xmit(skb, ipfragok); + return ip_queue_xmit(skb, 0); } static struct sctp_af sctp_af_inet; -- cgit v1.2.1 From 283d07ac201ee9f8aa6dc6f7519436b48760baff Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Sun, 3 Aug 2008 21:15:59 -0700 Subject: ipv6: Do not drop packet if skb->local_df is set to true The old code will drop IPv6 packet if ipfragok is not set, since ipfragok is obsoleted, will be instead by used skb->local_df, so this check must be changed to skb->local_df. This patch fix this problem and not drop packet if skb->local_df is set to true. Signed-off-by: Wei Yongjun Acked-by: Herbert Xu Signed-off-by: David S. Miller --- net/ipv6/ip6_output.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index a027003d69a4..a4402de425d9 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -269,7 +269,7 @@ int ip6_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl, skb->mark = sk->sk_mark; mtu = dst_mtu(dst); - if ((skb->len <= mtu) || ipfragok || skb_is_gso(skb)) { + if ((skb->len <= mtu) || skb->local_df || skb_is_gso(skb)) { IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_OUTREQUESTS); return NF_HOOK(PF_INET6, NF_INET_LOCAL_OUT, skb, NULL, dst->dev, -- cgit v1.2.1 From 6e583ce5242f32e925dcb198f7123256d0798370 Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Sun, 3 Aug 2008 21:29:57 -0700 Subject: net: eliminate refcounting in backlog queue Avoid the overhead of atomic increment/decrement on each received packet. This helps performance of non-NAPI devices (like loopback). Use cleanup function to walk queue on each cpu and clean out any left over packets. Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/core/dev.c | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index cbf80098980c..fc6c9881eca8 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1909,7 +1909,6 @@ int netif_rx(struct sk_buff *skb) if (queue->input_pkt_queue.qlen <= netdev_max_backlog) { if (queue->input_pkt_queue.qlen) { enqueue: - dev_hold(skb->dev); __skb_queue_tail(&queue->input_pkt_queue, skb); local_irq_restore(flags); return NET_RX_SUCCESS; @@ -2270,6 +2269,20 @@ out: return ret; } +/* Network device is going away, flush any packets still pending */ +static void flush_backlog(void *arg) +{ + struct net_device *dev = arg; + struct softnet_data *queue = &__get_cpu_var(softnet_data); + struct sk_buff *skb, *tmp; + + skb_queue_walk_safe(&queue->input_pkt_queue, skb, tmp) + if (skb->dev == dev) { + __skb_unlink(skb, &queue->input_pkt_queue); + kfree_skb(skb); + } +} + static int process_backlog(struct napi_struct *napi, int quota) { int work = 0; @@ -2279,7 +2292,6 @@ static int process_backlog(struct napi_struct *napi, int quota) napi->weight = weight_p; do { struct sk_buff *skb; - struct net_device *dev; local_irq_disable(); skb = __skb_dequeue(&queue->input_pkt_queue); @@ -2288,14 +2300,9 @@ static int process_backlog(struct napi_struct *napi, int quota) local_irq_enable(); break; } - local_irq_enable(); - dev = skb->dev; - netif_receive_skb(skb); - - dev_put(dev); } while (++work < quota && jiffies == start_time); return work; @@ -4169,6 +4176,8 @@ void netdev_run_todo(void) dev->reg_state = NETREG_UNREGISTERED; + on_each_cpu(flush_backlog, dev, 1); + netdev_wait_allrefs(dev); /* paranoia */ -- cgit v1.2.1 From 98f7dfd86cbbd377e2cbc293529681b914296f68 Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Fri, 18 Jul 2008 13:52:59 +0800 Subject: mac80211: pass dtim_period to low level driver This patch adds the dtim_period in ieee80211_bss_conf, this allows the low level driver to know the dtim_period, and to plan power save accordingly. Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Signed-off-by: Zhu Yi Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 1 + net/mac80211/mlme.c | 11 +++++++++++ 2 files changed, 12 insertions(+) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index a2e200f9811e..ec59345af65b 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -82,6 +82,7 @@ struct ieee80211_sta_bss { u8 bssid[ETH_ALEN]; u8 ssid[IEEE80211_MAX_SSID_LEN]; + u8 dtim_period; u16 capability; /* host byte order */ enum ieee80211_band band; int freq; diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index acb04133a95d..591e6331c427 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -551,6 +551,7 @@ static void ieee80211_set_associated(struct net_device *dev, /* set timing information */ sdata->bss_conf.beacon_int = bss->beacon_int; sdata->bss_conf.timestamp = bss->timestamp; + sdata->bss_conf.dtim_period = bss->dtim_period; changed |= ieee80211_handle_bss_capability(sdata, bss); @@ -2688,6 +2689,16 @@ static void ieee80211_rx_bss_info(struct net_device *dev, bss->beacon_int = le16_to_cpu(mgmt->u.beacon.beacon_int); bss->capability = le16_to_cpu(mgmt->u.beacon.capab_info); + if (elems->tim) { + struct ieee80211_tim_ie *tim_ie = + (struct ieee80211_tim_ie *)elems->tim; + bss->dtim_period = tim_ie->dtim_period; + } + + /* set default value for buggy APs */ + if (!elems->tim || bss->dtim_period == 0) + bss->dtim_period = 1; + bss->supp_rates_len = 0; if (elems->supp_rates) { clen = IEEE80211_MAX_SUPP_RATES - bss->supp_rates_len; -- cgit v1.2.1 From ea95bba41e69c616bb1512cf59d22f33266b8568 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Fri, 18 Jul 2008 13:53:00 +0800 Subject: mac80211: make listen_interval be limited by low level driver This patch makes possible for a driver to specify maximal listen interval The possibility for user to configure listen interval is not implemented yet, currently the maximum provided by the driver or 1 is used. Mac80211 uses config handler to set listen interval for to the driver. Signed-off-by: Tomas Winkler Signed-off-by: Emmanuel Grumbach Signed-off-by: Zhu Yi Signed-off-by: John W. Linville --- net/mac80211/main.c | 5 +++++ net/mac80211/mlme.c | 6 ++++-- 2 files changed, 9 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index a4c5b90de769..0c02c471bca2 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1689,6 +1689,11 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) if (local->hw.conf.beacon_int < 10) local->hw.conf.beacon_int = 100; + if (local->hw.max_listen_interval == 0) + local->hw.max_listen_interval = 1; + + local->hw.conf.listen_interval = local->hw.max_listen_interval; + local->wstats_flags |= local->hw.flags & (IEEE80211_HW_SIGNAL_UNSPEC | IEEE80211_HW_SIGNAL_DB | IEEE80211_HW_SIGNAL_DBM) ? diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 591e6331c427..779affd8b8fe 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -774,7 +774,8 @@ static void ieee80211_send_assoc(struct net_device *dev, mgmt->frame_control = IEEE80211_FC(IEEE80211_FTYPE_MGMT, IEEE80211_STYPE_REASSOC_REQ); mgmt->u.reassoc_req.capab_info = cpu_to_le16(capab); - mgmt->u.reassoc_req.listen_interval = cpu_to_le16(1); + mgmt->u.reassoc_req.listen_interval = + cpu_to_le16(local->hw.conf.listen_interval); memcpy(mgmt->u.reassoc_req.current_ap, ifsta->prev_bssid, ETH_ALEN); } else { @@ -782,7 +783,8 @@ static void ieee80211_send_assoc(struct net_device *dev, mgmt->frame_control = IEEE80211_FC(IEEE80211_FTYPE_MGMT, IEEE80211_STYPE_ASSOC_REQ); mgmt->u.assoc_req.capab_info = cpu_to_le16(capab); - mgmt->u.assoc_req.listen_interval = cpu_to_le16(1); + mgmt->u.reassoc_req.listen_interval = + cpu_to_le16(local->hw.conf.listen_interval); } /* SSID */ -- cgit v1.2.1 From 80693ceb78b08baa3b66a900d9225b2cf9c6f0ed Mon Sep 17 00:00:00 2001 From: Daniel Drake Date: Sat, 19 Jul 2008 23:31:17 +0100 Subject: mac80211: automatic IBSS channel selection When joining an ad-hoc network, the user is currently required to specify the channel. The network will not be joined otherwise, unless it happens to be sitting on the currently active channel. This patch implements automatic channel selection when the user has not locked the interface onto a specific channel. Signed-off-by: Daniel Drake Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 20 ++++++++++++++++---- 1 file changed, 16 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 779affd8b8fe..5358420d8796 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -3663,11 +3663,21 @@ static int ieee80211_sta_find_ibss(struct net_device *dev, "%s\n", print_mac(mac, bssid), print_mac(mac2, ifsta->bssid)); #endif /* CONFIG_MAC80211_IBSS_DEBUG */ - if (found && memcmp(ifsta->bssid, bssid, ETH_ALEN) != 0 && - (bss = ieee80211_rx_bss_get(dev, bssid, - local->hw.conf.channel->center_freq, - ifsta->ssid, ifsta->ssid_len))) { + + if (found && memcmp(ifsta->bssid, bssid, ETH_ALEN) != 0) { int ret; + int search_freq; + + if (ifsta->flags & IEEE80211_STA_AUTO_CHANNEL_SEL) + search_freq = bss->freq; + else + search_freq = local->hw.conf.channel->center_freq; + + bss = ieee80211_rx_bss_get(dev, bssid, search_freq, + ifsta->ssid, ifsta->ssid_len); + if (!bss) + goto dont_join; + printk(KERN_DEBUG "%s: Selected IBSS BSSID %s" " based on configured SSID\n", dev->name, print_mac(mac, bssid)); @@ -3675,6 +3685,8 @@ static int ieee80211_sta_find_ibss(struct net_device *dev, ieee80211_rx_bss_put(local, bss); return ret; } + +dont_join: #ifdef CONFIG_MAC80211_IBSS_DEBUG printk(KERN_DEBUG " did not try to join ibss\n"); #endif /* CONFIG_MAC80211_IBSS_DEBUG */ -- cgit v1.2.1 From 4c43e0d0ecd5196ed5c67f64ed2f1860770eed34 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Mon, 4 Aug 2008 16:00:39 +0800 Subject: iwlwifi: HW bug fixes This patch adds few HW bug fixes. Signed-off-by: Tomas Winkler Signed-off-by: Zhu Yi Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 5358420d8796..e1d11c9b6729 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -3920,7 +3920,7 @@ done: if (sdata->vif.type == IEEE80211_IF_TYPE_IBSS) { struct ieee80211_if_sta *ifsta = &sdata->u.sta; if (!(ifsta->flags & IEEE80211_STA_BSSID_SET) || - (!ifsta->state == IEEE80211_IBSS_JOINED && + (!(ifsta->state == IEEE80211_IBSS_JOINED) && !ieee80211_sta_active_ibss(dev))) ieee80211_sta_find_ibss(dev, ifsta); } -- cgit v1.2.1 From 378a2f090f7a478704a372a4869b8a9ac206234e Mon Sep 17 00:00:00 2001 From: Jarek Poplawski Date: Mon, 4 Aug 2008 22:31:03 -0700 Subject: net_sched: Add qdisc __NET_XMIT_STOLEN flag Patrick McHardy noticed: "The other problem that affects all qdiscs supporting actions is TC_ACT_QUEUED/TC_ACT_STOLEN getting mapped to NET_XMIT_SUCCESS even though the packet is not queued, corrupting upper qdiscs' qlen counters." and later explained: "The reason why it translates it at all seems to be to not increase the drops counter. Within a single qdisc this could be avoided by other means easily, upper qdiscs would still increase the counter when we return anything besides NET_XMIT_SUCCESS though. This means we need a new NET_XMIT return value to indicate this to the upper qdiscs. So I'd suggest to introduce NET_XMIT_STOLEN, return that to upper qdiscs and translate it to NET_XMIT_SUCCESS in dev_queue_xmit, similar to NET_XMIT_BYPASS." David Miller noticed: "Maybe these NET_XMIT_* values being passed around should be a set of bits. They could be composed of base meanings, combined with specific attributes. So you could say "NET_XMIT_DROP | __NET_XMIT_NO_DROP_COUNT" The attributes get masked out by the top-level ->enqueue() caller, such that the base meanings are the only thing that make their way up into the stack. If it's only about communication within the qdisc tree, let's simply code it that way." This patch is trying to realize these ideas. Signed-off-by: Jarek Poplawski Signed-off-by: David S. Miller --- net/sched/sch_atm.c | 12 +++++++----- net/sched/sch_cbq.c | 23 +++++++++++++++-------- net/sched/sch_dsmark.c | 8 +++++--- net/sched/sch_hfsc.c | 8 +++++--- net/sched/sch_htb.c | 18 +++++++++++------- net/sched/sch_netem.c | 3 ++- net/sched/sch_prio.c | 8 +++++--- net/sched/sch_red.c | 2 +- net/sched/sch_sfq.c | 2 +- net/sched/sch_tbf.c | 3 ++- 10 files changed, 54 insertions(+), 33 deletions(-) (limited to 'net') diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index 6b517b9dac5b..27dd773481bc 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -415,7 +415,7 @@ static int atm_tc_enqueue(struct sk_buff *skb, struct Qdisc *sch) case TC_ACT_QUEUED: case TC_ACT_STOLEN: kfree_skb(skb); - return NET_XMIT_SUCCESS; + return NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; case TC_ACT_SHOT: kfree_skb(skb); goto drop; @@ -432,9 +432,11 @@ static int atm_tc_enqueue(struct sk_buff *skb, struct Qdisc *sch) ret = qdisc_enqueue(skb, flow->q); if (ret != 0) { drop: __maybe_unused - sch->qstats.drops++; - if (flow) - flow->qstats.drops++; + if (net_xmit_drop_count(ret)) { + sch->qstats.drops++; + if (flow) + flow->qstats.drops++; + } return ret; } sch->bstats.bytes += qdisc_pkt_len(skb); @@ -530,7 +532,7 @@ static int atm_tc_requeue(struct sk_buff *skb, struct Qdisc *sch) if (!ret) { sch->q.qlen++; sch->qstats.requeues++; - } else { + } else if (net_xmit_drop_count(ret)) { sch->qstats.drops++; p->link.qstats.drops++; } diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 14954bf4a683..765ae5659000 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -256,7 +256,7 @@ cbq_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) switch (result) { case TC_ACT_QUEUED: case TC_ACT_STOLEN: - *qerr = NET_XMIT_SUCCESS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; case TC_ACT_SHOT: return NULL; case TC_ACT_RECLASSIFY: @@ -397,9 +397,11 @@ cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch) return ret; } - sch->qstats.drops++; - cbq_mark_toplevel(q, cl); - cl->qstats.drops++; + if (net_xmit_drop_count(ret)) { + sch->qstats.drops++; + cbq_mark_toplevel(q, cl); + cl->qstats.drops++; + } return ret; } @@ -430,8 +432,10 @@ cbq_requeue(struct sk_buff *skb, struct Qdisc *sch) cbq_activate_class(cl); return 0; } - sch->qstats.drops++; - cl->qstats.drops++; + if (net_xmit_drop_count(ret)) { + sch->qstats.drops++; + cl->qstats.drops++; + } return ret; } @@ -664,13 +668,15 @@ static int cbq_reshape_fail(struct sk_buff *skb, struct Qdisc *child) q->rx_class = NULL; if (cl && (cl = cbq_reclassify(skb, cl)) != NULL) { + int ret; cbq_mark_toplevel(q, cl); q->rx_class = cl; cl->q->__parent = sch; - if (qdisc_enqueue(skb, cl->q) == 0) { + ret = qdisc_enqueue(skb, cl->q); + if (ret == NET_XMIT_SUCCESS) { sch->q.qlen++; sch->bstats.packets++; sch->bstats.bytes += qdisc_pkt_len(skb); @@ -678,7 +684,8 @@ static int cbq_reshape_fail(struct sk_buff *skb, struct Qdisc *child) cbq_activate_class(cl); return 0; } - sch->qstats.drops++; + if (net_xmit_drop_count(ret)) + sch->qstats.drops++; return 0; } diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index a935676987e2..7170275d9f99 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -236,7 +236,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch) case TC_ACT_QUEUED: case TC_ACT_STOLEN: kfree_skb(skb); - return NET_XMIT_SUCCESS; + return NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; case TC_ACT_SHOT: goto drop; @@ -254,7 +254,8 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch) err = qdisc_enqueue(skb, p->q); if (err != NET_XMIT_SUCCESS) { - sch->qstats.drops++; + if (net_xmit_drop_count(err)) + sch->qstats.drops++; return err; } @@ -321,7 +322,8 @@ static int dsmark_requeue(struct sk_buff *skb, struct Qdisc *sch) err = p->q->ops->requeue(skb, p->q); if (err != NET_XMIT_SUCCESS) { - sch->qstats.drops++; + if (net_xmit_drop_count(err)) + sch->qstats.drops++; return err; } diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 0ae7d19dcba8..5cf9ae716118 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1166,7 +1166,7 @@ hfsc_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) switch (result) { case TC_ACT_QUEUED: case TC_ACT_STOLEN: - *qerr = NET_XMIT_SUCCESS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; case TC_ACT_SHOT: return NULL; } @@ -1586,8 +1586,10 @@ hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch) err = qdisc_enqueue(skb, cl->qdisc); if (unlikely(err != NET_XMIT_SUCCESS)) { - cl->qstats.drops++; - sch->qstats.drops++; + if (net_xmit_drop_count(err)) { + cl->qstats.drops++; + sch->qstats.drops++; + } return err; } diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 75a40951c4f2..538d79b489ae 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -221,7 +221,7 @@ static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, switch (result) { case TC_ACT_QUEUED: case TC_ACT_STOLEN: - *qerr = NET_XMIT_SUCCESS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; case TC_ACT_SHOT: return NULL; } @@ -572,9 +572,11 @@ static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch) kfree_skb(skb); return ret; #endif - } else