Add support for bpf_setsockopt and
Subsystems (3)
v2 cover Add support for bpf_setsockopt and
v2: Added details about the test in commit log
This patch contains support to set and get socket options from setsockopt
bpf program.
This enables us to set multiple socket option when the user changes a particular socket option.
Example use case, when the user sets the IPV6_TCLASS socket option we would also like to change the tcp-cc for that socket. We don't have any use case for calling bpf_setsockopt from supposedly read-only sys_getsockopt, so it is made available to BPF_CGROUP_SETSOCKOPT only.
Prankur gupta (2):
bpf: Add support for {set|get} socket options from setsockopt BPF selftests/bpf: Add test for {set|get} socket option from setsockopt
BPF program
kernel/bpf/cgroup.c | 8 +++ tools/testing/selftests/bpf/bpf_tcp_helpers.h | 18 +++++ .../bpf/prog_tests/sockopt_qos_to_cc.c | 70 +++++++++++++++++++ .../selftests/bpf/progs/sockopt_qos_to_cc.c | 39 +++++++++++ 4 files changed, 135 insertions(+) create mode 100644 tools/testing/selftests/bpf/prog_tests/sockopt_qos_to_cc.c create mode 100644 tools/testing/selftests/bpf/progs/sockopt_qos_to_cc.c
Signature · 1 line
2.30.2
PATCH v2 1/2 bpf: Add support for {set|get} socket options from setsockopt BPF
Add logic to call bpf_setsockopt and bpf_getsockopt from setsockopt BPF programs.
Example use case, when the user sets the IPV6_TCLASS socket option we would also like to change the tcp-cc for that socket. We don't have any use case for calling bpf_setsockopt from supposedly read-only sys_getsockopti, so it is made available to BPF_CGROUP_SETSOCKOPT only.
--- kernel/bpf/cgroup.c | 8 ++++++++ 1 file changed, 8 insertions(+)
kernel/bpf/cgroup.c +8 -0
@@ -1873,6 +1873,14 @@ cg_sockopt_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)return &bpf_sk_storage_get_proto;case BPF_FUNC_sk_storage_delete:return &bpf_sk_storage_delete_proto; +case BPF_FUNC_setsockopt: +if (prog->expected_attach_type ==BPF_CGROUP_SETSOCKOPT ) +return &bpf_sk_setsockopt_proto; +return NULL ; +case BPF_FUNC_getsockopt: +if (prog->expected_attach_type ==BPF_CGROUP_SETSOCKOPT ) +return &bpf_sk_getsockopt_proto; +return NULL ;#endif #ifdef CONFIG_INET case BPF_FUNC_tcp_sock:
Signature · 1 line
2.30.2
PATCH v2 2/2 selftests/bpf: Add test for {set|get} socket option from setsockopt BPF program
In reply to Prankur guptaAdding selftests for new added functionality to call bpf_setsockopt and bpf_getsockopt from setsockopt BPF programs
Test Details:
1. BPF Program
Checks for changes in IPV6_TCLASS(SOL_IPV6) via setsockopt If the cca for the socket is not cubic do nothing
If the newly set value for IPV6_TCLASS is 45 (0x2d) (as per our usecase) then change the cc from cubic to reno
2. User Space Program
Creates an AF_INET6 socket and set the cca for that to be "cubic" Attach the program and set the IPV6_TCLASS to 0x2d using setsockopt Verify the cca for the socket changed to reno
--- tools/testing/selftests/bpf/bpf_tcp_helpers.h | 18 +++++ .../bpf/prog_tests/sockopt_qos_to_cc.c | 70 +++++++++++++++++++ .../selftests/bpf/progs/sockopt_qos_to_cc.c | 39 +++++++++++ 3 files changed, 127 insertions(+) create mode 100644 tools/testing/selftests/bpf/prog_tests/sockopt_qos_to_cc.c create mode 100644 tools/testing/selftests/bpf/progs/sockopt_qos_to_cc.c
tools/testing/selftests/bpf/bpf_tcp_helpers.h +18 -0
@@ -12,6 +12,10 @@ SEC("struct_ops/"#name) \ BPF_PROG(name, args) +#ifndef SOL_TCP +#define SOL_TCP 6 +#endif +#define tcp_jiffies32 ((__u32)bpf_jiffies64())struct sock_common {
