OpenCloudOS-Kernel/include/net/sch_generic.h

895 lines
21 KiB
C
Raw Normal View History

#ifndef __NET_SCHED_GENERIC_H
#define __NET_SCHED_GENERIC_H
#include <linux/netdevice.h>
#include <linux/types.h>
#include <linux/rcupdate.h>
#include <linux/pkt_sched.h>
#include <linux/pkt_cls.h>
#include <linux/percpu.h>
qdisc: bulk dequeue support for qdiscs with TCQ_F_ONETXQUEUE Based on DaveM's recent API work on dev_hard_start_xmit(), that allows sending/processing an entire skb list. This patch implements qdisc bulk dequeue, by allowing multiple packets to be dequeued in dequeue_skb(). The optimization principle for this is two fold, (1) to amortize locking cost and (2) avoid expensive tailptr update for notifying HW. (1) Several packets are dequeued while holding the qdisc root_lock, amortizing locking cost over several packet. The dequeued SKB list is processed under the TXQ lock in dev_hard_start_xmit(), thus also amortizing the cost of the TXQ lock. (2) Further more, dev_hard_start_xmit() will utilize the skb->xmit_more API to delay HW tailptr update, which also reduces the cost per packet. One restriction of the new API is that every SKB must belong to the same TXQ. This patch takes the easy way out, by restricting bulk dequeue to qdisc's with the TCQ_F_ONETXQUEUE flag, that specifies the qdisc only have attached a single TXQ. Some detail about the flow; dev_hard_start_xmit() will process the skb list, and transmit packets individually towards the driver (see xmit_one()). In case the driver stops midway in the list, the remaining skb list is returned by dev_hard_start_xmit(). In sch_direct_xmit() this returned list is requeued by dev_requeue_skb(). To avoid overshooting the HW limits, which results in requeuing, the patch limits the amount of bytes dequeued, based on the drivers BQL limits. In-effect bulking will only happen for BQL enabled drivers. Small amounts for extra HoL blocking (2x MTU/0.24ms) were measured at 100Mbit/s, with bulking 8 packets, but the oscillating nature of the measurement indicate something, like sched latency might be causing this effect. More comparisons show, that this oscillation goes away occationally. Thus, we disregard this artifact completely and remove any "magic" bulking limit. For now, as a conservative approach, stop bulking when seeing TSO and segmented GSO packets. They already benefit from bulking on their own. A followup patch add this, to allow easier bisect-ability for finding regressions. Jointed work with Hannes, Daniel and Florian. Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com> Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-10-02 04:35:59 +08:00
#include <linux/dynamic_queue_limits.h>
#include <linux/list.h>
#include <linux/refcount.h>
#include <net/gen_stats.h>
#include <net/rtnetlink.h>
struct Qdisc_ops;
struct qdisc_walker;
struct tcf_walker;
struct module;
struct qdisc_rate_table {
struct tc_ratespec rate;
u32 data[256];
struct qdisc_rate_table *next;
int refcnt;
};
enum qdisc_state_t {
__QDISC_STATE_SCHED,
__QDISC_STATE_DEACTIVATED,
};
struct qdisc_size_table {
struct rcu_head rcu;
struct list_head list;
struct tc_sizespec szopts;
int refcnt;
u16 data[];
};
/* similar to sk_buff_head, but skb->prev pointer is undefined. */
struct qdisc_skb_head {
struct sk_buff *head;
struct sk_buff *tail;
__u32 qlen;
spinlock_t lock;
};
struct Qdisc {
int (*enqueue)(struct sk_buff *skb,
struct Qdisc *sch,
struct sk_buff **to_free);
struct sk_buff * (*dequeue)(struct Qdisc *sch);
unsigned int flags;
#define TCQ_F_BUILTIN 1
#define TCQ_F_INGRESS 2
#define TCQ_F_CAN_BYPASS 4
#define TCQ_F_MQROOT 8
#define TCQ_F_ONETXQUEUE 0x10 /* dequeue_skb() can assume all skbs are for
* q->dev_queue : It can test
* netif_xmit_frozen_or_stopped() before
* dequeueing next packet.
* Its true for MQ/MQPRIO slaves, or non
* multiqueue device.
*/
#define TCQ_F_WARN_NONWC (1 << 16)
#define TCQ_F_CPUSTATS 0x20 /* run using percpu statistics */
net_sched: fix qdisc_tree_decrease_qlen() races qdisc_tree_decrease_qlen() suffers from two problems on multiqueue devices. One problem is that it updates sch->q.qlen and sch->qstats.drops on the mq/mqprio root qdisc, while it should not : Daniele reported underflows errors : [ 681.774821] PAX: sch->q.qlen: 0 n: 1 [ 681.774825] PAX: size overflow detected in function qdisc_tree_decrease_qlen net/sched/sch_api.c:769 cicus.693_49 min, count: 72, decl: qlen; num: 0; context: sk_buff_head; [ 681.774954] CPU: 2 PID: 19 Comm: ksoftirqd/2 Tainted: G O 4.2.6.201511282239-1-grsec #1 [ 681.774955] Hardware name: ASUSTeK COMPUTER INC. X302LJ/X302LJ, BIOS X302LJ.202 03/05/2015 [ 681.774956] ffffffffa9a04863 0000000000000000 0000000000000000 ffffffffa990ff7c [ 681.774959] ffffc90000d3bc38 ffffffffa95d2810 0000000000000007 ffffffffa991002b [ 681.774960] ffffc90000d3bc68 ffffffffa91a44f4 0000000000000001 0000000000000001 [ 681.774962] Call Trace: [ 681.774967] [<ffffffffa95d2810>] dump_stack+0x4c/0x7f [ 681.774970] [<ffffffffa91a44f4>] report_size_overflow+0x34/0x50 [ 681.774972] [<ffffffffa94d17e2>] qdisc_tree_decrease_qlen+0x152/0x160 [ 681.774976] [<ffffffffc02694b1>] fq_codel_dequeue+0x7b1/0x820 [sch_fq_codel] [ 681.774978] [<ffffffffc02680a0>] ? qdisc_peek_dequeued+0xa0/0xa0 [sch_fq_codel] [ 681.774980] [<ffffffffa94cd92d>] __qdisc_run+0x4d/0x1d0 [ 681.774983] [<ffffffffa949b2b2>] net_tx_action+0xc2/0x160 [ 681.774985] [<ffffffffa90664c1>] __do_softirq+0xf1/0x200 [ 681.774987] [<ffffffffa90665ee>] run_ksoftirqd+0x1e/0x30 [ 681.774989] [<ffffffffa90896b0>] smpboot_thread_fn+0x150/0x260 [ 681.774991] [<ffffffffa9089560>] ? sort_range+0x40/0x40 [ 681.774992] [<ffffffffa9085fe4>] kthread+0xe4/0x100 [ 681.774994] [<ffffffffa9085f00>] ? kthread_worker_fn+0x170/0x170 [ 681.774995] [<ffffffffa95d8d1e>] ret_from_fork+0x3e/0x70 mq/mqprio have their own ways to report qlen/drops by folding stats on all their queues, with appropriate locking. A second problem is that qdisc_tree_decrease_qlen() calls qdisc_lookup() without proper locking : concurrent qdisc updates could corrupt the list that qdisc_match_from_root() parses to find a qdisc given its handle. Fix first problem adding a TCQ_F_NOPARENT qdisc flag that qdisc_tree_decrease_qlen() can use to abort its tree traversal, as soon as it meets a mq/mqprio qdisc children. Second problem can be fixed by RCU protection. Qdisc are already freed after RCU grace period, so qdisc_list_add() and qdisc_list_del() simply have to use appropriate rcu list variants. A future patch will add a per struct netdev_queue list anchor, so that qdisc_tree_decrease_qlen() can have more efficient lookups. Reported-by: Daniele Fucini <dfucini@gmail.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Cong Wang <cwang@twopensource.com> Cc: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-12-02 12:08:51 +08:00
#define TCQ_F_NOPARENT 0x40 /* root of its hierarchy :
* qdisc_tree_decrease_qlen() should stop.
