diff options
author | David S. Miller <davem@davemloft.net> | 2019-12-27 14:20:10 -0800 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2019-12-27 14:20:10 -0800 |
commit | 2bbc078f812d45b8decb55935dab21199bd21489 (patch) | |
tree | b217e030e0f80a26561cef679e8ae2643162b346 /samples | |
parent | 9e41fbf3dd38327d440a8f3ba0c234519dbb5280 (diff) | |
parent | 7c8dce4b166113743adad131b5a24c4acc12f92c (diff) | |
download | linux-stable-2bbc078f812d45b8decb55935dab21199bd21489.tar.gz linux-stable-2bbc078f812d45b8decb55935dab21199bd21489.tar.bz2 linux-stable-2bbc078f812d45b8decb55935dab21199bd21489.zip |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next
Daniel Borkmann says:
====================
pull-request: bpf-next 2019-12-27
The following pull-request contains BPF updates for your *net-next* tree.
We've added 127 non-merge commits during the last 17 day(s) which contain
a total of 110 files changed, 6901 insertions(+), 2721 deletions(-).
There are three merge conflicts. Conflicts and resolution looks as follows:
1) Merge conflict in net/bpf/test_run.c:
There was a tree-wide cleanup c593642c8be0 ("treewide: Use sizeof_field() macro")
which gets in the way with b590cb5f802d ("bpf: Switch to offsetofend in
BPF_PROG_TEST_RUN"):
<<<<<<< HEAD
if (!range_is_zero(__skb, offsetof(struct __sk_buff, priority) +
sizeof_field(struct __sk_buff, priority),
=======
if (!range_is_zero(__skb, offsetofend(struct __sk_buff, priority),
>>>>>>> 7c8dce4b166113743adad131b5a24c4acc12f92c
There are a few occasions that look similar to this. Always take the chunk with
offsetofend(). Note that there is one where the fields differ in here:
<<<<<<< HEAD
if (!range_is_zero(__skb, offsetof(struct __sk_buff, tstamp) +
sizeof_field(struct __sk_buff, tstamp),
=======
if (!range_is_zero(__skb, offsetofend(struct __sk_buff, gso_segs),
>>>>>>> 7c8dce4b166113743adad131b5a24c4acc12f92c
Just take the one with offsetofend() /and/ gso_segs. Latter is correct due to
850a88cc4096 ("bpf: Expose __sk_buff wire_len/gso_segs to BPF_PROG_TEST_RUN").
2) Merge conflict in arch/riscv/net/bpf_jit_comp.c:
(I'm keeping Bjorn in Cc here for a double-check in case I got it wrong.)
<<<<<<< HEAD
if (is_13b_check(off, insn))
return -1;
emit(rv_blt(tcc, RV_REG_ZERO, off >> 1), ctx);
=======
emit_branch(BPF_JSLT, RV_REG_T1, RV_REG_ZERO, off, ctx);
>>>>>>> 7c8dce4b166113743adad131b5a24c4acc12f92c
Result should look like:
emit_branch(BPF_JSLT, tcc, RV_REG_ZERO, off, ctx);
3) Merge conflict in arch/riscv/include/asm/pgtable.h:
<<<<<<< HEAD
=======
#define VMALLOC_SIZE (KERN_VIRT_SIZE >> 1)
#define VMALLOC_END (PAGE_OFFSET - 1)
#define VMALLOC_START (PAGE_OFFSET - VMALLOC_SIZE)
#define BPF_JIT_REGION_SIZE (SZ_128M)
#define BPF_JIT_REGION_START (PAGE_OFFSET - BPF_JIT_REGION_SIZE)
#define BPF_JIT_REGION_END (VMALLOC_END)
/*
* Roughly size the vmemmap space to be large enough to fit enough
* struct pages to map half the virtual address space. Then
* position vmemmap directly below the VMALLOC region.