if (qdisc_enqueue(skb, cl->un.leaf.q) != NET_XMIT_SUCCESS) { - sch->qstats.drops++; - cl->qstats.drops++; + } else if ((ret = qdisc_enqueue(skb, cl->un.leaf.q)) != NET_XMIT_SUCCESS) { + if (net_xmit_drop_count(ret)) { + sch->qstats.drops++; + cl->qstats.drops++; + } return NET_XMIT_DROP; } else { cl->bstats.packets += @@ -615,10 +617,12 @@ static int htb_requeue(struct sk_buff *skb, struct Qdisc *sch) kfree_skb(skb); return ret; #endif - } else if (cl->un.leaf.q->ops->requeue(skb, cl->un.leaf.q) != + } else if ((ret = cl->un.leaf.q->ops->requeue(skb, cl->un.leaf.q)) != NET_XMIT_SUCCESS) { - sch->qstats.drops++; - cl->qstats.drops++; + if (net_xmit_drop_count(ret)) { + sch->qstats.drops++; + cl->qstats.drops++; + } return NET_XMIT_DROP; } else htb_activate(q, cl); diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index a59085700678..6cd6f2bc749e 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -240,8 +240,9 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) sch->q.qlen++; sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; - } else + } else if (net_xmit_drop_count(ret)) { sch->qstats.drops++; + } pr_debug("netem: enqueue ret %d\n", ret); return ret; diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index f849243eb095..adb1a52b77d3 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -45,7 +45,7 @@ prio_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) switch (err) { case TC_ACT_STOLEN: case TC_ACT_QUEUED: - *qerr = NET_XMIT_SUCCESS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; case TC_ACT_SHOT: return NULL; } @@ -88,7 +88,8 @@ prio_enqueue(struct sk_buff *skb, struct Qdisc *sch) sch->q.qlen++; return NET_XMIT_SUCCESS; } - sch->qstats.drops++; + if (net_xmit_drop_count(ret)) + sch->qstats.drops++; return ret; } @@ -114,7 +115,8 @@ prio_requeue(struct sk_buff *skb, struct Qdisc* sch) sch->qstats.requeues++; return 0; } - sch->qstats.drops++; + if (net_xmit_drop_count(ret)) + sch->qstats.drops++; return NET_XMIT_DROP; } diff --git a/net/sched/sch_red.c b/net/sched/sch_red.c index 3f2d1d7f3bbd..5da05839e225 100644 --- a/net/sched/sch_red.c +++ b/net/sched/sch_red.c @@ -97,7 +97,7 @@ static int red_enqueue(struct sk_buff *skb, struct Qdisc* sch) sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; sch->q.qlen++; - } else { + } else if (net_xmit_drop_count(ret)) { q->stats.pdrop++; sch->qstats.drops++; } diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c index 8589da666568..3a456e1b829a 100644 --- a/net/sched/sch_sfq.c +++ b/net/sched/sch_sfq.c @@ -178,7 +178,7 @@ static unsigned int sfq_classify(struct sk_buff *skb, struct Qdisc *sch, switch (result) { case TC_ACT_STOLEN: case TC_ACT_QUEUED: - *qerr = NET_XMIT_SUCCESS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; case TC_ACT_SHOT: return 0; } diff --git a/net/sched/sch_tbf.c b/net/sched/sch_tbf.c index b296672f7632..7d3b7ff3bf07 100644 --- a/net/sched/sch_tbf.c +++ b/net/sched/sch_tbf.c @@ -135,7 +135,8 @@ static int tbf_enqueue(struct sk_buff *skb, struct Qdisc* sch) ret = qdisc_enqueue(skb, q->qdisc); if (ret != 0) { - sch->qstats.drops++; + if (net_xmit_drop_count(ret)) + sch->qstats.drops++; return ret; } -- cgit v1.2.1 From c27f339af90bb874a7a9c680b17abfd32d4a727b Mon Sep 17 00:00:00 2001 From: Jarek Poplawski Date: Mon, 4 Aug 2008 22:39:11 -0700 Subject: net_sched: Add qdisc __NET_XMIT_BYPASS flag Patrick McHardy noticed that it would be nice to handle NET_XMIT_BYPASS by NET_XMIT_SUCCESS with an internal qdisc flag __NET_XMIT_BYPASS and to remove the mapping from dev_queue_xmit(). David Miller spotted a serious bug in the first version of this patch. Signed-off-by: Jarek Poplawski Signed-off-by: David S. Miller --- net/core/dev.c | 1 - net/sched/sch_atm.c | 2 +- net/sched/sch_cbq.c | 4 ++-- net/sched/sch_dsmark.c | 2 +- net/sched/sch_hfsc.c | 4 ++-- net/sched/sch_htb.c | 6 +++--- net/sched/sch_netem.c | 2 +- net/sched/sch_prio.c | 6 +++--- net/sched/sch_sfq.c | 6 +++--- 9 files changed, 16 insertions(+), 17 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index fc6c9881eca8..01993ad74e76 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1805,7 +1805,6 @@ gso: spin_unlock(root_lock); - rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; goto out; } diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index 27dd773481bc..43d37256c15e 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -457,7 +457,7 @@ drop: __maybe_unused return 0; } tasklet_schedule(&p->task); - return NET_XMIT_BYPASS; + return NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; } /* diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 765ae5659000..4e261ce62f48 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -230,7 +230,7 @@ cbq_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) (cl = cbq_class_lookup(q, prio)) != NULL) return cl; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; for (;;) { int result = 0; defmap = head->defaults; @@ -377,7 +377,7 @@ cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch) q->rx_class = cl; #endif if (cl == NULL) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index 7170275d9f99..edd1298f85f6 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -268,7 +268,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch) drop: kfree_skb(skb); sch->qstats.drops++; - return NET_XMIT_BYPASS; + return NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; } static struct sk_buff *dsmark_dequeue(struct Qdisc *sch) diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 5cf9ae716118..c2b8d9cce3d2 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1159,7 +1159,7 @@ hfsc_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) if (cl->level == 0) return cl; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; tcf = q->root.filter_list; while (tcf && (result = tc_classify(skb, tcf, &res)) >= 0) { #ifdef CONFIG_NET_CLS_ACT @@ -1578,7 +1578,7 @@ hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch) cl = hfsc_classify(skb, sch, &err); if (cl == NULL) { - if (err == NET_XMIT_BYPASS) + if (err & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return err; diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 538d79b489ae..be35422711a3 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -214,7 +214,7 @@ static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, if ((cl = htb_find(skb->priority, sch)) != NULL && cl->level == 0) return cl; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; tcf = q->filter_list; while (tcf && (result = tc_classify(skb, tcf, &res)) >= 0) { #ifdef CONFIG_NET_CLS_ACT @@ -567,7 +567,7 @@ static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch) } #ifdef CONFIG_NET_CLS_ACT } else if (!cl) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; @@ -612,7 +612,7 @@ static int htb_requeue(struct sk_buff *skb, struct Qdisc *sch) } #ifdef CONFIG_NET_CLS_ACT } else if (!cl) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 6cd6f2bc749e..fb0294d0b55e 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -176,7 +176,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) if (count == 0) { sch->qstats.drops++; kfree_skb(skb); - return NET_XMIT_BYPASS; + return NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; } skb_orphan(skb); diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index adb1a52b77d3..eac197610edf 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -38,7 +38,7 @@ prio_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) struct tcf_result res; int err; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; if (TC_H_MAJ(skb->priority) != sch->handle) { err = tc_classify(skb, q->filter_list, &res); #ifdef CONFIG_NET_CLS_ACT @@ -74,7 +74,7 @@ prio_enqueue(struct sk_buff *skb, struct Qdisc *sch) #ifdef CONFIG_NET_CLS_ACT if (qdisc == NULL) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; @@ -103,7 +103,7 @@ prio_requeue(struct sk_buff *skb, struct Qdisc* sch) qdisc = prio_classify(skb, sch, &ret); #ifdef CONFIG_NET_CLS_ACT if (qdisc == NULL) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c index 3a456e1b829a..6e041d10dbdb 100644 --- a/net/sched/sch_sfq.c +++ b/net/sched/sch_sfq.c @@ -171,7 +171,7 @@ static unsigned int sfq_classify(struct sk_buff *skb, struct Qdisc *sch, if (!q->filter_list) return sfq_hash(q, skb) + 1; - *qerr = NET_XMIT_BYPASS; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; result = tc_classify(skb, q->filter_list, &res); if (result >= 0) { #ifdef CONFIG_NET_CLS_ACT @@ -285,7 +285,7 @@ sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch) hash = sfq_classify(skb, sch, &ret); if (hash == 0) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; @@ -339,7 +339,7 @@ sfq_requeue(struct sk_buff *skb, struct Qdisc *sch) hash = sfq_classify(skb, sch, &ret); if (hash == 0) { - if (ret == NET_XMIT_BYPASS) + if (ret & __NET_XMIT_BYPASS) sch->qstats.drops++; kfree_skb(skb); return ret; -- cgit v1.2.1 From ad619800e4e034cad44299b2a22df9eebb043ac3 Mon Sep 17 00:00:00 2001 From: Rami Rosen Date: Tue, 5 Aug 2008 01:21:22 -0700 Subject: bridge: fix compile warning in net/bridge/br_netfilter.c This patch fixes the following warning due to incompatible pointer assignment: net/bridge/br_netfilter.c: In function 'br_netfilter_rtable_init': net/bridge/br_netfilter.c:116: warning: assignment from incompatible pointer type This warning is due to commit 4adf0af6818f3ea52421dc0bae836cfaf20ef72a from July 30 (send correct MTU value in PMTU (revised)). Signed-off-by: Rami Rosen Signed-off-by: David S. Miller --- net/bridge/br_netfilter.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c index 6e280a8a31ee..6a9a6cd74b1e 100644 --- a/net/bridge/br_netfilter.c +++ b/net/bridge/br_netfilter.c @@ -113,7 +113,7 @@ void br_netfilter_rtable_init(struct net_bridge *br) struct rtable *rt = &br->fake_rtable; atomic_set(&rt->u.dst.__refcnt, 1); - rt->u.dst.dev = &br->dev; + rt->u.dst.dev = br->dev; rt->u.dst.path = &rt->u.dst; rt->u.dst.metrics[RTAX_MTU - 1] = 1500; rt->u.dst.flags = DST_NOXFRM; -- cgit v1.2.1 From ef647f1300d69adb8223d970554d59d7e244db6d Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Tue, 5 Aug 2008 18:42:51 -0700 Subject: bridge: Eliminate unnecessary forward delay From: Stephen Hemminger Based upon original patch by Herbert Xu, which contained the following problem description: -------------------- When the forward delay is set to zero, we still delay the setting of the forwarding state by one or possibly two timers depending on whether STP is enabled. This could either turn out to be instantaneous, or horribly slow depending on the load of the machine. As there is nothing preventing us from enabling forwarding straight away, this patch eliminates this potential delay by executing the code directly if the forward delay is zero. The effect of this problem is that immediately after the carrier comes on a port, the bridge will drop all packets received from that port until it enters forwarding mode, thus causing unnecessary packet loss. Note that this patch doesn't fully remove the delay due to the link watcher. We should also check the carrier state when we are about to drop an incoming packet because the port is disabled. But that's for another patch. -------------------- This version of the fix takes a different approach, in that it just does the state change directly. Signed-off-by: David S. Miller --- net/bridge/br_stp.c | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/bridge/br_stp.c b/net/bridge/br_stp.c index 921bbe5cb94a..6e63ec3f1fcf 100644 --- a/net/bridge/br_stp.c +++ b/net/bridge/br_stp.c @@ -368,14 +368,25 @@ static void br_make_blocking(struct net_bridge_port *p) /* called under bridge lock */ static void br_make_forwarding(struct net_bridge_port *p) { - if (p->state == BR_STATE_BLOCKING) { - if (p->br->stp_enabled == BR_KERNEL_STP) - p->state = BR_STATE_LISTENING; - else - p->state = BR_STATE_LEARNING; + struct net_bridge *br = p->br; - br_log_state(p); - mod_timer(&p->forward_delay_timer, jiffies + p->br->forward_delay); } + if (p->state != BR_STATE_BLOCKING) + return; + + if (br->forward_delay == 0) { + p->state = BR_STATE_FORWARDING; + br_topology_change_detection(br); + del_timer(&p->forward_delay_timer); + } + else if (p->br->stp_enabled == BR_KERNEL_STP) + p->state = BR_STATE_LISTENING; + else + p->state = BR_STATE_LEARNING; + + br_log_state(p); + + if (br->forward_delay != 0) + mod_timer(&p->forward_delay_timer, jiffies + br->forward_delay); } /* called under bridge lock */ -- cgit v1.2.1 From 1211a64554065316e02b3c62b320088ad4f63260 Mon Sep 17 00:00:00 2001 From: Robert Olsson Date: Tue, 5 Aug 2008 18:44:26 -0700 Subject: pktgen: random flow Random flow generation has not worked. This fixes it. Signed-off-by: Robert Olsson Signed-off-by: David S. Miller --- net/core/pktgen.