*/
#define TCQ_F_INVISIBLE 0x80 /* invisible by default in dump */
u32 limit;
const struct Qdisc_ops *ops;
struct qdisc_size_table __rcu *stab;
struct hlist_node hash;
u32 handle;
u32 parent;
struct netdev_queue *dev_queue;
struct net_rate_estimator __rcu *rate_est;
net: sched: fix panic in rate estimators Doing the following commands on a non idle network device panics the box instantly, because cpu_bstats gets overwritten by stats. tc qdisc add dev eth0 root <your_favorite_qdisc> ... some traffic (one packet is enough) ... tc qdisc replace dev eth0 root est 1sec 4sec <your_favorite_qdisc> [ 325.355596] BUG: unable to handle kernel paging request at ffff8841dc5a074c [ 325.362609] IP: [<ffffffff81541c9e>] __gnet_stats_copy_basic+0x3e/0x90 [ 325.369158] PGD 1fa7067 PUD 0 [ 325.372254] Oops: 0000 [#1] SMP [ 325.375514] Modules linked in: ... [ 325.398346] CPU: 13 PID: 14313 Comm: tc Not tainted 3.19.0-smp-DEV #1163 [ 325.412042] task: ffff8800793ab5d0 ti: ffff881ff2fa4000 task.ti: ffff881ff2fa4000 [ 325.419518] RIP: 0010:[<ffffffff81541c9e>] [<ffffffff81541c9e>] __gnet_stats_copy_basic+0x3e/0x90 [ 325.428506] RSP: 0018:ffff881ff2fa7928 EFLAGS: 00010286 [ 325.433824] RAX: 000000000000000c RBX: ffff881ff2fa796c RCX: 000000000000000c [ 325.440988] RDX: ffff8841dc5a0744 RSI: 0000000000000060 RDI: 0000000000000060 [ 325.448120] RBP: ffff881ff2fa7948 R08: ffffffff81cd4f80 R09: 0000000000000000 [ 325.455268] R10: ffff883ff223e400 R11: 0000000000000000 R12: 000000015cba0744 [ 325.462405] R13: ffffffff81cd4f80 R14: ffff883ff223e460 R15: ffff883feea0722c [ 325.469536] FS: 00007f2ee30fa700(0000) GS:ffff88407fa20000(0000) knlGS:0000000000000000 [ 325.477630] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 325.483380] CR2: ffff8841dc5a074c CR3: 0000003feeae9000 CR4: 00000000001407e0 [ 325.490510] Stack: [ 325.492524] ffff883feea0722c ffff883fef719dc0 ffff883feea0722c ffff883ff223e4a0 [ 325.499990] ffff881ff2fa79a8 ffffffff815424ee ffff883ff223e49c 000000015cba0744 [ 325.507460] 00000000f2fa7978 0000000000000000 ffff881ff2fa79a8 ffff883ff223e4a0 [ 325.514956] Call Trace: [ 325.517412] [<ffffffff815424ee>] gen_new_estimator+0x8e/0x230 [ 325.523250] [<ffffffff815427aa>] gen_replace_estimator+0x4a/0x60 [ 325.529349] [<ffffffff815718ab>] tc_modify_qdisc+0x52b/0x590 [ 325.535117] [<ffffffff8155edd0>] rtnetlink_rcv_msg+0xa0/0x240 [ 325.540963] [<ffffffff8155ed30>] ? __rtnl_unlock+0x20/0x20 [ 325.546532] [<ffffffff8157f811>] netlink_rcv_skb+0xb1/0xc0 [ 325.552145] [<ffffffff8155b355>] rtnetlink_rcv+0x25/0x40 [ 325.557558] [<ffffffff8157f0d8>] netlink_unicast+0x168/0x220 [ 325.563317] [<ffffffff8157f47c>] netlink_sendmsg+0x2ec/0x3e0 Lets play safe and not use an union : percpu 'pointers' are mostly read anyway, and we have typically few qdiscs per host. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: John Fastabend <john.fastabend@gmail.com> Fixes: 22e0f8b9322c ("net: sched: make bstats per cpu and estimator RCU safe") Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-30 09:30:12 +08:00
struct gnet_stats_basic_cpu __percpu *cpu_bstats;
struct gnet_stats_queue __percpu *cpu_qstats;
/*
* For performance sake on SMP, we put highly modified fields at the end
*/
struct sk_buff *gso_skb ____cacheline_aligned_in_smp;
struct qdisc_skb_head q;
net: sched: fix panic in rate estimators Doing the following commands on a non idle network device panics the box instantly, because cpu_bstats gets overwritten by stats. tc qdisc add dev eth0 root <your_favorite_qdisc> ... some traffic (one packet is enough) ... tc qdisc replace dev eth0 root est 1sec 4sec <your_favorite_qdisc> [ 325.355596] BUG: unable to handle kernel paging request at ffff8841dc5a074c [ 325.362609] IP: [<ffffffff81541c9e>] __gnet_stats_copy_basic+0x3e/0x90 [ 325.369158] PGD 1fa7067 PUD 0 [ 325.372254] Oops: 0000 [#1] SMP [ 325.375514] Modules linked in: ... [ 325.398346] CPU: 13 PID: 14313 Comm: tc Not tainted 3.19.0-smp-DEV #1163 [ 325.412042] task: ffff8800793ab5d0 ti: ffff881ff2fa4000 task.ti: ffff881ff2fa4000 [ 325.419518] RIP: 0010:[<ffffffff81541c9e>] [<ffffffff81541c9e>] __gnet_stats_copy_basic+0x3e/0x90 [ 325.428506] RSP: 0018:ffff881ff2fa7928 EFLAGS: 00010286 [ 325.433824] RAX: 000000000000000c RBX: ffff881ff2fa796c RCX: 000000000000000c [ 325.440988] RDX: ffff8841dc5a0744 RSI: 0000000000000060 RDI: 0000000000000060 [ 325.448120] RBP: ffff881ff2fa7948 R08: ffffffff81cd4f80 R09: 0000000000000000 [ 325.455268] R10: ffff883ff223e400 R11: 0000000000000000 R12: 000000015cba0744 [ 325.462405] R13: ffffffff81cd4f80 R14: ffff883ff223e460 R15: ffff883feea0722c [ 325.469536] FS: 00007f2ee30fa700(0000) GS:ffff88407fa20000(0000) knlGS:0000000000000000 [ 325.477630] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 325.483380] CR2: ffff8841dc5a074c CR3: 0000003feeae9000 CR4: 00000000001407e0 [ 325.490510] Stack: [ 325.492524] ffff883feea0722c ffff883fef719dc0 ffff883feea0722c ffff883ff223e4a0 [ 325.499990] ffff881ff2fa79a8 ffffffff815424ee ffff883ff223e49c 000000015cba0744 [ 325.507460] 00000000f2fa7978 0000000000000000 ffff881ff2fa79a8 ffff883ff223e4a0 [ 325.514956] Call Trace: [ 325.517412] [<ffffffff815424ee>] gen_new_estimator+0x8e/0x230 [ 325.523250] [<ffffffff815427aa>] gen_replace_estimator+0x4a/0x60 [ 325.529349] [<ffffffff815718ab>] tc_modify_qdisc+0x52b/0x590 [ 325.535117] [<ffffffff8155edd0>] rtnetlink_rcv_msg+0xa0/0x240 [ 325.540963] [<ffffffff8155ed30>] ? __rtnl_unlock+0x20/0x20 [ 325.546532] [<ffffffff8157f811>] netlink_rcv_skb+0xb1/0xc0 [ 325.552145] [<ffffffff8155b355>] rtnetlink_rcv+0x25/0x40 [ 325.557558] [<ffffffff8157f0d8>] netlink_unicast+0x168/0x220 [ 325.563317] [<ffffffff8157f47c>] netlink_sendmsg+0x2ec/0x3e0 Lets play safe and not use an union : percpu 'pointers' are mostly read anyway, and we have typically few qdiscs per host. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: John Fastabend <john.fastabend@gmail.com> Fixes: 22e0f8b9322c ("net: sched: make bstats per cpu and estimator RCU safe") Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-30 09:30:12 +08:00