*/
#define VMEMMAP_SHIFT \
(CONFIG_VA_BITS - PAGE_SHIFT - 1 + STRUCT_PAGE_MAX_SHIFT)
#define VMEMMAP_SIZE BIT(VMEMMAP_SHIFT)
#define VMEMMAP_END (VMALLOC_START - 1)
#define VMEMMAP_START (VMALLOC_START - VMEMMAP_SIZE)
#define vmemmap ((struct page *)VMEMMAP_START)
>>>>>>> 7c8dce4b166113743adad131b5a24c4acc12f92c
Only take the BPF_* defines from there and move them higher up in the
same file. Remove the rest from the chunk. The VMALLOC_* etc defines
got moved via 01f52e16b868 ("riscv: define vmemmap before pfn_to_page
calls"). Result:
[...]
#define __S101 PAGE_READ_EXEC
#define __S110 PAGE_SHARED_EXEC
#define __S111 PAGE_SHARED_EXEC
#define VMALLOC_SIZE (KERN_VIRT_SIZE >> 1)
#define VMALLOC_END (PAGE_OFFSET - 1)
#define VMALLOC_START (PAGE_OFFSET - VMALLOC_SIZE)
#define BPF_JIT_REGION_SIZE (SZ_128M)
#define BPF_JIT_REGION_START (PAGE_OFFSET - BPF_JIT_REGION_SIZE)
#define BPF_JIT_REGION_END (VMALLOC_END)
/*
* Roughly size the vmemmap space to be large enough to fit enough
* struct pages to map half the virtual address space. Then
* position vmemmap directly below the VMALLOC region.
*/
#define VMEMMAP_SHIFT \
(CONFIG_VA_BITS - PAGE_SHIFT - 1 + STRUCT_PAGE_MAX_SHIFT)
#define VMEMMAP_SIZE BIT(VMEMMAP_SHIFT)
#define VMEMMAP_END (VMALLOC_START - 1)
#define VMEMMAP_START (VMALLOC_START - VMEMMAP_SIZE)
[...]
Let me know if there are any other issues.
Anyway, the main changes are:
1) Extend bpftool to produce a struct (aka "skeleton") tailored and specific
to a provided BPF object file. This provides an alternative, simplified API
compared to standard libbpf interaction. Also, add libbpf extern variable
resolution for .kconfig section to import Kconfig data, from Andrii Nakryiko.
2) Add BPF dispatcher for XDP which is a mechanism to avoid indirect calls by
generating a branch funnel as discussed back in bpfconf'19 at LSF/MM. Also,
add various BPF riscv JIT improvements, from Björn Töpel.
3) Extend bpftool to allow matching BPF programs and maps by name,
from Paul Chaignon.
4) Support for replacing cgroup BPF programs attached with BPF_F_ALLOW_MULTI
flag for allowing updates without service interruption, from Andrey Ignatov.
5) Cleanup and simplification of ring access functions for AF_XDP with a
bonus of 0-5% performance improvement, from Magnus Karlsson.
6) Enable BPF JITs for x86-64 and arm64 by default. Also, final version of
audit support for BPF, from Daniel Borkmann and latter with Jiri Olsa.
7) Move and extend test_select_reuseport into BPF program tests under
BPF selftests, from Jakub Sitnicki.
8) Various BPF sample improvements for xdpsock for customizing parameters
to set up and benchmark AF_XDP, from Jay Jayatheerthan.
9) Improve libbpf to provide a ulimit hint on permission denied errors.
Also change XDP sample programs to attach in driver mode by default,
from Toke Høiland-Jørgensen.
10) Extend BPF test infrastructure to allow changing skb mark from tc BPF
programs, from Nikita V. Shirokov.
11) Optimize prologue code sequence in BPF arm32 JIT, from Russell King.
12) Fix xdp_redirect_cpu BPF sample to manually attach to tracepoints after
libbpf conversion, from Jesper Dangaard Brouer.
