Skip to contentSkip to conversation

BPF packet scheduler

PATCH Superseded Geliang Tang 4 participants 18 messages Apr 20, 2022 – Apr 21, 2022

Subsystems (6)

v12 of 18 Superseded Latest is v18 →

lore
Share & tools
lore
#0 lore raw

v12 cover BPF packet scheduler

v12:
- add call_me_again flag.

base-commit: export/20220419T055400

v11:
- add retrans argment for get_subflow()

base-commit: export/20220408T100826

v10:
- patch 5: keep msk->last_snd setting in get_subflow().
- patch 6: add bpf_mptcp_sched_btf_struct_access().
- patch 8: use MIN() in sys/param.h, instead of defining a new one.
- update commit logs.

base-commit: export/20220406T054706

v9:
- patch 2: add the missing mptcp_sched_init() invoking in mptcp_proto_init().
- patch 5: set last_snd after invoking get_subflow().
- patch 7: merge the squash-to patch.

v8:
- use global sched_list instead of pernet sched_list.
- drop synchronize_rcu() in mptcp_unregister_scheduler().
- update mptcp_init_sched and mptcp_release_sched as Mat and Florian suggested.
- fix the build break in patch 8.
- depends on: "add skc_to_mptcp_sock" v14.
- export/20220325T055307

v7:
- add bpf_try_module_get in mptcp_init_sched.
- add bpf_module_put in mptcp_release_sched.
- rename bpf_first to mptcp_bpf_first.
- update commit logs.

v6:
- still use pernet sched_list, use current->nsproxy->net_ns in BPF context instead of using init_net.
- patch 1:
- use rcu_read_lock instead of spin_lock in mptcp_sched_find as Florian suggested.
- drop synchronize_rcu in sched_exit_net as Florian suggested.
- keep synchronize_rcu in mptcp_unregister_scheduler, otherwise, got a workqueue lockup in my test.
- update Makefile as Mat suggested.
- patch 2:
- add mptcp_sched_data_init to register default sched, instead of registering it in init_net.
- patch 5:
- move mptcp_sched_get_subflow to protocol.h as Mat suggested.
- patch 6:
- use current->nsproxy->net_ns instead of init_net.
- patch 8:
- add send_data to send more data, instead of send_byte.

v5:
- patch 1: define per-namespace sched_list (but only used init_net namespace. It is difficult to get 'net' in bpf_mptcp_sched_reg and bpf_mptcp_sched_unreg. I need some suggestions here.)
- patch 2: skip mptcp_sched_default in mptcp_unregister_scheduler.
- patch 8: add tests into mptcp.c, instead of bpf_tcp_ca.c.

v4:
- set msk->sched to &mptcp_sched_default when the sched argument is NULL
in mptcp_init_sched().

v3:
- add mptcp_release_sched helper in patch 4.
- rename mptcp_set_sched to mptcp_init_sched in patch 4.
- add mptcp_sched_first_release in patch 7.
- do some cleanups.

v2:
- split into more small patches.
- change all parameters of mptcp_sched_ops from sk to msk:

       void (*init)(struct mptcp_sock *msk);
       void (*release)(struct mptcp_sock *msk);
       struct sock *   (*get_subflow)(struct mptcp_sock *msk);

- add tests in bpf_tcp_ca.c, instead of adding a new one.

v1:
- Addressed to the commends in the RFC version.

Geliang Tang (10):
mptcp: add struct mptcp_sched_ops
mptcp: register default scheduler
mptcp: add a new sysctl scheduler
mptcp: add sched in mptcp_sock
mptcp: add get_subflow wrapper
mptcp: add bpf_mptcp_sched_ops
mptcp: add last_snd write access
mptcp: add call_me_again flag
selftests: bpf: add bpf_first scheduler
selftests: bpf: add bpf_first test

 Documentation/networking/mptcp-sysctl.rst     |   8 +
 include/net/mptcp.h                           |  13 ++
 kernel/bpf/bpf_struct_ops_types.h             |   4 +
 net/mptcp/Makefile                            |   2 +-
 net/mptcp/bpf.c                               | 148 ++++++++++++++++++
 net/mptcp/ctrl.c                              |  14 ++
 net/mptcp/protocol.c                          |  24 ++-
 net/mptcp/protocol.h                          |  18 +++
 net/mptcp/sched.c                             | 105 +++++++++++++
 .../testing/selftests/bpf/bpf_mptcp_helpers.h |  15 ++
 .../testing/selftests/bpf/prog_tests/mptcp.c  | 113 +++++++++++++
 .../selftests/bpf/progs/mptcp_bpf_first.c     |  34 ++++
 12 files changed, 493 insertions(+), 5 deletions(-)
 create mode 100644 net/mptcp/sched.c
 create mode 100644 tools/testing/selftests/bpf/progs/mptcp_bpf_first.c
Signature · 1 line
2.34.1
#1 lore raw

PATCH v12 1/10 mptcp: add struct mptcp_sched_ops

This patch defines struct mptcp_sched_ops, which has three struct members, name, owner and list, and three function pointers, init, release and get_subflow.

Add the scheduler registering, unregistering and finding functions to add, delete and find a packet scheduler on the global list mptcp_sched_list.

---
 include/net/mptcp.h                           | 13 +++++
 net/mptcp/Makefile                            |  2 +-
 net/mptcp/protocol.h                          |  3 +
 net/mptcp/sched.c                             | 56 +++++++++++++++++++
 .../testing/selftests/bpf/bpf_mptcp_helpers.h | 12 ++++
 5 files changed, 85 insertions(+), 1 deletion(-)
 create mode 100644 net/mptcp/sched.c
include/net/mptcp.h +13 -0
index 877077b53200..78a5cb44fccd 100644
@@ -95,6 +95,19 @@ struct mptcp_out_options {
 #endif
 };
 