struct gnet_stats_basic_packed bstats;
seqcount_t running;
net: sched: fix panic in rate estimators Doing the following commands on a non idle network device panics the box instantly, because cpu_bstats gets overwritten by stats. tc qdisc add dev eth0 root <your_favorite_qdisc> ... some traffic (one packet is enough) ... tc qdisc replace dev eth0 root est 1sec 4sec <your_favorite_qdisc> [ 325.355596] BUG: unable to handle kernel paging request at ffff8841dc5a074c [ 325.362609] IP: [<ffffffff81541c9e>] __gnet_stats_copy_basic+0x3e/0x90 [ 325.369158] PGD 1fa7067 PUD 0 [ 325.372254] Oops: 0000 [#1] SMP [ 325.375514] Modules linked in: ... [ 325.398346] CPU: 13 PID: 14313 Comm: tc Not tainted 3.19.0-smp-DEV #1163 [ 325.412042] task: ffff8800793ab5d0 ti: ffff881ff2fa4000 task.ti: ffff881ff2fa4000 [ 325.419518] RIP: 0010:[<ffffffff81541c9e>] [<ffffffff81541c9e>] __gnet_stats_copy_basic+0x3e/0x90 [ 325.428506] RSP: 0018:ffff881ff2fa7928 EFLAGS: 00010286 [ 325.433824] RAX: 000000000000000c RBX: ffff881ff2fa796c RCX: 000000000000000c [ 325.440988] RDX: ffff8841dc5a0744 RSI: 0000000000000060 RDI: 0000000000000060 [ 325.448120] RBP: ffff881ff2fa7948 R08: ffffffff81cd4f80 R09: 0000000000000000 [ 325.455268] R10: ffff883ff223e400 R11: 0000000000000000 R12: 000000015cba0744 [ 325.462405] R13: ffffffff81cd4f80 R14: ffff883ff223e460 R15: ffff883feea0722c [ 325.469536] FS: 00007f2ee30fa700(0000) GS:ffff88407fa20000(0000) knlGS:0000000000000000 [ 325.477630] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 325.483380] CR2: ffff8841dc5a074c CR3: 0000003feeae9000 CR4: 00000000001407e0 [ 325.490510] Stack: [ 325.492524] ffff883feea0722c ffff883fef719dc0 ffff883feea0722c ffff883ff223e4a0 [ 325.499990] ffff881ff2fa79a8 ffffffff815424ee ffff883ff223e49c 000000015cba0744 [ 325.507460] 00000000f2fa7978 0000000000000000 ffff881ff2fa79a8 ffff883ff223e4a0 [ 325.514956] Call Trace: [ 325.517412] [<ffffffff815424ee>] gen_new_estimator+0x8e/0x230 [ 325.523250] [<ffffffff815427aa>] gen_replace_estimator+0x4a/0x60 [ 325.529349] [<ffffffff815718ab>] tc_modify_qdisc+0x52b/0x590 [ 325.535117] [<ffffffff8155edd0>] rtnetlink_rcv_msg+0xa0/0x240 [ 325.540963] [<ffffffff8155ed30>] ? __rtnl_unlock+0x20/0x20 [ 325.546532] [<ffffffff8157f811>] netlink_rcv_skb+0xb1/0xc0 [ 325.552145] [<ffffffff8155b355>] rtnetlink_rcv+0x25/0x40 [ 325.557558] [<ffffffff8157f0d8>] netlink_unicast+0x168/0x220 [ 325.563317] [<ffffffff8157f47c>] netlink_sendmsg+0x2ec/0x3e0 Lets play safe and not use an union : percpu 'pointers' are mostly read anyway, and we have typically few qdiscs per host. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: John Fastabend <john.fastabend@gmail.com> Fixes: 22e0f8b9322c ("net: sched: make bstats per cpu and estimator RCU safe") Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-30 09:30:12 +08:00
struct gnet_stats_queue qstats;
unsigned long state;
struct Qdisc *next_sched;
struct sk_buff *skb_bad_txq;
int padded;
refcount_t refcnt;
spinlock_t busylock ____cacheline_aligned_in_smp;
};
net_sched: fix a refcount_t issue with noop_qdisc syzkaller reported a refcount_t warning [1] Issue here is that noop_qdisc refcnt was never really considered as a true refcount, since qdisc_destroy() found TCQ_F_BUILTIN set : if (qdisc->flags & TCQ_F_BUILTIN || !refcount_dec_and_test(&qdisc->refcnt))) return; Meaning that all atomic_inc() we did on noop_qdisc.refcnt were not really needed, but harmless until refcount_t came. To fix this problem, we simply need to not increment noop_qdisc.refcnt, since we never decrement it. [1] refcount_t: increment on 0; use-after-free. ------------[ cut here ]------------ WARNING: CPU: 0 PID: 21754 at lib/refcount.c:152 refcount_inc+0x47/0x50 lib/refcount.c:152 Kernel panic - not syncing: panic_on_warn set ... CPU: 0 PID: 21754 Comm: syz-executor7 Not tainted 4.13.0-rc6+ #20 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:16 [inline] dump_stack+0x194/0x257 lib/dump_stack.c:52 panic+0x1e4/0x417 kernel/panic.c:180 __warn+0x1c4/0x1d9 kernel/panic.c:541 report_bug+0x211/0x2d0 lib/bug.c:183 fixup_bug+0x40/0x90 arch/x86/kernel/traps.c:190 do_trap_no_signal arch/x86/kernel/traps.c:224 [inline] do_trap+0x260/0x390 arch/x86/kernel/traps.c:273 do_error_trap+0x120/0x390 arch/x86/kernel/traps.c:310 do_invalid_op+0x1b/0x20 arch/x86/kernel/traps.c:323 invalid_op+0x1e/0x30 arch/x86/entry/entry_64.S:846 RIP: 0010:refcount_inc+0x47/0x50 lib/refcount.c:152 RSP: 0018:ffff8801c43477a0 EFLAGS: 00010282 RAX: 000000000000002b RBX: ffffffff86093c14 RCX: 0000000000000000 RDX: 000000000000002b RSI: ffffffff8159314e RDI: ffffed0038868ee8 RBP: ffff8801c43477a8 R08: 0000000000000001 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000000 R12: ffffffff86093ac0 R13: 0000000000000001 R14: ffff8801d0f3bac0 R15: dffffc0000000000 attach_default_qdiscs net/sched/sch_generic.c:792 [inline] dev_activate+0x7d3/0xaa0 net/sched/sch_generic.c:833 __dev_open+0x227/0x330 net/core/dev.c:1380 __dev_change_flags+0x695/0x990 net/core/dev.c:6726 dev_change_flags+0x88/0x140 net/core/dev.c:6792 dev_ifsioc+0x5a6/0x930 net/core/dev_ioctl.c:256 dev_ioctl+0x2bc/0xf90 net/core/dev_ioctl.c:554 sock_do_ioctl+0x94/0xb0 net/socket.c:968 sock_ioctl+0x2c2/0x440 net/socket.c:1058 vfs_ioctl fs/ioctl.c:45 [inline] do_vfs_ioctl+0x1b1/0x1520 fs/ioctl.c:685 SYSC_ioctl fs/ioctl.c:700 [inline] SyS_ioctl+0x8f/0xc0 fs/ioctl.c:691 Fixes: 7b9364050246 ("net, sched: convert Qdisc.refcnt from atomic_t to refcount_t") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: Dmitry Vyukov <dvyukov@google.com> Cc: Reshetova, Elena <elena.reshetova@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-08-25 12:12:28 +08:00
