From 8393b811f38acdf7fd8da2028708edad3e68ce1f Mon Sep 17 00:00:00 2001 From: Gabriele Mazzotta Date: Sat, 25 Apr 2015 19:52:36 +0200 Subject: libata: Add helper to determine when PHY events should be ignored This is a preparation commit that will allow to add other criteria according to which PHY events should be dropped. Signed-off-by: Gabriele Mazzotta Signed-off-by: Tejun Heo Cc: stable@vger.kernel.org --- include/linux/libata.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/libata.h b/include/linux/libata.h index 8dad4a307bb8..6138d87277af 100644 --- a/include/linux/libata.h +++ b/include/linux/libata.h @@ -1201,6 +1201,7 @@ extern struct ata_device *ata_dev_pair(struct ata_device *adev); extern int ata_do_set_mode(struct ata_link *link, struct ata_device **r_failed_dev); extern void ata_scsi_port_error_handler(struct Scsi_Host *host, struct ata_port *ap); extern void ata_scsi_cmd_error_handler(struct Scsi_Host *host, struct ata_port *ap, struct list_head *eh_q); +extern bool sata_lpm_ignore_phy_events(struct ata_link *link); extern int ata_cable_40wire(struct ata_port *ap); extern int ata_cable_80wire(struct ata_port *ap); -- cgit v1.2.3-71-gd317 From 09c5b4803a80a5451d950d6a539d2eb311dc0fb1 Mon Sep 17 00:00:00 2001 From: Gabriele Mazzotta Date: Sat, 25 Apr 2015 19:52:37 +0200 Subject: libata: Ignore spurious PHY event on LPM policy change When the LPM policy is set to ATA_LPM_MAX_POWER, the device might generate a spurious PHY event that cuases errors on the link. Ignore this event if it occured within 10s after the policy change. The timeout was chosen observing that on a Dell XPS13 9333 these spurious events can occur up to roughly 6s after the policy change. Link: http://lkml.kernel.org/g/3352987.ugV1Ipy7Z5@xps13 Signed-off-by: Gabriele Mazzotta Signed-off-by: Tejun Heo Cc: stable@vger.kernel.org --- drivers/ata/libata-core.c | 15 ++++++++++++++- drivers/ata/libata-eh.c | 3 +++ include/linux/libata.h | 9 +++++++++ 3 files changed, 26 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c index 12adcf78f94b..85e659945c12 100644 --- a/drivers/ata/libata-core.c +++ b/drivers/ata/libata-core.c @@ -6766,8 +6766,21 @@ u32 ata_wait_register(struct ata_port *ap, void __iomem *reg, u32 mask, u32 val, */ bool sata_lpm_ignore_phy_events(struct ata_link *link) { + unsigned long lpm_timeout = link->last_lpm_change + + msecs_to_jiffies(ATA_TMOUT_SPURIOUS_PHY); + /* if LPM is enabled, PHYRDY doesn't mean anything */ - return !!(link->lpm_policy > ATA_LPM_MAX_POWER); + if (link->lpm_policy > ATA_LPM_MAX_POWER) + return true; + + /* ignore the first PHY event after the LPM policy changed + * as it is might be spurious + */ + if ((link->flags & ATA_LFLAG_CHANGED) && + time_before(jiffies, lpm_timeout)) + return true; + + return false; } EXPORT_SYMBOL_GPL(sata_lpm_ignore_phy_events); diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c index 07f41be38fbe..cf0022ec07f2 100644 --- a/drivers/ata/libata-eh.c +++ b/drivers/ata/libata-eh.c @@ -3597,6 +3597,9 @@ static int ata_eh_set_lpm(struct ata_link *link, enum ata_lpm_policy policy, } } + link->last_lpm_change = jiffies; + link->flags |= ATA_LFLAG_CHANGED; + return 0; fail: diff --git a/include/linux/libata.h b/include/linux/libata.h index 6138d87277af..28aeae46f355 100644 --- a/include/linux/libata.h +++ b/include/linux/libata.h @@ -205,6 +205,7 @@ enum { ATA_LFLAG_SW_ACTIVITY = (1 << 7), /* keep activity stats */ ATA_LFLAG_NO_LPM = (1 << 8), /* disable LPM on this link */ ATA_LFLAG_RST_ONCE = (1 << 9), /* limit recovery to one reset */ + ATA_LFLAG_CHANGED = (1 << 10), /* LPM state changed on this link */ /* struct ata_port flags */ ATA_FLAG_SLAVE_POSS = (1 << 0), /* host supports slave dev */ @@ -309,6 +310,12 @@ enum { */ ATA_TMOUT_PMP_SRST_WAIT = 5000, + /* When the LPM policy is set to ATA_LPM_MAX_POWER, there might + * be a spurious PHY event, so ignore the first PHY event that + * occurs within 10s after the policy change. + */ + ATA_TMOUT_SPURIOUS_PHY = 10000, + /* ATA bus states */ BUS_UNKNOWN = 0, BUS_DMA = 1, @@ -788,6 +795,8 @@ struct ata_link { struct ata_eh_context eh_context; struct ata_device device[ATA_MAX_DEVICES]; + + unsigned long last_lpm_change; /* when last LPM change happened */ }; #define ATA_LINK_CLEAR_BEGIN offsetof(struct ata_link, active_tag) #define ATA_LINK_CLEAR_END offsetof(struct ata_link, device[0]) -- cgit v1.2.3-71-gd317 From 0df48c26d8418c5c9fba63fac15b660d70ca2f1c Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 28 Apr 2015 15:28:17 -0700 Subject: tcp: add tcpi_bytes_acked to tcp_info This patch tracks total number of bytes acked for a TCP socket. This is the sum of all changes done to tp->snd_una, and allows for precise tracking of delivered data. RFC4898 named this : tcpEStatsAppHCThruOctetsAcked This is a 64bit field, and can be fetched both from TCP_INFO getsockopt() if one has a handle on a TCP socket, or from inet_diag netlink facility (iproute2/ss patch will follow) Note that tp->bytes_acked was placed near tp->snd_una for best data locality and minimal performance impact. Signed-off-by: Eric Dumazet Acked-by: Yuchung Cheng Cc: Matt Mathis Cc: Eric Salo Cc: Martin Lau Cc: Chris Rapier Signed-off-by: David S. Miller --- include/linux/tcp.h | 4 ++++ include/net/tcp.h | 2 +- include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 6 +++++- net/ipv4/tcp_input.c | 13 +++++++++++-- 5 files changed, 22 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/include/linux/tcp.h b/include/linux/tcp.h index 0caa3a2d4106..0f73b43171da 100644 --- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -150,6 +150,10 @@ struct tcp_sock { u32 rcv_wup; /* rcv_nxt on last window update sent */ u32 snd_nxt; /* Next sequence we send */ + u64 bytes_acked; /* RFC4898 tcpEStatsAppHCThruOctetsAcked + * sum(delta(snd_una)), or how many bytes + * were acked. + */ u32 snd_una; /* First byte we want an ack for */ u32 snd_sml; /* Last byte of the most recently transmitted small packet */ u32 rcv_tstamp; /* timestamp of last received ACK (for keepalives) */ diff --git a/include/net/tcp.h b/include/net/tcp.h index 051dc5c2802d..dd7b4ea6a10c 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -576,7 +576,7 @@ static inline int tcp_bound_to_half_wnd(struct tcp_sock *tp, int pktsize) } /* tcp.c */ -void tcp_get_info(const struct sock *, struct tcp_info *); +void tcp_get_info(struct sock *, struct tcp_info *); /* Read 'sendfile()'-style from a TCP socket */ typedef int (*sk_read_actor_t)(read_descriptor_t *, struct sk_buff *, diff --git a/include/uapi/linux/tcp.h b/include/uapi/linux/tcp.h index 3b9718328d8b..6666e98a0af9 100644 --- a/include/uapi/linux/tcp.h +++ b/include/uapi/linux/tcp.h @@ -189,6 +189,7 @@ struct tcp_info { __u64 tcpi_pacing_rate; __u64 tcpi_max_pacing_rate; + __u64 tcpi_bytes_acked; /* RFC4898 tcpEStatsAppHCThruOctetsAcked */ }; /* for TCP_MD5SIG socket option */ diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 8c5cd9efebbc..4bf0e8ca7b5b 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -2592,7 +2592,7 @@ EXPORT_SYMBOL(compat_tcp_setsockopt); #endif /* Return information about state of tcp endpoint in API format. */ -void tcp_get_info(const struct sock *sk, struct tcp_info *info) +void tcp_get_info(struct sock *sk, struct tcp_info *info) { const struct tcp_sock *tp = tcp_sk(sk); const struct inet_connection_sock *icsk = inet_csk(sk); @@ -2663,6 +2663,10 @@ void tcp_get_info(const struct sock *sk, struct tcp_info *info) rate = READ_ONCE(sk->sk_max_pacing_rate); info->tcpi_max_pacing_rate = rate != ~0U ? rate : ~0ULL; + + spin_lock_bh(&sk->sk_lock.slock); + info->tcpi_bytes_acked = tp->bytes_acked; + spin_unlock_bh(&sk->sk_lock.slock); } EXPORT_SYMBOL_GPL(tcp_get_info); diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 3a4d9b34bed4..378d3f4d4dc3 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3280,6 +3280,15 @@ static inline bool tcp_may_update_window(const struct tcp_sock *tp, (ack_seq == tp->snd_wl1 && nwin > tp->snd_wnd); } +/* If we update tp->snd_una, also update tp->bytes_acked */ +static void tcp_snd_una_update(struct tcp_sock *tp, u32 ack) +{ + u32 delta = ack - tp->snd_una; + + tp->bytes_acked += delta; + tp->snd_una = ack; +} + /* Update our send window. * * Window update algorithm, described in RFC793/RFC1122 (used in linux-2.2 @@ -3315,7 +3324,7 @@ static int tcp_ack_update_window(struct sock *sk, const struct sk_buff *skb, u32 } } - tp->snd_una = ack; + tcp_snd_una_update(tp, ack); return flag; } @@ -3497,7 +3506,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) * Note, we use the fact that SND.UNA>=SND.WL2. */ tcp_update_wl(tp, ack_seq); - tp->snd_una = ack; + tcp_snd_una_update(tp, ack); flag |= FLAG_WIN_UPDATE; tcp_in_ack_event(sk, CA_ACK_WIN_UPDATE); -- cgit v1.2.3-71-gd317 From bdd1f9edacb5f5835d1e6276571bbbe5b88ded48 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 28 Apr 2015 15:28:18 -0700 Subject: tcp: add tcpi_bytes_received to tcp_info This patch tracks total number of payload bytes received on a TCP socket. This is the sum of all changes done to tp->rcv_nxt RFC4898 named this : tcpEStatsAppHCThruOctetsReceived This is a 64bit field, and can be fetched both from TCP_INFO getsockopt() if one has a handle on a TCP socket, or from inet_diag netlink facility (iproute2/ss patch will follow) Note that tp->bytes_received was placed near tp->rcv_nxt for best data locality and minimal performance impact. Signed-off-by: Eric Dumazet Cc: Yuchung Cheng Cc: Matt Mathis Cc: Eric Salo Cc: Martin Lau Cc: Chris Rapier Acked-by: Yuchung Cheng Signed-off-by: David S. Miller --- include/linux/tcp.h | 4 ++++ include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 1 + net/ipv4/tcp_fastopen.c | 1 + net/ipv4/tcp_input.c | 17 +++++++++++++---- 5 files changed, 20 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/include/linux/tcp.h b/include/linux/tcp.h index 0f73b43171da..3b2911502a8c 100644 --- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -145,6 +145,10 @@ struct tcp_sock { * read the code and the spec side by side (and laugh ...) * See RFC793 and RFC1122. The RFC writes these in capitals. */ + u64 bytes_received; /* RFC4898 tcpEStatsAppHCThruOctetsReceived + * sum(delta(rcv_nxt)), or how many bytes + * were acked. + */ u32 rcv_nxt; /* What we want to receive next */ u32 copied_seq; /* Head of yet unread data */ u32 rcv_wup; /* rcv_nxt on last window update sent */ diff --git a/include/uapi/linux/tcp.h b/include/uapi/linux/tcp.h index 6666e98a0af9..a48f93f3207b 100644 --- a/include/uapi/linux/tcp.h +++ b/include/uapi/linux/tcp.h @@ -190,6 +190,7 @@ struct tcp_info { __u64 tcpi_pacing_rate; __u64 tcpi_max_pacing_rate; __u64 tcpi_bytes_acked; /* RFC4898 tcpEStatsAppHCThruOctetsAcked */ + __u64 tcpi_bytes_received; /* RFC4898 tcpEStatsAppHCThruOctetsReceived */ }; /* for TCP_MD5SIG socket option */ diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 4bf0e8ca7b5b..99fcc0b22c92 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -2666,6 +2666,7 @@ void tcp_get_info(struct sock *sk, struct tcp_info *info) spin_lock_bh(&sk->sk_lock.slock); info->tcpi_bytes_acked = tp->bytes_acked; + info->tcpi_bytes_received = tp->bytes_received; spin_unlock_bh(&sk->sk_lock.slock); } EXPORT_SYMBOL_GPL(tcp_get_info); diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c index e3d87aca6be8..3c673d5e6cff 100644 --- a/net/ipv4/tcp_fastopen.c +++ b/net/ipv4/tcp_fastopen.c @@ -206,6 +206,7 @@ static bool tcp_fastopen_create_child(struct sock *sk, skb_set_owner_r(skb2, child); __skb_queue_tail(&child->sk_receive_queue, skb2); tp->syn_data_acked = 1; + tp->bytes_received = end_seq - TCP_SKB_CB(skb)->seq - 1; } else { end_seq = TCP_SKB_CB(skb)->seq + 1; } diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 378d3f4d4dc3..7e6962bcfc30 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3289,6 +3289,15 @@ static void tcp_snd_una_update(struct tcp_sock *tp, u32 ack) tp->snd_una = ack; } +/* If we update tp->rcv_nxt, also update tp->bytes_received */ +static void tcp_rcv_nxt_update(struct tcp_sock *tp, u32 seq) +{ + u32 delta = seq - tp->rcv_nxt; + + tp->bytes_received += delta; + tp->rcv_nxt = seq; +} + /* Update our send window. * * Window update algorithm, described in RFC793/RFC1122 (used in linux-2.2 @@ -4245,7 +4254,7 @@ static void tcp_ofo_queue(struct sock *sk) tail = skb_peek_tail(&sk->sk_receive_queue); eaten = tail && tcp_try_coalesce(sk, tail, skb, &fragstolen); - tp->rcv_nxt = TCP_SKB_CB(skb)->end_seq; + tcp_rcv_nxt_update(tp, TCP_SKB_CB(skb)->end_seq); if (!eaten) __skb_queue_tail(&sk->sk_receive_queue, skb); if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN) @@ -4413,7 +4422,7 @@ static int __must_check tcp_queue_rcv(struct sock *sk, struct sk_buff *skb, int __skb_pull(skb, hdrlen); eaten = (tail && tcp_try_coalesce(sk, tail, skb, fragstolen)) ? 