xfrm: interface: Add unstable helpers for XFRM metadata
Subsystems (7)
v6 cover xfrm: interface: Add unstable helpers for XFRM metadata
This patch series adds xfrm metadata helpers using the unstable kfunc call interface for the TC-BPF hooks.
This allows steering traffic towards different IPsec connections based on logic implemented in bpf programs.
The helpers are integrated into the xfrm_interface module. For this purpose the main functionality of this module is moved to xfrm_interface_core.c.
---
changes in v6: fix sparse warning in patch 2 changes in v5:
- avoid cleanup of percpu dsts as detailed in patch 2
changes in v3:
- tag bpf-next tree instead of ipsec-next
- add IFLA_XFRM_COLLECT_METADATA sync patch
Eyal Birger (4):
xfrm: interface: rename xfrm_interface.c to xfrm_interface_core.c
xfrm: interface: Add unstable helpers for setting/getting XFRM
metadata from TC-BPF
tools: add IFLA_XFRM_COLLECT_METADATA to uapi/linux/if_link.h selftests/bpf: add xfrm_info tests
include/net/dst_metadata.h | 1 +
include/net/xfrm.h | 17 +
net/core/dst.c | 8 +-
net/core/filter.c | 9 +
net/xfrm/Makefile | 8 +
net/xfrm/xfrm_interface_bpf.c | 115 ++++++
...xfrm_interface.c => xfrm_interface_core.c} | 14 +
tools/include/uapi/linux/if_link.h | 1 +
tools/testing/selftests/bpf/DENYLIST.s390x | 1 +
tools/testing/selftests/bpf/config | 2 +
.../selftests/bpf/prog_tests/xfrm_info.c | 365 ++++++++++++++++++
.../selftests/bpf/progs/bpf_tracing_net.h | 3 +
tools/testing/selftests/bpf/progs/xfrm_info.c | 35 ++
13 files changed, 577 insertions(+), 2 deletions(-)
create mode 100644 net/xfrm/xfrm_interface_bpf.c
rename net/xfrm/{xfrm_interface.c => xfrm_interface_core.c} (98%)
create mode 100644 tools/testing/selftests/bpf/prog_tests/xfrm_info.c
create mode 100644 tools/testing/selftests/bpf/progs/xfrm_info.cSignature · 1 line
2.34.1
PATCH v6 1/4 xfrm: interface: rename xfrm_interface.c to xfrm_interface_core.c
This change allows adding additional files to the xfrm_interface module.
---
net/xfrm/Makefile | 2 ++
net/xfrm/{xfrm_interface.c => xfrm_interface_core.c} | 0
2 files changed, 2 insertions(+)
rename net/xfrm/{xfrm_interface.c => xfrm_interface_core.c} (100%)net/xfrm/Makefile +2 -0
@@ -3,6 +3,8 @@ # Makefile for the XFRM subsystem. # +xfrm_interface-$(CONFIG_XFRM_INTERFACE) += xfrm_interface_core.o + obj-$(CONFIG_XFRM) := xfrm_policy.o xfrm_state.o xfrm_hash.o \ xfrm_input.o xfrm_output.o \ xfrm_sysctl.o xfrm_replay.o xfrm_device.o
net/xfrm/xfrm_interface_core.c renamed from net/xfrm/xfrm_interface.c +0 -0
Signature · 1 line
2.34.1
PATCH v6 2/4 xfrm: interface: Add unstable helpers for setting/getting XFRM metadata from TC-BPF
In reply to Eyal BirgerThis change adds xfrm metadata helpers using the unstable kfunc call interface for the TC-BPF hooks. This allows steering traffic towards different IPsec connections based on logic implemented in bpf programs.
This object is built based on the availability of BTF debug info.
When setting the xfrm metadata, percpu metadata dsts are used in order to avoid allocating a metadata dst per packet.
In order to guarantee safe module unload, the percpu dsts are allocated on first use and never freed. The percpu pointer is stored in net/core/filter.c so that it can be reused on module reload.
The metadata percpu dsts take ownership of the original skb dsts so that they may be used as part of the xfrm transmission logic - e.g. for MTU calculations.