+#define MPTCP_SCHED_NAME_MAX 16
+
+struct mptcp_sched_ops {
+	struct sock *(*get_subflow)(struct mptcp_sock *msk, bool retrans);
+
+	char			name[MPTCP_SCHED_NAME_MAX];
+	struct module		*owner;
+	struct list_head	list;
+
+	void (*init)(struct mptcp_sock *msk);
+	void (*release)(struct mptcp_sock *msk);
+} ____cacheline_aligned_in_smp;
+
 #ifdef CONFIG_MPTCP
 extern struct request_sock_ops mptcp_subflow_request_sock_ops;
 
net/mptcp/Makefile +1 -1
index 4004347db47e..702b86e8ecb0 100644
@@ -2,7 +2,7 @@
 obj-$(CONFIG_MPTCP) += mptcp.o
 
 mptcp-y := protocol.o subflow.o options.o token.o crypto.o ctrl.o pm.o diag.o \
-	   mib.o pm_netlink.o sockopt.o pm_userspace.o
+	   mib.o pm_netlink.o sockopt.o pm_userspace.o sched.o
 
 obj-$(CONFIG_SYN_COOKIES) += syncookies.o
 obj-$(CONFIG_INET_MPTCP_DIAG) += mptcp_diag.o
net/mptcp/protocol.h +3 -0
index f542aeaa5b09..18f8739bbf9c 100644
@@ -621,6 +621,9 @@ int mptcp_subflow_create_socket(struct sock *sk, struct socket **new_sock);
 void mptcp_info2sockaddr(const struct mptcp_addr_info *info,
 			 struct sockaddr_storage *addr,
 			 unsigned short family);
+struct mptcp_sched_ops *mptcp_sched_find(const char *name);
+int mptcp_register_scheduler(struct mptcp_sched_ops *sched);
+void mptcp_unregister_scheduler(struct mptcp_sched_ops *sched);
 
 static inline bool __mptcp_subflow_active(struct mptcp_subflow_context *subflow)
 {
net/mptcp/sched.c new file +56 -0

56 lines not on this page · raw message

tools/testing/selftests/bpf/bpf_mptcp_helpers.h +12 -0
index b5a43b108982..8d4e4f08d327 100644
@@ -6,6 +6,18 @@
 
 #include "bpf_tcp_helpers.h"
 
+#define MPTCP_SCHED_NAME_MAX 16
+
+struct mptcp_sched_ops {
+	char name[MPTCP_SCHED_NAME_MAX];
+
+	void (*init)(struct mptcp_sock *msk);
+	void (*release)(struct mptcp_sock *msk);
+
+	struct sock *(*get_subflow)(struct mptcp_sock *msk, bool retrans);
+	void *owner;
+};
+
 struct mptcp_sock {
 	struct inet_connection_sock	sk;
 
Signature · 1 line
2.34.1
#2 lore raw

Re: [PATCH mptcp-next v12 01/10] mptcp: add struct mptcp_sched_ops

Hi Geliang,

Quoted 6 lines · Geliang Tang

This patch defines struct mptcp_sched_ops, which has three struct members, name, owner and list, and three function pointers, init, release and get_subflow.

Add the scheduler registering, unregistering and finding functions to add, delete and find a packet scheduler on the global list mptcp_sched_list.

Thank you for looking at that!

Quoted 16 lines
---
 include/net/mptcp.h                           | 13 +++++
 net/mptcp/Makefile                            |  2 +-
 net/mptcp/protocol.h                          |  3 +
 net/mptcp/sched.c                             | 56 +++++++++++++++++++
 .../testing/selftests/bpf/bpf_mptcp_helpers.h | 12 ++++
 5 files changed, 85 insertions(+), 1 deletion(-)
 create mode 100644 net/mptcp/sched.c
include/net/mptcp.h +0 -0
index 877077b53200..78a5cb44fccd 100644

include/net/mptcp.h struct mptcp_out_options { · line 101

Quoted 7 lines
include/net/mptcp.h +4 -0
@@ -95,6 +95,19 @@ struct mptcp_out_options {
 #endif
 };
 
+#define MPTCP_SCHED_NAME_MAX 16
+
+struct mptcp_sched_ops {
+	struct sock *(*get_subflow)(struct mptcp_sock *msk, bool retrans);

May I suggest using "reinject" name instead of "retrans". To avoid confusions, the RFC uses "reinjection" word when a packet was previously sent to a different subflow and the goal is to reinject it to another one. It is different from retransmitting the same packet over the same
subflow: TCP level retransmission with the same TCP seq num.

(same below + in other new functions, e.g. mptcp_get_subflow_default(), mptcp_sched_get_subflow(), etc.)

Cheers,
Matt

Signature · 2 lines
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net
#3 lore raw

PATCH v12 2/10 mptcp: register default scheduler

In reply to Geliang Tang

This patch defines the default packet scheduler mptcp_sched_default, register it in mptcp_sched_init(), which is invoked in mptcp_proto_init(). Skip deleting this default scheduler in mptcp_unregister_scheduler().

---
 net/mptcp/protocol.c |  9 +++++++++
 net/mptcp/protocol.h |  2 ++
 net/mptcp/sched.c    | 14 ++++++++++++++
 3 files changed, 25 insertions(+)
net/mptcp/protocol.c +9 -0
index 6d59bfdd6bbd..8679289a8394 100644
@@ -2251,6 +2251,14 @@ static struct sock *mptcp_subflow_get_retrans(struct mptcp_sock *msk)
 	return min_stale_count > 1 ? backup : NULL;
 }
 
+struct sock *mptcp_get_subflow_default(struct mptcp_sock *msk, bool retrans)
+{
+	if (retrans)
+		return mptcp_subflow_get_retrans(msk);
+
+	return mptcp_subflow_get_send(msk);
+}
+
 static void mptcp_dispose_initial_subflow(struct mptcp_sock *msk)
 {
 	if (msk->subflow) {
@@ -3805,6 +3813,7 @@ void __init mptcp_proto_init(void)
 
 	mptcp_subflow_init();
 	mptcp_pm_init();
+	mptcp_sched_init();
 	mptcp_token_init();
 
 	if (proto_register(&mptcp_prot, MPTCP_USE_SLAB) != 0)
net/mptcp/protocol.h +2 -0
index 18f8739bbf9c..9fd4c689e423 100644
@@ -624,6 +624,8 @@ void mptcp_info2sockaddr(const struct mptcp_addr_info *info,
 struct mptcp_sched_ops *mptcp_sched_find(const char *name);
 int mptcp_register_scheduler(struct mptcp_sched_ops *sched);
 void mptcp_unregister_scheduler(struct mptcp_sched_ops *sched);
+struct sock *mptcp_get_subflow_default(struct mptcp_sock *msk, bool retrans);
+void mptcp_sched_init(void);
 
 static inline bool __mptcp_subflow_active(struct mptcp_subflow_context *subflow)
 {
net/mptcp/sched.c +14 -0
index c5d3bbafba71..bd0beff8cac8 100644
@@ -13,6 +13,12 @@
 #include <linux/spinlock.h>
 #include "protocol.h"
 
