From: Vedang Patel <vedang.patel@intel.com> To: netdev@vger.kernel.org Cc: jeffrey.t.kirsher@intel.com, davem@davemloft.net, jhs@mojatatu.com, xiyou.wangcong@gmail.com, jiri@resnulli.us, intel-wired-lan@lists.osuosl.org, vinicius.gomes@intel.com, l@dorileo.org, jakub.kicinski@netronome.com, m-karicheri2@ti.com, sergei.shtylyov@cogentembedded.com, eric.dumazet@gmail.com, aaron.f.brown@intel.com, Vedang Patel <vedang.patel@intel.com> Subject: [PATCH net-next v5 7/7] taprio: Adjust timestamps for TCP packets Date: Fri, 21 Jun 2019 10:28:28 -0700 [thread overview] Message-ID: <1561138108-12943-8-git-send-email-vedang.patel@intel.com> (raw) In-Reply-To: <1561138108-12943-1-git-send-email-vedang.patel@intel.com> When the taprio qdisc is running in "txtime offload" mode, it will set the launchtime value (in skb->tstamp) for all the packets which do not have the SO_TXTIME socket option. But, the TCP packets already have this value set and it indicates the earliest departure time represented in CLOCK_MONOTONIC clock. We need to respect the timestamp set by the TCP subsystem. So, convert this time to the clock which taprio is using and ensure that the packet is not transmitted before the deadline set by TCP. Signed-off-by: Vedang Patel <vedang.patel@intel.com> --- net/sched/sch_taprio.c | 41 ++++++++++++++++++++++++++++++++++++++++- 1 file changed, 40 insertions(+), 1 deletion(-) diff --git a/net/sched/sch_taprio.c b/net/sched/sch_taprio.c index 44540c30887e..36cad8d68883 100644 --- a/net/sched/sch_taprio.c +++ b/net/sched/sch_taprio.c @@ -22,6 +22,7 @@ #include <net/pkt_cls.h> #include <net/sch_generic.h> #include <net/sock.h> +#include <net/tcp.h> static LIST_HEAD(taprio_list); static DEFINE_SPINLOCK(taprio_list_lock); @@ -277,6 +278,41 @@ static ktime_t get_cycle_start(struct sched_gate_list *sched, return ktime_sub(time, cycle_elapsed); } +/* This returns the tstamp value set by TCP in terms of the set clock. */ +static ktime_t get_tcp_tstamp(struct taprio_sched *q, struct sk_buff *skb) +{ + unsigned int offset = skb_network_offset(skb); + const struct ipv6hdr *ipv6h; + const struct iphdr *iph; + struct ipv6hdr _ipv6h; + + ipv6h = skb_header_pointer(skb, offset, sizeof(_ipv6h), &_ipv6h); + if (!ipv6h) + return 0; + + if (ipv6h->version == 4) { + iph = (struct iphdr *)ipv6h; + offset += iph->ihl * 4; + + /* special-case 6in4 tunnelling, as that is a common way to get + * v6 connectivity in the home + */ + if (iph->protocol == IPPROTO_IPV6) { + ipv6h = skb_header_pointer(skb, offset, + sizeof(_ipv6h), &_ipv6h); + + if (!ipv6h || ipv6h->nexthdr != IPPROTO_TCP) + return 0; + } else if (iph->protocol != IPPROTO_TCP) { + return 0; + } + } else if (ipv6h->version == 6 && ipv6h->nexthdr != IPPROTO_TCP) { + return 0; + } + + return ktime_mono_to_any(skb->skb_mstamp_ns, q->tk_offset); +} + /* There are a few scenarios where we will have to modify the txtime from * what is read from next_txtime in sched_entry. They are: * 1. If txtime is in the past, @@ -294,7 +330,7 @@ static ktime_t get_cycle_start(struct sched_gate_list *sched, */ static long get_packet_txtime(struct sk_buff *skb, struct Qdisc *sch) { - ktime_t