1 : 0; - tcp_sk(sk)->rcv_nxt = TCP_SKB_CB(skb)->end_seq; + tcp_rcv_nxt_update(tcp_sk(sk), TCP_SKB_CB(skb)->end_seq); if (!eaten) { __skb_queue_tail(&sk->sk_receive_queue, skb); skb_set_owner_r(skb, sk); @@ -4506,7 +4515,7 @@ queue_and_out: eaten = tcp_queue_rcv(sk, skb, 0, &fragstolen); } - tp->rcv_nxt = TCP_SKB_CB(skb)->end_seq; + tcp_rcv_nxt_update(tp, TCP_SKB_CB(skb)->end_seq); if (skb->len) tcp_event_data_recv(sk, skb); if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN) @@ -5254,7 +5263,7 @@ void tcp_rcv_established(struct sock *sk, struct sk_buff *skb, tcp_rcv_rtt_measure_ts(sk, skb); __skb_pull(skb, tcp_header_len); - tp->rcv_nxt = TCP_SKB_CB(skb)->end_seq; + tcp_rcv_nxt_update(tp, TCP_SKB_CB(skb)->end_seq); NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPHPHITSTOUSER); eaten = 1; } -- cgit v1.2.3-71-gd317 From 64f40ff5bbdb1b679fb3c4dbc8230d6517d2b8dc Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 28 Apr 2015 16:23:48 -0700 Subject: tcp: prepare CC get_info() access from getsockopt() We would like that optional info provided by Congestion Control modules using netlink can also be read using getsockopt() This patch changes get_info() to put this information in a buffer, instead of skb, like tcp_get_info(), so that following patch can reuse this common infrastructure. Signed-off-by: Eric Dumazet Cc: Yuchung Cheng Cc: Neal Cardwell Acked-by: Neal Cardwell Acked-by: Daniel Borkmann Acked-by: Yuchung Cheng Signed-off-by: David S. Miller --- include/net/tcp.h | 5 ++++- include/uapi/linux/inet_diag.h | 4 ++++ net/ipv4/inet_diag.c | 8 +++++--- net/ipv4/tcp_dctcp.c | 20 ++++++++++---------- net/ipv4/tcp_illinois.c | 21 +++++++++++---------- net/ipv4/tcp_vegas.c | 19 ++++++++++--------- net/ipv4/tcp_vegas.h | 3 ++- 7 files changed, 46 insertions(+), 34 deletions(-) (limited to 'include') diff --git a/include/net/tcp.h b/include/net/tcp.h index dd7b4ea6a10c..6d204f3f9df8 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -804,6 +804,8 @@ enum tcp_ca_ack_event_flags { /* Requires ECN/ECT set on all packets */ #define TCP_CONG_NEEDS_ECN 0x2 +union tcp_cc_info; + struct tcp_congestion_ops { struct list_head list; u32 key; @@ -829,7 +831,8 @@ struct tcp_congestion_ops { /* hook for packet ack accounting (optional) */ void (*pkts_acked)(struct sock *sk, u32 num_acked, s32 rtt_us); /* get info for inet_diag (optional) */ - int (*get_info)(struct sock *sk, u32 ext, struct sk_buff *skb); + size_t (*get_info)(struct sock *sk, u32 ext, int *attr, + union tcp_cc_info *info); char name[TCP_CA_NAME_MAX]; struct module *owner; diff --git a/include/uapi/linux/inet_diag.h b/include/uapi/linux/inet_diag.h index d65c0a09efd3..c7093c75bdd6 100644 --- a/include/uapi/linux/inet_diag.h +++ b/include/uapi/linux/inet_diag.h @@ -143,4 +143,8 @@ struct tcp_dctcp_info { __u32 dctcp_ab_tot; }; +union tcp_cc_info { + struct tcpvegas_info vegas; + struct tcp_dctcp_info dctcp; +}; #endif /* _UAPI_INET_DIAG_H_ */ diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c index bb77ebdae3b3..4d32262c7502 100644 --- a/net/ipv4/inet_diag.c +++ b/net/ipv4/inet_diag.c @@ -224,14 +224,16 @@ int inet_sk_diag_fill(struct sock *sk, struct inet_connection_sock *icsk, handler->idiag_get_info(sk, r, info); if (sk->sk_state < TCP_TIME_WAIT) { - int err = 0; + union tcp_cc_info info; + size_t sz = 0; + int attr; rcu_read_lock(); ca_ops = READ_ONCE(icsk->icsk_ca_ops); if (ca_ops && ca_ops->get_info) - err = ca_ops->get_info(sk, ext, skb); + sz = ca_ops->get_info(sk, ext, &attr, &info); rcu_read_unlock(); - if (err < 0) + if (sz && nla_put(skb, attr, sz, &info) < 0) goto errout; } diff --git a/net/ipv4/tcp_dctcp.c b/net/ipv4/tcp_dctcp.c index 4376016f7fa5..4c41c1287197 100644 --- a/net/ipv4/tcp_dctcp.c +++ b/net/ipv4/tcp_dctcp.c @@ -277,7 +277,8 @@ static void dctcp_cwnd_event(struct sock *sk, enum tcp_ca_event ev) } } -static int dctcp_get_info(struct sock *sk, u32 ext, struct sk_buff *skb) +static size_t dctcp_get_info(struct sock *sk, u32 ext, int *attr, + union tcp_cc_info *info) { const struct dctcp *ca = inet_csk_ca(sk); @@ -286,18 +287,17 @@ static int dctcp_get_info(struct sock *sk, u32 ext, struct sk_buff *skb) */ if (ext & (1 << (INET_DIAG_DCTCPINFO - 1)) || ext & (1 << (INET_DIAG_VEGASINFO - 1))) { - struct tcp_dctcp_info info; - - memset(&info, 0, sizeof(info)); + memset(info, 0, sizeof(struct tcp_dctcp_info)); if (inet_csk(sk)->icsk_ca_ops != &dctcp_reno) { - info.dctcp_enabled = 1; - info.dctcp_ce_state = (u16) ca->ce_state; - info.dctcp_alpha = ca->dctcp_alpha; - info.dctcp_ab_ecn = ca->acked_bytes_ecn; - info.dctcp_ab_tot = ca->acked_bytes_total; + info->dctcp.dctcp_enabled = 1; + info->dctcp.dctcp_ce_state = (u16) ca->ce_state; + info->dctcp.dctcp_alpha = ca->dctcp_alpha; + info->dctcp.dctcp_ab_ecn = ca->acked_bytes_ecn; + info->dctcp.dctcp_ab_tot = ca->acked_bytes_total; } - return nla_put(skb, INET_DIAG_DCTCPINFO, sizeof(info), &info); + *attr = INET_DIAG_DCTCPINFO; + return sizeof(*info); } return 0; } diff --git a/net/ipv4/tcp_illinois.c b/net/ipv4/tcp_illinois.c index 67476f085e48..f71002e4db0b 100644 --- a/net/ipv4/tcp_illinois.c +++ b/net/ipv4/tcp_illinois.c @@ -300,24 +300,25 @@ static u32 tcp_illinois_ssthresh(struct sock *sk) } /* Extract info for Tcp socket info provided via netlink. */ -static int tcp_illinois_info(struct sock *sk, u32 ext, struct sk_buff *skb) +static size_t tcp_illinois_info(struct sock *sk, u32 ext, int *attr, + union tcp_cc_info *info) { const struct illinois *ca = inet_csk_ca(sk); if (ext & (1 << (INET_DIAG_VEGASINFO - 1))) { - struct tcpvegas_info info = { - .tcpv_enabled = 1, - .tcpv_rttcnt = ca->cnt_rtt, - .tcpv_minrtt = ca->base_rtt, - }; + info->vegas.tcpv_enabled = 1; + info->vegas.tcpv_rttcnt = ca->cnt_rtt; + info->vegas.tcpv_minrtt = ca->base_rtt; + info->vegas.tcpv_rtt = 0; - if (info.tcpv_rttcnt > 0) { + if (info->vegas.tcpv_rttcnt > 0) { u64 t = ca->sum_rtt; - do_div(t, info.tcpv_rttcnt); - info.tcpv_rtt = t; + do_div(t, info->vegas.tcpv_rttcnt); + info->vegas.tcpv_rtt = t; } - return nla_put(skb, INET_DIAG_VEGASINFO, sizeof(info), &info); + *attr = INET_DIAG_VEGASINFO; + return sizeof(struct tcpvegas_info); } return 0; } diff --git a/net/ipv4/tcp_vegas.c b/net/ipv4/tcp_vegas.c index c71a1b8f7bde..a6cea1d5e20d 100644 --- a/net/ipv4/tcp_vegas.c +++ b/net/ipv4/tcp_vegas.c @@ -286,18 +286,19 @@ static void tcp_vegas_cong_avoid(struct sock *sk, u32 ack, u32 acked) } /* Extract info for Tcp socket info provided via netlink. */ -int tcp_vegas_get_info(struct sock *sk, u32 ext, struct sk_buff *skb) +size_t tcp_vegas_get_info(struct sock *sk, u32 ext, int *attr, + union tcp_cc_info *info) { const struct vegas *ca = inet_csk_ca(sk); + if (ext & (1 << (INET_DIAG_VEGASINFO - 1))) { - struct tcpvegas_info info = { - .tcpv_enabled = ca->doing_vegas_now, - .tcpv_rttcnt = ca->cntRTT, - .tcpv_rtt = ca->baseRTT, - .tcpv_minrtt = ca->minRTT, - }; - - return nla_put(skb, INET_DIAG_VEGASINFO, sizeof(info), &info); + info->vegas.tcpv_enabled = ca->doing_vegas_now, + info->vegas.tcpv_rttcnt = ca->cntRTT, + info->vegas.tcpv_rtt = ca->baseRTT, + info->vegas.tcpv_minrtt = ca->minRTT, + + *attr = INET_DIAG_VEGASINFO; + return sizeof(struct tcpvegas_info); } return 0; } diff --git a/net/ipv4/tcp_vegas.h b/net/ipv4/tcp_vegas.h index e8a6b33cc61d..ef9da5306c68 100644 --- a/net/ipv4/tcp_vegas.h +++ b/net/ipv4/tcp_vegas.h @@ -19,6 +19,7 @@ void tcp_vegas_init(struct sock *sk); void tcp_vegas_state(struct sock *sk, u8 ca_state); void tcp_vegas_pkts_acked(struct sock *sk, u32 cnt, s32 rtt_us); void tcp_vegas_cwnd_event(struct sock *sk, enum tcp_ca_event event); -int tcp_vegas_get_info(struct sock *sk, u32 ext, struct sk_buff *skb); +size_t tcp_vegas_get_info(struct sock *sk, u32 ext, int *attr, + union tcp_cc_info *info); #endif /* __TCP_VEGAS_H */ -- cgit v1.2.3-71-gd317 From 6e9250f59ef9efb932c84850cd221f22c2a03c4a Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 28 Apr 2015 16:23:49 -0700 Subject: tcp: add TCP_CC_INFO socket option Some Congestion Control modules can provide per flow information, but current way to get this information is to use netlink. Like TCP_INFO, let's add TCP_CC_INFO so that applications can issue a getsockopt() if they have a socket file descriptor, instead of playing complex netlink games. Sample usage would be : union tcp_cc_info info; socklen_t len = sizeof(info); if (getsockopt(fd, SOL_TCP, TCP_CC_INFO, &info, &len) == -1) Signed-off-by: Eric Dumazet Cc: Yuchung Cheng Cc: Neal Cardwell Acked-by: Neal Cardwell Acked-by: Daniel Borkmann Acked-by: Yuchung Cheng Signed-off-by: David S. Miller --- include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 21 +++++++++++++++++++++ 2 files changed, 22 insertions(+) (limited to 'include') diff --git a/include/uapi/linux/tcp.h b/include/uapi/linux/tcp.h index a48f93f3207b..faa72f4fa547 100644 --- a/include/uapi/linux/tcp.h +++ b/include/uapi/linux/tcp.h @@ -112,6 +112,7 @@ enum { #define TCP_FASTOPEN 23 /* Enable FastOpen on listeners */ #define TCP_TIMESTAMP 24 #define TCP_NOTSENT_LOWAT 25 /* limit number of unsent bytes in write queue */ +#define TCP_CC_INFO 26 /* Get Congestion Control (optional) info */ struct tcp_repair_opt { __u32 opt_code; diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 99fcc0b22c92..46efa03d2b11 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -252,6 +252,7 @@ #include #include #include +#include #include #include #include @@ -2739,6 +2740,26 @@ static int do_tcp_getsockopt(struct sock *sk, int level, return -EFAULT; return 0; } + case TCP_CC_INFO: { + const struct tcp_congestion_ops *ca_ops; + union tcp_cc_info info; + size_t sz = 0; + int attr; + + if (get_user(len, optlen)) + return -EFAULT; + + ca_ops = icsk->icsk_ca_ops; + if (ca_ops && ca_ops->get_info) + sz = ca_ops->get_info(sk, ~0U, &attr, &info); + + len = min_t(unsigned int, len, sz); + if (put_user(len, optlen)) + return -EFAULT; + if (copy_to_user(optval, &info, len)) + return -EFAULT; + return 0; + } case TCP_QUICKACK: val = !icsk->icsk_ack.pingpong; break; -- cgit v1.2.3-71-gd317 From 42fb23e2f57accbed69804e6d72ea3a88b8a84c4 Mon Sep 17 00:00:00 2001 From: Varka Bhadram Date: Thu, 30 Apr 2015 17:44:52 +0200 Subject: mac802154: add description to mac802154 APIs This patch adds the proper description to the mac802154 core APIs. Signed-off-by: Varka Bhadram Acked-by: Alexander Aring Signed-off-by: Alexander Aring Signed-off-by: Marcel Holtmann --- include/net/mac802154.h | 94 +++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 92 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/net/mac802154.h b/include/net/mac802154.h index e18e7fd43f47..7df28a4c23f9 100644 --- a/include/net/mac802154.h +++ b/include/net/mac802154.h @@ -247,19 +247,109 @@ static inline void ieee802154_le64_to_be64(void *be64_dst, const void *le64_src) __put_unaligned_memmove64(swab64p(le64_src), be64_dst); } -/* Basic interface to register ieee802154 device */ +/** + * ieee802154_alloc_hw - Allocate a new hardware device + * + * This must be called once for each hardware device. The returned pointer + * must be used to refer to this device when calling other functions. + * mac802154 allocates a private data area for the driver pointed to by + * @priv in &struct ieee802154_hw, the size of this area is given as + * @priv_data_len. + * + * @priv_data_len: length of private data + * @ops: callbacks for this device + * + * Return: A pointer to the new hardware device, or %NULL on error. + */ struct ieee802154_hw * ieee802154_alloc_hw(size_t priv_data_len, const struct ieee802154_ops *ops); + +/** + * ieee802154_free_hw - free hardware descriptor + * + * This function frees everything that was allocated, including the + * private data for the driver. You must call ieee802154_unregister_hw() + * before calling this function. + * + * @hw: the hardware to free + */ void ieee802154_free_hw(struct ieee802154_hw *hw); + +/** + * ieee802154_register_hw - Register hardware device + * + * You must call this function before any other functions in + * mac802154. Note that before a hardware can be registered, you + * need to fill the contained wpan_phy's information. + * + * @hw: the device to register as returned by ieee802154_alloc_hw() + * + * Return: 0 on success. An error code otherwise. + */ int ieee802154_register_hw(struct ieee802154_hw *hw); + +/** + * ieee802154_unregister_hw - Unregister a hardware device + * + * This function instructs mac802154 to free allocated resources + * and unregister netdevices from the networking subsystem. + * + * @hw: the hardware to unregister + */ void ieee802154_unregister_hw(struct ieee802154_hw *hw); +/** + * ieee802154_rx - receive frame + * + * Use this function to hand received frames to mac802154. The receive + * buffer in @skb must start with an IEEE 802.15.4 header. In case of a + * paged @skb is used, the driver is recommended to put the ieee802154 + * header of the frame on the linear part of the @skb to avoid memory + * allocation and/or memcpy by the stack. + * + * This function may not be called in IRQ context. Calls to this function + * for a single hardware must be synchronized against each other. + * + * @hw: the hardware this frame came in on + * @skb: the buffer to receive, owned by mac802154 after this call + */ void ieee802154_rx(struct ieee802154_hw *hw, struct sk_buff *skb); + +/** + * ieee802154_rx_irqsafe - receive frame + * + * Like ieee802154_rx() but can be called in IRQ context + * (internally defers to a tasklet.) + * + * @hw: the hardware this frame came in on + * @skb: the buffer to receive, owned by mac802154 after this call + * @lqi: link quality indicator + */ void ieee802154_rx_irqsafe(struct ieee802154_hw *hw, struct sk_buff *skb, u8 lqi); - +/** + * ieee802154_wake_queue - wake ieee802154 queue + * @hw: pointer as obtained from ieee802154_alloc_hw(). + * + * Drivers should use this function instead of netif_wake_queue. + */ void ieee802154_wake_queue(struct ieee802154_hw *hw); + +/** + * ieee802154_stop_queue - stop ieee802154 queue + * @hw: pointer as obtained from ieee802154_alloc_hw(). + * + * Drivers should use this function instead of netif_stop_queue. + */ void ieee802154_stop_queue(struct ieee802154_hw *hw); + +/** + * ieee802154_xmit_complete - frame transmission complete + * + * @hw: pointer as obtained