+static struct mptcp_sched_ops mptcp_sched_default = {
+	.get_subflow    = mptcp_get_subflow_default,
+	.name           = "default",
+	.owner          = THIS_MODULE,
+};
+
 static DEFINE_SPINLOCK(mptcp_sched_list_lock);
 static LIST_HEAD(mptcp_sched_list);
 
@@ -50,7 +56,15 @@ int mptcp_register_scheduler(struct mptcp_sched_ops *sched)
 
 void mptcp_unregister_scheduler(struct mptcp_sched_ops *sched)
 {
+	if (sched == &mptcp_sched_default)
+		return;
+
 	spin_lock(&mptcp_sched_list_lock);
 	list_del_rcu(&sched->list);
 	spin_unlock(&mptcp_sched_list_lock);
 }
+
+void mptcp_sched_init(void)
+{
+	mptcp_register_scheduler(&mptcp_sched_default);
+}
Signature · 1 line
2.34.1
#4 lore raw

PATCH v12 3/10 mptcp: add a new sysctl scheduler

In reply to Geliang Tang

This patch adds a new sysctl, named scheduler, to support for selection of different schedulers. Export mptcp_get_scheduler helper to get this sysctl.

---
 Documentation/networking/mptcp-sysctl.rst |  8 ++++++++
 net/mptcp/ctrl.c                          | 14 ++++++++++++++
 net/mptcp/protocol.h                      |  1 +
 3 files changed, 23 insertions(+)
Documentation/networking/mptcp-sysctl.rst +8 -0
index e263dfcc4b40..d9e69fdc7ea3 100644
@@ -75,3 +75,11 @@ stale_loss_cnt - INTEGER
 	This is a per-namespace sysctl.
 
 	Default: 4
+
+scheduler - STRING
+	Select the scheduler of your choice.
+
+	Support for selection of different schedulers. This is a per-namespace
+	sysctl.
+
+	Default: "default"
net/mptcp/ctrl.c +14 -0
index ae20b7d92e28..c46c22a84d23 100644
@@ -32,6 +32,7 @@ struct mptcp_pernet {
 	u8 checksum_enabled;
 	u8 allow_join_initial_addr_port;
 	u8 pm_type;
+	char scheduler[MPTCP_SCHED_NAME_MAX];
 };
 
 static struct mptcp_pernet *mptcp_get_pernet(const struct net *net)
@@ -69,6 +70,11 @@ int mptcp_get_pm_type(const struct net *net)
 	return mptcp_get_pernet(net)->pm_type;
 }
 
+const char *mptcp_get_scheduler(const struct net *net)
+{
+	return mptcp_get_pernet(net)->scheduler;
+}
+
 static void mptcp_pernet_set_defaults(struct mptcp_pernet *pernet)
 {
 	pernet->mptcp_enabled = 1;
@@ -77,6 +83,7 @@ static void mptcp_pernet_set_defaults(struct mptcp_pernet *pernet)
 	pernet->allow_join_initial_addr_port = 1;
 	pernet->stale_loss_cnt = 4;
 	pernet->pm_type = MPTCP_PM_TYPE_KERNEL;
+	strcpy(pernet->scheduler, "default");
 }
 
 #ifdef CONFIG_SYSCTL
@@ -128,6 +135,12 @@ static struct ctl_table mptcp_sysctl_table[] = {
 		.extra1       = SYSCTL_ZERO,
 		.extra2       = &mptcp_pm_type_max
 	},
+	{
+		.procname = "scheduler",
+		.maxlen	= MPTCP_SCHED_NAME_MAX,
+		.mode = 0644,
+		.proc_handler = proc_dostring,
+	},
 	{}
 };
 
@@ -149,6 +162,7 @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
 	table[3].data = &pernet->allow_join_initial_addr_port;
 	table[4].data = &pernet->stale_loss_cnt;
 	table[5].data = &pernet->pm_type;
+	table[6].data = &pernet->scheduler;
 
 	hdr = register_net_sysctl(net, MPTCP_SYSCTL_PATH, table);
 	if (!hdr)
net/mptcp/protocol.h +1 -0
index 9fd4c689e423..d05dd2e2d05c 100644
@@ -596,6 +596,7 @@ int mptcp_is_checksum_enabled(const struct net *net);
 int mptcp_allow_join_id0(const struct net *net);
 unsigned int mptcp_stale_loss_cnt(const struct net *net);
 int mptcp_get_pm_type(const struct net *net);
+const char *mptcp_get_scheduler(const struct net *net);
 void mptcp_subflow_fully_established(struct mptcp_subflow_context *subflow,
 				     struct mptcp_options_received *mp_opt);
 bool __mptcp_retransmit_pending_data(struct sock *sk);
Signature · 1 line
2.34.1
#5 lore raw

PATCH v12 4/10 mptcp: add sched in mptcp_sock

In reply to Geliang Tang

This patch adds a new struct member sched in struct mptcp_sock. And two helpers mptcp_init_sched() and mptcp_release_sched() to init and release it.

Init it with the sysctl scheduler in mptcp_init_sock(), copy the scheduler from the parent in mptcp_sk_clone(), and release it in __mptcp_destroy_sock().