13) Minor misc improvements from various others.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'samples')
-rw-r--r-- | samples/bpf/Makefile | 5 | ||||
-rw-r--r-- | samples/bpf/xdp1_user.c | 5 | ||||
-rw-r--r-- | samples/bpf/xdp_adjust_tail_user.c | 5 | ||||
-rw-r--r-- | samples/bpf/xdp_fwd_user.c | 17 | ||||
-rw-r--r-- | samples/bpf/xdp_redirect_cpu_user.c | 63 | ||||
-rw-r--r-- | samples/bpf/xdp_redirect_map_user.c | 5 | ||||
-rw-r--r-- | samples/bpf/xdp_redirect_user.c | 5 | ||||
-rw-r--r-- | samples/bpf/xdp_router_ipv4_user.c | 3 | ||||
-rw-r--r-- | samples/bpf/xdp_rxq_info_user.c | 4 | ||||
-rw-r--r-- | samples/bpf/xdp_sample_pkts_user.c | 12 | ||||
-rw-r--r-- | samples/bpf/xdp_tx_iptunnel_user.c | 5 | ||||
-rw-r--r-- | samples/bpf/xdpsock_user.c | 431 |
12 files changed, 504 insertions, 56 deletions
diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile index 1fc42ad8ff49..5b89c0370f33 100644 --- a/samples/bpf/Makefile +++ b/samples/bpf/Makefile @@ -38,6 +38,8 @@ tprogs-y += tc_l2_redirect tprogs-y += lwt_len_hist tprogs-y += xdp_tx_iptunnel tprogs-y += test_map_in_map +tprogs-y += per_socket_stats_example +tprogs-y += xdp_redirect tprogs-y += xdp_redirect_map tprogs-y += xdp_redirect_cpu tprogs-y += xdp_monitor @@ -196,7 +198,7 @@ endif TPROGCFLAGS_bpf_load.o += -Wno-unused-variable -TPROGS_LDLIBS += $(LIBBPF) -lelf +TPROGS_LDLIBS += $(LIBBPF) -lelf -lz TPROGLDLIBS_tracex4 += -lrt TPROGLDLIBS_trace_output += -lrt TPROGLDLIBS_map_perf_test += -lrt @@ -234,6 +236,7 @@ BTF_LLVM_PROBE := $(shell echo "int main() { return 0; }" | \ readelf -S ./llvm_btf_verify.o | grep BTF; \ /bin/rm -f ./llvm_btf_verify.o) +BPF_EXTRA_CFLAGS += -fno-stack-protector ifneq ($(BTF_LLVM_PROBE),) BPF_EXTRA_CFLAGS += -g else diff --git a/samples/bpf/xdp1_user.c b/samples/bpf/xdp1_user.c index 3e553eed95a7..38a8852cb57f 100644 --- a/samples/bpf/xdp1_user.c +++ b/samples/bpf/xdp1_user.c @@ -98,7 +98,7 @@ int main(int argc, char **argv) xdp_flags |= XDP_FLAGS_SKB_MODE; break; case 'N': - xdp_flags |= XDP_FLAGS_DRV_MODE; + /* default, set below */ break; case 'F': xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST; @@ -109,6 +109,9 @@ int main(int argc, char **argv) } } + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + if (optind == argc) { usage(basename(argv[0])); return 1; diff --git a/samples/bpf/xdp_adjust_tail_user.c b/samples/bpf/xdp_adjust_tail_user.c index d86e9ad0356b..008789eb6ada 100644 --- a/samples/bpf/xdp_adjust_tail_user.c +++ b/samples/bpf/xdp_adjust_tail_user.c @@ -120,7 +120,7 @@ int main(int argc, char **argv) xdp_flags |= XDP_FLAGS_SKB_MODE; break; case 'N': - xdp_flags |= XDP_FLAGS_DRV_MODE; + /* default, set below */ break; case 'F': xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST; @@ -132,6 +132,9 @@ int main(int argc, char **argv) opt_flags[opt] = 0; } + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + for (i = 0; i < strlen(optstr); i++) { if (opt_flags[(unsigned int)optstr[i]]) { fprintf(stderr, "Missing argument -%c\n", optstr[i]); diff --git a/samples/bpf/xdp_fwd_user.c b/samples/bpf/xdp_fwd_user.c index 97ff1dad7669..c30f9acfdb84 