transmit_end_time, interval_end, interval_start; + ktime_t transmit_end_time, interval_end, interval_start, tcp_tstamp; struct taprio_sched *q = qdisc_priv(sch); struct sched_gate_list *sched, *admin; ktime_t minimum_time, now, txtime; @@ -305,6 +341,9 @@ static long get_packet_txtime(struct sk_buff *skb, struct Qdisc *sch) now = taprio_get_time(q); minimum_time = ktime_add_ns(now, q->txtime_delay); + tcp_tstamp = get_tcp_tstamp(q, skb); + minimum_time = max_t(ktime_t, minimum_time, tcp_tstamp); + rcu_read_lock(); admin = rcu_dereference(q->admin_sched); sched = rcu_dereference(q->oper_sched); -- 2.7.3
WARNING: multiple messages have this Message-ID (diff)
From: Vedang Patel <vedang.patel@intel.com> To: intel-wired-lan@osuosl.org Subject: [Intel-wired-lan] [PATCH net-next v5 7/7] taprio: Adjust timestamps for TCP packets Date: Fri, 21 Jun 2019 10:28:28 -0700 [thread overview] Message-ID: <1561138108-12943-8-git-send-email-vedang.patel@intel.com> (raw) In-Reply-To: <1561138108-12943-1-git-send-email-vedang.patel@intel.com> When the taprio qdisc is running in "txtime offload" mode, it will set the launchtime value (in skb->tstamp) for all the packets which do not have the SO_TXTIME socket option. But, the TCP packets already have this value set and it indicates the earliest departure time represented in CLOCK_MONOTONIC clock. We need to respect the timestamp set by the TCP subsystem. So, convert this time to the clock which taprio is using and ensure that the packet is not transmitted before the deadline set by TCP. Signed-off-by: Vedang Patel <vedang.patel@intel.com> --- net/sched/sch_taprio.c | 41 ++++++++++++++++++++++++++++++++++++++++- 1 file changed, 40 insertions(+), 1 deletion(-) diff --git a/net/sched/sch_taprio.c b/net/sched/sch_taprio.c index 44540c30887e..36cad8d68883 100644 --- a/net/sched/sch_taprio.c +++ b/net/sched/sch_taprio.c @@ -22,6 +22,7 @@ #include <net/pkt_cls.h> #include <net/sch_generic.h> #include <net/sock.h> +#include <net/tcp.h> static LIST_HEAD(taprio_list); static DEFINE_SPINLOCK(taprio_list_lock); @@ -277,6 +278,41 @@ static ktime_t get_cycle_start(struct sched_gate_list *sched, return ktime_sub(time, cycle_elapsed); } +/* This returns the tstamp value set by TCP in terms of the set clock. */ +static ktime_t get_tcp_tstamp(struct taprio_sched *q, struct sk_buff *skb) +{ + unsigned int offset = skb_network_offset(skb); + const struct ipv6hdr *ipv6h; + const struct iphdr *iph; + struct ipv6hdr _ipv6h; + + ipv6h = skb_header_pointer(skb, offset, sizeof(_ipv6h), &_ipv6h); + if (!ipv6h) + return 0; + + if (ipv6h->version == 4) { + iph = (struct iphdr *)ipv6h; + offset += iph->ihl * 4; + + /* special-case 6in4 tunnelling, as that is a common way to get + * v6 connectivity in the home + */ + if (iph->protocol == IPPROTO_IPV6) { + ipv6h = skb_header_pointer(skb, offset, + sizeof(_ipv6h), &_ipv6h); + + if (!ipv6h || ipv6h->nexthdr != IPPROTO_TCP) + return 0; + } else if (iph->protocol != IPPROTO_TCP) { + return 0; + } + } else if (ipv6h->version == 6 && ipv6h->nexthdr != IPPROTO_TCP) { + return 0; + } + + return ktime_mono_to_any(skb->skb_mstamp_ns, q->tk_offset); +} + /* There