---
v6:
fix sparse warning:
"symbol 'xfrm_bpf_md_dst' was not declared. Should it be static?" by moving the xfrm_bpf_md_dst declaration to xfrm.h
v5:
- remove cleanup of percpu dsts as suggested by Martin KaFai Lau
- remove redundant extern in header file
v4: changes requested by Martin KaFai Lau:
- add kfunc documentation
- remove redundant memset
- minor coding
v3:
- remove redunant memset() as suggested by Martin KaFai Lau
- remove __exit annotation from cleanup() function as it's used from an __init function
v2: changed added following points raised by Martin KaFai Lau:
- make sure dst is refcounted prior to caching
- free dst_orig regardless of CONFIG_DST_CACHE
- call xfrm interface bpf cleanup in case of kfunc registration errors
--- include/net/dst_metadata.h | 1 + include/net/xfrm.h | 17 +++++ net/core/dst.c | 8 ++- net/core/filter.c | 9 +++ net/xfrm/Makefile | 6 ++ net/xfrm/xfrm_interface_bpf.c | 115 +++++++++++++++++++++++++++++++++ net/xfrm/xfrm_interface_core.c | 14 ++++ 7 files changed, 168 insertions(+), 2 deletions(-) create mode 100644 net/xfrm/xfrm_interface_bpf.c
include/net/dst_metadata.h +1 -0
@@ -26,6 +26,7 @@ struct macsec_info {struct xfrm_md_info {u32 if_id;int link; +struct dst_entry *dst_orig; };struct metadata_dst {
include/net/xfrm.h +17 -0
@@ -2086,4 +2086,21 @@ static inline bool xfrm6_local_dontfrag(const struct sock *sk)return false; }#endif + +#if (IS_BUILTIN (CONFIG_XFRM_INTERFACE ) &&IS_ENABLED (CONFIG_DEBUG_INFO_BTF )) || \ + (IS_MODULE (CONFIG_XFRM_INTERFACE ) &&IS_ENABLED (CONFIG_DEBUG_INFO_BTF_MODULES )) + +extern struct metadata_dst __percpu *xfrm_bpf_md_dst; + +int register_xfrm_interface_bpf (void ); + +#else + +static inline int register_xfrm_interface_bpf (void ) +{ +return 0 ; +} + +#endif +#endif /* _NET_XFRM_H */
net/core/dst.c +6 -2
@@ -316,6 +316,8 @@ void metadata_dst_free(struct metadata_dst *md_dst)if (md_dst->type ==METADATA_IP_TUNNEL )dst_cache_destroy (&md_dst->u.tun_info.dst_cache);#endif +if (md_dst->type ==METADATA_XFRM ) +dst_release (md_dst->u.xfrm_info.dst_orig);kfree (md_dst); }EXPORT_SYMBOL_GPL (metadata_dst_free);
@@ -340,16 +342,18 @@ EXPORT_SYMBOL_GPL(metadata_dst_alloc_percpu);void metadata_dst_free_percpu (struct metadata_dst __percpu *md_dst) { -#ifdef CONFIG_DST_CACHE int cpu;for_each_possible_cpu (cpu) {struct metadata_dst *one_md_dst =per_cpu_ptr (md_dst, cpu); +#ifdef CONFIG_DST_CACHE if (one_md_dst->type ==METADATA_IP_TUNNEL )dst_cache_destroy (&one_md_dst->u.tun_info.dst_cache); - }#endif +if (one_md_dst->type ==METADATA_XFRM ) +dst_release (one_md_dst->u.xfrm_info.dst_orig); + }free_percpu (md_dst); }EXPORT_SYMBOL_GPL (metadata_dst_free_percpu);