from ieee802154_alloc_hw(). + * @skb: buffer for transmission + * @ifs_handling: indicate interframe space handling + */ void ieee802154_xmit_complete(struct ieee802154_hw *hw, struct sk_buff *skb, bool ifs_handling); -- cgit v1.2.3-71-gd317 From 5b4a10390460cccf17a9fac739e153d68cf25ef5 Mon Sep 17 00:00:00 2001 From: Varka Bhadram Date: Thu, 30 Apr 2015 17:44:57 +0200 Subject: cfg802154: pass name_assign_type to rdev_add_virtual_intf() This code is based on commit 6bab2e19c5ffd ("cfg80211: pass name_assign_type to rdev_add_virtual_intf()") This will expose in sysfs whether the ifname of a IEEE-802.15.4 device is set by userspace or generated by the kernel. We are using two types of name_assign_types o NET_NAME_ENUM: Default interface name provided by kernel o NET_NAME_USER: Interface name provided by user. Signed-off-by: Varka Bhadram Signed-off-by: Alexander Aring Signed-off-by: Marcel Holtmann --- include/net/cfg802154.h | 2 ++ net/ieee802154/nl-phy.c | 5 ++++- net/ieee802154/nl802154.c | 2 +- net/ieee802154/rdev-ops.h | 10 +++++++--- net/mac802154/cfg.c | 9 ++++++--- net/mac802154/ieee802154_i.h | 3 ++- net/mac802154/iface.c | 5 +++-- net/mac802154/main.c | 3 ++- 8 files changed, 27 insertions(+), 12 deletions(-) (limited to 'include') diff --git a/include/net/cfg802154.h b/include/net/cfg802154.h index eeda67652766..6ea16c84293b 100644 --- a/include/net/cfg802154.h +++ b/include/net/cfg802154.h @@ -30,11 +30,13 @@ struct wpan_phy_cca; struct cfg802154_ops { struct net_device * (*add_virtual_intf_deprecated)(struct wpan_phy *wpan_phy, const char *name, + unsigned char name_assign_type, int type); void (*del_virtual_intf_deprecated)(struct wpan_phy *wpan_phy, struct net_device *dev); int (*add_virtual_intf)(struct wpan_phy *wpan_phy, const char *name, + unsigned char name_assign_type, enum nl802154_iftype type, __le64 extended_addr); int (*del_virtual_intf)(struct wpan_phy *wpan_phy, diff --git a/net/ieee802154/nl-phy.c b/net/ieee802154/nl-phy.c index 1b9d25f6e898..346c6665d25e 100644 --- a/net/ieee802154/nl-phy.c +++ b/net/ieee802154/nl-phy.c @@ -175,6 +175,7 @@ int ieee802154_add_iface(struct sk_buff *skb, struct genl_info *info) int rc = -ENOBUFS; struct net_device *dev; int type = __IEEE802154_DEV_INVALID; + unsigned char name_assign_type; pr_debug("%s\n", __func__); @@ -190,8 +191,10 @@ int ieee802154_add_iface(struct sk_buff *skb, struct genl_info *info) if (devname[nla_len(info->attrs[IEEE802154_ATTR_DEV_NAME]) - 1] != '\0') return -EINVAL; /* phy name should be null-terminated */ + name_assign_type = NET_NAME_USER; } else { devname = "wpan%d"; + name_assign_type = NET_NAME_ENUM; } if (strlen(devname) >= IFNAMSIZ) @@ -221,7 +224,7 @@ int ieee802154_add_iface(struct sk_buff *skb, struct genl_info *info) } dev = rdev_add_virtual_intf_deprecated(wpan_phy_to_rdev(phy), devname, - type); + name_assign_type, type); if (IS_ERR(dev)) { rc = PTR_ERR(dev); goto nla_put_failure; diff --git a/net/ieee802154/nl802154.c b/net/ieee802154/nl802154.c index a4daf91b8d0a..f3c12f6a4a39 100644 --- a/net/ieee802154/nl802154.c +++ b/net/ieee802154/nl802154.c @@ -589,7 +589,7 @@ static int nl802154_new_interface(struct sk_buff *skb, struct genl_info *info) return rdev_add_virtual_intf(rdev, nla_data(info->attrs[NL802154_ATTR_IFNAME]), - type, extended_addr); + NET_NAME_USER, type, extended_addr); } static int nl802154_del_interface(struct sk_buff *skb, struct genl_info *info) diff --git a/net/ieee802154/rdev-ops.h b/net/ieee802154/rdev-ops.h index 624413ee095f..7b5a9dd94fe5 100644 --- a/net/ieee802154/rdev-ops.h +++ b/net/ieee802154/rdev-ops.h @@ -8,10 +8,12 @@ static inline struct net_device * rdev_add_virtual_intf_deprecated(struct cfg802154_registered_device *rdev, - const char *name, int type) + const char *name, + unsigned char name_assign_type, + int type) { return rdev->ops->add_virtual_intf_deprecated(&rdev->wpan_phy, name, - type); + name_assign_type, type); } static inline void @@ -23,13 +25,15 @@ rdev_del_virtual_intf_deprecated(struct cfg802154_registered_device *rdev, static inline int rdev_add_virtual_intf(struct cfg802154_registered_device *rdev, char *name, + unsigned char name_assign_type, enum nl802154_iftype type, __le64 extended_addr) { int ret; trace_802154_rdev_add_virtual_intf(&rdev->wpan_phy, name, type, extended_addr); - ret = rdev->ops->add_virtual_intf(&rdev->wpan_phy, name, type, + ret = rdev->ops->add_virtual_intf(&rdev->wpan_phy, name, + name_assign_type, type, extended_addr); trace_802154_rdev_return_int(&rdev->wpan_phy, ret); return ret; diff --git a/net/mac802154/cfg.c b/net/mac802154/cfg.c index 5d9f68c75e5f..70be9c799f8a 100644 --- a/net/mac802154/cfg.c +++ b/net/mac802154/cfg.c @@ -22,13 +22,14 @@ static struct net_device * ieee802154_add_iface_deprecated(struct wpan_phy *wpan_phy, - const char *name, int type) + const char *name, + unsigned char name_assign_type, int type) { struct ieee802154_local *local = wpan_phy_priv(wpan_phy); struct net_device *dev; rtnl_lock(); - dev = ieee802154_if_add(local, name, type, + dev = ieee802154_if_add(local, name, name_assign_type, type, cpu_to_le64(0x0000000000000000ULL)); rtnl_unlock(); @@ -45,12 +46,14 @@ static void ieee802154_del_iface_deprecated(struct wpan_phy *wpan_phy, static int ieee802154_add_iface(struct wpan_phy *phy, const char *name, + unsigned char name_assign_type, enum nl802154_iftype type, __le64 extended_addr) { struct ieee802154_local *local = wpan_phy_priv(phy); struct net_device *err; - err = ieee802154_if_add(local, name, type, extended_addr); + err = ieee802154_if_add(local, name, name_assign_type, type, + extended_addr); return PTR_ERR_OR_ZERO(err); } diff --git a/net/mac802154/ieee802154_i.h b/net/mac802154/ieee802154_i.h index bebd70ffc7a3..127ba18386fc 100644 --- a/net/mac802154/ieee802154_i.h +++ b/net/mac802154/ieee802154_i.h @@ -182,7 +182,8 @@ void ieee802154_iface_exit(void); void ieee802154_if_remove(struct ieee802154_sub_if_data *sdata); struct net_device * ieee802154_if_add(struct ieee802154_local *local, const char *name, - enum nl802154_iftype type, __le64 extended_addr); + unsigned char name_assign_type, enum nl802154_iftype type, + __le64 extended_addr); void ieee802154_remove_interfaces(struct ieee802154_local *local); #endif /* __IEEE802154_I_H */ diff --git a/net/mac802154/iface.c b/net/mac802154/iface.c index 38b56f9d9386..91b75abbd1a1 100644 --- a/net/mac802154/iface.c +++ b/net/mac802154/iface.c @@ -522,7 +522,8 @@ ieee802154_setup_sdata(struct ieee802154_sub_if_data *sdata, struct net_device * ieee802154_if_add(struct ieee802154_local *local, const char *name, - enum nl802154_iftype type, __le64 extended_addr) + unsigned char name_assign_type, enum nl802154_iftype type, + __le64 extended_addr) { struct net_device *ndev = NULL; struct ieee802154_sub_if_data *sdata = NULL; @@ -531,7 +532,7 @@ ieee802154_if_add(struct ieee802154_local *local, const char *name, ASSERT_RTNL(); ndev = alloc_netdev(sizeof(*sdata) + local->hw.vif_data_size, name, - NET_NAME_UNKNOWN, ieee802154_if_setup); + name_assign_type, ieee802154_if_setup); if (!ndev) return ERR_PTR(-ENOMEM); diff --git a/net/mac802154/main.c b/net/mac802154/main.c index 8500378c8318..68b9667323ec 100644 --- a/net/mac802154/main.c +++ b/net/mac802154/main.c @@ -161,7 +161,8 @@ int ieee802154_register_hw(struct ieee802154_hw *hw) rtnl_lock(); - dev = ieee802154_if_add(local, "wpan%d", NL802154_IFTYPE_NODE, + dev = ieee802154_if_add(local, "wpan%d", NET_NAME_ENUM, + NL802154_IFTYPE_NODE, cpu_to_le64(0x0000000000000000ULL)); if (IS_ERR(dev)) { rtnl_unlock(); -- cgit v1.2.3-71-gd317 From a5d28090405038ca1f40c13f38d6d4285456efee Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 30 Apr 2015 09:40:40 -0700 Subject: codel: fix maxpacket/mtu confusion Under presence of TSO/GSO/GRO packets, codel at low rates can be quite useless. In following example, not a single packet was ever dropped, while average delay in codel queue is ~100 ms ! qdisc codel 0: parent 1:12 limit 16000p target 5.0ms interval 100.0ms Sent 134376498 bytes 88797 pkt (dropped 0, overlimits 0 requeues 0) backlog 13626b 3p requeues 0 count 0 lastcount 0 ldelay 96.9ms drop_next 0us maxpacket 9084 ecn_mark 0 drop_overlimit 0 This comes from a confusion of what should be the minimal backlog. It is pretty clear it is not 64KB or whatever max GSO packet ever reached the qdisc. codel intent was to use MTU of the device. After the fix, we finally drop some packets, and rtt/cwnd of my single TCP flow are meeting our expectations. qdisc codel 0: parent 1:12 limit 16000p target 5.0ms interval 100.0ms Sent 102798497 bytes 67912 pkt (dropped 1365, overlimits 0 requeues 0) backlog 6056b 3p requeues 0 count 1 lastcount 1 ldelay 36.3ms drop_next 0us maxpacket 10598 ecn_mark 0 drop_overlimit 0 Signed-off-by: Eric Dumazet Cc: Kathleen Nichols Cc: Dave Taht Cc: Van Jacobson Signed-off-by: David S. Miller --- include/net/codel.h | 10 +++++++--- net/sched/sch_codel.c | 2 +- net/sched/sch_fq_codel.c | 2 +- 3 files changed, 9 insertions(+), 5 deletions(-) (limited to 'include') diff --git a/include/net/codel.h b/include/net/codel.h index aeee28081245..1e18005f7f65 100644 --- a/include/net/codel.h +++ b/include/net/codel.h @@ -120,11 +120,13 @@ static inline u32 codel_time_to_us(codel_time_t val) * struct codel_params - contains codel parameters * @target: target queue size (in time units) * @interval: width of moving time window + * @mtu: device mtu, or minimal queue backlog in bytes. * @ecn: is Explicit Congestion Notification enabled */ struct codel_params { codel_time_t target; codel_time_t interval; + u32 mtu; bool ecn; }; @@ -166,10 +168,12 @@ struct codel_stats { u32 ecn_mark; }; -static void codel_params_init(struct codel_params *params) +static void codel_params_init(struct codel_params *params, + const struct Qdisc *sch) { params->interval = MS2TIME(100); params->target = MS2TIME(5); + params->mtu = psched_mtu(qdisc_dev(sch)); params->ecn = false; } @@ -180,7 +184,7 @@ static void codel_vars_init(struct codel_vars *vars) static void codel_stats_init(struct codel_stats *stats) { - stats->maxpacket = 256; + stats->maxpacket = 0; } /* @@ -234,7 +238,7 @@ static bool codel_should_drop(const struct sk_buff *skb, stats->maxpacket = qdisc_pkt_len(skb); if (codel_time_before(vars->ldelay, params->target) || - sch->qstats.backlog <= stats->maxpacket) { + sch->qstats.backlog <= params->mtu) { /* went below - stay below for at least interval */ vars->first_above_time = 0; return false; diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c index de28f8e968e8..7a0bdb16ac92 100644 --- a/net/sched/sch_codel.c +++ b/net/sched/sch_codel.c @@ -164,7 +164,7 @@ static int codel_init(struct Qdisc *sch, struct nlattr *opt) sch->limit = DEFAULT_CODEL_LIMIT; - codel_params_init(&q->params); + codel_params_init(&q->params, sch); codel_vars_init(&q->vars); codel_stats_init(&q->stats); diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c index 1e52decb7b59..c244c45b78d7 100644 --- a/net/sched/sch_fq_codel.c +++ b/net/sched/sch_fq_codel.c @@ -391,7 +391,7 @@ static int fq_codel_init(struct Qdisc *sch, struct nlattr *opt) q->perturbation = prandom_u32(); INIT_LIST_HEAD(&q->new_flows); INIT_LIST_HEAD(&q->old_flows); - codel_params_init(&q->cparams); + codel_params_init(&q->cparams, sch); codel_stats_init(&q->cstats); q->cparams.ecn = true; -- cgit v1.2.3-71-gd317 From ff419b3f95ab7a97c5f72876b53f12a249dacc2a Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Sun, 26 Apr 2015 20:13:34 -0700 Subject: mac80211: fix 90 kernel-doc warnings Eliminate 90 of these warnings: Warning(..//include/net/mac80211.h:1682): No description found for parameter 'drv_priv[0]' Signed-off-by: Randy Dunlap Signed-off-by: Johannes Berg --- include/net/mac80211.