---
 net/mptcp/protocol.c                          |  7 ++++
 net/mptcp/protocol.h                          |  4 +++
 net/mptcp/sched.c                             | 34 +++++++++++++++++++
 .../testing/selftests/bpf/bpf_mptcp_helpers.h |  1 +
 4 files changed, 46 insertions(+)
net/mptcp/protocol.c +7 -0
index 8679289a8394..3d50c35c0b33 100644
@@ -2665,6 +2665,11 @@ static int mptcp_init_sock(struct sock *sk)
 	if (ret)
 		return ret;
 
+	ret = mptcp_init_sched(mptcp_sk(sk),
+			       mptcp_sched_find(mptcp_get_scheduler(net)));
+	if (ret)
+		return ret;
+
 	/* fetch the ca name; do it outside __mptcp_init_sock(), so that clone will
 	 * propagate the correct value
 	 */
@@ -2824,6 +2829,7 @@ static void __mptcp_destroy_sock(struct sock *sk)
 	sk_stop_timer(sk, &sk->sk_timer);
 	mptcp_data_unlock(sk);
 	msk->pm.status = 0;
+	mptcp_release_sched(msk);
 
 	/* clears msk->subflow, allowing the following loop to close
 	 * even the initial subflow
@@ -3001,6 +3007,7 @@ struct sock *mptcp_sk_clone(const struct sock *sk,
 	msk->snd_una = msk->write_seq;
 	msk->wnd_end = msk->snd_nxt + req->rsk_rcv_wnd;
 	msk->setsockopt_seq = mptcp_sk(sk)->setsockopt_seq;
+	mptcp_init_sched(msk, mptcp_sk(sk)->sched);
 
 	if (mp_opt->suboptions & OPTIONS_MPTCP_MPC) {
 		msk->can_ack = true;
net/mptcp/protocol.h +4 -0
index d05dd2e2d05c..b06836d1a899 100644
@@ -298,6 +298,7 @@ struct mptcp_sock {
 	struct socket	*subflow; /* outgoing connect/listener/!mp_capable */
 	struct sock	*first;
 	struct mptcp_pm_data	pm;
+	struct mptcp_sched_ops	*sched;
 	struct {
 		u32	space;	/* bytes copied in last measurement window */
 		u32	copied; /* bytes copied in this measurement window */
@@ -627,6 +628,9 @@ int mptcp_register_scheduler(struct mptcp_sched_ops *sched);
 void mptcp_unregister_scheduler(struct mptcp_sched_ops *sched);
 struct sock *mptcp_get_subflow_default(struct mptcp_sock *msk, bool retrans);
 void mptcp_sched_init(void);
+int mptcp_init_sched(struct mptcp_sock *msk,
+		     struct mptcp_sched_ops *sched);
+void mptcp_release_sched(struct mptcp_sock *msk);
 
 static inline bool __mptcp_subflow_active(struct mptcp_subflow_context *subflow)
 {
net/mptcp/sched.c +34 -0
index bd0beff8cac8..8025dc51fbe9 100644
@@ -68,3 +68,37 @@ void mptcp_sched_init(void)
 {
 	mptcp_register_scheduler(&mptcp_sched_default);
 }
+
+int mptcp_init_sched(struct mptcp_sock *msk,
+		     struct mptcp_sched_ops *sched)
+{
+	struct mptcp_sched_ops *sched_init = &mptcp_sched_default;
+
+	if (sched)
+		sched_init = sched;
+
+	if (!bpf_try_module_get(sched_init, sched_init->owner))
+		return -EBUSY;
+
+	msk->sched = sched_init;
+	if (msk->sched->init)
+		msk->sched->init(msk);
+
+	pr_debug("sched=%s", msk->sched->name);
+
+	return 0;
+}
+
+void mptcp_release_sched(struct mptcp_sock *msk)
+{
+	struct mptcp_sched_ops *sched = msk->sched;
+
+	if (!sched)
+		return;
+
+	msk->sched = NULL;
+	if (sched->release)
+		sched->release(msk);
+
+	bpf_module_put(sched, sched->owner);
+}
tools/testing/selftests/bpf/bpf_mptcp_helpers.h +1 -0
index 8d4e4f08d327..eb51fc1f54d5 100644
@@ -23,6 +23,7 @@ struct mptcp_sock {
 
 	__u32		token;
 	struct sock	*first;
+	struct mptcp_sched_ops *sched;
 	char		ca_name[TCP_CA_NAME_MAX];
 } __attribute__((preserve_access_index));
 
Signature · 1 line
2.34.1
#6 lore raw

PATCH v12 5/10 mptcp: add get_subflow wrapper

In reply to Geliang Tang

This patch defines a new wrapper mptcp_sched_get_subflow(), invoke get_subflow() of msk->sched in it. Use the wrapper instead of using mptcp_subflow_get_send() directly.

---
 net/mptcp/protocol.c | 8 ++++----
 net/mptcp/protocol.h | 7 +++++++
 2 files changed, 11 insertions(+), 4 deletions(-)
net/mptcp/protocol.c +4 -4
index 3d50c35c0b33..c1c23450eff1 100644
@@ -1567,7 +1567,7 @@ void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 			int ret = 0;
 
 			prev_ssk = ssk;
-			ssk = mptcp_subflow_get_send(msk);
+			ssk = mptcp_sched_get_subflow(msk, false);
 
 			/* First check. If the ssk has changed since
 			 * the last round, release prev_ssk
@@ -1636,7 +1636,7 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk)
 			 * check for a different subflow usage only after
 			 * spooling the first chunk of data
 			 */
-			xmit_ssk = first ? ssk : mptcp_subflow_get_send(mptcp_sk(sk));
+			xmit_ssk = first ? ssk : mptcp_sched_get_subflow(mptcp_sk(sk), false);
 			if (!xmit_ssk)
 				goto out;
 			if (xmit_ssk != ssk) {
@@ -2481,7 +2481,7 @@ static void __mptcp_retrans(struct sock *sk)
 	mptcp_clean_una_wakeup(sk);
 