@@ -203,6 +207,20 @@ static __always_inline bool tcp_is_cwnd_limited(const struct sock *sk)return !!BPF_CORE_READ_BITFIELD (tp, is_cwnd_limited); } +static __always_inline bool tcp_cc_eq(const char *a, const char *b) +{ +int i; + +for (i =0 ; i <TCP_CA_NAME_MAX ; i++) { +if (a[i] != b[i]) +return false; +if (!a[i]) +break ; + } + +return true; +} + extern __u32 tcp_slow_start(struct tcp_sock *tp, __u32 acked) __ksym; extern void tcp_cong_avoid_ai(struct tcp_sock *tp, __u32 w, __u32 acked) __ksym;
tools/testing/selftests/bpf/prog_tests/sockopt_qos_to_cc.c new file +70 -0
70 lines not on this page · raw message
tools/testing/selftests/bpf/progs/sockopt_qos_to_cc.c new file +39 -0
@@ -0,0 +1,39 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2021 Facebook */ +#include <string.h> +#include <linux/tcp.h> +#include <netinet/in.h> +#include <linux/bpf.h> +#include <bpf/bpf_helpers.h> +#include "bpf_tcp_helpers.h" + +char _license[] SEC("license") = "GPL"; + +SEC("cgroup/setsockopt") +int sockopt_qos_to_cc (struct bpf_sockopt *ctx) +{ +void *optval_end = ctx->optval_end; +int *optval = ctx->optval; +char buf[TCP_CA_NAME_MAX ]; +char cc_reno[TCP_CA_NAME_MAX ] ="reno" ; +char cc_cubic[TCP_CA_NAME_MAX ] ="cubic" ; + +if (ctx->level !=SOL_IPV6 || ctx->optname !=IPV6_TCLASS ) +return 1 ; + +if (optval +1 > optval_end) +return 0 ;/* EPERM, bounds check */ + +if (bpf_getsockopt (ctx->sk,SOL_TCP ,TCP_CONGESTION , &buf,sizeof (buf))) +return 0 ; + +if (!tcp_cc_eq (buf, cc_cubic)) +return 0 ; + +if (*optval ==0x2d ) { +if (bpf_setsockopt (ctx->sk,SOL_TCP ,TCP_CONGESTION , &cc_reno, +sizeof (cc_reno))) +return 0 ; + } +return 1 ; +}
Signature · 1 line
2.30.2
Re: [PATCH bpf-next 2/2] selftests/bpf: Add test for {set|get} socket option from setsockopt BPF program
Quoted 11 lines · Prankur gupta
Adding selftests for new added functionality to call
bpf_setsockoptandbpf_getsockoptfrom setsockopt BPF programsTest Details:
1. BPF Program
Checks for changes inIPV6_TCLASS(SOL_IPV6)via setsockopt If the cca for the socket is not cubic do nothing
If the newly set value forIPV6_TCLASSis 45 (0x2d) (as per our usecase) then change the cc from cubic to reno
Quoted 6 lines
2. User Space Program
Creates anAF_INET6socket and set the cca for that to be "cubic" Attach the program and set theIPV6_TCLASSto0x2dusing setsockopt Verify the cca for the socket changed to renoSigned-off-by Prankur gupta <[email protected]>
This series was applied to bpf/bpf-next.git (refs/heads/master):
Hello:
This series was applied to bpf/bpf-next.git (refs/heads/master):
Quoted 12 lines · On Tue, 17 Aug 2021 15:42:19 -0700 you wrote:
v2: Added details about the test in commit log
This patch contains support to set and get socket options from setsockopt
bpf program.
This enables us to set multiple socket option when the user changes a particular socket option.
Example use case, when the user sets theIPV6_TCLASSsocket option we would also like to change the tcp-cc for that socket. We don't have any use case for callingbpf_setsockoptfrom supposedly read-onlysys_getsockopt, so it is made available toBPF_CGROUP_SETSOCKOPTonly.
Quoted 1 line
[...]
Here is the summary with links:
- [bpf-next,1/2] bpf: Add support for {set|get} socket options from setsockopt BPF
https://git.kernel.org/bpf/bpf-next/c/2c531639deb5 - [bpf-next,2/2] selftests/bpf: Add test for {set|get} socket option from setsockopt BPF program https://git.kernel.org/bpf/bpf-next/c/f2a6ee924d26
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html