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/core/pktgen.c b/net/core/pktgen.c index 3284605f2ec7..9c87320fdf3f 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -2085,15 +2085,19 @@ static inline int f_pick(struct pktgen_dev *pkt_dev) if (pkt_dev->flows[flow].count >= pkt_dev->lflow) { /* reset time */ pkt_dev->flows[flow].count = 0; + pkt_dev->flows[flow].flags = 0; pkt_dev->curfl += 1; if (pkt_dev->curfl >= pkt_dev->cflows) pkt_dev->curfl = 0; /*reset */ } } else { flow = random32() % pkt_dev->cflows; + pkt_dev->curfl = flow; - if (pkt_dev->flows[flow].count > pkt_dev->lflow) + if (pkt_dev->flows[flow].count > pkt_dev->lflow) { pkt_dev->flows[flow].count = 0; + pkt_dev->flows[flow].flags = 0; + } } return pkt_dev->curfl; -- cgit v1.2.1 From ff2a79a5a934fe0dbb136ffad61f79b5c6385614 Mon Sep 17 00:00:00 2001 From: Robert Olsson Date: Tue, 5 Aug 2008 18:45:05 -0700 Subject: pktgen: mac count dst_mac_count and src_mac_count patch from Eneas Hunguana We have sent one mac address to much. Signed-off-by: Robert Olsson Signed-off-by: David S. Miller --- net/core/pktgen.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/core/pktgen.c b/net/core/pktgen.c index 9c87320fdf3f..2498cdaf8cbe 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -2166,7 +2166,7 @@ static void mod_cur_headers(struct pktgen_dev *pkt_dev) mc = random32() % pkt_dev->src_mac_count; else { mc = pkt_dev->cur_src_mac_offset++; - if (pkt_dev->cur_src_mac_offset > + if (pkt_dev->cur_src_mac_offset >= pkt_dev->src_mac_count) pkt_dev->cur_src_mac_offset = 0; } @@ -2193,7 +2193,7 @@ static void mod_cur_headers(struct pktgen_dev *pkt_dev) else { mc = pkt_dev->cur_dst_mac_offset++; - if (pkt_dev->cur_dst_mac_offset > + if (pkt_dev->cur_dst_mac_offset >= pkt_dev->dst_mac_count) { pkt_dev->cur_dst_mac_offset = 0; } -- cgit v1.2.1 From ffb208479bd62ab26c29a242faeb1de1c6d5fcdc Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Tue, 5 Aug 2008 18:46:57 -0700 Subject: AX.25: Fix sysctl registration if !CONFIG_AX25_DAMA_SLAVE Since 49ffcf8f99e8d33ec8afb450956804af518fd788 ("sysctl: update sysctl_check_table") setting struct ctl_table.procname = NULL does no longer work as it used to the way the AX.25 code is expecting it to resulting in the AX.25 sysctl registration code to break if CONFIG_AX25_DAMA_SLAVE was not set as in some distribution kernels. Kernel releases from 2.6.24 are affected. Signed-off-by: Ralf Baechle Signed-off-by: David S. Miller --- net/ax25/sysctl_net_ax25.c | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/ax25/sysctl_net_ax25.c b/net/ax25/sysctl_net_ax25.c index f597987b2424..f288fc4aef9b 100644 --- a/net/ax25/sysctl_net_ax25.c +++ b/net/ax25/sysctl_net_ax25.c @@ -36,6 +36,7 @@ static struct ctl_path ax25_path[] = { { .procname = "ax25", .ctl_name = NET_AX25, }, { } }; + static const ctl_table ax25_param_table[] = { { .ctl_name = NET_AX25_IP_DEFAULT_MODE, @@ -167,6 +168,7 @@ static const ctl_table ax25_param_table[] = { .extra1 = &min_proto, .extra2 = &max_proto }, +#ifdef CONFIG_AX25_DAMA_SLAVE { .ctl_name = NET_AX25_DAMA_SLAVE_TIMEOUT, .procname = "dama_slave_timeout", @@ -177,6 +179,8 @@ static const ctl_table ax25_param_table[] = { .extra1 = &min_ds_timeout, .extra2 = &max_ds_timeout }, +#endif + { .ctl_name = 0 } /* that's all, folks! */ }; @@ -210,16 +214,6 @@ void ax25_register_sysctl(void) ax25_table[n].procname = ax25_dev->dev->name; ax25_table[n].mode = 0555; -#ifndef CONFIG_AX25_DAMA_SLAVE - /* - * We do not wish to have a representation of this parameter - * in /proc/sys/ when configured *not* to include the - * AX.25 DAMA slave code, do we? - */ - - child[AX25_VALUES_DS_TIMEOUT].procname = NULL; -#endif - child[AX25_MAX_VALUES].ctl_name = 0; /* just in case... */ for (k = 0; k < AX25_MAX_VALUES; k++) -- cgit v1.2.1