 	/* first check ssk: need to kick "stale" logic */
-	ssk = mptcp_subflow_get_retrans(msk);
+	ssk = mptcp_sched_get_subflow(msk, true);
 	dfrag = mptcp_rtx_head(sk);
 	if (!dfrag) {
 		if (mptcp_data_fin_enabled(msk)) {
@@ -3146,7 +3146,7 @@ void __mptcp_check_push(struct sock *sk, struct sock *ssk)
 		return;
 
 	if (!sock_owned_by_user(sk)) {
-		struct sock *xmit_ssk = mptcp_subflow_get_send(mptcp_sk(sk));
+		struct sock *xmit_ssk = mptcp_sched_get_subflow(mptcp_sk(sk), false);
 
 		if (xmit_ssk == ssk)
 			__mptcp_subflow_push_pending(sk, ssk);
net/mptcp/protocol.h +7 -0
index b06836d1a899..d831effe2dfe 100644
@@ -632,6 +632,13 @@ int mptcp_init_sched(struct mptcp_sock *msk,
 		     struct mptcp_sched_ops *sched);
 void mptcp_release_sched(struct mptcp_sock *msk);
 
+static inline struct sock *mptcp_sched_get_subflow(struct mptcp_sock *msk, bool retrans)
+{
+	return msk->sched ? INDIRECT_CALL_INET_1(msk->sched->get_subflow,
+						 mptcp_get_subflow_default, msk, retrans) :
+			    mptcp_get_subflow_default(msk, retrans);
+}
+
 static inline bool __mptcp_subflow_active(struct mptcp_subflow_context *subflow)
 {
 	struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
Signature · 1 line
2.34.1
#7 lore raw

PATCH v12 6/10 mptcp: add bpf_mptcp_sched_ops

In reply to Geliang Tang

This patch implements a new struct bpf_struct_ops, bpf_mptcp_sched_ops. Register and unregister the bpf scheduler in .reg and .unreg.

This MPTCP BPF scheduler implementation is similar to BPF TCP CC. And net/ipv4/bpf_tcp_ca.c is a frame of reference for this patch.

---
 kernel/bpf/bpf_struct_ops_types.h |   4 +
 net/mptcp/bpf.c                   | 129 ++++++++++++++++++++++++++++++
 2 files changed, 133 insertions(+)
kernel/bpf/bpf_struct_ops_types.h +4 -0
index 5678a9ddf817..5a6b0c0d8d3d 100644
@@ -8,5 +8,9 @@ BPF_STRUCT_OPS_TYPE(bpf_dummy_ops)
 #ifdef CONFIG_INET
 #include <net/tcp.h>
 BPF_STRUCT_OPS_TYPE(tcp_congestion_ops)
+#ifdef CONFIG_MPTCP
+#include <net/mptcp.h>
+BPF_STRUCT_OPS_TYPE(mptcp_sched_ops)
+#endif
 #endif
 #endif
net/mptcp/bpf.c +129 -0

137 lines not on this page · raw message

Signature · 1 line
2.34.1
#8 lore raw

PATCH v12 7/10 mptcp: add last_snd write access

In reply to Geliang Tang

This patch exports the member last_snd of struct mptcp_sock in bpf_mptcp_helpers.h, and adds BPF write access to it.

---
 net/mptcp/bpf.c                                 | 16 ++++++++++++++++
 tools/testing/selftests/bpf/bpf_mptcp_helpers.h |  1 +
 2 files changed, 17 insertions(+)
net/mptcp/bpf.c +16 -0
index e849fc3fb6c5..bd3c50b07ab2 100644
@@ -40,6 +40,7 @@ static int bpf_mptcp_sched_btf_struct_access(struct bpf_verifier_log *log,
 {
 	const struct btf_type *state;
 	u32 type_id;
+	size_t end;
 
 	if (atype == BPF_READ)
 		return btf_struct_access(log, btf, t, off, size, atype,
@@ -55,6 +56,21 @@ static int bpf_mptcp_sched_btf_struct_access(struct bpf_verifier_log *log,
 		return -EACCES;
 	}
 
+	switch (off) {
+	case offsetof(struct mptcp_sock, last_snd):
+		end = offsetofend(struct mptcp_sock, last_snd);
+		break;
+	default:
+		bpf_log(log, "no write support to mptcp_sock at off %d\n", off);
+		return -EACCES;
+	}
+
+	if (off + size > end) {
+		bpf_log(log, "access beyond mptcp_sock at off %u size %u ended at %lu",
+			off, size, end);
+		return -EACCES;
+	}
+
 	return NOT_INIT;
 }
 
tools/testing/selftests/bpf/bpf_mptcp_helpers.h +1 -0
index eb51fc1f54d5..90accd94e947 100644
@@ -21,6 +21,7 @@ struct mptcp_sched_ops {
 struct mptcp_sock {
 	struct inet_connection_sock	sk;
 
+	struct sock	*last_snd;
 	__u32		token;
 	struct sock	*first;
 	struct mptcp_sched_ops *sched;
Signature · 1 line
2.34.1
#9 lore raw

Re: [PATCH mptcp-next v12 07/10] mptcp: add last_snd write access

Hi Geliang,

Quoted 2 lines · Geliang Tang

This patch exports the member last_snd of struct mptcp_sock in bpf_mptcp_helpers.h, and adds BPF write access to it.

Do we really need the have the different schedulers modifying it? Could we not let the caller of "sched->get_subflow()" setting it?

I mean: 'last_snd' should always be the same as what get_subflow() will return, no? If yes, why not setting it in mptcp_sched_get_subflow()?

Cheers,
Matt

Signature · 2 lines
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net
#10 lore raw

Re: [PATCH mptcp-next v12 07/10] mptcp: add last_snd write access

In reply to Geliang Tang

Hi Geliang,

Quoted 27 lines · Geliang Tang

This patch exports the member last_snd of struct mptcp_sock in bpf_mptcp_helpers.h, and adds BPF write access to it.