static inline void qdisc_refcount_inc(struct Qdisc *qdisc)
{
if (qdisc->flags & TCQ_F_BUILTIN)
return;
refcount_inc(&qdisc->refcnt);
}
static inline bool qdisc_is_running(const struct Qdisc *qdisc)
{
return (raw_read_seqcount(&qdisc->running) & 1) ? true : false;
}
static inline bool qdisc_run_begin(struct Qdisc *qdisc)
{
if (qdisc_is_running(qdisc))
return false;
/* Variant of write_seqcount_begin() telling lockdep a trylock
* was attempted.
*/
raw_write_seqcount_begin(&qdisc->running);
seqcount_acquire(&qdisc->running.dep_map, 0, 1, _RET_IP_);
return true;
}
static inline void qdisc_run_end(struct Qdisc *qdisc)
{
write_seqcount_end(&qdisc->running);
}
qdisc: bulk dequeue support for qdiscs with TCQ_F_ONETXQUEUE Based on DaveM's recent API work on dev_hard_start_xmit(), that allows sending/processing an entire skb list. This patch implements qdisc bulk dequeue, by allowing multiple packets to be dequeued in dequeue_skb(). The optimization principle for this is two fold, (1) to amortize locking cost and (2) avoid expensive tailptr update for notifying HW. (1) Several packets are dequeued while holding the qdisc root_lock, amortizing locking cost over several packet. The dequeued SKB list is processed under the TXQ lock in dev_hard_start_xmit(), thus also amortizing the cost of the TXQ lock. (2) Further more, dev_hard_start_xmit() will utilize the skb->xmit_more API to delay HW tailptr update, which also reduces the cost per packet. One restriction of the new API is that every SKB must belong to the same TXQ. This patch takes the easy way out, by restricting bulk dequeue to qdisc's with the TCQ_F_ONETXQUEUE flag, that specifies the qdisc only have attached a single TXQ. Some detail about the flow; dev_hard_start_xmit() will process the skb list, and transmit packets individually towards the driver (see xmit_one()). In case the driver stops midway in the list, the remaining skb list is returned by dev_hard_start_xmit(). In sch_direct_xmit() this returned list is requeued by dev_requeue_skb(). To avoid overshooting the HW limits, which results in requeuing, the patch limits the amount of bytes dequeued, based on the drivers BQL limits. In-effect bulking will only happen for BQL enabled drivers. Small amounts for extra HoL blocking (2x MTU/0.24ms) were measured at 100Mbit/s, with bulking 8 packets, but the oscillating nature of the measurement indicate something, like sched latency might be causing this effect. More comparisons show, that this oscillation goes away occationally. Thus, we disregard this artifact completely and remove any "magic" bulking limit. For now, as a conservative approach, stop bulking when seeing TSO and segmented GSO packets. They already benefit from bulking on their own. A followup patch add this, to allow easier bisect-ability for finding regressions. Jointed work with Hannes, Daniel and Florian. Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com> Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-10-02 04:35:59 +08:00
static inline bool qdisc_may_bulk(const struct Qdisc *qdisc)
{
return qdisc->flags & TCQ_F_ONETXQUEUE;
}
static inline int qdisc_avail_bulklimit(const struct netdev_queue *txq)
{
#ifdef CONFIG_BQL
/* Non-BQL migrated drivers will return 0, too. */
return dql_avail(&txq->dql);
#else
return 0;
#endif
}
struct Qdisc_class_ops {
/* Child qdisc manipulation */
struct netdev_queue * (*select_queue)(struct Qdisc *, struct tcmsg *);
int (*graft)(struct Qdisc *, unsigned long cl,
struct Qdisc *, struct Qdisc **);
struct Qdisc * (*leaf)(struct Qdisc *, unsigned long cl);
void (*qlen_notify)(struct Qdisc *, unsigned long);
/* Class manipulation routines */
unsigned long (*find)(struct Qdisc *, u32 classid);
int (*change)(struct Qdisc *, u32, u32,
struct nlattr **, unsigned long *);
int (*delete)(struct Qdisc *, unsigned long);
void (*walk)(struct Qdisc *, struct qdisc_walker * arg);
/* Filter manipulation */
struct tcf_block * (*tcf_block)(struct Qdisc *, unsigned long);
unsigned long (*bind_tcf)(struct Qdisc *, unsigned long,
u32 classid);
void (*unbind_tcf)(struct Qdisc *, unsigned long);
/* rtnetlink specific */
int (*dump)(struct Qdisc *, unsigned long,
struct sk_buff *skb, struct tcmsg*);
int (*dump_stats)(struct Qdisc *, unsigned long,
struct gnet_dump *);
};
struct Qdisc_ops {
struct Qdisc_ops *next;
const struct Qdisc_class_ops *cl_ops;
char id[IFNAMSIZ];
int priv_size;
int (*enqueue)(struct sk_buff *skb,
struct Qdisc *sch,
struct sk_buff **to_free);
struct sk_buff * (*dequeue)(struct Qdisc *);
struct sk_buff * (*peek)(struct Qdisc *);
int (*init)(struct Qdisc *, struct nlattr *arg);
void (*reset)(struct Qdisc *);
void (*destroy)(struct Qdisc *);
int (*change)(struct Qdisc *, struct nlattr *arg);
void (*attach)(struct Qdisc *);
int (*dump)(struct Qdisc *, struct sk_buff *);
int (*dump_stats)(struct Qdisc *, struct gnet_dump *);
struct module *owner;
};
struct tcf_result {
union {
struct {
unsigned long class;
u32 classid;
};
const struct tcf_proto *goto_tp;
};
};
struct tcf_proto_ops {
struct list_head head;
char kind[IFNAMSIZ];