are a few scenarios where we will have to modify the txtime from * what is read from next_txtime in sched_entry. They are: * 1. If txtime is in the past, @@ -294,7 +330,7 @@ static ktime_t get_cycle_start(struct sched_gate_list *sched, */ static long get_packet_txtime(struct sk_buff *skb, struct Qdisc *sch) { - ktime_t transmit_end_time, interval_end, interval_start; + ktime_t transmit_end_time, interval_end, interval_start, tcp_tstamp; struct taprio_sched *q = qdisc_priv(sch); struct sched_gate_list *sched, *admin; ktime_t minimum_time, now, txtime; @@ -305,6 +341,9 @@ static long get_packet_txtime(struct sk_buff *skb, struct Qdisc *sch) now = taprio_get_time(q); minimum_time = ktime_add_ns(now, q->txtime_delay); + tcp_tstamp = get_tcp_tstamp(q, skb); + minimum_time = max_t(ktime_t, minimum_time, tcp_tstamp); + rcu_read_lock(); admin = rcu_dereference(q->admin_sched); sched = rcu_dereference(q->oper_sched); -- 2.7.3
next prev parent reply other threads:[~2019-06-21 17:29 UTC|newest] Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-06-21 17:28 [PATCH net-next v5 0/7] net/sched: Add txtime-assist support for taprio Vedang Patel 2019-06-21 17:28 ` [Intel-wired-lan] " Vedang Patel 2019-06-21 17:28 ` [PATCH net-next v5 1/7] igb: clear out skb->tstamp after reading the txtime Vedang Patel 2019-06-21 17:28 ` [Intel-wired-lan] " Vedang Patel 2019-06-21 17:28 ` [PATCH net-next v5 2/7] etf: Add skip_sock_check Vedang Patel 2019-06-21 17:28 ` [Intel-wired-lan] " Vedang Patel 2019-06-24 21:40 ` Jakub Kicinski 2019-06-24 21:40 ` [Intel-wired-lan] " Jakub Kicinski 2019-06-24 23:34 ` Patel, Vedang 2019-06-24 23:34 ` [Intel-wired-lan] " Patel, Vedang 2019-06-21 17:28 ` [PATCH net-next v5 3/7] taprio: calculate cycle_time when schedule is installed Vedang Patel 2019-06-21 17:28 ` [Intel-wired-lan] " Vedang Patel 2019-06-21 17:28 ` [PATCH net-next v5 4/7] taprio: Remove inline directive Vedang Patel 2019-06-21 17:28 ` [Intel-wired-lan] " Vedang Patel 2019-06-21 17:28 ` [PATCH net-next v5 5/7] taprio: Add support for txtime-assist mode Vedang Patel 2019-06-21 17:28 ` [Intel-wired-lan] " Vedang Patel 2019-06-23 3:29 ` kbuild test robot 2019-06-23 3:29 ` [Intel-wired-lan] " kbuild test robot 2019-06-21 17:28 ` [PATCH net-next v5 6/7] taprio: make clock reference conversions easier Vedang Patel 2019-06-21 17:28 ` [Intel-wired-lan] " Vedang Patel 2019-06-21 17:28 ` Vedang Patel [this message] 2019-06-21 17:28 ` [Intel-wired-lan] [PATCH net-next v5 7/7] taprio: Adjust timestamps for TCP packets Vedang Patel
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=1561138108-12943-8-git-send-email-vedang.patel@intel.com \ --to=vedang.patel@intel.com \ --cc=aaron.f.brown@intel.com \ --cc=davem@davemloft.net \ --cc=eric.dumazet@gmail.com \ --cc=intel-wired-lan@lists.osuosl.org \ --cc=jakub.kicinski@netronome.com \ --cc=jeffrey.t.kirsher@intel.com \ --cc=jhs@mojatatu.com \ --cc=jiri@resnulli.us \ --cc=l@dorileo.org \ --cc=m-karicheri2@ti.com \ --cc=netdev@vger.kernel.org \ --cc=sergei.shtylyov@cogentembedded.com \ --cc=vinicius.gomes@intel.com \ --cc=xiyou.wangcong@gmail.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.