---
 net/mptcp/bpf.c                                 | 16 ++++++++++++++++
 tools/testing/selftests/bpf/bpf_mptcp_helpers.h |  1 +
 2 files changed, 17 insertions(+)
net/mptcp/bpf.c +3 -0
index e849fc3fb6c5..bd3c50b07ab2 100644
@@ -40,6 +40,7 @@ static int bpf_mptcp_sched_btf_struct_access(struct bpf_verifier_log *log,
 {
 	const struct btf_type *state;
 	u32 type_id;
+	size_t end;
 
 	if (atype == BPF_READ)
 		return btf_struct_access(log, btf, t, off, size, atype,
@@ -55,6 +56,21 @@ static int bpf_mptcp_sched_btf_struct_access(struct bpf_verifier_log *log,
 		return -EACCES;
 	}
 
+	switch (off) {
+	case offsetof(struct mptcp_sock, last_snd):

net/mptcp/bpf.c static int bpf_mptcp_sched_btf_struct_access(struct bpf_verifier_log *log, · line 70

Quoted 10 lines
net/mptcp/bpf.c +10 -0
+		end = offsetofend(struct mptcp_sock, last_snd);
+		break;
+	default:
+		bpf_log(log, "no write support to mptcp_sock at off %d\n", off);
+		return -EACCES;
+	}
+
+	if (off + size > end) {
+		bpf_log(log, "access beyond mptcp_sock at off %u size %u ended at %lu",
+			off, size, end);

A small note about this line: the CI[1] seems to complain about it when compiling in i386:

net/mptcp/bpf.c static int bpf_mptcp_sched_btf_struct_access(struct bpf_verifier_log *log,

Quoted 12 lines
net/mptcp/bpf.c +0 -0
  net/mptcp/bpf.c: In function 'bpf_mptcp_sched_btf_struct_access':
  net/mptcp/bpf.c:69:71: error: format '%lu' expects argument of type 'long unsigned int', but argument 5 has type 'size_t' {aka 'unsigned int'} [-Werror=format=]
     69 |   bpf_log(log, "access beyond mptcp_sock at off %u size %u ended at %lu",
        |                                                                     ~~^
        |                                                                       |
        |                                                                       long unsigned int
        |                                                                     %u
     70 |    off, size, end);
        |               ~~~                                                      
        |               |
        |               size_t {aka unsigned int}
  cc1: all warnings being treated as errors

I think you need to use %zu.

https://github.com/multipath-tcp/mptcp_net-next/runs/6089328349?check_suite_focus=true

Cheers,
Matt

Signature · 2 lines
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net
#11 lore raw

Re: [PATCH mptcp-next v12 07/10] mptcp: add last_snd write access

In reply to Geliang Tang
Thank you for the patch! Perhaps something to improve:

Hi Geliang,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on mptcp/export]
[cannot apply to bpf-next/master bpf/master linus/master v5.18-rc3 next-20220420] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/intel-lab-lkp/linux/commits/Geliang-Tang/BPF-packet-scheduler/20220420-130015
base:   https://github.com/multipath-tcp/mptcp_net-next.git export
config: i386-randconfig-m021 (https://download.01.org/0day-ci/archive/20220421/[email protected]/config)
compiler: gcc-11 (Debian 11.2.0-20) 11.2.0
reproduce (this is a W=1 build):
        # https://github.com/intel-lab-lkp/linux/commit/c711e7d26efa699f1cfe5e2c9619bbf5cc98a7a7
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Geliang-Tang/BPF-packet-scheduler/20220420-130015
        git checkout c711e7d26efa699f1cfe5e2c9619bbf5cc98a7a7
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        make W=1 O=build_dir ARCH=i386 SHELL=/bin/bash net/mptcp/

If you fix the issue, kindly add following tag as appropriate

All warnings (new ones prefixed by >>):

net/mptcp/bpf.c: In function 'bpf_mptcp_sched_btf_struct_access':

>> net/mptcp/bpf.c:69:85: warning: format '%lu' expects argument of type 'long unsigned int', but argument 5 has type 'size_t' {aka 'unsigned int'} [-Wformat=]
      69 |                 bpf_log(log, "access beyond mptcp_sock at off %u size %u ended at %lu",
         |                                                                                   ~~^
         |                                                                                     |
         |                                                                                     long unsigned int
         |                                                                                   %u
      70 |                         off, size, end);
         |                                    ~~~                                               
         |                                    |
         |                                    size_t {aka unsigned int}

vim +69 net/mptcp/bpf.c

    33	
    34	static int bpf_mptcp_sched_btf_struct_access(struct bpf_verifier_log *log,
    35						     const struct btf *btf,
    36						     const struct btf_type *t, int off,
    37						     int size, enum bpf_access_type atype,
    38						     u32 *next_btf_id,
    39						     enum bpf_type_flag *flag)
    40	{
    41		const struct btf_type *state;
    42		u32 type_id;
    43		size_t end;
    44	
    45		if (atype == BPF_READ)
    46			return btf_struct_access(log, btf, t, off, size, atype,
    47						 next_btf_id, flag);
    48	
    49		type_id = btf_find_by_name_kind(btf, "mptcp_sock", BTF_KIND_STRUCT);
    50		if (type_id < 0)
    51			return -EINVAL;
    52	
    53		state = btf_type_by_id(btf, type_id);
    54		if (t != state) {
    55			bpf_log(log, "only read is supported\n");
    56			return -EACCES;
    57		}
    58	
    59		switch (off) {
    60		case offsetof(struct mptcp_sock, last_snd):
    61			end = offsetofend(struct mptcp_sock, last_snd);
    62			break;
    63		default:
    64			bpf_log(log, "no write support to mptcp_sock at off %d\n", off);
    65			return -EACCES;
    66		}
    67	
    68		if (off + size > end) {
  > 69			bpf_log(log, "access beyond mptcp_sock at off %u size %u ended at %lu",
    70				off, size, end);
    71			return -EACCES;
    72		}
    73	
    74		return NOT_INIT;
    75	}
    76	
Signature · 2 lines
0-DAY CI Kernel Test Service
https://01.org/lkp
#12 lore raw

PATCH v12 8/10 mptcp: add call_me_again flag

In reply to Geliang Tang

For supporting a "redundant" packet scheduler in the future, this patch adds a flag named call_me_again to indicate that get_subflow() function needs to be called again.