int (*classify)(struct sk_buff *,
const struct tcf_proto *,
struct tcf_result *);
int (*init)(struct tcf_proto*);
void (*destroy)(struct tcf_proto*);
void* (*get)(struct tcf_proto*, u32 handle);
int (*change)(struct net *net, struct sk_buff *,
struct tcf_proto*, unsigned long,
u32 handle, struct nlattr **,
void **, bool);
int (*delete)(struct tcf_proto*, void *, bool*);
void (*walk)(struct tcf_proto*, struct tcf_walker *arg);
void (*bind_class)(void *, u32, unsigned long);
/* rtnetlink specific */
int (*dump)(struct net*, struct tcf_proto*, void *,
struct sk_buff *skb, struct tcmsg*);
struct module *owner;
};
struct tcf_proto {
/* Fast access part */
struct tcf_proto __rcu *next;
void __rcu *root;
int (*classify)(struct sk_buff *,
const struct tcf_proto *,
struct tcf_result *);
__be16 protocol;
/* All the rest */
u32 prio;
u32 classid;
struct Qdisc *q;
void *data;
const struct tcf_proto_ops *ops;
struct tcf_chain *chain;
struct rcu_head rcu;
};
struct qdisc_skb_cb {
unsigned int pkt_len;
u16 slave_dev_queue_mapping;
u16 tc_classid;
#define QDISC_CB_PRIV_LEN 20
unsigned char data[QDISC_CB_PRIV_LEN];
};
struct tcf_chain {
struct tcf_proto __rcu *filter_chain;
struct tcf_proto __rcu **p_filter_chain;
struct list_head list;
struct tcf_block *block;
u32 index; /* chain index */
unsigned int refcnt;
};
struct tcf_block {
struct list_head chain_list;
struct net *net;
struct Qdisc *q;
};
static inline void qdisc_cb_private_validate(const struct sk_buff *skb, int sz)
{
struct qdisc_skb_cb *qcb;
bonding: Fix corrupted queue_mapping In the transmit path of the bonding driver, skb->cb is used to stash the skb->queue_mapping so that the bonding device can set its own queue mapping. This value becomes corrupted since the skb->cb is also used in __dev_xmit_skb. When transmitting through bonding driver, bond_select_queue is called from dev_queue_xmit. In bond_select_queue the original skb->queue_mapping is copied into skb->cb (via bond_queue_mapping) and skb->queue_mapping is overwritten with the bond driver queue. Subsequently in dev_queue_xmit, __dev_xmit_skb is called which writes the packet length into skb->cb, thereby overwriting the stashed queue mappping. In bond_dev_queue_xmit (called from hard_start_xmit), the queue mapping for the skb is set to the stashed value which is now the skb length and hence is an invalid queue for the slave device. If we want to save skb->queue_mapping into skb->cb[], best place is to add a field in struct qdisc_skb_cb, to make sure it wont conflict with other layers (eg : Qdiscc, Infiniband...) This patchs also makes sure (struct qdisc_skb_cb)->data is aligned on 8 bytes : netem qdisc for example assumes it can store an u64 in it, without misalignment penalty. Note : we only have 20 bytes left in (struct qdisc_skb_cb)->data[]. The largest user is CHOKe and it fills it. Based on a previous patch from Tom Herbert. Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: Tom Herbert <therbert@google.com> Cc: John Fastabend <john.r.fastabend@intel.com> Cc: Roland Dreier <roland@kernel.org> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-12 14:03:51 +08:00
BUILD_BUG_ON(sizeof(skb->cb) < offsetof(struct qdisc_skb_cb, data) + sz);
BUILD_BUG_ON(sizeof(qcb->data) < sz);
}
static inline int qdisc_qlen(const struct Qdisc *q)
net: Avoid enqueuing skb for default qdiscs dev_queue_xmit enqueue's a skb and calls qdisc_run which dequeue's the skb and xmits it. In most cases, the skb that is enqueue'd is the same one that is dequeue'd (unless the queue gets stopped or multiple cpu's write to the same queue and ends in a race with qdisc_run). For default qdiscs, we can remove the redundant enqueue/dequeue and simply xmit the skb since the default qdisc is work-conserving. The patch uses a new flag - TCQ_F_CAN_BYPASS to identify the default fast queue. The controversial part of the patch is incrementing qlen when a skb is requeued - this is to avoid checks like the second line below: + } else if ((q->flags & TCQ_F_CAN_BYPASS) && !qdisc_qlen(q) && >> !q->gso_skb && + !test_and_set_bit(__QDISC_STATE_RUNNING, &q->state)) { Results of a 2 hour testing for multiple netperf sessions (1, 2, 4, 8, 12 sessions on a 4 cpu system-X). The BW numbers are aggregate Mb/s across iterations tested with this version on System-X boxes with Chelsio 10gbps cards: ---------------------------------- Size | ORG BW NEW BW | ---------------------------------- 128K | 156964 159381 | 256K | 158650 162042 | ---------------------------------- Changes from ver1: 1. Move sch_direct_xmit declaration from sch_generic.h to pkt_sched.h 2. Update qdisc basic statistics for direct xmit path. 3. Set qlen to zero in qdisc_reset. 4. Changed some function names to more meaningful ones. Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-06 09:44:21 +08:00
{
return q->q.qlen;
}
static inline struct qdisc_skb_cb *qdisc_skb_cb(const struct sk_buff *skb)
{
return (struct qdisc_skb_cb *)skb->cb;
}
static inline spinlock_t *qdisc_lock(struct Qdisc *qdisc)
{
return &qdisc->q.lock;
}
static inline struct Qdisc *qdisc_root(const struct Qdisc *qdisc)
{
struct Qdisc *q = rcu_dereference_rtnl(qdisc->dev_queue->qdisc);
return q;
}
static inline struct Qdisc *qdisc_root_sleeping(const struct Qdisc *qdisc)
{
return qdisc->dev_queue->qdisc_sleeping;
}
/* The qdisc root lock is a mechanism by which to top level
* of a qdisc tree can be locked from any qdisc node in the
* forest. This allows changing the configuration of some
* aspect of the qdisc tree while blocking out asynchronous
* qdisc access in the packet processing paths.