net/core/filter.c +9 -0
@@ -5630,6 +5630,15 @@ static const struct bpf_func_proto bpf_bind_proto = { };#ifdef CONFIG_XFRM + +#if (IS_BUILTIN (CONFIG_XFRM_INTERFACE ) &&IS_ENABLED (CONFIG_DEBUG_INFO_BTF )) || \ + (IS_MODULE (CONFIG_XFRM_INTERFACE ) &&IS_ENABLED (CONFIG_DEBUG_INFO_BTF_MODULES )) + +struct metadata_dst __percpu *xfrm_bpf_md_dst; +EXPORT_SYMBOL_GPL (xfrm_bpf_md_dst); + +#endif + BPF_CALL_5(bpf_skb_get_xfrm_state, struct sk_buff *, skb, u32, index, struct bpf_xfrm_state *, to, u32, size, u64, flags) {
net/xfrm/Makefile +6 -0
@@ -5,6 +5,12 @@ xfrm_interface-$(CONFIG_XFRM_INTERFACE) += xfrm_interface_core.o +ifeq ($(CONFIG_XFRM_INTERFACE),m) +xfrm_interface-$(CONFIG_DEBUG_INFO_BTF_MODULES) += xfrm_interface_bpf.o +else ifeq ($(CONFIG_XFRM_INTERFACE),y) +xfrm_interface-$(CONFIG_DEBUG_INFO_BTF) += xfrm_interface_bpf.o +endif + obj-$(CONFIG_XFRM) := xfrm_policy.o xfrm_state.o xfrm_hash.o \ xfrm_input.o xfrm_output.o \ xfrm_sysctl.o xfrm_replay.o xfrm_device.o
net/xfrm/xfrm_interface_bpf.c new file +115 -0
115 lines not on this page · raw message
net/xfrm/xfrm_interface_core.c +14 -0
@@ -396,6 +396,14 @@ xfrmi_xmit2(struct sk_buff *skb, struct net_device *dev, struct flowi *fl) if_id = md_info->if_id; fl->flowi_oif = md_info->link; +if (md_info->dst_orig) { +struct dst_entry *tmp_dst = dst; + + dst = md_info->dst_orig; +skb_dst_set (skb, dst); + md_info->dst_orig =NULL ; +dst_release (tmp_dst); + } }else { if_id = xi->p.if_id; }
@@ -1162,12 +1170,18 @@ static int __init xfrmi_init(void)if (err <0 ) goto rtnl_link_failed; + err =register_xfrm_interface_bpf (); +if (err <0 ) + goto kfunc_failed; +lwtunnel_encap_add_ops (&xfrmi_encap_ops,LWTUNNEL_ENCAP_XFRM );xfrm_if_register_cb (&xfrm_if_cb);return err; +kfunc_failed: +rtnl_link_unregister (&xfrmi_link_ops); rtnl_link_failed:xfrmi6_fini (); xfrmi6_failed:
Signature · 1 line
2.34.1
PATCH v6 3/4 tools: add IFLA_XFRM_COLLECT_METADATA to uapi/linux/if_link.h
In reply to Eyal BirgerNeeded for XFRM metadata tests.
--- tools/include/uapi/linux/if_link.h | 1 + 1 file changed, 1 insertion(+)
tools/include/uapi/linux/if_link.h +1 -0
@@ -673,6 +673,7 @@ enum {IFLA_XFRM_UNSPEC ,IFLA_XFRM_LINK ,IFLA_XFRM_IF_ID , +IFLA_XFRM_COLLECT_METADATA , __IFLA_XFRM_MAX };
Signature · 1 line
2.34.1
PATCH v6 4/4 selftests/bpf: add xfrm_info tests
In reply to Eyal BirgerTest the xfrm_info kfunc helpers.
The test setup creates three name spaces - NS0, NS1, NS2.
XFRM tunnels are setup between NS0 and the two other NSs.
The kfunc helpers are used to steer traffic from NS0 to the other NSs based on a userspace populated bpf global variable and validate that the return traffic had arrived from the desired NS.