100644 --- a/samples/bpf/xdp_fwd_user.c +++ b/samples/bpf/xdp_fwd_user.c @@ -27,11 +27,13 @@ #include "libbpf.h" #include <bpf/bpf.h> +static __u32 xdp_flags = XDP_FLAGS_UPDATE_IF_NOEXIST; + static int do_attach(int idx, int prog_fd, int map_fd, const char *name) { int err; - err = bpf_set_link_xdp_fd(idx, prog_fd, 0); + err = bpf_set_link_xdp_fd(idx, prog_fd, xdp_flags); if (err < 0) { printf("ERROR: failed to attach program to %s\n", name); return err; @@ -49,7 +51,7 @@ static int do_detach(int idx, const char *name) { int err; - err = bpf_set_link_xdp_fd(idx, -1, 0); + err = bpf_set_link_xdp_fd(idx, -1, xdp_flags); if (err < 0) printf("ERROR: failed to detach program from %s\n", name); @@ -83,11 +85,17 @@ int main(int argc, char **argv) int attach = 1; int ret = 0; - while ((opt = getopt(argc, argv, ":dD")) != -1) { + while ((opt = getopt(argc, argv, ":dDSF")) != -1) { switch (opt) { case 'd': attach = 0; break; + case 'S': + xdp_flags |= XDP_FLAGS_SKB_MODE; + break; + case 'F': + xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST; + break; case 'D': prog_name = "xdp_fwd_direct"; break; @@ -97,6 +105,9 @@ int main(int argc, char **argv) } } + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + if (optind == argc) { usage(basename(argv[0])); return 1; diff --git a/samples/bpf/xdp_redirect_cpu_user.c b/samples/bpf/xdp_redirect_cpu_user.c index 0da6e9e7132e..79a2fb7d16cb 100644 --- a/samples/bpf/xdp_redirect_cpu_user.c +++ b/samples/bpf/xdp_redirect_cpu_user.c @@ -16,6 +16,10 @@ static const char *__doc__ = #include <getopt.h> #include <net/if.h> #include <time.h> +#include <linux/limits.h> + +#define __must_check +#include <linux/err.h> #include <arpa/inet.h> #include <linux/if_link.h> @@ -46,6 +50,10 @@ static int cpus_count_map_fd; static int cpus_iterator_map_fd; static int exception_cnt_map_fd; +#define NUM_TP 5 +struct bpf_link *tp_links[NUM_TP] = { 0 }; +static int tp_cnt = 0; + /* Exit return codes */ #define EXIT_OK 0 #define EXIT_FAIL 1 @@ -88,6 +96,10 @@ static void int_exit(int sig) printf("program on interface changed, not removing\n"); } } + /* Detach tracepoints */ + while (tp_cnt) + bpf_link__destroy(tp_links[--tp_cnt]); + exit(EXIT_OK); } @@ -588,23 +600,61 @@ static void stats_poll(int interval, bool use_separators, char *prog_name, free_stats_record(prev); } +static struct bpf_link * attach_tp(struct bpf_object *obj, + const char *tp_category, + const char* tp_name) +{ + struct bpf_program *prog; + struct bpf_link *link; + char sec_name[PATH_MAX]; + int len; + + len = snprintf(sec_name, PATH_MAX, "tracepoint/%s/%s", + tp_category, tp_name); + if (len < 0) + exit(EXIT_FAIL); + + prog = bpf_object__find_program_by_title(obj, sec_name); + if (!prog) { + fprintf(stderr, "ERR: finding progsec: %s\n", sec_name); + exit(EXIT_FAIL_BPF); + } + + link = bpf_program__attach_tracepoint(prog, tp_category, tp_name); + if (IS_ERR(link)) + exit(EXIT_FAIL_BPF); + + return link; +} + +static void init_tracepoints(struct bpf_object *obj) { + tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_redirect_err"); + tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_redirect_map_err"); + tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_exception"); + tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_cpumap_enqueue"); + tp_links[tp_cnt++] = attach_tp(obj, "xdp", "xdp_cpumap_kthread"); +} + static int init_map_fds(struct bpf_object *obj) { - cpu_map_fd = bpf_object__find_map_fd_by_name(obj, "cpu_map"); - rx_cnt_map_fd = bpf_object__find_map_fd_by_name(obj, "rx_cnt"); + /* Maps updated by tracepoints */ redirect_err_cnt_map_fd = bpf_object__find_map_fd_by_name(obj, "redirect_err_cnt"); + exception_cnt_map_fd = + bpf_object__find_map_fd_by_name(obj, "exception_cnt"); cpumap_enqueue_cnt_map_fd = bpf_object__find_map_fd_by_name(obj, "cpumap_enqueue_cnt"); cpumap_kthread_cnt_map_fd = bpf_object__find_map_fd_by_name(obj, "cpumap_kthread_cnt"); + + /* Maps used by XDP */ + rx_cnt_map_fd = bpf_object__find_map_fd_by_name(obj, "rx_cnt"); + cpu_map_fd = bpf_object__find_map_fd_by_name(obj, "cpu_map"); cpus_available_map_fd = bpf_object__find_map_fd_by_name(obj, "cpus_available"); cpus_count_map_fd = bpf_object__find_map_fd_by_name(obj, "cpus_count"); cpus_iterator_map_fd = bpf_object__find_map_fd_by_name(obj, "cpus_iterator"); - exception_cnt_map_fd = - bpf_object__find_map_fd_by_name(obj, "exception_cnt"); if (cpu_map_fd < 0 || rx_cnt_map_fd < 0 || redirect_err_cnt_map_fd < 0 || cpumap_enqueue_cnt_map_fd < 0 || @@ -662,6 +712,7 @@ int main(int argc, char **argv) strerror(errno)); return EXIT_FAIL; } + init_tracepoints(obj); if (init_map_fds(obj) < 0) { fprintf(stderr, "bpf_object__find_map_fd_by_name failed\n"); return EXIT_FAIL; @@ -728,6 +779,10 @@ int main(int argc, char **argv) return EXIT_FAIL_OPTION; } } + + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + /* Required option */ if (ifindex == -1) { fprintf(stderr, "ERR: required option --dev missing\n"); diff --git a/samples/bpf/xdp_redirect_map_user.c b/samples/bpf/xdp_redirect_map_user.c index f70ee33907fd..cc840661faab 100644 --- a/samples/bpf/xdp_redirect_map_user.c +++ b/samples/bpf/xdp_redirect_map_user.c @@ -116,7 +116,7 @@ int main(int argc, char **argv) xdp_flags |= XDP_FLAGS_SKB_MODE; break; case 'N': - xdp_flags |= XDP_FLAGS_DRV_MODE; + /* default, set below */ break; case 'F': xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST; @@ -127,6 +127,9 @@ int main(int argc, char **argv) } } + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + if (optind == argc) { printf("usage: %s <IFNAME|IFINDEX>_IN <IFNAME|IFINDEX>_OUT\n", argv[0]); return 1; diff --git a/samples/bpf/xdp_redirect_user.c b/samples/bpf/xdp_redirect_user.c index 5440cd620607..71dff8e3382a 100644 --- a/samples/bpf/xdp_redirect_user.c +++ b/samples/bpf/xdp_redirect_user.c @@ -117,7 +117,7 @@ int main(int argc, char **argv) xdp_flags |= XDP_FLAGS_SKB_MODE; break; case 'N': - xdp_flags |= XDP_FLAGS_DRV_MODE; + /* default, set below */ break; case 'F': xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST; @@ -128,6 +128,9 @@ int main(int argc, char **argv) } } + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + if (optind == argc) { printf("usage: %s <IFNAME|IFINDEX>_IN <IFNAME|IFINDEX>_OUT\n", argv[0]); return 1; diff --git a/samples/bpf/xdp_router_ipv4_user.c b/samples/bpf/xdp_router_ipv4_user.c index 1469b66ebad1..fef286c5add2 100644 --- a/samples/bpf/xdp_router_ipv4_user.c +++ b/samples/bpf/xdp_router_ipv4_user.c @@ -662,6 +662,9 @@ int main(int ac, char **argv) } } + if (!