---
 net/mptcp/bpf.c                                 | 3 +++
 net/mptcp/protocol.h                            | 1 +
 net/mptcp/sched.c                               | 1 +
 tools/testing/selftests/bpf/bpf_mptcp_helpers.h | 1 +
 4 files changed, 6 insertions(+)
net/mptcp/bpf.c +3 -0
index bd3c50b07ab2..e203653860fa 100644
@@ -60,6 +60,9 @@ static int bpf_mptcp_sched_btf_struct_access(struct bpf_verifier_log *log,
 	case offsetof(struct mptcp_sock, last_snd):
 		end = offsetofend(struct mptcp_sock, last_snd);
 		break;
+	case offsetofend(struct mptcp_sock, sched):
+		end = offsetofend(struct mptcp_sock, sched) + sizeof(u8);
+		break;
 	default:
 		bpf_log(log, "no write support to mptcp_sock at off %d\n", off);
 		return -EACCES;
net/mptcp/protocol.h +1 -0
index d831effe2dfe..58b980660d20 100644
@@ -299,6 +299,7 @@ struct mptcp_sock {
 	struct sock	*first;
 	struct mptcp_pm_data	pm;
 	struct mptcp_sched_ops	*sched;
+	u8		call_me_again:1;
 	struct {
 		u32	space;	/* bytes copied in last measurement window */
 		u32	copied; /* bytes copied in this measurement window */
net/mptcp/sched.c +1 -0
index 8025dc51fbe9..05ab45505f88 100644
@@ -83,6 +83,7 @@ int mptcp_init_sched(struct mptcp_sock *msk,
 	msk->sched = sched_init;
 	if (msk->sched->init)
 		msk->sched->init(msk);
+	msk->call_me_again = 0;
 
 	pr_debug("sched=%s", msk->sched->name);
 
tools/testing/selftests/bpf/bpf_mptcp_helpers.h +1 -0
index 90accd94e947..f02d75f91e37 100644
@@ -25,6 +25,7 @@ struct mptcp_sock {
 	__u32		token;
 	struct sock	*first;
 	struct mptcp_sched_ops *sched;
+	__u8		call_me_again:1;
 	char		ca_name[TCP_CA_NAME_MAX];
 } __attribute__((preserve_access_index));
 
Signature · 1 line
2.34.1
#13 lore raw

Re: [PATCH mptcp-next v12 08/10] mptcp: add call_me_again flag

Hi Geliang,

Quoted 3 lines · Geliang Tang

For supporting a "redundant" packet scheduler in the future, this patch adds a flag named call_me_again to indicate that get_subflow() function needs to be called again.

If I'm not mistaken, in mptcp.org, there are two functions that can be modified by a packets scheduler:

- get_subflow(): return an ssk to where an skb passed in argument will
be sent. A boolean is also passed to inform if it is a zero window test

- next_segment(): return the next skb to be sent. A parameter called
"reinject" can be set to 3 values: from reinject queue (1), from sent queue (0), meta-level retransmission for optimisations purposes (-1).

In short, a redundant scheduler will then make sure the same skb is returned multiple times in next_segment(). (After the first time, 'reinject' is set to -1.)

I understand it is quite different from what we currently have here in upstream where we don't really have a "next_segment()" that can be modified by the different schedulers. Having this 'call_me_again' variable should do the job and it is quite "flexible". At least more than setting a function parameter. But I don't know if it is a proper way and if it will fix all situations. I didn't follow the development on all different packet schedulers in mptcp.org. Maybe we can check with Christoph if we need to "fix an API".

Cheers,
Matt

Signature · 2 lines
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net
#14 lore raw

PATCH v12 9/10 selftests: bpf: add bpf_first scheduler

In reply to Geliang Tang

This patch implements the simplest MPTCP scheduler, named bpf_first, which always picks the first subflow to send data. It's a sample of MPTCP BPF scheduler implementations.

---
 .../selftests/bpf/progs/mptcp_bpf_first.c     | 34 +++++++++++++++++++
 1 file changed, 34 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/progs/mptcp_bpf_first.c
tools/testing/selftests/bpf/progs/mptcp_bpf_first.c new file +34 -0
new file mode 100644
index 000000000000..425b89c654b2
@@ -0,0 +1,34 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2022, SUSE. */
+
+#include <linux/bpf.h>
+#include "bpf_mptcp_helpers.h"
+
+char _license[] SEC("license") = "GPL";
+
+SEC("struct_ops/mptcp_sched_first_init")
+void BPF_PROG(mptcp_sched_first_init, struct mptcp_sock *msk)
+{
+}
+
+SEC("struct_ops/mptcp_sched_first_release")
+void BPF_PROG(mptcp_sched_first_release, struct mptcp_sock *msk)
+{
+}
+
+struct sock *BPF_STRUCT_OPS(bpf_first_get_subflow, struct mptcp_sock *msk, bool retrans)
+{
+	struct sock *ssk = msk->first;
+
+	msk->last_snd = ssk;
+	msk->call_me_again = 0;
+	return ssk;
+}
+
+SEC(".struct_ops")
+struct mptcp_sched_ops first = {
+	.init		= (void *)mptcp_sched_first_init,
+	.release	= (void *)mptcp_sched_first_release,
+	.get_subflow	= (void *)bpf_first_get_subflow,
+	.name		= "bpf_first",
+};
Signature · 1 line
2.34.1
#15 lore raw

PATCH v12 10/10 selftests: bpf: add bpf_first test

In reply to Geliang Tang

This patch expends the MPTCP test base to support MPTCP packet scheduler tests. Add the bpf_first scheduler test in it. Use sysctl to set net.mptcp.scheduler to use this sched.

Some code in send_data() is from prog_tests/bpf_tcp_ca.c.

---
 .../testing/selftests/bpf/prog_tests/mptcp.c  | 113 ++++++++++++++++++
 1 file changed, 113 insertions(+)
tools/testing/selftests/bpf/prog_tests/mptcp.c +113 -0
index 7e704f5aab05..377ebc0fbcbe 100644
@@ -1,9 +1,11 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020, Tessares SA. */
 
+#include <sys/param.h>
 #include <test_progs.h>
 #include "cgroup_helpers.h"
 #include "network_helpers.h"
+#include "mptcp_bpf_first.skel.h"
 
 #ifndef TCP_CA_NAME_MAX
 #define TCP_CA_NAME_MAX	16
@@ -19,6 +21,8 @@ struct mptcp_storage {
 };
 
 static char monitor_log_path[64];
+static const unsigned int total_bytes = 10 * 1024 * 1024;
+static int stop, duration;
 
 static int verify_tsk(int map_fd, int client_fd)
 {
@@ -251,8 +255,117 @@ void test_base(void)
 	close(cgroup_fd);
 }
 