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/net/mac80211.h b/include/net/mac80211.h index b4bef1152c05..8e3668b44c29 100644 --- a/include/net/mac80211.h +++ b/include/net/mac80211.h @@ -1666,6 +1666,8 @@ struct ieee80211_tx_control { * @sta: station table entry, %NULL for per-vif queue * @tid: the TID for this queue (unused for per-vif queue) * @ac: the AC for this queue + * @drv_priv: data area for driver use, will always be aligned to + * sizeof(void *). * * The driver can obtain packets from this queue by calling * ieee80211_tx_dequeue(). -- cgit v1.2.3-71-gd317 From c967a0873a7836c7a77bf611f1c7d3f47c554c45 Mon Sep 17 00:00:00 2001 From: Tom Herbert Date: Tue, 5 May 2015 09:06:30 -0700 Subject: mpls: Move reserved label definitions Move to include/uapi/linux/mpls.h to be externally visibile. Signed-off-by: Tom Herbert Signed-off-by: David S. Miller --- include/uapi/linux/mpls.h | 10 ++++++++++ net/mpls/af_mpls.c | 18 +++++++++--------- net/mpls/internal.h | 10 ---------- 3 files changed, 19 insertions(+), 19 deletions(-) (limited to 'include') diff --git a/include/uapi/linux/mpls.h b/include/uapi/linux/mpls.h index bc9abfe88c9a..0fe6ea5a41d5 100644 --- a/include/uapi/linux/mpls.h +++ b/include/uapi/linux/mpls.h @@ -31,4 +31,14 @@ struct mpls_label { #define MPLS_LS_TTL_MASK 0x000000FF #define MPLS_LS_TTL_SHIFT 0 +/* Reserved labels */ +#define MPLS_LABEL_IPV4_EXPLICIT_NULL 0 /* RFC3032 */ +#define MPLS_LABEL_ROUTER_ALERT 1 /* RFC3032 */ +#define MPLS_LABEL_IPV6_EXPLICIT_NULL 2 /* RFC3032 */ +#define MPLS_LABEL_IMPLICIT_NULL 3 /* RFC3032 */ +#define MPLS_LABEL_ENTROPY_INDICATOR 7 /* RFC6790 */ +#define MPLS_LABEL_GAL 13 /* RFC5586 */ +#define MPLS_LABEL_OAM_ALERT 14 /* RFC3429 */ +#define MPLS_LABEL_EXTENSION 15 /* RFC7274 */ + #endif /* _UAPI_MPLS_H */ diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index 954810c76a86..b6eb7615960a 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -647,7 +647,7 @@ int nla_get_labels(const struct nlattr *nla, return -EINVAL; switch (dec.label) { - case LABEL_IMPLICIT_NULL: + case MPLS_LABEL_IMPLICIT_NULL: /* RFC3032: This is a label that an LSR may * assign and distribute, but which never * actually appears in the encapsulation. @@ -935,7 +935,7 @@ static int resize_platform_label_table(struct net *net, size_t limit) } /* In case the predefined labels need to be populated */ - if (limit > LABEL_IPV4_EXPLICIT_NULL) { + if (limit > MPLS_LABEL_IPV4_EXPLICIT_NULL) { struct net_device *lo = net->loopback_dev; rt0 = mpls_rt_alloc(lo->addr_len); if (!rt0) @@ -945,7 +945,7 @@ static int resize_platform_label_table(struct net *net, size_t limit) rt0->rt_via_table = NEIGH_LINK_TABLE; memcpy(rt0->rt_via, lo->dev_addr, lo->addr_len); } - if (limit > LABEL_IPV6_EXPLICIT_NULL) { + if (limit > MPLS_LABEL_IPV6_EXPLICIT_NULL) { struct net_device *lo = net->loopback_dev; rt2 = mpls_rt_alloc(lo->addr_len); if (!rt2) @@ -973,15 +973,15 @@ static int resize_platform_label_table(struct net *net, size_t limit) memcpy(labels, old, cp_size); /* If needed set the predefined labels */ - if ((old_limit <= LABEL_IPV6_EXPLICIT_NULL) && - (limit > LABEL_IPV6_EXPLICIT_NULL)) { - RCU_INIT_POINTER(labels[LABEL_IPV6_EXPLICIT_NULL], rt2); + if ((old_limit <= MPLS_LABEL_IPV6_EXPLICIT_NULL) && + (limit > MPLS_LABEL_IPV6_EXPLICIT_NULL)) { + RCU_INIT_POINTER(labels[MPLS_LABEL_IPV6_EXPLICIT_NULL], rt2); rt2 = NULL; } - if ((old_limit <= LABEL_IPV4_EXPLICIT_NULL) && - (limit > LABEL_IPV4_EXPLICIT_NULL)) { - RCU_INIT_POINTER(labels[LABEL_IPV4_EXPLICIT_NULL], rt0); + if ((old_limit <= MPLS_LABEL_IPV4_EXPLICIT_NULL) && + (limit > MPLS_LABEL_IPV4_EXPLICIT_NULL)) { + RCU_INIT_POINTER(labels[MPLS_LABEL_IPV4_EXPLICIT_NULL], rt0); rt0 = NULL; } diff --git a/net/mpls/internal.h b/net/mpls/internal.h index 693877d69606..b064c345042c 100644 --- a/net/mpls/internal.h +++ b/net/mpls/internal.h @@ -1,16 +1,6 @@ #ifndef MPLS_INTERNAL_H #define MPLS_INTERNAL_H -#define LABEL_IPV4_EXPLICIT_NULL 0 /* RFC3032 */ -#define LABEL_ROUTER_ALERT_LABEL 1 /* RFC3032 */ -#define LABEL_IPV6_EXPLICIT_NULL 2 /* RFC3032 */ -#define LABEL_IMPLICIT_NULL 3 /* RFC3032 */ -#define LABEL_ENTROPY_INDICATOR 7 /* RFC6790 */ -#define LABEL_GAL 13 /* RFC5586 */ -#define LABEL_OAM_ALERT 14 /* RFC3429 */ -#define LABEL_EXTENSION 15 /* RFC7274 */ - - struct mpls_shim_hdr { __be32 label_stack_entry; }; -- cgit v1.2.3-71-gd317 From 0782e63bc6fe7e2d3408d250df11d388b7799c6b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 5 May 2015 19:49:49 +0200 Subject: sched: Handle priority boosted tasks proper in setscheduler() Ronny reported that the following scenario is not handled correctly: T1 (prio = 10) lock(rtmutex); T2 (prio = 20) lock(rtmutex) boost T1 T1 (prio = 20) sys_set_scheduler(prio = 30) T1 prio = 30 .... sys_set_scheduler(prio = 10) T1 prio = 30 The last step is wrong as T1 should now be back at prio 20. Commit c365c292d059 ("sched: Consider pi boosting in setscheduler()") only handles the case where a boosted tasks tries to lower its priority. Fix it by taking the new effective priority into account for the decision whether a change of the priority is required. Reported-by: Ronny Meeus Tested-by: Steven Rostedt Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Steven Rostedt Cc: Cc: Borislav Petkov Cc: H. Peter Anvin Cc: Mike Galbraith Fixes: c365c292d059 ("sched: Consider pi boosting in setscheduler()") Link: http://lkml.kernel.org/r/alpine.DEB.2.11.1505051806060.4225@nanos Signed-off-by: Ingo Molnar --- include/linux/sched/rt.h | 7 ++++--- kernel/locking/rtmutex.c | 12 +++++++----- kernel/sched/core.c | 26 ++++++++++++++------------ 3 files changed, 25 insertions(+), 20 deletions(-) (limited to 'include') diff --git a/include/linux/sched/rt.h b/include/linux/sched/rt.h index 6341f5be6e24..a30b172df6e1 100644 --- a/include/linux/sched/rt.h +++ b/include/linux/sched/rt.h @@ -18,7 +18,7 @@ static inline int rt_task(struct task_struct *p) #ifdef CONFIG_RT_MUTEXES extern int rt_mutex_getprio(struct task_struct *p); extern void rt_mutex_setprio(struct task_struct *p, int prio); -extern int rt_mutex_check_prio(struct task_struct *task, int newprio); +extern int rt_mutex_get_effective_prio(struct task_struct *task, int newprio); extern struct task_struct *rt_mutex_get_top_task(struct task_struct *task); extern void rt_mutex_adjust_pi(struct task_struct *p); static inline bool tsk_is_pi_blocked(struct task_struct *tsk) @@ -31,9 +31,10 @@ static inline int rt_mutex_getprio(struct task_struct *p) return p->normal_prio; } -static inline int rt_mutex_check_prio(struct task_struct *task, int newprio) +static inline int rt_mutex_get_effective_prio(struct task_struct *task, + int newprio) { - return 0; + return newprio; } static inline struct task_struct *rt_mutex_get_top_task(struct task_struct *task) diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c index b73279367087..b025295f4966 100644 --- a/kernel/locking/rtmutex.c +++ b/kernel/locking/rtmutex.c @@ -265,15 +265,17 @@ struct task_struct *rt_mutex_get_top_task(struct task_struct *task) } /* - * Called by sched_setscheduler() to check whether the priority change - * is overruled by a possible priority boosting. + * Called by sched_setscheduler() to get the priority which will be + * effective after the change. */ -int rt_mutex_check_prio(struct task_struct *task, int newprio) +int rt_mutex_get_effective_prio(struct task_struct *task, int newprio) { if (!task_has_pi_waiters(task)) - return 0; + return newprio; - return task_top_pi_waiter(task)->task->prio <= newprio; + if (task_top_pi_waiter(task)->task->prio <= newprio) + return task_top_pi_waiter(task)->task->prio; + return newprio; } /* diff --git a/kernel/sched/core.c b/kernel/sched/core.c index fe22f7510bce..34db9bf892a3 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -3300,15 +3300,18 @@ static void __setscheduler_params(struct task_struct *p, /* Actually do priority change: must hold pi & rq lock. */ static void __setscheduler(struct rq *rq, struct task_struct *p, - const struct sched_attr *attr) + const struct sched_attr *attr, bool keep_boost) { __setscheduler_params(p, attr); /* - * If we get here, there was no pi waiters boosting the - * task. It is safe to use the normal prio. + * Keep a potential priority boosting if called from + * sched_setscheduler(). */ - p->prio = normal_prio(p); + if (keep_boost) + p->prio = rt_mutex_get_effective_prio(p, normal_prio(p)); + else + p->prio = normal_prio(p); if (dl_prio(p->prio)) p->sched_class = &dl_sched_class; @@ -3408,7 +3411,7 @@ static int __sched_setscheduler(struct task_struct *p, int newprio = dl_policy(attr->sched_policy) ? MAX_DL_PRIO - 1 : MAX_RT_PRIO - 1 - attr->sched_priority; int retval, oldprio, oldpolicy = -1, queued, running; - int policy = attr->sched_policy; + int new_effective_prio, policy = attr->sched_policy; unsigned long flags; const struct sched_class *prev_class; struct rq *rq; @@ -3590,15 +3593,14 @@ change: oldprio = p->prio; /* - * Special case for priority boosted tasks. - * - * If the new priority is lower or equal (user space view) - * than the current (boosted) priority, we just store the new + * Take priority boosted tasks into account. If the new + * effective priority is unchanged, we just store the new * normal parameters and do not touch the scheduler class and * the runqueue. This will be done when the task deboost * itself. */ - if (rt_mutex_check_prio(p, newprio)) { + new_effective_prio = rt_mutex_get_effective_prio(p, newprio); + if (new_effective_prio == oldprio) { __setscheduler_params(p, attr); task_rq_unlock(rq, p, &flags); return 0; @@ -3612,7 +3614,7 @@ change: put_prev_task(rq, p); prev_class = p->sched_class; - __setscheduler(rq, p, attr); + __setscheduler(rq, p, attr, true); if (running) p->sched_class->set_curr_task(rq); @@ -7346,7 +7348,7 @@ static void normalize_task(struct rq *rq, struct task_struct *p) queued = task_on_rq_queued(p); if (queued) dequeue_task(rq, p, 0); - __setscheduler(rq, p, &attr); + __setscheduler(rq, p, &attr, false); if (queued) { enqueue_task(rq, p, 0); resched_curr(rq); -- cgit v1.2.3-71-gd317 From 0cd3be6e9a46f84ef7a42e1a5645d32ad547b12e Mon Sep 17 00:00:00 2001 From: Sebastian Hesselbarth Date: Mon, 4 May 2015 23:04:16 +0200 Subject: clk: si5351: Do not pass struct clk in platform_data When registering clk-si5351 by platform_data, we should not pass struct clk for the reference clocks. Drop struct clk from platform_data and rework the driver to use devm_clk_get of named clock references. While at it, check for at least one valid input clock and properly prepare/ enable valid reference clocks. Signed-off-by: Sebastian Hesselbarth Reported-by: Michael Welling Reported-by: Jean-Francois Moine Reported-by: Russell King Tested-by: Michael Welling Tested-by: Jean-Francois Moine Signed-off-by: Michael Turquette --- drivers/clk/clk-si5351.c | 63 +++++++++++++++++++++++++----------- include/linux/platform_data/si5351.h | 4 --- 2 files changed, 45 insertions(+), 22 deletions(-) (limited to 'include') diff --git a/drivers/clk/clk-si5351.c b/drivers/clk/clk-si5351.c index 44ea107cfc67..30335d3b99af 100644 --- a/drivers/clk/clk-si5351.c +++ b/drivers/clk/clk-si5351.c @@ -1128,13 +1128,6 @@ static int si5351_dt_parse(struct i2c_client *client, if (!pdata) return -ENOMEM; - pdata->clk_xtal = of_clk_get(np, 0); - if (!IS_ERR(pdata->clk_xtal)) - clk_put(pdata->clk_xtal); - pdata->clk_clkin = of_clk_get(np, 1); - if (!IS_ERR(pdata->clk_clkin)) - clk_put(pdata->clk_clkin); - /* * property silabs,pll-source : , [<..>] * allow to selectively set pll source @@ -1328,8 +1321,22 @@ static int si5351_i2c_probe(struct i2c_client *client, i2c_set_clientdata(client, drvdata); drvdata->client = client; drvdata->variant = variant; - drvdata->pxtal = pdata->clk_xtal; - drvdata->pclkin = pdata->clk_clkin; + drvdata->pxtal = devm_clk_get(&client->dev, "xtal"); + drvdata->pclkin = devm_clk_get(&client->dev, "clkin"); + + if (PTR_ERR(drvdata->pxtal) == -EPROBE_DEFER || + PTR_ERR(drvdata->pclkin) == -EPROBE_DEFER) + return -EPROBE_DEFER; + + /* + * Check for valid parent clock: VARIANT_A and VARIANT_B need XTAL, + * VARIANT_C can have CLKIN instead. + */ + if (IS_ERR(drvdata->pxtal) && + (drvdata->variant != SI5351_VARIANT_C || IS_ERR(drvdata->pclkin))) { + dev_err(&client->dev, "missing parent clock\n"); + return -EINVAL; + } drvdata->regmap = devm_regmap_init_i2c(client, &si5351_regmap_config); if (IS_ERR(drvdata->regmap)) { @@ -1393,6 +1400,11 @@ static int si5351_i2c_probe(struct i2c_client *client, } } + if (!IS_ERR(drvdata->pxtal)) + clk_prepare_enable(drvdata->pxtal); + if (!IS_ERR(drvdata->pclkin)) + clk_prepare_enable(drvdata->pclkin); + /* register xtal input clock gate */ memset(&init, 0, sizeof(init)); init.name = si5351_input_names[0]; @@ -1407,7 +1419,8 @@ static int si5351_i2c_probe(struct i2c_client *client, clk = devm_clk_register(&client->dev, &drvdata->xtal); if (IS_ERR(clk)) { dev_err(&client->dev, "unable to register %s\n", init.name); - return PTR_ERR(clk); + ret = PTR_ERR(clk); + goto err_clk; } /* register clkin input clock gate */ @@ -1425,7 +1438,8 @@ static int si5351_i2c_probe(struct i2c_client *client, if (IS_ERR(clk)) { dev_err(&client->dev, "unable to register %s\n", init.name); - return PTR_ERR(clk); + ret = PTR_ERR(clk); + goto err_clk; } } @@ -1447,7 +1461,8 @@ static int si5351_i2c_probe(struct i2c_client *client, clk = devm_clk_register(&client->dev, &drvdata->pll[0].hw); if (IS_ERR(clk)) { dev_err(&client->dev, "unable to register %s\n", init.name); - return -EINVAL; + ret = PTR_ERR(clk); + goto err_clk; } /* register PLLB or VXCO (Si5351B) */ @@ -1471,7 +1486,8 @@ static int si5351_i2c_probe(struct i2c_client *client, clk = devm_clk_register(&client->dev, &drvdata->pll[1].hw); if (IS_ERR(clk)) { dev_err(&client->dev, "unable to register %s\n", init.name); - return -EINVAL; + ret = PTR_ERR(clk); + goto err_clk; } /* register clk multisync and clk out divider */ @@ -1492,8 +1508,10 @@ static int si5351_i2c_probe(struct i2c_client *client, num_clocks * sizeof(*drvdata->onecell.clks), GFP_KERNEL); if (WARN_ON(!drvdata->msynth || !drvdata->clkout || - !drvdata->onecell.clks)) - return -ENOMEM; + !drvdata->onecell.clks)) { + ret = -ENOMEM; + goto err_clk; + } for (n = 0; n < num_clocks; n++) { drvdata->msynth[n].num = n; @@ -1511,7 +1529,8 @@ static int si5351_i2c_probe(struct i2c_client *client, if (IS_ERR(clk)) { dev_err(&client->dev, "unable to register %s\n", init.name); - return -EINVAL; + ret = PTR_ERR(clk); + goto err_clk; } } @@ -1538,7 +1557,8 @@ static int si5351_i2c_probe(struct i2c_client *client, if (IS_ERR(clk)) { dev_err(&client->dev, "unable to register %s\n", init.name); - return -EINVAL; + ret = PTR_ERR(clk); + goto err_clk; } drvdata->onecell.clks[n] = clk; @@ -1557,10 +1577,17 @@ static int si5351_i2c_probe(struct i2c_client *client, &drvdata->onecell); if (ret) { dev_err(&client->dev, "unable to add clk provider\n"); - return ret; + goto err_clk; } return 0; + +err_clk: + if (!IS_ERR(drvdata->pxtal)) + clk_disable_unprepare(drvdata->pxtal); + if (!IS_ERR(drvdata->pclkin)) + clk_disable_unprepare(drvdata->pclkin); + return ret; } static const struct i2c_device_id si5351_i2c_ids[] = { diff --git a/include/linux/platform_data/si5351.h b/include/linux/platform_data/si5351.h index a947ab8b441a..533d9807e543 100644 --- a/include/linux/platform_data/si5351.h +++ b/include/linux/platform_data/si5351.h @@ -5,8 +5,6 @@ #ifndef __LINUX_PLATFORM_DATA_SI5351_H__ #define __LINUX_PLATFORM_DATA_SI5351_H__ -struct clk; - /** * enum si5351_pll_src - Si5351 pll clock source * @SI5351_PLL_SRC_DEFAULT: default, do not change eeprom config @@ -107,8 +105,6 @@ struct si5351_clkout_config { * @clkout: array of clkout configuration */ struct si5351_platform_data { - struct clk *clk_xtal; - struct clk *clk_clkin; enum si5351_pll_src pll_src[2]; struct si5351_clkout_config clkout[8]; }; -- cgit v1.2.3-71-gd317 From 78f5b899195019f71f7593c604d75ca61658eae3 Mon Sep 17 00:00:00 2001 From: Tom Herbert Date: Thu, 7 May 2015 08:08:51 -0700 Subject: mpls: Change reserved label names to be consistent with netbsd Since these are now visible to userspace it is nice to be consistent with BSD (sys/netmpls/mpls.h in netBSD). Signed-off-by: Tom Herbert Signed-off-by: David S. Miller --- include/uapi/linux/mpls.h | 12 ++++++------ net/mpls/af_mpls.c | 18 +++++++++--------- 2 files changed, 15 insertions(+), 15 deletions(-) (limited to 'include') diff --git a/include/uapi/linux/mpls.h b/include/uapi/linux/mpls.h index 0fe6ea5a41d5..139d4dd1cab8 100644 --- a/include/uapi/linux/mpls.h +++ b/include/uapi/linux/mpls.h @@ -32,13 +32,13 @@ struct mpls_label { #define MPLS_LS_TTL_SHIFT 0 /* Reserved labels */ -#define MPLS_LABEL_IPV4_EXPLICIT_NULL 0 /* RFC3032 */ -#define MPLS_LABEL_ROUTER_ALERT 1 /* RFC3032 */ -#define MPLS_LABEL_IPV6_EXPLICIT_NULL 2 /* RFC3032 */ -#define MPLS_LABEL_IMPLICIT_NULL 3 /* RFC3032 */ -#define MPLS_LABEL_ENTROPY_INDICATOR 7 /* RFC6790 */ +#define MPLS_LABEL_IPV4NULL 0 /* RFC3032 */ +#define MPLS_LABEL_RTALERT 1 /* RFC3032 */ +#define MPLS_LABEL_IPV6NULL 2 /* RFC3032 */ +#define MPLS_LABEL_IMPLNULL 3 /* RFC3032 */ +#define MPLS_LABEL_ENTROPY 7 /* RFC6790 */ #define MPLS_LABEL_GAL 13 /* RFC5586 */ -#define MPLS_LABEL_OAM_ALERT 14 /* RFC3429 */ +#define MPLS_LABEL_OAMALERT 14 /* RFC3429 */ #define MPLS_LABEL_EXTENSION 15 /* RFC7274 */ #endif /* _UAPI_MPLS_H */ diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index b6eb7615960a..7b3f732269e4 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -647,7 +647,7 @@ int nla_get_labels(const struct nlattr *nla, return -EINVAL; switch (dec.label) { - case MPLS_LABEL_IMPLICIT_NULL: + case MPLS_LABEL_IMPLNULL: /* RFC3032: This is a label that an LSR may * assign and distribute, but which never * actually appears in the encapsulation. @@ -935,7 +935,7 @@ static int resize_platform_label_table(struct net *net, size_t limit) } /* In case the predefined labels need to be populated */ - if (limit > MPLS_LABEL_IPV4_EXPLICIT_NULL) { + if (limit > MPLS_LABEL_IPV4NULL) { struct net_device *lo = net->loopback_dev; rt0 = mpls_rt_alloc(lo->addr_len); if (!rt0) @@ -945,7 +945,7 @@ static int resize_platform_label_table(struct net *net, size_t limit) rt0->rt_via_table = NEIGH_LINK_TABLE; memcpy(rt0->rt_via, lo->dev_addr, lo->addr_len); } - if (limit > MPLS_LABEL_IPV6_EXPLICIT_NULL) { + if (limit > MPLS_LABEL_IPV6NULL) { struct net_device *lo = net->loopback_dev; rt2 = mpls_rt_alloc(lo->addr_len); if (!rt2) @@ -973,15 +973,15 @@ static int resize_platform_label_table(struct net *net, size_t limit) memcpy(labels, old, cp_size); /* If needed set the predefined labels */ - if ((old_limit <= MPLS_LABEL_IPV6_EXPLICIT_NULL) && - (limit > MPLS_LABEL_IPV6_EXPLICIT_NULL)) { - RCU_INIT_POINTER(labels[MPLS_LABEL_IPV6_EXPLICIT_NULL], rt2); + if ((old_limit <= MPLS_LABEL_IPV6NULL) && + (limit > MPLS_LABEL_IPV6NULL)) { + RCU_INIT_POINTER(labels[MPLS_LABEL_IPV6NULL], rt2); rt2 = NULL; } - if ((old_limit <= MPLS_LABEL_IPV4_EXPLICIT_NULL) && - (limit > MPLS_LABEL_IPV4_EXPLICIT_NULL)) { - RCU_INIT_POINTER(labels[MPLS_LABEL_IPV4_EXPLICIT_NULL], rt0); + if ((old_limit <= MPLS_LABEL_IPV4NULL) && + (limit > MPLS_LABEL_IPV4NULL)) { + RCU_INIT_POINTER(labels[MPLS_LABEL_IPV4NULL], rt0); rt0 = NULL; } -- cgit v1.2.3-71-gd317 From 1a48632ffed61352a7810ce089dc5a8bcd505a60 Mon Sep 17 00:00:00 2001 From: Peter Hurley Date: Mon, 13 Apr 2015 13:24:34 -0400 Subject: pty: Fix input race when closing A read() from a pty master may mistakenly indicate EOF (errno == -EIO) after the pty slave has closed, even though input data remains to be read. For example, pty slave | input worker | pty master | | | | n_tty_read() pty_write() | | input avail? no add data | | sleep schedule worker --->| | . |---> flush_to_ldisc() | . pty_close() | fill read buffer | . wait for worker | wakeup reader --->| . | read buffer full? |---> input avail ? yes |<--- yes - exit worker | copy 4096 bytes to user TTY_OTHER_CLOSED <---| |<--- kick worker | | **** New read() before worker starts **** | | n_tty_read() | | input avail? no | | TTY_OTHER_CLOSED? yes | | return -EIO Several conditions are required to trigger this race: 1. the ldisc read buffer must become full so the input worker exits 2. the read() count parameter must be >= 4096 so the ldisc read buffer is empty 3. the subsequent read() occurs before the kicked worker has processed more input However, the underlying cause of the race is that data is pipelined, while tty state is not; ie., data already written by the pty slave end is not yet visible to the pty master end, but state changes by the pty slave end are visible to the pty master end immediately. Pipeline the TTY_OTHER_CLOSED state through input worker to the reader. 1. Introduce TTY_OTHER_DONE which is set by the input worker when TTY_OTHER_CLOSED is set and either the input buffers are flushed or input processing has completed. Readers/polls are woken when TTY_OTHER_DONE is set. 2. Reader/poll checks TTY_OTHER_DONE instead of TTY_OTHER_CLOSED. 3. A new input worker is started from pty_close() after setting TTY_OTHER_CLOSED, which ensures the TTY_OTHER_DONE state will be set if the last input worker is already finished (or just about to exit). Remove tty_flush_to_ldisc(); no in-tree callers. Fixes: 52bce7f8d4fc ("pty, n_tty: Simplify input processing on final close") Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=96311 BugLink: http://bugs.launchpad.net/bugs/1429756 Cc: # 3.19+ Reported-by: Andy Whitcroft Reported-by: H.J. Lu Signed-off-by: Peter Hurley Signed-off-by: Greg Kroah-Hartman --- Documentation/serial/tty.txt | 3 +++ drivers/tty/n_hdlc.c | 4 ++-- drivers/tty/n_tty.c | 22 ++++++++++++++++++---- drivers/tty/pty.c | 5 +++-- drivers/tty/tty_buffer.c | 41 +++++++++++++++++++++++++++-------------- include/linux/tty.h | 2 +- 6 files changed, 54 insertions(+), 23 deletions(-) (limited to 'include') diff --git a/Documentation/serial/tty.txt b/Documentation/serial/tty.txt index 1e52d67d0abf..dbe6623fed1c 100644 --- a/Documentation/serial/tty.txt +++ b/Documentation/serial/tty.txt @@ -198,6 +198,9 @@ TTY_IO_ERROR If set, causes all subsequent userspace read/write TTY_OTHER_CLOSED Device is a pty and the other side has closed. +TTY_OTHER_DONE Device is a pty and the other side has closed and + all pending input processing has been completed. + TTY_NO_WRITE_SPLIT Prevent driver from splitting up writes into smaller chunks. diff --git a/drivers/tty/n_hdlc.c b/drivers/tty/n_hdlc.c index 644ddb841d9f..bbc4ce66c2c1 100644 --- a/drivers/tty/n_hdlc.c +++ b/drivers/tty/n_hdlc.c @@ -600,7 +600,7 @@ static ssize_t n_hdlc_tty_read(struct tty_struct *tty, struct file *file, add_wait_queue(&tty->read_wait, &wait); for (;;) { - if (test_bit(TTY_OTHER_CLOSED, &tty->flags)) { + if (test_bit(TTY_OTHER_DONE, &tty->flags)) { ret = -EIO; break; } @@ -828,7 +828,7 @@ static unsigned int n_hdlc_tty_poll(struct tty_struct *tty, struct file *filp, /* set bits for operations that won't block */ if (n_hdlc->rx_buf_list.head) mask |= POLLIN | POLLRDNORM; /* readable */ - if (test_bit(TTY_OTHER_CLOSED, &tty->flags)) + if (test_bit(TTY_OTHER_DONE, &tty->flags)) mask |= POLLHUP; if (tty_hung_up_p(filp)) mask |= POLLHUP; diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c index cf6e0f2e1331..cc57a3a6b02b 100644 --- a/drivers/tty/n_tty.c +++ b/drivers/tty/n_tty.c @@ -1949,6 +1949,18 @@ static inline int input_available_p(struct tty_struct *tty, int poll) return ldata->commit_head - ldata->read_tail >= amt; } +static inline int check_other_done(struct tty_struct *tty) +{ + int done = test_bit(TTY_OTHER_DONE, &tty->flags); + if (done) { + /* paired with cmpxchg() in check_other_closed(); ensures + * read buffer head index is not stale + */ + smp_mb__after_atomic(); + } + return done; +} + /** * copy_from_read_buf - copy read data directly * @tty: terminal device @@ -2167,7 +2179,7 @@ static ssize_t n_tty_read(struct tty_struct *tty, struct file *file, struct n_tty_data *ldata = tty->disc_data; unsigned char __user *b = buf; DEFINE_WAIT_FUNC(wait, woken_wake_function); - int c; + int c, done; int minimum, time; ssize_t retval = 0; long timeout; @@ -2235,8 +2247,10 @@ static ssize_t n_tty_read(struct tty_struct *tty, struct file *file, ((minimum - (b - buf)) >= 1)) ldata->minimum_to_wake = (minimum - (b - buf)); + done = check_other_done(tty); + if (!input_available_p(tty, 0)) { - if (test_bit(TTY_OTHER_CLOSED, &tty->flags)) { + if (done) { retval = -EIO; break; } @@ -2443,12 +2457,12 @@ static unsigned int n_tty_poll(struct tty_struct *tty, struct file *file, poll_wait(file, &tty->read_wait, wait); poll_wait(file, &tty->write_wait, wait); + if (check_other_done(tty)) + mask |= POLLHUP; if (input_available_p(tty, 1)) mask |= POLLIN | POLLRDNORM; if (tty->packet && tty->link->ctrl_status) mask |= POLLPRI | POLLIN | POLLRDNORM; - if (test_bit(TTY_OTHER_CLOSED, &tty->flags)) - mask |= POLLHUP; if (tty_hung_up_p(file)) mask |= POLLHUP; if (!(mask & (POLLHUP | POLLIN | POLLRDNORM))) { diff --git a/drivers/tty/pty.c b/drivers/tty/pty.c index e72ee629cead..4d5e8409769c 100644 --- a/drivers/tty/pty.c +++ b/drivers/tty/pty.c @@ -53,9 +53,8 @@ static void pty_close(struct tty_struct *tty, struct file *filp) /* Review - krefs on tty_link ?? */ if (!tty->link) return; - tty_flush_to_ldisc(tty->link); set_bit(TTY_OTHER_CLOSED, &tty->link->flags); - wake_up_interruptible(&tty->link->read_wait); + tty_flip_buffer_push(tty->link->port); wake_up_interruptible(&tty->link->write_wait); if (tty->driver->subtype == PTY_TYPE_MASTER) { set_bit(TTY_OTHER_CLOSED, &tty->flags); @@ -243,7 +242,9 @@ static int pty_open(struct tty_struct *tty, struct file *filp) goto out; clear_bit(TTY_IO_ERROR, &tty->flags); + /* TTY_OTHER_CLOSED must be cleared before TTY_OTHER_DONE */ clear_bit(TTY_OTHER_CLOSED, &tty->link->flags); + clear_bit(TTY_OTHER_DONE, &tty->link->flags); set_bit(TTY_THROTTLED, &tty->flags); return 0; diff --git a/drivers/tty/tty_buffer.c b/drivers/tty/tty_buffer.c index 75661641f5fe..2f78b77f0f81 100644 --- a/drivers/tty/tty_buffer.c +++ b/drivers/tty/tty_buffer.c @@ -37,6 +37,28 @@ #define TTY_BUFFER_PAGE (((PAGE_SIZE - sizeof(struct tty_buffer)) / 2) & ~0xFF) +/* + * If all tty flip buffers have been processed by flush_to_ldisc() or + * dropped by tty_buffer_flush(), check if the linked pty has been closed. + * If so, wake the reader/poll to process + */ +static inline void check_other_closed(struct tty_struct *tty) +{ + unsigned long flags, old; + + /* transition from TTY_OTHER_CLOSED => TTY_OTHER_DONE must be atomic */ + for (flags = ACCESS_ONCE(tty->flags); + test_bit(TTY_OTHER_CLOSED, &flags); + ) { + old = flags; + __set_bit(TTY_OTHER_DONE, &flags); + flags = cmpxchg(&tty->flags, old, flags); + if (old == flags) { + wake_up_interruptible(&tty->read_wait); + break; + } + } +} /** * tty_buffer_lock_exclusive - gain exclusive access to buffer @@ -229,6 +251,8 @@ void tty_buffer_flush(struct tty_struct *tty, struct tty_ldisc *ld) if (ld && ld->ops->flush_buffer) ld->ops->flush_buffer(tty); + check_other_closed(tty); + atomic_dec(&buf->priority); mutex_unlock(&buf->lock); } @@ -471,8 +495,10 @@ static void flush_to_ldisc(struct work_struct *work) smp_rmb(); count = head->commit - head->read; if (!count) { - if (next == NULL) + if (next == NULL) { + check_other_closed(tty); break; + } buf->head = next; tty_buffer_free(port, head); continue; @@ -488,19 +514,6 @@ static void flush_to_ldisc(struct work_struct *work) tty_ldisc_deref(disc); } -/** - * tty_flush_to_ldisc - * @tty: tty to push - * - * Push the terminal flip buffers to the line discipline. - * - * Must not be called from IRQ context. - */ -void tty_flush_to_ldisc(struct tty_struct *tty) -{ - flush_work(&tty->port->buf.work); -} - /** * tty_flip_buffer_push - terminal * @port: tty port to push diff --git a/include/linux/tty.h b/include/linux/tty.h index fe5623c9af71..d76631f615c2 100644 --- a/include/linux/tty.h +++ b/include/linux/tty.h @@ -339,6 +339,7 @@ struct tty_file_private { #define TTY_EXCLUSIVE 3 /* Exclusive open mode */ #define TTY_DEBUG 4 /* Debugging */ #define TTY_DO_WRITE_WAKEUP 5 /* Call write_wakeup after queuing new */ +#define TTY_OTHER_DONE 6 /* Closed pty has completed input processing */ #define TTY_LDISC_OPEN 11 /* Line discipline is open */ #define TTY_PTY_LOCK 16 /* pty private */ #define TTY_NO_WRITE_SPLIT 17 /* Preserve write boundaries to driver */ @@ -462,7 +463,6 @@ extern int tty_hung_up_p(struct file *filp); extern void do_SAK(struct tty_struct *tty); extern void __do_SAK(struct tty_struct *tty); extern void no_tty(void); -extern void tty_flush_to_ldisc(struct tty_struct *tty); extern void tty_buffer_free_all(struct tty_port *port); extern void tty_buffer_flush(struct tty_struct *tty, struct tty_ldisc *ld); extern void tty_buffer_init(struct tty_port *port); -- cgit v1.2.3-71-gd317 From 2d94e5224e81c58986a8cf44a3bf4830ce5cb96e Mon