(flags & XDP_FLAGS_SKB_MODE)) + flags |= XDP_FLAGS_DRV_MODE; + if (optind == ac) { usage(basename(argv[0])); return 1; diff --git a/samples/bpf/xdp_rxq_info_user.c b/samples/bpf/xdp_rxq_info_user.c index 8fc3ad01de72..fc4983fd6959 100644 --- a/samples/bpf/xdp_rxq_info_user.c +++ b/samples/bpf/xdp_rxq_info_user.c @@ -551,6 +551,10 @@ int main(int argc, char **argv) return EXIT_FAIL_OPTION; } } + + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + /* Required option */ if (ifindex == -1) { fprintf(stderr, "ERR: required option --dev missing\n"); diff --git a/samples/bpf/xdp_sample_pkts_user.c b/samples/bpf/xdp_sample_pkts_user.c index a5760e8bf2c4..8c1af1b7372d 100644 --- a/samples/bpf/xdp_sample_pkts_user.c +++ b/samples/bpf/xdp_sample_pkts_user.c @@ -52,13 +52,13 @@ static int do_detach(int idx, const char *name) __u32 curr_prog_id = 0; int err = 0; - err = bpf_get_link_xdp_id(idx, &curr_prog_id, 0); + err = bpf_get_link_xdp_id(idx, &curr_prog_id, xdp_flags); if (err) { printf("bpf_get_link_xdp_id failed\n"); return err; } if (prog_id == curr_prog_id) { - err = bpf_set_link_xdp_fd(idx, -1, 0); + err = bpf_set_link_xdp_fd(idx, -1, xdp_flags); if (err < 0) printf("ERROR: failed to detach prog from %s\n", name); } else if (!curr_prog_id) { @@ -115,7 +115,7 @@ int main(int argc, char **argv) .prog_type = BPF_PROG_TYPE_XDP, }; struct perf_buffer_opts pb_opts = {}; - const char *optstr = "F"; + const char *optstr = "FS"; int prog_fd, map_fd, opt; struct bpf_object *obj; struct bpf_map *map; @@ -127,12 +127,18 @@ int main(int argc, char **argv) case 'F': xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST; break; + case 'S': + xdp_flags |= XDP_FLAGS_SKB_MODE; + break; default: usage(basename(argv[0])); return 1; } } + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + if (optind == argc) { usage(basename(argv[0])); return 1; diff --git a/samples/bpf/xdp_tx_iptunnel_user.c b/samples/bpf/xdp_tx_iptunnel_user.c index 2fe4c7f5ffe5..5f33b5530032 100644 --- a/samples/bpf/xdp_tx_iptunnel_user.c +++ b/samples/bpf/xdp_tx_iptunnel_user.c @@ -231,7 +231,7 @@ int main(int argc, char **argv) xdp_flags |= XDP_FLAGS_SKB_MODE; break; case 'N': - xdp_flags |= XDP_FLAGS_DRV_MODE; + /* default, set below */ break; case 'F': xdp_flags &= ~XDP_FLAGS_UPDATE_IF_NOEXIST; @@ -243,6 +243,9 @@ int main(int argc, char **argv) opt_flags[opt] = 0; } + if (!