*
* It is only legal to do this when the root will not change
* on us. Otherwise we'll potentially lock the wrong qdisc
* root. This is enforced by holding the RTNL semaphore, which
* all users of this lock accessor must do.
*/
static inline spinlock_t *qdisc_root_lock(const struct Qdisc *qdisc)
{
struct Qdisc *root = qdisc_root(qdisc);
ASSERT_RTNL();
return qdisc_lock(root);
}
static inline spinlock_t *qdisc_root_sleeping_lock(const struct Qdisc *qdisc)
{
struct Qdisc *root = qdisc_root_sleeping(qdisc);
ASSERT_RTNL();
return qdisc_lock(root);
}
static inline seqcount_t *qdisc_root_sleeping_running(const struct Qdisc *qdisc)
{
struct Qdisc *root = qdisc_root_sleeping(qdisc);
ASSERT_RTNL();
return &root->running;
}
static inline struct net_device *qdisc_dev(const struct Qdisc *qdisc)
{
return qdisc->dev_queue->dev;
}
static inline void sch_tree_lock(const struct Qdisc *q)
{
spin_lock_bh(qdisc_root_sleeping_lock(q));
}
static inline void sch_tree_unlock(const struct Qdisc *q)
{
spin_unlock_bh(qdisc_root_sleeping_lock(q));
}
extern struct Qdisc noop_qdisc;
extern struct Qdisc_ops noop_qdisc_ops;
extern struct Qdisc_ops pfifo_fast_ops;
extern struct Qdisc_ops mq_qdisc_ops;
extern struct Qdisc_ops noqueue_qdisc_ops;
extern const struct Qdisc_ops *default_qdisc_ops;
static inline const struct Qdisc_ops *
get_default_qdisc_ops(const struct net_device *dev, int ntx)
{
return ntx < dev->real_num_tx_queues ?
default_qdisc_ops : &pfifo_fast_ops;
}
struct Qdisc_class_common {
u32 classid;
struct hlist_node hnode;
};
struct Qdisc_class_hash {
struct hlist_head *hash;
unsigned int hashsize;
unsigned int hashmask;
unsigned int hashelems;
};
static inline unsigned int qdisc_class_hash(u32 id, u32 mask)
{
id ^= id >> 8;
id ^= id >> 4;
return id & mask;
}
static inline struct Qdisc_class_common *
qdisc_class_find(const struct Qdisc_class_hash *hash, u32 id)
{
struct Qdisc_class_common *cl;
unsigned int h;
if (!id)
return NULL;
h = qdisc_class_hash(id, hash->hashmask);
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 09:06:00 +08:00
hlist_for_each_entry(cl, &hash->hash[h], hnode) {
if (cl->classid == id)
return cl;
}
return NULL;
}
int qdisc_class_hash_init(struct Qdisc_class_hash *);
void qdisc_class_hash_insert(struct Qdisc_class_hash *,
struct Qdisc_class_common *);
void qdisc_class_hash_remove(struct Qdisc_class_hash *,
struct Qdisc_class_common *);
void qdisc_class_hash_grow(struct Qdisc *, struct Qdisc_class_hash *);
void qdisc_class_hash_destroy(struct Qdisc_class_hash *);
void dev_init_scheduler(struct net_device *dev);
void dev_shutdown(struct net_device *dev);
void dev_activate(struct net_device *dev);
void dev_deactivate(struct net_device *dev);
void dev_deactivate_many(struct list_head *head);
struct Qdisc *dev_graft_qdisc(struct netdev_queue *dev_queue,
struct Qdisc *qdisc);
void qdisc_reset(struct Qdisc *qdisc);
void qdisc_destroy(struct Qdisc *qdisc);
void qdisc_tree_reduce_backlog(struct Qdisc *qdisc, unsigned int n,
unsigned int len);
struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
const struct Qdisc_ops *ops);
struct Qdisc *qdisc_create_dflt(struct netdev_queue *dev_queue,
const struct Qdisc_ops *ops, u32 parentid);
void __qdisc_calculate_pkt_len(struct sk_buff *skb,
const struct qdisc_size_table *stab);
2015-09-16 14:05:43 +08:00
int skb_do_redirect(struct sk_buff *);
static inline void skb_reset_tc(struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
skb->tc_redirected = 0;
#endif
}
static inline bool skb_at_tc_ingress(const struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
return skb->tc_at_ingress;
#else
return false;
#endif
}
static inline bool skb_skip_tc_classify(struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ACT
if (skb->tc_skip_classify) {
skb->tc_skip_classify = 0;
return true;
}
#endif
return false;
}
/* Reset all TX qdiscs greater then index of a device. */
static inline void qdisc_reset_all_tx_gt(struct net_device *dev, unsigned int i)
{
struct Qdisc *qdisc;
for (; i < dev->num_tx_queues; i++) {
qdisc = rtnl_dereference(netdev_get_tx_queue(dev, i)->qdisc);
if (qdisc) {
spin_lock_bh(qdisc_lock(qdisc));
qdisc_reset(qdisc);
spin_unlock_bh(qdisc_lock(qdisc));
}
}
}
static inline void qdisc_reset_all_tx(struct net_device *dev)
{
qdisc_reset_all_tx_gt(dev, 0);
}
/* Are all TX queues of the device empty? */
static inline bool qdisc_all_tx_empty(const struct net_device *dev)
{
unsigned int i;
rcu_read_lock();
for (i = 0; i < dev->num_tx_queues; i++) {
struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
const struct Qdisc *q = rcu_dereference(txq->qdisc);
if (q->q.qlen) {
rcu_read_unlock();
return false;
}
}
rcu_read_unlock();
return true;
}
/* Are any of the TX qdiscs changing? */
static inline bool qdisc_tx_changing(const struct net_device *dev)
{
unsigned int i;
for (i = 0; i < dev->num_tx_queues; i++) {
struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
if (rcu_access_pointer(txq->qdisc) != txq->qdisc_sleeping)
return true;
}
return false;
}
/* Is the device using the noop qdisc on all queues? */
static inline bool qdisc_tx_is_noop(const struct net_device *dev)
{
unsigned int i;
for (i = 0; i < dev->num_tx_queues; i++) {
struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
if (rcu_access_pointer(txq->qdisc) != &noop_qdisc)
return false;
}
return true;
}
static inline unsigned int qdisc_pkt_len(const struct sk_buff *skb)
{