---
v4: changes suggested by Martin KaFai Lau:
- minor coding and wording changes
- use vmlinux.h and bpf_tracing_net.h in test program
- avoid running on s390x in CI as kfuncs aren't supported there
v3: changes suggested by Martin KaFai Lau:
- rename test_xfrm_info{,_kern}.c -> xfrm_info.c
- avoid running in a separate thread
- simplify test setup
- verify underlay/overlay success
- create NS0 xfrmi external device using netlink to avoid dependency on newer iproute2
- remove use of bpf_trace_printk() in test programs
- add "preserve_access_index" attribute annotation in test program struct definition
- use bss variables access instead of a map for userspace/bpf interface
v2:
- use an lwt route in NS1 for testing that flow as well
- indendation fix
--- tools/testing/selftests/bpf/DENYLIST.s390x | 1 + tools/testing/selftests/bpf/config | 2 + .../selftests/bpf/prog_tests/xfrm_info.c | 365 ++++++++++++++++++ .../selftests/bpf/progs/bpf_tracing_net.h | 3 + tools/testing/selftests/bpf/progs/xfrm_info.c | 35 ++ 5 files changed, 406 insertions(+) create mode 100644 tools/testing/selftests/bpf/prog_tests/xfrm_info.c create mode 100644 tools/testing/selftests/bpf/progs/xfrm_info.c
tools/testing/selftests/bpf/DENYLIST.s390x +1 -0
@@ -85,3 +85,4 @@ xdp_bonding # failed to auto-attach program 'trace_ xdp_bpf2bpf # failed to auto-attach program 'trace_on_entry': -524 (trampoline) xdp_do_redirect # prog_run_max_size unexpected error: -22 (errno 22) xdp_synproxy # JIT does not support calling kernel function (kfunc) +xfrm_info # JIT does not support calling kernel function (kfunc)
tools/testing/selftests/bpf/config +2 -0
@@ -23,6 +23,7 @@ CONFIG_IKCONFIG_PROC=y CONFIG_IMA=y CONFIG_IMA_READ_POLICY=y CONFIG_IMA_WRITE_POLICY=y +CONFIG_INET_ESP=y CONFIG_IP_NF_FILTER=y CONFIG_IP_NF_RAW=y CONFIG_IP_NF_TARGET_SYNPROXY=y
@@ -74,3 +75,4 @@ CONFIG_TEST_BPF=y CONFIG_USERFAULTFD=y CONFIG_VXLAN=y CONFIG_XDP_SOCKETS=y +CONFIG_XFRM_INTERFACE=y
tools/testing/selftests/bpf/prog_tests/xfrm_info.c new file +365 -0
365 lines not on this page · raw message
tools/testing/selftests/bpf/progs/bpf_tracing_net.h +3 -0
@@ -25,6 +25,9 @@#define IPV6_TCLASS 67#define IPV6_AUTOFLOWLABEL 70 +#define TC_ACT_UNSPEC (-1) +#define TC_ACT_SHOT 2 +#define SOL_TCP 6#define TCP_NODELAY 1#define TCP_MAXSEG 2
tools/testing/selftests/bpf/progs/xfrm_info.c new file +35 -0
@@ -0,0 +1,35 @@ +// SPDX-License-Identifier: GPL-2.0 +#include "vmlinux.h" +#include "bpf_tracing_net.h" +#include <bpf/bpf_helpers.h> + +__u32 req_if_id; +__u32 resp_if_id; + +int bpf_skb_set_xfrm_info(struct __sk_buff *skb_ctx, + const struct bpf_xfrm_info *from) __ksym; +int bpf_skb_get_xfrm_info(struct __sk_buff *skb_ctx, + struct bpf_xfrm_info *to) __ksym; + +SEC("tc") +int set_xfrm_info (struct __sk_buff *skb) +{ +struct bpf_xfrm_info info = { .if_id = req_if_id }; + +return bpf_skb_set_xfrm_info (skb, &info) ?TC_ACT_SHOT :TC_ACT_UNSPEC ; +} + +SEC("tc") +int get_xfrm_info (struct __sk_buff *skb) +{ +struct bpf_xfrm_info info = {}; + +if (bpf_skb_get_xfrm_info (skb, &info) <0 ) +return TC_ACT_SHOT ; + + resp_if_id = info.if_id; + +return TC_ACT_UNSPEC ; +} + +char _license[] SEC("license") = "GPL";
Signature · 1 line
2.34.1
Martin KaFai Lau
reviewerRe: [PATCH bpf-next,v6 4/4] selftests/bpf: add xfrm_info tests
Quoted 1 line · Eyal Birger
+#define
PING_ARGS"-i 0.01 -c 3 -w 10 -q"
Applied with a few changes.
PING_ARGS is removed because it is unused.