+static void *server(void *arg)
+{
+	int lfd = (int)(long)arg, err = 0, fd;
+	ssize_t nr_sent = 0, bytes = 0;
+	char batch[1500];
+
+	fd = accept(lfd, NULL, NULL);
+	while (fd == -1) {
+		if (errno == EINTR)
+			continue;
+		err = -errno;
+		goto done;
+	}
+
+	if (settimeo(fd, 0)) {
+		err = -errno;
+		goto done;
+	}
+
+	while (bytes < total_bytes && !READ_ONCE(stop)) {
+		nr_sent = send(fd, &batch,
+			       MIN(total_bytes - bytes, sizeof(batch)), 0);
+		if (nr_sent == -1 && errno == EINTR)
+			continue;
+		if (nr_sent == -1) {
+			err = -errno;
+			break;
+		}
+		bytes += nr_sent;
+	}
+
+	CHECK(bytes != total_bytes, "send", "%zd != %u nr_sent:%zd errno:%d\n",
+	      bytes, total_bytes, nr_sent, errno);
+
+done:
+	if (fd >= 0)
+		close(fd);
+	if (err) {
+		WRITE_ONCE(stop, 1);
+		return ERR_PTR(err);
+	}
+	return NULL;
+}
+
+static void send_data(int lfd, int fd)
+{
+	ssize_t nr_recv = 0, bytes = 0;
+	pthread_t srv_thread;
+	void *thread_ret;
+	char batch[1500];
+	int err;
+
+	WRITE_ONCE(stop, 0);
+
+	err = pthread_create(&srv_thread, NULL, server, (void *)(long)lfd);
+	if (CHECK(err != 0, "pthread_create", "err:%d errno:%d\n", err, errno))
+		return;
+
+	/* recv total_bytes */
+	while (bytes < total_bytes && !READ_ONCE(stop)) {
+		nr_recv = recv(fd, &batch,
+			       MIN(total_bytes - bytes, sizeof(batch)), 0);
+		if (nr_recv == -1 && errno == EINTR)
+			continue;
+		if (nr_recv == -1)
+			break;
+		bytes += nr_recv;
+	}
+
+	CHECK(bytes != total_bytes, "recv", "%zd != %u nr_recv:%zd errno:%d\n",
+	      bytes, total_bytes, nr_recv, errno);
+
+	WRITE_ONCE(stop, 1);
+
+	pthread_join(srv_thread, &thread_ret);
+	CHECK(IS_ERR(thread_ret), "pthread_join", "thread_ret:%ld",
+	      PTR_ERR(thread_ret));
+}
+
+static void test_first(void)
+{
+	struct mptcp_bpf_first *first_skel;
+	int server_fd, client_fd;
+	struct bpf_link *link;
+
+	first_skel = mptcp_bpf_first__open_and_load();
+	if (CHECK(!first_skel, "bpf_first__open_and_load", "failed\n"))
+		return;
+
+	link = bpf_map__attach_struct_ops(first_skel->maps.first);
+	if (!ASSERT_OK_PTR(link, "bpf_map__attach_struct_ops")) {
+		mptcp_bpf_first__destroy(first_skel);
+		return;
+	}
+
+	system("sysctl -q net.mptcp.scheduler=bpf_first");
+	server_fd = start_mptcp_server(AF_INET, NULL, 0, 0);
+	client_fd = connect_to_mptcp_fd(server_fd, 0);
+
+	send_data(server_fd, client_fd);
+
+	close(client_fd);
+	close(server_fd);
+	bpf_link__destroy(link);
+	mptcp_bpf_first__destroy(first_skel);
+}
+
 void test_mptcp(void)
 {
 	if (test__start_subtest("base"))
 		test_base();
+	if (test__start_subtest("first"))
+		test_first();
 }
Signature · 1 line
2.34.1
#16 lore raw

Re: selftests: bpf: add bpf_first test: Build Failure

Hi Geliang,

Thank you for your modifications, that's great!

But sadly, our CI spotted some issues with it when trying to build it.

You can find more details there:

https://patchwork.kernel.org/project/mptcp/patch/7072b5ca4b8e6d710506a789c4e962c9e33cc54d.1650430389.git.geliang.tang@suse.com/
https://github.com/multipath-tcp/mptcp_net-next/actions/runs/2193665679

Status: failure
Initiator: MPTCPimporter
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/cd43e2aef03f

Feel free to reply to this email if you cannot access logs, if you need some support to fix the error, if this doesn't seem to be caused by your modifications or if the error is a false positive one.

Cheers,
MPTCP GH Action bot
Bot operated by Matthieu Baerts (Tessares)

#17 lore raw

Re: selftests: bpf: add bpf_first test: Tests Results

In reply to Geliang Tang

Hi Geliang,

Thank you for your modifications, that's great!

Our CI did some validations and here is its report:

- KVM Validation: normal:
- Success! ✅:
- Task: https://cirrus-ci.com/task/6322146069708800
- Summary: https://api.cirrus-ci.com/v1/artifact/task/6322146069708800/summary/summary.txt

- KVM Validation: debug:
- Unstable: 1 failed test(s): selftest_diag - Critical: KMemLeak ❌:
- Task: https://cirrus-ci.com/task/4914771186155520
- Summary: https://api.cirrus-ci.com/v1/artifact/task/4914771186155520/summary/summary.txt

Initiator: Patchew Applier
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/cd43e2aef03f

If there are some issues, you can reproduce them using the same environment as the one used by the CI thanks to a docker image, e.g.:

    $ cd [kernel source code]
    $ docker run -v "${PWD}:${PWD}:rw" -w "${PWD}" --privileged --rm -it \
        --pull always mptcp/mptcp-upstream-virtme-docker:latest \
        auto-debug

For more details:

https://github.com/multipath-tcp/mptcp-upstream-virtme-docker

Please note that despite all the efforts that have been already done to have a stable tests suite when executed on a public CI like here, it is possible some reported issues are not due to your modifications. Still, do not hesitate to help us improve that ;-)

Cheers,
MPTCP GH Action bot
Bot operated by Matthieu Baerts (Tessares)