return qdisc_skb_cb(skb)->pkt_len;
}
/* additional qdisc xmit flags (NET_XMIT_MASK in linux/netdevice.h) */
enum net_xmit_qdisc_t {
__NET_XMIT_STOLEN = 0x00010000,
__NET_XMIT_BYPASS = 0x00020000,
};
#ifdef CONFIG_NET_CLS_ACT
#define net_xmit_drop_count(e) ((e) & __NET_XMIT_STOLEN ? 0 : 1)
#else
#define net_xmit_drop_count(e) (1)
#endif
static inline void qdisc_calculate_pkt_len(struct sk_buff *skb,
const struct Qdisc *sch)
{
#ifdef CONFIG_NET_SCHED
struct qdisc_size_table *stab = rcu_dereference_bh(sch->stab);
if (stab)
__qdisc_calculate_pkt_len(skb, stab);
#endif
}
static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
struct sk_buff **to_free)
{
qdisc_calculate_pkt_len(skb, sch);
return sch->enqueue(skb, sch, to_free);
}
static inline bool qdisc_is_percpu_stats(const struct Qdisc *q)
{
return q->flags & TCQ_F_CPUSTATS;
}
static inline void _bstats_update(struct gnet_stats_basic_packed *bstats,
__u64 bytes, __u32 packets)
{
bstats->bytes += bytes;
bstats->packets += packets;
}
static inline void bstats_update(struct gnet_stats_basic_packed *bstats,
const struct sk_buff *skb)
{
_bstats_update(bstats,
qdisc_pkt_len(skb),
skb_is_gso(skb) ? skb_shinfo(skb)->gso_segs : 1);
}
static inline void _bstats_cpu_update(struct gnet_stats_basic_cpu *bstats,
__u64 bytes, __u32 packets)
{
u64_stats_update_begin(&bstats->syncp);
_bstats_update(&bstats->bstats, bytes, packets);
u64_stats_update_end(&bstats->syncp);
}
static inline void bstats_cpu_update(struct gnet_stats_basic_cpu *bstats,
const struct sk_buff *skb)
{
u64_stats_update_begin(&bstats->syncp);
bstats_update(&bstats->bstats, skb);
u64_stats_update_end(&bstats->syncp);
}
static inline void qdisc_bstats_cpu_update(struct Qdisc *sch,
const struct sk_buff *skb)
{
bstats_cpu_update(this_cpu_ptr(sch->cpu_bstats), skb);
}
static inline void qdisc_bstats_update(struct Qdisc *sch,
const struct sk_buff *skb)
net: Avoid enqueuing skb for default qdiscs dev_queue_xmit enqueue's a skb and calls qdisc_run which dequeue's the skb and xmits it. In most cases, the skb that is enqueue'd is the same one that is dequeue'd (unless the queue gets stopped or multiple cpu's write to the same queue and ends in a race with qdisc_run). For default qdiscs, we can remove the redundant enqueue/dequeue and simply xmit the skb since the default qdisc is work-conserving. The patch uses a new flag - TCQ_F_CAN_BYPASS to identify the default fast queue. The controversial part of the patch is incrementing qlen when a skb is requeued - this is to avoid checks like the second line below: + } else if ((q->flags & TCQ_F_CAN_BYPASS) && !qdisc_qlen(q) && >> !q->gso_skb && + !test_and_set_bit(__QDISC_STATE_RUNNING, &q->state)) { Results of a 2 hour testing for multiple netperf sessions (1, 2, 4, 8, 12 sessions on a 4 cpu system-X). The BW numbers are aggregate Mb/s across iterations tested with this version on System-X boxes with Chelsio 10gbps cards: ---------------------------------- Size | ORG BW NEW BW | ---------------------------------- 128K | 156964 159381 | 256K | 158650 162042 | ---------------------------------- Changes from ver1: 1. Move sch_direct_xmit declaration from sch_generic.h to pkt_sched.h 2. Update qdisc basic statistics for direct xmit path. 3. Set qlen to zero in qdisc_reset. 4. Changed some function names to more meaningful ones. Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-06 09:44:21 +08:00
{
bstats_update(&sch->bstats, skb);
net: Avoid enqueuing skb for default qdiscs dev_queue_xmit enqueue's a skb and calls qdisc_run which dequeue's the skb and xmits it. In most cases, the skb that is enqueue'd is the same one that is dequeue'd (unless the queue gets stopped or multiple cpu's write to the same queue and ends in a race with qdisc_run). For default qdiscs, we can remove the redundant enqueue/dequeue and simply xmit the skb since the default qdisc is work-conserving. The patch uses a new flag - TCQ_F_CAN_BYPASS to identify the default fast queue. The controversial part of the patch is incrementing qlen when a skb is requeued - this is to avoid checks like the second line below: + } else if ((q->flags & TCQ_F_CAN_BYPASS) && !qdisc_qlen(q) && >> !q->gso_skb && + !test_and_set_bit(__QDISC_STATE_RUNNING, &q->state)) { Results of a 2 hour testing for multiple netperf sessions (1, 2, 4, 8, 12 sessions on a 4 cpu system-X). The BW numbers are aggregate Mb/s across iterations tested with this version on System-X boxes with Chelsio 10gbps cards: ---------------------------------- Size | ORG BW NEW BW | ---------------------------------- 128K | 156964 159381 | 256K | 158650 162042 | ---------------------------------- Changes from ver1: 1. Move sch_direct_xmit declaration from sch_generic.h to pkt_sched.h 2. Update qdisc basic statistics for direct xmit path. 3. Set qlen to zero in qdisc_reset. 4. Changed some function names to more meaningful ones. Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-06 09:44:21 +08:00
}
static inline void qdisc_qstats_backlog_dec(struct Qdisc *sch,
const struct sk_buff *skb)
{
sch->qstats.backlog -= qdisc_pkt_len(skb);
}
static inline void qdisc_qstats_backlog_inc(struct Qdisc *sch,
const struct sk_buff *skb)
{
sch->qstats.backlog += qdisc_pkt_len(skb);
}
static inline void __qdisc_qstats_drop(struct Qdisc *sch, int count)
{
sch->qstats.drops += count;
}
static inline void qstats_drop_inc(struct gnet_stats_queue *qstats)
{
qstats->drops++;
}