[ ... ]
Quoted 43 lines
tools/testing/selftests/bpf/prog_tests/xfrm_info.c new file +43
-0+static int test_xfrm_ping (struct xfrm_info *skel,u32 if_id) +{ + skel->bss->req_if_id = if_id; + +SYS ("ping -i 0.01 -c 3 -w 10 -q 192.168.1.200 > /dev/null" ); + +if (!ASSERT_EQ (skel->bss->resp_if_id, if_id,"if_id" )) + goto fail; + +return 0 ; +fail: +return -1 ; +} + +static void _test_xfrm_info (void ) +{ + LIBBPF_OPTS(bpf_tc_hook, tc_hook, .attach_point = BPF_TC_INGRESS); +int get_xfrm_info_prog_fd, set_xfrm_info_prog_fd; +struct xfrm_info *skel =NULL ; +struct nstoken *nstoken =NULL ; +int ifindex; + +/* load and attach bpf progs to ipsec dev tc hook point */ + skel =xfrm_info__open_and_load (); +if (!ASSERT_OK_PTR (skel,"xfrm_info__open_and_load" )) + goto done; + nstoken =open_netns (NS0 ); +if (!ASSERT_OK_PTR (nstoken,"setns " NS0 )) + goto done; + ifindex =if_nametoindex ("ipsec0" ); +if (!ASSERT_NEQ (ifindex,0 ,"ipsec0 ifindex" )) + goto done; + tc_hook.ifindex = ifindex; + set_xfrm_info_prog_fd =bpf_program__fd (skel->progs.set_xfrm_info); + get_xfrm_info_prog_fd =bpf_program__fd (skel->progs.get_xfrm_info); +if (!ASSERT_GE (set_xfrm_info_prog_fd,0 ,"bpf_program__fd" )) + goto done; +if (!ASSERT_GE (get_xfrm_info_prog_fd,0 ,"bpf_program__fd" )) + goto done; +if (attach_tc_prog (&tc_hook, get_xfrm_info_prog_fd, + set_xfrm_info_prog_fd)) + goto done; +
tools/testing/selftests/bpf/prog_tests/xfrm_info.c line 347
Quoted 10 lines
tools/testing/selftests/bpf/prog_tests/xfrm_info.c new file +10
-0+/* perform test */ +if (!ASSERT_EQ (test_xfrm_ping (skel,IF_ID_0_TO_1 ),0 ,"ping " NS1 )) + goto done; +if (!ASSERT_EQ (test_xfrm_ping (skel,IF_ID_0_TO_2 ),0 ,"ping " NS2 )) + goto done; + +done: +if (nstoken) +close_netns (nstoken); +if (skel)
NULL check is removed. xfrm_info__destroy() can handle NULL.
Quoted 2 lines
(quoted diff) +2
-0+ xfrm_info__destroy(skel); +}
This series was applied to bpf/bpf-next.git (master) by Martin KaFai Lau <[email protected]>:
Hello:
This series was applied to bpf/bpf-next.git (master) by Martin KaFai Lau <[email protected]>:
Quoted 11 lines · On Sat, 3 Dec 2022 10:46:55 +0200 you wrote:
This patch series adds xfrm metadata helpers using the unstable kfunc call interface for the TC-BPF hooks.
This allows steering traffic towards different IPsec connections based on logic implemented in bpf programs.
The helpers are integrated into the
xfrm_interfacemodule. For this purpose the main functionality of this module is moved toxfrm_interface_core.c.[...]
Here is the summary with links:
- [bpf-next,v6,1/4] xfrm: interface: rename xfrm_interface.c to xfrm_interface_core.c
https://git.kernel.org/bpf/bpf-next/c/ee9a113ab634 - [bpf-next,v6,2/4] xfrm: interface: Add unstable helpers for setting/getting XFRM metadata from TC-BPF https://git.kernel.org/bpf/bpf-next/c/94151f5aa966 - [bpf-next,v6,3/4] tools: add IFLA_XFRM_COLLECT_METADATA to uapi/linux/if_link.h https://git.kernel.org/bpf/bpf-next/c/4f4ac4d9106e - [bpf-next,v6,4/4] selftests/bpf: add xfrm_info tests https://git.kernel.org/bpf/bpf-next/c/90a3a05eb33f
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html