Sep 17 00:00:00 2001 From: Srinivas Pandruvada Date: Thu, 7 May 2015 14:26:34 -0700 Subject: HID: hid-sensor-hub: Fix debug lock warning When CONFIG_DEBUG_LOCK_ALLOC is defined, mutex magic is compared and warned for (l->magic != l), here l is the address of mutex pointer. In hid-sensor-hub as part of hsdev creation, a per hsdev mutex is initialized during MFD cell creation. This hsdev, which contains, mutex is part of platform data for the a cell. But platform_data is copied in platform_device_add_data() in platform.c. This copy will copy the whole hsdev structure including mutex. But once copied the magic will no longer match. So when client driver call sensor_hub_input_attr_get_raw_value, this will trigger mutex warning. So to avoid this allocate mutex dynamically. This will be same even after copy. Signed-off-by: Srinivas Pandruvada Signed-off-by: Jiri Kosina --- drivers/hid/hid-sensor-hub.c | 13 ++++++++++--- include/linux/hid-sensor-hub.h | 4 ++-- 2 files changed, 12 insertions(+), 5 deletions(-) (limited to 'include') diff --git a/drivers/hid/hid-sensor-hub.c b/drivers/hid/hid-sensor-hub.c index c3f6f1e311ea..090a1ba0abb6 100644 --- a/drivers/hid/hid-sensor-hub.c +++ b/drivers/hid/hid-sensor-hub.c @@ -294,7 +294,7 @@ int sensor_hub_input_attr_get_raw_value(struct hid_sensor_hub_device *hsdev, if (!report) return -EINVAL; - mutex_lock(&hsdev->mutex); + mutex_lock(hsdev->mutex_ptr); if (flag == SENSOR_HUB_SYNC) { memset(&hsdev->pending, 0, sizeof(hsdev->pending)); init_completion(&hsdev->pending.ready); @@ -328,7 +328,7 @@ int sensor_hub_input_attr_get_raw_value(struct hid_sensor_hub_device *hsdev, kfree(hsdev->pending.raw_data); hsdev->pending.status = false; } - mutex_unlock(&hsdev->mutex); + mutex_unlock(hsdev->mutex_ptr); return ret_val; } @@ -667,7 +667,14 @@ static int sensor_hub_probe(struct hid_device *hdev, hsdev->vendor_id = hdev->vendor; hsdev->product_id = hdev->product; hsdev->usage = collection->usage; - mutex_init(&hsdev->mutex); + hsdev->mutex_ptr = devm_kzalloc(&hdev->dev, + sizeof(struct mutex), + GFP_KERNEL); + if (!hsdev->mutex_ptr) { + ret = -ENOMEM; + goto err_stop_hw; + } + mutex_init(hsdev->mutex_ptr); hsdev->start_collection_index = i; if (last_hsdev) last_hsdev->end_collection_index = i; diff --git a/include/linux/hid-sensor-hub.h b/include/linux/hid-sensor-hub.h index 0408421d885f..0042bf330b99 100644 --- a/include/linux/hid-sensor-hub.h +++ b/include/linux/hid-sensor-hub.h @@ -74,7 +74,7 @@ struct sensor_hub_pending { * @usage: Usage id for this hub device instance. * @start_collection_index: Starting index for a phy type collection * @end_collection_index: Last index for a phy type collection - * @mutex: synchronizing mutex. + * @mutex_ptr: synchronizing mutex pointer. * @pending: Holds information of pending sync read request. */ struct hid_sensor_hub_device { @@ -84,7 +84,7 @@ struct hid_sensor_hub_device { u32 usage; int start_collection_index; int end_collection_index; - struct mutex mutex; + struct mutex *mutex_ptr; struct sensor_hub_pending pending; }; -- cgit v1.2.3-71-gd317 From fcf3b54282e4c5a95a1f45f67558bc105acdbc6a Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 12 May 2015 12:51:38 -0400 Subject: drm/radeon: add new bonaire pci id Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- include/drm/drm_pciids.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/drm/drm_pciids.h b/include/drm/drm_pciids.h index 2dd405c9be78..45c39a37f924 100644 --- a/include/drm/drm_pciids.h +++ b/include/drm/drm_pciids.h @@ -186,6 +186,7 @@ {0x1002, 0x6658, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BONAIRE|RADEON_NEW_MEMMAP}, \ {0x1002, 0x665c, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BONAIRE|RADEON_NEW_MEMMAP}, \ {0x1002, 0x665d, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BONAIRE|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x665f, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BONAIRE|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6660, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6663, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6664, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ -- cgit v1.2.3-71-gd317 From 336b7e1f230912cd8df2497be8dd7be4647d8fc8 Mon Sep 17 00:00:00 2001 From: Mike Snitzer Date: Mon, 11 May 2015 14:06:32 -0400 Subject: block: remove export for blk_queue_bio With commit ff36ab345 ("dm: remove request-based logic from make_request_fn wrapper") DM no longer calls blk_queue_bio() directly, so remove its export. Doing so required a forward declaration in blk-core.c. Signed-off-by: Mike Snitzer Signed-off-by: Jens Axboe --- block/blk-core.c | 5 +++-- include/linux/blkdev.h | 2 -- 2 files changed, 3 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/block/blk-core.c b/block/blk-core.c index 7871603f0a29..03b5f8d77f37 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -734,6 +734,8 @@ blk_init_queue_node(request_fn_proc *rfn, spinlock_t *lock, int node_id) } EXPORT_SYMBOL(blk_init_queue_node); +static void blk_queue_bio(struct request_queue *q, struct bio *bio); + struct request_queue * blk_init_allocated_queue(struct request_queue *q, request_fn_proc *rfn, spinlock_t *lock) @@ -1578,7 +1580,7 @@ void init_request_from_bio(struct request *req, struct bio *bio) blk_rq_bio_prep(req->q, req, bio); } -void blk_queue_bio(struct request_queue *q, struct bio *bio) +static void blk_queue_bio(struct request_queue *q, struct bio *bio) { const bool sync = !!(bio->bi_rw & REQ_SYNC); struct blk_plug *plug; @@ -1686,7 +1688,6 @@ out_unlock: spin_unlock_irq(q->queue_lock); } } -EXPORT_SYMBOL_GPL(blk_queue_bio); /* for device mapper only */ /* * If bio->bi_dev is a partition, remap the location diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 7f9a516f24de..5d93a6645e88 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -821,8 +821,6 @@ extern int scsi_cmd_ioctl(struct request_queue *, struct gendisk *, fmode_t, extern int sg_scsi_ioctl(struct request_queue *, struct gendisk *, fmode_t, struct scsi_ioctl_command __user *); -extern void blk_queue_bio(struct request_queue *q, struct bio *bio); - /* * A queue has just exitted congestion. Note this in the global counter of * congested queues, and wake up anyone who was waiting for requests to be -- cgit v1.2.3-71-gd317 From 01d460dd70adc858e15307332832183c622bee50 Mon Sep 17 00:00:00 2001 From: David Ahern Date: Tue, 12 May 2015 09:37:00 -0600 Subject: net: Remove remaining remnants of pm_qos from netdevice.h Commit e2c6544829f removed pm_qos from struct net_device but left the comment and header file. Remove those. Signed-off-by: David Ahern Cc: Thomas Graf Signed-off-by: David S. Miller --- include/linux/netdevice.h | 3 --- 1 file changed, 3 deletions(-) (limited to 'include') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 1899c74a7127..05b9a694e213 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -25,7 +25,6 @@ #ifndef _LINUX_NETDEVICE_H #define _LINUX_NETDEVICE_H -#include #include #include #include @@ -1499,8 +1498,6 @@ enum netdev_priv_flags { * * @qdisc_tx_busylock: XXX: need comments on this one * - * @pm_qos_req: Power Management QoS object - * * FIXME: cleanup struct net_device such that network protocol info * moves out. */ -- cgit v1.2.3-71-gd317 From f7bcb70ebae0dcdb5a2d859b09e4465784d99029 Mon Sep 17 00:00:00 2001 From: John Stultz Date: Fri, 8 May 2015 13:47:23 -0700 Subject: ktime: Fix ktime_divns to do signed division It was noted that the 32bit implementation of ktime_divns() was doing unsigned division and didn't properly handle negative values. And when a ktime helper was changed to utilize ktime_divns, it caused a regression on some IR blasters. See the following bugzilla for details: https://bugzilla.redhat.com/show_bug.cgi?id=1200353 This patch fixes the problem in ktime_divns by checking and preserving the sign bit, and then reapplying it if appropriate after the division, it also changes the return type to a s64 to make it more obvious this is expected. Nicolas also pointed out that negative dividers would cause infinite loops on 32bit systems, negative dividers is unlikely for users of this function, but out of caution this patch adds checks for negative dividers for both 32-bit (BUG_ON) and 64-bit(WARN_ON) versions to make sure no such use cases creep in. [ tglx: Hand an u64 to do_div() to avoid the compiler warning ] Fixes: 166afb64511e 'ktime: Sanitize ktime_to_us/ms conversion' Reported-and-tested-by: Trevor Cordes Signed-off-by: John Stultz Acked-by: Nicolas Pitre Cc: Ingo Molnar Cc: Josh Boyer Cc: One Thousand Gnomes Cc: Link: http://lkml.kernel.org/r/1431118043-23452-1-git-send-email-john.stultz@linaro.org Signed-off-by: Thomas Gleixner --- include/linux/ktime.h | 27 +++++++++++++++++++++------ kernel/time/hrtimer.c | 14 ++++++++------ 2 files changed, 29 insertions(+), 12 deletions(-) (limited to 'include') diff --git a/include/linux/ktime.h b/include/linux/ktime.h index 5fc3d1083071..2b6a204bd8d4 100644 --- a/include/linux/ktime.h +++ b/include/linux/ktime.h @@ -166,19 +166,34 @@ static inline bool ktime_before(const ktime_t cmp1, const ktime_t cmp2) } #if BITS_PER_LONG < 64 -extern u64 __ktime_divns(const ktime_t kt, s64 div); -static inline u64 ktime_divns(const ktime_t kt, s64 div) +extern s64 __ktime_divns(const ktime_t kt, s64 div); +static inline s64 ktime_divns(const ktime_t kt, s64 div) { + /* + * Negative divisors could cause an inf loop, + * so bug out here. + */ + BUG_ON(div < 0); if (__builtin_constant_p(div) && !(div >> 32)) { - u64 ns = kt.tv64; - do_div(ns, div); - return ns; + s64 ns = kt.tv64; + u64 tmp = ns < 0 ? -ns : ns; + + do_div(tmp, div); + return ns < 0 ? -tmp : tmp; } else { return __ktime_divns(kt, div); } } #else /* BITS_PER_LONG < 64 */ -# define ktime_divns(kt, div) (u64)((kt).tv64 / (div)) +static inline s64 ktime_divns(const ktime_t kt, s64 div) +{ + /* + * 32-bit implementation cannot handle negative divisors, + * so catch them on 64bit as well. + */ + WARN_ON(div < 0); + return kt.tv64 / div; +} #endif static inline s64 ktime_to_us(const ktime_t kt) diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c index 76d4bd962b19..93ef7190bdea 100644 --- a/kernel/time/hrtimer.c +++ b/kernel/time/hrtimer.c @@ -266,21 +266,23 @@ lock_hrtimer_base(const struct hrtimer *timer, unsigned long *flags) /* * Divide a ktime value by a nanosecond value */ -u64 __ktime_divns(const ktime_t kt, s64 div) +s64 __ktime_divns(const ktime_t kt, s64 div) { - u64 dclc; int sft = 0; + s64 dclc; + u64 tmp; dclc = ktime_to_ns(kt); + tmp = dclc < 0 ? -dclc : dclc; + /* Make sure the divisor is less than 2^32: */ while (div >> 32) { sft++; div >>= 1; } - dclc >>= sft; - do_div(dclc, (unsigned long) div); - - return dclc; + tmp >>= sft; + do_div(tmp, (unsigned long) div); + return dclc < 0 ? -tmp : tmp; } EXPORT_SYMBOL_GPL(__ktime_divns); #endif /* BITS_PER_LONG >= 64 */ -- cgit v1.2.3-71-gd317 From 8f4fc071b1926d0b20336e2b3f8ab85c94c734c5 Mon Sep 17 00:00:00 2001 From: Vladimir Davydov Date: Thu, 14 May 2015 15:16:55 -0700 Subject: gfp: add __GFP_NOACCOUNT Not all kmem allocations should be accounted to memcg. The following patch gives an example when accounting of a certain type of allocations to memcg can effectively result in a memory leak. This patch adds the __GFP_NOACCOUNT flag which if passed to kmalloc and friends will force the allocation to go through the root cgroup. It will be used by the next patch. Note, since in case of kmemleak enabled each kmalloc implies yet another allocation from the kmemleak_object cache, we add __GFP_NOACCOUNT to gfp_kmemleak_mask. Alternatively, we could introduce a per kmem cache flag disabling accounting for all allocations of a particular kind, but (a) we would not be able to bypass accounting for kmalloc then and (b) a kmem cache with this flag set could not be merged with a kmem cache without this flag, which would increase the number of global caches and therefore fragmentation even if the memory cgroup controller is not used. Despite its generic name, currently __GFP_NOACCOUNT disables accounting only for kmem allocations while user page allocations are always charged. To catch abusing of this flag, a warning is issued on an attempt of passing it to mem_cgroup_try_charge. Signed-off-by: Vladimir Davydov Cc: Tejun Heo Cc: Johannes Weiner Cc: Michal Hocko Cc: Christoph Lameter Cc: Pekka Enberg Cc: David Rientjes Cc: Joonsoo Kim Cc: Greg Thelen Cc: Greg Kroah-Hartman Cc: [4.0.x] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/gfp.h | 2 ++ include/linux/memcontrol.h | 4 ++++ mm/kmemleak.c | 3 ++- 3 files changed, 8 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/gfp.h b/include/linux/gfp.h index 97a9373e61e8..15928f0647e4 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -30,6 +30,7 @@ struct vm_area_struct; #define ___GFP_HARDWALL 0x20000u #define ___GFP_THISNODE 0x40000u #define ___GFP_RECLAIMABLE 0x80000u +#define ___GFP_NOACCOUNT 0x100000u #define ___GFP_NOTRACK 0x200000u #define ___GFP_NO_KSWAPD 0x400000u #define ___GFP_OTHER_NODE 0x800000u @@ -87,6 +88,7 @@ struct vm_area_struct; #define __GFP_HARDWALL ((__force gfp_t)___GFP_HARDWALL) /* Enforce hardwall cpuset memory allocs */ #define __GFP_THISNODE ((__force gfp_t)___GFP_THISNODE)/* No fallback, no policies */ #define __GFP_RECLAIMABLE ((__force gfp_t)___GFP_RECLAIMABLE) /* Page is reclaimable */ +#define __GFP_NOACCOUNT ((__force gfp_t)___GFP_NOACCOUNT) /* Don't account to kmemcg */ #define __GFP_NOTRACK ((__force gfp_t)___GFP_NOTRACK) /* Don't track with kmemcheck */ #define __GFP_NO_KSWAPD ((__force gfp_t)___GFP_NO_KSWAPD) diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 72dff5fb0d0c..6c8918114804 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -463,6 +463,8 @@ memcg_kmem_newpage_charge(gfp_t gfp, struct mem_cgroup **memcg, int order) if (!memcg_kmem_enabled()) return true; + if (gfp & __GFP_NOACCOUNT) + return true; /* * __GFP_NOFAIL allocations will move on even if charging is not * possible. Therefore we don't even try, and have this allocation @@ -522,6 +524,8 @@ memcg_kmem_get_cache(struct kmem_cache *cachep, gfp_t gfp) { if (!memcg_kmem_enabled()) return cachep; + if (gfp & __GFP_NOACCOUNT) + return cachep; if (gfp & __GFP_NOFAIL) return cachep; if (in_interrupt() || (!current->mm) || (current->flags & PF_KTHREAD)) diff --git a/mm/kmemleak.c b/mm/kmemleak.c index 5405aff5a590..f0fe4f2c1fa7 100644 --- a/mm/kmemleak.c +++ b/mm/kmemleak.c @@ -115,7 +115,8 @@ #define BYTES_PER_POINTER sizeof(void *) /* GFP bitmask for kmemleak internal allocations */ -#define gfp_kmemleak_mask(gfp) (((gfp) & (GFP_KERNEL | GFP_ATOMIC)) | \ +#define gfp_kmemleak_mask(gfp) (((gfp) & (GFP_KERNEL | GFP_ATOMIC | \ + __GFP_NOACCOUNT)) | \ __GFP_NORETRY | __GFP_NOMEMALLOC | \ __GFP_NOWARN) -- cgit v1.2.3-71-gd317 From 929aa5b250bfc59aca492d3213c7f3a53e2a5247 Mon Sep 17 00:00:00 2001 From: Josh Triplett Date: Thu, 14 May 2015 15:17:01 -0700 Subject: uidgid: make uid_valid and gid_valid work with !CONFIG_MULTIUSER {u,g}id_valid call {u,g}id_eq, which calls __k{u,g}id_val on both arguments and compares. With !CONFIG_MULTIUSER, __k{u,g}id_val return a constant 0, which makes {u,g}id_valid always return false. Change {u,g}id_valid to compare their argument against -1 instead. That produces identical results in the normal CONFIG_MULTIUSER=y case, but with !CONFIG_MULTIUSER will make {u,g}id_valid constant-fold into "return true;" rather than "return false;". This fixes uses of devpts without CONFIG_MULTIUSER. Signed-off-by: Josh Triplett Reported-by: Fengguang Wu , Cc: Peter Hurley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/uidgid.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/uidgid.h b/include/linux/uidgid.h index 0ee05da38899..03835522dfcb 100644 --- a/include/linux/uidgid.h +++ b/include/linux/uidgid.h @@ -109,12 +109,12 @@ static inline bool gid_lte(kgid_t left, kgid_t right) static inline bool uid_valid(kuid_t uid) { - return !uid_eq(uid, INVALID_UID); + return __kuid_val(uid) != (uid_t) -1; } static inline bool gid_valid(kgid_t gid) { - return !gid_eq(gid, INVALID_GID); + return __kgid_val(gid) != (gid_t) -1; } #ifdef CONFIG_USER_NS -- cgit v1.2.3-71-gd317 From eea39946a1f36e8a5a47c86e7ecfca6076868505 Mon Sep 17 00:00:00 2001 From: Roopa Prabhu Date: Wed, 13 May 2015 21:17:41 -0700 Subject: rename RTNH_F_EXTERNAL to RTNH_F_OFFLOAD RTNH_F_EXTERNAL today is printed as "offload" in iproute2 output. This patch renames the flag to be consistent with what the user sees. Signed-off-by: Roopa Prabhu Signed-off-by: David S. Miller --- include/uapi/linux/rtnetlink.h | 2 +- net/ipv4/fib_trie.c | 2 +- net/switchdev/switchdev.c | 6 +++--- 3 files changed, 5 insertions(+), 5 deletions(-) (limited to 'include') diff --git a/include/uapi/linux/rtnetlink.h b/include/uapi/linux/rtnetlink.h index 974db03f7b1a..17fb02f488da 100644 --- a/include/uapi/linux/rtnetlink.h +++ b/include/uapi/linux/rtnetlink.h @@ -337,7 +337,7 @@ struct rtnexthop { #define RTNH_F_DEAD 1 /* Nexthop is dead (used by multipath) */ #define RTNH_F_PERVASIVE 2 /* Do recursive gateway lookup */ #define RTNH_F_ONLINK 4 /* Gateway is forced on link */ -#define RTNH_F_EXTERNAL 8 /* Route installed externally */ +#define RTNH_F_OFFLOAD 8 /* offloaded route */ /* Macros to handle hexthops */ diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index e13fcc602da2..64c2076ced54 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -1764,7 +1764,7 @@ void fib_table_flush_external(struct fib_table *tb) /* record local slen */ slen = fa->fa_slen; - if (!fi || !(fi->fib_flags & RTNH_F_EXTERNAL)) + if (!fi || !(fi->fib_flags & RTNH_F_OFFLOAD)) continue; netdev_switch_fib_ipv4_del(n->key, diff --git a/net/switchdev/switchdev.c b/net/switchdev/switchdev.c index 46568b85c333..055453d48668 100644 --- a/net/switchdev/switchdev.c +++ b/net/switchdev/switchdev.c @@ -338,7 +338,7 @@ int netdev_switch_fib_ipv4_add(u32 dst, int dst_len, struct fib_info *fi, fi, tos, type, nlflags, tb_id); if (!err) - fi->fib_flags |= RTNH_F_EXTERNAL; + fi->fib_flags |= RTNH_F_OFFLOAD; } return err; @@ -364,7 +364,7 @@ int netdev_switch_fib_ipv4_del(u32 dst, int dst_len, struct fib_info *fi, const struct swdev_ops *ops; int err = 0; - if (!(fi->fib_flags & RTNH_F_EXTERNAL)) + if (!(fi->fib_flags & RTNH_F_OFFLOAD)) return 0; dev = netdev_switch_get_dev_by_nhs(fi); @@ -376,7 +376,7 @@ int netdev_switch_fib_ipv4_del(u32 dst, int dst_len, struct fib_info *fi, err = ops->swdev_fib_ipv4_del(dev, htonl(dst), dst_len, fi, tos, type, tb_id); if (!err) - fi->fib_flags &= ~RTNH_F_EXTERNAL; + fi->fib_flags &= ~RTNH_F_OFFLOAD; } return err; -- cgit v1.2.3-71-gd317 From b3cad287d13b5f6695c6b4aab72969cd64bf0171 Mon Sep 17 00:00:00 2001 From: Jesper Dangaard Brouer Date: Thu, 7 May 2015 14:54:16 +0200 Subject: conntrack: RFC5961 challenge ACK confuse conntrack LAST-ACK transition In compliance with RFC5961, the network stack send challenge ACK in response to spurious SYN packets, since commit 0c228e833c88 ("tcp: Restore RFC5961-compliant behavior for SYN packets"). This pose a problem for netfilter conntrack in state LAST_ACK, because this challenge ACK is (falsely) seen as ACKing last FIN, causing a false state transition (into TIME_WAIT). The challenge ACK is hard to distinguish from real last ACK. Thus, solution introduce a flag that tracks the potential for seeing a challenge ACK, in case a SYN packet is let through and current state is LAST_ACK. When conntrack transition LAST_ACK to TIME_WAIT happens, this flag is used for determining if we are expecting a challenge ACK. Scapy based reproducer script avail here: https://github.com/netoptimizer/network-testing/blob/master/scapy/tcp_hacks_3WHS_LAST_ACK.py Fixes: 0c228e833c88 ("tcp: Restore RFC5961-compliant behavior for SYN packets") Signed-off-by: Jesper Dangaard Brouer Acked-by: Jozsef Kadlecsik Signed-off-by: Pablo Neira Ayuso --- include/uapi/linux/netfilter/nf_conntrack_tcp.h | 3 +++ net/netfilter/nf_conntrack_proto_tcp.c | 35 ++++++++++++++++++++++--- 2 files changed, 35 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/uapi/linux/netfilter/nf_conntrack_tcp.h b/include/uapi/linux/netfilter/nf_conntrack_tcp.h index 9993a421201c..ef9f80f0f529 100644 --- a/include/uapi/linux/netfilter/nf_conntrack_tcp.h +++ b/include/uapi/linux/netfilter/nf_conntrack_tcp.h @@ -42,6 +42,9 @@ enum tcp_conntrack { /* The field td_maxack has been set */ #define IP_CT_TCP_FLAG_MAXACK_SET 0x20 +/* Marks possibility for expected RFC5961 challenge ACK */ +#define IP_CT_EXP_CHALLENGE_ACK 0x40 + struct nf_ct_tcp_flags { __u8 flags; __u8 mask; diff --git a/net/netfilter/nf_conntrack_proto_tcp.c b/net/netfilter/nf_conntrack_proto_tcp.c index 5caa0c41bf26..70383de72054 100644 --- a/net/netfilter/nf_conntrack_proto_tcp.c +++ b/net/netfilter/nf_conntrack_proto_tcp.c @@ -202,7 +202,7 @@ static const u8 tcp_conntracks[2][6][TCP_CONNTRACK_MAX] = { * sES -> sES :-) * sFW -> sCW Normal close request answered by ACK. * sCW -> sCW - * sLA -> sTW Last ACK detected. + * sLA -> sTW Last ACK detected (RFC5961 challenged) * sTW -> sTW Retransmitted last ACK. Remain in the same state. * sCL -> sCL */ @@ -261,7 +261,7 @@ static const u8 tcp_conntracks[2][6][TCP_CONNTRACK_MAX] = { * sES -> sES :-) * sFW -> sCW Normal close request answered by ACK. * sCW -> sCW - * sLA -> sTW Last ACK detected. + * sLA -> sTW Last ACK detected (RFC5961 challenged) * sTW -> sTW Retransmitted last ACK. * sCL -> sCL */ @@ -906,6 +906,7 @@ static int tcp_packet(struct nf_conn *ct, 1 : ct->proto.tcp.last_win; ct->proto.tcp.seen[ct->proto.tcp.last_dir].td_scale = ct->proto.tcp.last_wscale; + ct->proto.tcp.last_flags &= ~IP_CT_EXP_CHALLENGE_ACK; ct->proto.tcp.seen[ct->proto.tcp.last_dir].flags = ct->proto.tcp.last_flags; memset(&ct->proto.tcp.seen[dir], 0, @@ -923,7 +924,9 @@ static int tcp_packet(struct nf_conn *ct, * may be in sync but we are not. In that case, we annotate * the TCP options and let the packet go through. If it is a * valid SYN packet, the server will reply with a SYN/ACK, and - * then we'll get in sync. Otherwise, the server ignores it. */ + * then we'll get in sync. Otherwise, the server potentially + * responds with a challenge ACK if implementing RFC5961. + */ if (index == TCP_SYN_SET && dir == IP_CT_DIR_ORIGINAL) { struct ip_ct_tcp_state seen = {}; @@ -939,6 +942,13 @@ static int tcp_packet(struct nf_conn *ct, ct->proto.tcp.last_flags |= IP_CT_TCP_FLAG_SACK_PERM; } + /* Mark the potential for RFC5961 challenge ACK, + * this pose a special problem for LAST_ACK state + * as ACK is intrepretated as ACKing last FIN. + */ + if (old_state == TCP_CONNTRACK_LAST_ACK) + ct->proto.tcp.last_flags |= + IP_CT_EXP_CHALLENGE_ACK; } spin_unlock_bh(&ct->lock); if (LOG_INVALID(net, IPPROTO_TCP)) @@ -970,6 +980,25 @@ static int tcp_packet(struct nf_conn *ct, nf_log_packet(net, pf, 0, skb, NULL, NULL, NULL, "nf_ct_tcp: invalid state "); return -NF_ACCEPT; + case TCP_CONNTRACK_TIME_WAIT: + /* RFC5961 compliance cause stack to send "challenge-ACK" + * e.g. in response to spurious SYNs. Conntrack MUST + * not believe this ACK is acking last FIN. + */ + if (old_state == TCP_CONNTRACK_LAST_ACK && + index == TCP_ACK_SET && + ct->proto.tcp.last_dir != dir && + ct->proto.tcp.last_index == TCP_SYN_SET && + (ct->proto.tcp.last_flags & IP_CT_EXP_CHALLENGE_ACK)) { + /* Detected RFC5961 challenge ACK */ + ct->proto.tcp.last_flags &= ~IP_CT_EXP_CHALLENGE_ACK; + spin_unlock_bh(&ct->lock); + if (LOG_INVALID(net, IPPROTO_TCP)) + nf_log_packet(net, pf, 0, skb, NULL, NULL, NULL, + "nf_ct_tcp: challenge-ACK ignored "); + return NF_ACCEPT; /* Don't change state */ + } + break; case TCP_CONNTRACK_CLOSE: if (index == TCP_RST_SET && (ct->proto.tcp.seen[!dir].flags & IP_CT_TCP_FLAG_MAXACK_SET) -- cgit v1.2.3-71-gd317 From 07ee0722bf941960fb3888f9c9b5839473372fd1 Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Fri, 15 May 2015 11:30:47 +0800 Subject: rhashtable: Add cap on number of elements in hash table We currently have no limit on the number of elements in a hash table. This is a problem because some users (tipc) set a ceiling on the maximum table size and when that is reached the hash table may degenerate. Others may encounter OOM when growing and if we allow insertions when that happens the hash table perofrmance may also suffer. This patch adds a new paramater insecure_max_entries which becomes the cap on the table. If unset it defaults to max_size * 2. If it is also zero it means that there is no cap on the number of elements in the table. However, the table will grow whenever the utilisation hits 100% and if that growth fails, you will get ENOMEM on insertion. As allowing oversubscription is potentially dangerous, the name contains the word insecure. Note that the cap is not a hard limit. This is done for performance reasons as enforcing a hard limit will result in use of atomic ops that are heavier than the ones we currently use. The reasoning is that we're only guarding against a gross over- subscription of the table, rather than a small breach of the limit. Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- include/linux/rhashtable.h | 19 +++++++++++++++++++ lib/rhashtable.c | 11 +++++++++++ 2 files changed, 30 insertions(+) (limited to 'include') diff --git a/include/linux/rhashtable.h b/include/linux/rhashtable.h index dbcbcc59aa92..843ceca9a21e 100644 --- a/include/linux/rhashtable.h +++ b/include/linux/rhashtable.h @@ -17,6 +17,7 @@ #ifndef _LINUX_RHASHTABLE_H #define _LINUX_RHASHTABLE_H +#include #include #include #include @@ -100,6 +101,7 @@ struct rhashtable; * @key_len: Length of key * @key_offset: Offset of key in struct to be hashed * @head_offset: Offset of rhash_head in struct to be hashed + * @insecure_max_entries: Maximum number of entries (may be exceeded) * @max_size: Maximum size while expanding * @min_size: Minimum size while shrinking * @nulls_base: Base value to generate nulls marker @@ -115,6 +117,7 @@ struct rhashtable_params { size_t key_len; size_t key_offset; size_t head_offset; + unsigned int insecure_max_entries; unsigned int max_size; unsigned int min_size; u32 nulls_base; @@ -286,6 +289,18 @@ static inline bool rht_grow_above_100(const struct rhashtable *ht, (!ht->p.max_size || tbl->size < ht->p.max_size); } +/** + * rht_grow_above_max - returns true if table is above maximum + * @ht: hash table + * @tbl: current table + */ +static inline bool rht_grow_above_max(const struct rhashtable *ht, + const struct bucket_table *tbl) +{ + return ht->p.insecure_max_entries && + atomic_read(&ht->nelems) >= ht->p.insecure_max_entries; +} + /* The bucket lock is selected based on the hash and protects mutations * on a group of hash buckets. * @@ -589,6 +604,10 @@ restart: goto out; } + err = -E2BIG; + if (unlikely(rht_grow_above_max(ht, tbl))) + goto out; + if (unlikely(rht_grow_above_100(ht, tbl))) { slow_path: spin_unlock_bh(lock); diff --git a/lib/rhashtable.c b/lib/rhashtable.c index b28df4019ade..4396434e4715 100644 --- a/lib/rhashtable.c +++ b/lib/rhashtable.c @@ -14,6 +14,7 @@ * published by the Free Software Foundation. */ +#include #include #include #include @@ -446,6 +447,10 @@ int rhashtable_insert_slow(struct rhashtable *ht, const void *key, if (key && rhashtable_lookup_fast(ht, key, ht->p)) goto exit; + err = -E2BIG; + if (unlikely(rht_grow_above_max(ht, tbl))) + goto exit; + err = -EAGAIN; if (rhashtable_check_elasticity(ht, tbl, hash) || rht_grow_above_100(ht, tbl)) @@ -738,6 +743,12 @@ int rhashtable_init(struct rhashtable *ht, if (params->max_size) ht->p.max_size = rounddown_pow_of_two(params->max_size); + if (params->insecure_max_entries) + ht->p.insecure_max_entries = + rounddown_pow_of_two(params->insecure_max_entries); + else + ht->p.insecure_max_entries = ht->p.max_size * 2; + ht->p.min_size = max(ht->p.min_size, HASH_MIN_SIZE); /* The maximum (not average) chain length grows with the -- cgit v1.2.3-71-gd317 From b5d721d761fccf491f92eefc611e318561683d0a Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 18 May 2015 17:06:14 -0700 Subject: inet: properly align icsk_ca_priv tcp_illinois and upcoming tcp_cdg require 64bit alignment of icsk_ca_priv x86 does not care, but other architectures might. Fixes: 05cbc0db03e82 ("ipv4: Create probe timer for tcp PMTU as per RFC4821") Signed-off-by: Eric Dumazet Cc: Fan Du Acked-by: Fan Du Signed-off-by: David S. Miller --- include/net/inet_connection_sock.