static inline void qstats_overlimit_inc(struct gnet_stats_queue *qstats)
{
qstats->overlimits++;
}
static inline void qdisc_qstats_drop(struct Qdisc *sch)
{
qstats_drop_inc(&sch->qstats);
}
static inline void qdisc_qstats_cpu_drop(struct Qdisc *sch)
{
this_cpu_inc(sch->cpu_qstats->drops);
}
static inline void qdisc_qstats_overlimit(struct Qdisc *sch)
{
sch->qstats.overlimits++;
}
static inline void qdisc_skb_head_init(struct qdisc_skb_head *qh)
{
qh->head = NULL;
qh->tail = NULL;
qh->qlen = 0;
}
static inline int __qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch,
struct qdisc_skb_head *qh)
{
struct sk_buff *last = qh->tail;
if (last) {
skb->next = NULL;
last->next = skb;
qh->tail = skb;
} else {
qh->tail = skb;
qh->head = skb;
}
qh->qlen++;
qdisc_qstats_backlog_inc(sch, skb);
return NET_XMIT_SUCCESS;
}
static inline int qdisc_enqueue_tail(struct sk_buff *skb, struct Qdisc *sch)
{
return __qdisc_enqueue_tail(skb, sch, &sch->q);
}
static inline struct sk_buff *__qdisc_dequeue_head(struct qdisc_skb_head *qh)
{
struct sk_buff *skb = qh->head;
if (likely(skb != NULL)) {
qh->head = skb->next;
qh->qlen--;
if (qh->head == NULL)
qh->tail = NULL;
skb->next = NULL;
}
return skb;
}
static inline struct sk_buff *qdisc_dequeue_head(struct Qdisc *sch)
{
struct sk_buff *skb = __qdisc_dequeue_head(&sch->q);
if (likely(skb != NULL)) {
qdisc_qstats_backlog_dec(sch, skb);
qdisc_bstats_update(sch, skb);
}
return skb;
}
/* Instead of calling kfree_skb() while root qdisc lock is held,
* queue the skb for future freeing at end of __dev_xmit_skb()
*/
static inline void __qdisc_drop(struct sk_buff *skb, struct sk_buff **to_free)
{
skb->next = *to_free;
*to_free = skb;
}
static inline unsigned int __qdisc_queue_drop_head(struct Qdisc *sch,
struct qdisc_skb_head *qh,
struct sk_buff **to_free)
{
struct sk_buff *skb = __qdisc_dequeue_head(qh);
if (likely(skb != NULL)) {
unsigned int len = qdisc_pkt_len(skb);
qdisc_qstats_backlog_dec(sch, skb);
__qdisc_drop(skb, to_free);
return len;
}
return 0;
}
static inline unsigned int qdisc_queue_drop_head(struct Qdisc *sch,
struct sk_buff **to_free)
{
return __qdisc_queue_drop_head(sch, &sch->q, to_free);
}
static inline struct sk_buff *qdisc_peek_head(struct Qdisc *sch)
{
const struct qdisc_skb_head *qh = &sch->q;
return qh->head;
}
/* generic pseudo peek method for non-work-conserving qdisc */
static inline struct sk_buff *qdisc_peek_dequeued(struct Qdisc *sch)
{
/* we can reuse ->gso_skb because peek isn't called for root qdiscs */
if (!sch->gso_skb) {
sch->gso_skb = sch->dequeue(sch);
if (sch->gso_skb) {
/* it's still part of the queue */
qdisc_qstats_backlog_inc(sch, sch->gso_skb);
sch->q.qlen++;
}
}
return sch->gso_skb;
}
/* use instead of qdisc->dequeue() for all qdiscs queried with ->peek() */
static inline struct sk_buff *qdisc_dequeue_peeked(struct Qdisc *sch)
{
struct sk_buff *skb = sch->gso_skb;
if (skb) {
sch->gso_skb = NULL;
qdisc_qstats_backlog_dec(sch, skb);
sch->q.qlen--;
} else {
skb = sch->dequeue(sch);
}
return skb;
}
static inline void __qdisc_reset_queue(struct qdisc_skb_head *qh)
{
/*
* We do not know the backlog in bytes of this list, it
* is up to the caller to correct it
*/
ASSERT_RTNL();
if (qh->qlen) {
rtnl_kfree_skbs(qh->head, qh->tail);
qh->head = NULL;
qh->tail = NULL;
qh->qlen = 0;
}
}
static inline void qdisc_reset_queue(struct Qdisc *sch)
{
__qdisc_reset_queue(&sch->q);
sch->qstats.backlog = 0;
}
static inline struct Qdisc *qdisc_replace(struct Qdisc *sch, struct Qdisc *new,
struct Qdisc **pold)
{
struct Qdisc *old;
sch_tree_lock(sch);
old = *pold;
*pold = new;
if (old != NULL) {
unsigned int qlen = old->q.qlen;
unsigned int backlog = old->qstats.backlog;
qdisc_reset(old);
qdisc_tree_reduce_backlog(old, qlen, backlog);
}
sch_tree_unlock(sch);
return old;
}
static inline void rtnl_qdisc_drop(struct sk_buff *skb, struct Qdisc *sch)
{
rtnl_kfree_skbs(skb, skb);
qdisc_qstats_drop(sch);
}
static inline int qdisc_drop(struct sk_buff *skb, struct Qdisc *sch,
struct sk_buff **to_free)
{
__qdisc_drop(skb, to_free);
qdisc_qstats_drop(sch);
return NET_XMIT_DROP;
}
/* Length to Time (L2T) lookup in a qdisc_rate_table, to determine how
long it will take to send a packet given its size.
*/
static inline u32 qdisc_l2t(struct qdisc_rate_table* rtab, unsigned int pktlen)
{
int slot = pktlen + rtab->rate.cell_align + rtab->rate.overhead;
if (slot < 0)
slot = 0;
slot >>= rtab->rate.cell_log;
if (slot > 255)
return rtab->data[255]*(slot >> 8) + rtab->data[slot & 0xFF];
return rtab->data[slot];
}
struct psched_ratecfg {
u64 rate_bytes_ps; /* bytes per second */
u32 mult;
u16 overhead;
u8 linklayer;
u8 shift;
};
static inline u64 psched_l2t_ns(const struct psched_ratecfg *r,
unsigned int len)
{
len += r->overhead;
if (unlikely(r->linklayer == TC_LINKLAYER_ATM))
return ((u64)(DIV_ROUND_UP(len,48)*53) * r->mult) >> r->shift;
return ((u64)len * r->mult) >> r->shift;
}
void psched_ratecfg_precompute(struct psched_ratecfg *r,
const struct tc_ratespec *conf,
u64 rate64);
static inline void psched_ratecfg_getrate(struct tc_ratespec *res,
const struct psched_ratecfg *r)
{
memset(res, 0, sizeof(*res));
/* legacy struct tc_ratespec has a 32bit @rate field
* Qdisc using 64bit rate should add new attributes
* in order to maintain compatibility.
*/
res->rate = min_t(u64, r->rate_bytes_ps, ~0U);
res->overhead = r->overhead;
res->linklayer = (r->linklayer & TC_LINKLAYER_MASK);
}
#endif