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h index 48a815823587..497bc14cdb85 100644 --- a/include/net/inet_connection_sock.h +++ b/include/net/inet_connection_sock.h @@ -129,9 +129,10 @@ struct inet_connection_sock { u32 probe_timestamp; } icsk_mtup; - u32 icsk_ca_priv[16]; u32 icsk_user_timeout; -#define ICSK_CA_PRIV_SIZE (16 * sizeof(u32)) + + u64 icsk_ca_priv[64 / sizeof(u64)]; +#define ICSK_CA_PRIV_SIZE (8 * sizeof(u64)) }; #define ICSK_TIME_RETRANS 1 /* Retransmit timer */ -- cgit v1.2.3-71-gd317 From 77bb3dfdc0d554befad58fdefbc41be5bc3ed38a Mon Sep 17 00:00:00 2001 From: David Vrabel Date: Tue, 19 May 2015 18:40:49 +0100 Subject: xen/events: don't bind non-percpu VIRQs with percpu chip A non-percpu VIRQ (e.g., VIRQ_CONSOLE) may be freed on a different VCPU than it is bound to. This can result in a race between handle_percpu_irq() and removing the action in __free_irq() because handle_percpu_irq() does not take desc->lock. The interrupt handler sees a NULL action and oopses. Only use the percpu chip/handler for per-CPU VIRQs (like VIRQ_TIMER). # cat /proc/interrupts | grep virq 40: 87246 0 xen-percpu-virq timer0 44: 0 0 xen-percpu-virq debug0 47: 0 20995 xen-percpu-virq timer1 51: 0 0 xen-percpu-virq debug1 69: 0 0 xen-dyn-virq xen-pcpu 74: 0 0 xen-dyn-virq mce 75: 29 0 xen-dyn-virq hvc_console Signed-off-by: David Vrabel Cc: --- drivers/tty/hvc/hvc_xen.c | 2 +- drivers/xen/events/events_base.c | 12 ++++++++---- include/xen/events.h | 2 +- 3 files changed, 10 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/drivers/tty/hvc/hvc_xen.c b/drivers/tty/hvc/hvc_xen.c index 5bab1c684bb1..7a3d146a5f0e 100644 --- a/drivers/tty/hvc/hvc_xen.c +++ b/drivers/tty/hvc/hvc_xen.c @@ -289,7 +289,7 @@ static int xen_initial_domain_console_init(void) return -ENOMEM; } - info->irq = bind_virq_to_irq(VIRQ_CONSOLE, 0); + info->irq = bind_virq_to_irq(VIRQ_CONSOLE, 0, false); info->vtermno = HVC_COOKIE; spin_lock(&xencons_lock); diff --git a/drivers/xen/events/events_base.c b/drivers/xen/events/events_base.c index 2b8553bd8715..38387950490e 100644 --- a/drivers/xen/events/events_base.c +++ b/drivers/xen/events/events_base.c @@ -957,7 +957,7 @@ unsigned xen_evtchn_nr_channels(void) } EXPORT_SYMBOL_GPL(xen_evtchn_nr_channels); -int bind_virq_to_irq(unsigned int virq, unsigned int cpu) +int bind_virq_to_irq(unsigned int virq, unsigned int cpu, bool percpu) { struct evtchn_bind_virq bind_virq; int evtchn, irq, ret; @@ -971,8 +971,12 @@ int bind_virq_to_irq(unsigned int virq, unsigned int cpu) if (irq < 0) goto out; - irq_set_chip_and_handler_name(irq, &xen_percpu_chip, - handle_percpu_irq, "virq"); + if (percpu) + irq_set_chip_and_handler_name(irq, &xen_percpu_chip, + handle_percpu_irq, "virq"); + else + irq_set_chip_and_handler_name(irq, &xen_dynamic_chip, + handle_edge_irq, "virq"); bind_virq.virq = virq; bind_virq.vcpu = cpu; @@ -1062,7 +1066,7 @@ int bind_virq_to_irqhandler(unsigned int virq, unsigned int cpu, { int irq, retval; - irq = bind_virq_to_irq(virq, cpu); + irq = bind_virq_to_irq(virq, cpu, irqflags & IRQF_PERCPU); if (irq < 0) return irq; retval = request_irq(irq, handler, irqflags, devname, dev_id); diff --git a/include/xen/events.h b/include/xen/events.h index 5321cd9636e6..7d95fdf9cf3e 100644 --- a/include/xen/events.h +++ b/include/xen/events.h @@ -17,7 +17,7 @@ int bind_evtchn_to_irqhandler(unsigned int evtchn, irq_handler_t handler, unsigned long irqflags, const char *devname, void *dev_id); -int bind_virq_to_irq(unsigned int virq, unsigned int cpu); +int bind_virq_to_irq(unsigned int virq, unsigned int cpu, bool percpu); int bind_virq_to_irqhandler(unsigned int virq, unsigned int cpu, irq_handler_t handler, unsigned long irqflags, const char *devname, -- cgit v1.2.3-71-gd317 From faecbb45ebefb20260ad4a631e011e93c896cb73 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 20 May 2015 13:42:25 +0200 Subject: Revert "netfilter: bridge: query conntrack about skb dnat" This reverts commit c055d5b03bb4cb69d349d787c9787c0383abd8b2. There are two issues: 'dnat_took_place' made me think that this is related to -j DNAT/MASQUERADE. But thats only one part of the story. This is also relevant for SNAT when we undo snat translation in reverse/reply direction. Furthermore, I originally wanted to do this mainly to avoid storing ipv6 addresses once we make DNAT/REDIRECT work for ipv6 on bridges. However, I forgot about SNPT/DNPT which is stateless. So we can't escape storing address for ipv6 anyway. Might as well do it for ipv4 too. Reported-and-tested-by: Bernhard Thaler Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- include/linux/skbuff.h | 1 + net/bridge/br_netfilter.c | 27 +++++++++------------------ 2 files changed, 10 insertions(+), 18 deletions(-) (limited to 'include') diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 66e374d62f64..f15154a879c7 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -176,6 +176,7 @@ struct nf_bridge_info { struct net_device *physindev; struct net_device *physoutdev; char neigh_header[8]; + __be32 ipv4_daddr; }; #endif diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c index ab55e2472beb..60ddfbeb47f5 100644 --- a/net/bridge/br_netfilter.c +++ b/net/bridge/br_netfilter.c @@ -37,10 +37,6 @@ #include #include -#if IS_ENABLED(CONFIG_NF_CONNTRACK) -#include -#endif - #include #include "br_private.h" #ifdef CONFIG_SYSCTL @@ -350,24 +346,15 @@ free_skb: return 0; } -static bool dnat_took_place(const struct sk_buff *skb) +static bool daddr_was_changed(const struct sk_buff *skb, + const struct nf_bridge_info *nf_bridge) { -#if IS_ENABLED(CONFIG_NF_CONNTRACK) - enum ip_conntrack_info ctinfo; - struct nf_conn *ct; - - ct = nf_ct_get(skb, &ctinfo); - if (!ct || nf_ct_is_untracked(ct)) - return false; - - return test_bit(IPS_DST_NAT_BIT, &ct->status); -#else - return false; -#endif + return ip_hdr(skb)->daddr != nf_bridge->ipv4_daddr; } /* This requires some explaining. If DNAT has taken place, * we will need to fix up the destination Ethernet address. + * This is also true when SNAT takes place (for the reply direction). * * There are two cases to consider: * 1. The packet was DNAT'ed to a device in the same bridge @@ -421,7 +408,7 @@ static int br_nf_pre_routing_finish(struct sock *sk, struct sk_buff *skb) nf_bridge->pkt_otherhost = false; } nf_bridge->mask ^= BRNF_NF_BRIDGE_PREROUTING; - if (dnat_took_place(skb)) { + if (daddr_was_changed(skb, nf_bridge)) { if ((err = ip_route_input(skb, iph->daddr, iph->saddr, iph->tos, dev))) { struct in_device *in_dev = __in_dev_get_rcu(dev); @@ -632,6 +619,7 @@ static unsigned int br_nf_pre_routing(const struct nf_hook_ops *ops, struct sk_buff *skb, const struct nf_hook_state *state) { + struct nf_bridge_info *nf_bridge; struct net_bridge_port *p; struct net_bridge *br; __u32 len = nf_bridge_encap_header_len(skb); @@ -669,6 +657,9 @@ static unsigned int br_nf_pre_routing(const struct nf_hook_ops *ops, if (!setup_pre_routing(skb)) return NF_DROP; + nf_bridge = nf_bridge_info_get(skb); + nf_bridge->ipv4_daddr = ip_hdr(skb)->daddr; + skb->protocol = htons(ETH_P_IP); NF_HOOK(NFPROTO_IPV4, NF_INET_PRE_ROUTING, state->sk, skb, -- cgit v1.2.3-71-gd317 From d654976cbf852ee20612ee10dbe57cdacda9f452 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 21 May 2015 21:51:19 -0700 Subject: tcp: fix a potential deadlock in tcp_get_info() Taking socket spinlock in tcp_get_info() can deadlock, as inet_diag_dump_icsk() holds the &hashinfo->ehash_locks[i], while packet processing can use the reverse locking order. We could avoid this locking for TCP_LISTEN states, but lockdep would certainly get confused as all TCP sockets share same lockdep classes. [ 523.722504] ====================================================== [ 523.728706] [ INFO: possible circular locking dependency detected ] [ 523.734990] 4.1.0-dbg-DEV #1676 Not tainted [ 523.739202] ------------------------------------------------------- [ 523.745474] ss/18032 is trying to acquire lock: [ 523.750002] (slock-AF_INET){+.-...}, at: [] tcp_get_info+0x2c4/0x360 [ 523.758129] [ 523.758129] but task is already holding lock: [ 523.763968] (&(&hashinfo->ehash_locks[i])->rlock){+.-...}, at: [] inet_diag_dump_icsk+0x1d5/0x6c0 [ 523.774661] [ 523.774661] which lock already depends on the new lock. [ 523.774661] [ 523.782850] [ 523.782850] the existing dependency chain (in reverse order) is: [ 523.790326] -> #1 (&(&hashinfo->ehash_locks[i])->rlock){+.-...}: [ 523.796599] [] lock_acquire+0xbb/0x270 [ 523.802565] [] _raw_spin_lock+0x38/0x50 [ 523.808628] [] __inet_hash_nolisten+0x78/0x110 [ 523.815273] [] tcp_v4_syn_recv_sock+0x24b/0x350 [ 523.822067] [] tcp_check_req+0x3c1/0x500 [ 523.828199] [] tcp_v4_do_rcv+0x239/0x3d0 [ 523.834331] [] tcp_v4_rcv+0xa8e/0xc10 [ 523.840202] [] ip_local_deliver_finish+0x133/0x3e0 [ 523.847214] [] ip_local_deliver+0xaa/0xc0 [ 523.853440] [] ip_rcv_finish+0x168/0x5c0 [ 523.859624] [] ip_rcv+0x307/0x420 Lets use u64_sync infrastructure instead. As a bonus, 64bit arches get optimized, as these are nop for them. Fixes: 0df48c26d841 ("tcp: add tcpi_bytes_acked to tcp_info") Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- include/linux/tcp.h | 2 ++ net/ipv4/tcp.c | 11 +++++++---- net/ipv4/tcp_fastopen.c | 4 ++++ net/ipv4/tcp_input.c | 4 ++++ 4 files changed, 17 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/include/linux/tcp.h b/include/linux/tcp.h index 3b2911502a8c..e8bbf403618f 100644 --- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -158,6 +158,8 @@ struct tcp_sock { * sum(delta(snd_una)), or how many bytes * were acked. */ + struct u64_stats_sync syncp; /* protects 64bit vars (cf tcp_get_info()) */ + u32 snd_una; /* First byte we want an ack for */ u32 snd_sml; /* Last byte of the most recently transmitted small packet */ u32 rcv_tstamp; /* timestamp of last received ACK (for keepalives) */ diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 46efa03d2b11..f1377f2a0472 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -402,6 +402,7 @@ void tcp_init_sock(struct sock *sk) tp->snd_ssthresh = TCP_INFINITE_SSTHRESH; tp->snd_cwnd_clamp = ~0; tp->mss_cache = TCP_MSS_DEFAULT; + u64_stats_init(&tp->syncp); tp->reordering = sysctl_tcp_reordering; tcp_enable_early_retrans(tp); @@ -2598,6 +2599,7 @@ void tcp_get_info(struct sock *sk, struct tcp_info *info) const struct tcp_sock *tp = tcp_sk(sk); const struct inet_connection_sock *icsk = inet_csk(sk); u32 now = tcp_time_stamp; + unsigned int start; u32 rate; memset(info, 0, sizeof(*info)); @@ -2665,10 +2667,11 @@ void tcp_get_info(struct sock *sk, struct tcp_info *info) rate = READ_ONCE(sk->sk_max_pacing_rate); info->tcpi_max_pacing_rate = rate != ~0U ? rate : ~0ULL; - spin_lock_bh(&sk->sk_lock.slock); - info->tcpi_bytes_acked = tp->bytes_acked; - info->tcpi_bytes_received = tp->bytes_received; - spin_unlock_bh(&sk->sk_lock.slock); + do { + start = u64_stats_fetch_begin_irq(&tp->syncp); + info->tcpi_bytes_acked = tp->bytes_acked; + info->tcpi_bytes_received = tp->bytes_received; + } while (u64_stats_fetch_retry_irq(&tp->syncp, start)); } EXPORT_SYMBOL_GPL(tcp_get_info); diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c index 3c673d5e6cff..46b087a27503 100644 --- a/net/ipv4/tcp_fastopen.c +++ b/net/ipv4/tcp_fastopen.c @@ -206,6 +206,10 @@ static bool tcp_fastopen_create_child(struct sock *sk, skb_set_owner_r(skb2, child); __skb_queue_tail(&child->sk_receive_queue, skb2); tp->syn_data_acked = 1; + + /* u64_stats_update_begin(&tp->syncp) not needed here, + * as we certainly are not changing upper 32bit value (0) + */ tp->bytes_received = end_seq - TCP_SKB_CB(skb)->seq - 1; } else { end_seq = TCP_SKB_CB(skb)->seq + 1; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 243d674b3ef5..c9ab964189a0 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3286,7 +3286,9 @@ static void tcp_snd_una_update(struct tcp_sock *tp, u32 ack) { u32 delta = ack - tp->snd_una; + u64_stats_update_begin(&tp->syncp); tp->bytes_acked += delta; + u64_stats_update_end(&tp->syncp); tp->snd_una = ack; } @@ -3295,7 +3297,9 @@ static void tcp_rcv_nxt_update(struct tcp_sock *tp, u32 seq) { u32 delta = seq - tp->rcv_nxt; + u64_stats_update_begin(&tp->syncp); tp->bytes_received += delta; + u64_stats_update_end(&tp->syncp); tp->rcv_nxt = seq; } -- cgit v1.2.3-71-gd317