(xdp_flags & XDP_FLAGS_SKB_MODE)) + xdp_flags |= XDP_FLAGS_DRV_MODE; + for (i = 0; i < strlen(optstr); i++) { if (opt_flags[(unsigned int)optstr[i]]) { fprintf(stderr, "Missing argument -%c\n", optstr[i]); diff --git a/samples/bpf/xdpsock_user.c b/samples/bpf/xdpsock_user.c index a15480010828..d74c4c83fc93 100644 --- a/samples/bpf/xdpsock_user.c +++ b/samples/bpf/xdpsock_user.c @@ -10,6 +10,9 @@ #include <linux/if_link.h> #include <linux/if_xdp.h> #include <linux/if_ether.h> +#include <linux/ip.h> +#include <linux/udp.h> +#include <arpa/inet.h> #include <locale.h> #include <net/ethernet.h> #include <net/if.h> @@ -45,12 +48,14 @@ #endif #define NUM_FRAMES (4 * 1024) -#define BATCH_SIZE 64 +#define MIN_PKT_SIZE 64 #define DEBUG_HEXDUMP 0 typedef __u64 u64; typedef __u32 u32; +typedef __u16 u16; +typedef __u8 u8; static unsigned long prev_time; @@ -65,6 +70,13 @@ static u32 opt_xdp_flags = XDP_FLAGS_UPDATE_IF_NOEXIST; static const char *opt_if = ""; static int opt_ifindex; static int opt_queue; +static unsigned long opt_duration; +static unsigned long start_time; +static bool benchmark_done; +static u32 opt_batch_size = 64; +static int opt_pkt_count; +static u16 opt_pkt_size = MIN_PKT_SIZE; +static u32 opt_pkt_fill_pattern = 0x12345678; static int opt_poll; static int opt_interval = 1; static u32 opt_xdp_bind_flags = XDP_USE_NEED_WAKEUP; @@ -167,10 +179,21 @@ static void dump_stats(void) } } +static bool is_benchmark_done(void) +{ + if (opt_duration > 0) { + unsigned long dt = (get_nsecs() - start_time); + + if (dt >= opt_duration) + benchmark_done = true; + } + return benchmark_done; +} + static void *poller(void *arg) { (void)arg; - for (;;) { + while (!is_benchmark_done()) { sleep(opt_interval); dump_stats(); } @@ -196,6 +219,11 @@ static void remove_xdp_program(void) static void int_exit(int sig) { + benchmark_done = true; +} + +static void xdpsock_cleanup(void) +{ struct xsk_umem *umem = xsks[0]->umem->umem; int i; @@ -204,8 +232,6 @@ static void int_exit(int sig) xsk_socket__delete(xsks[i]->xsk); (void)xsk_umem__delete(umem); remove_xdp_program(); - - exit(EXIT_SUCCESS); } static void __exit_with_error(int error, const char *file, const char *func, @@ -220,13 +246,6 @@ static void __exit_with_error(int error, const char *file, const char *func, #define exit_with_error(error) __exit_with_error(error, __FILE__, __func__, \ __LINE__) - -static const char pkt_data[] = - "\x3c\xfd\xfe\x9e\x7f\x71\xec\xb1\xd7\x98\x3a\xc0\x08\x00\x45\x00" - "\x00\x2e\x00\x00\x00\x00\x40\x11\x88\x97\x05\x08\x07\x08\xc8\x14" - "\x1e\x04\x10\x92\x10\x92\x00\x1a\x6d\xa3\x34\x33\x1f\x69\x40\x6b" - "\x54\x59\xb6\x14\x2d\x11\x44\xbf\xaf\xd9\xbe\xaa"; - static void swap_mac_addresses(void *data) { struct ether_header *eth = (struct ether_header *)data; @@ -274,11 +293,243 @@ static void hex_dump(void *pkt, size_t length, u64 addr) printf("\n"); } -static size_t gen_eth_frame(struct xsk_umem_info *umem, u64 addr) +static void *memset32_htonl(void *dest, u32 val, u32 size) +{ + u32 *ptr = (u32 *)dest; + int i; + + val = htonl(val); + + for (i = 0; i < (size & (~0x3)); i += 4) + ptr[i >> 2] = val; + + for (; i < size; i++) + ((char *)dest)[i] = ((char *)&val)[i & 3]; + + return dest; +} + +/* + * This function code has been taken from + * Linux kernel lib/checksum.c + */ +static inline unsigned short from32to16(unsigned int x) +{ + /* add up 16-bit and 16-bit for 16+c bit */ + x = (x & 0xffff) + (x >> 16); + /* add up carry.. */ + x = (x & 0xffff) + (x >> 16); + return x; +} + +/* + * This function code has been taken from + * Linux kernel lib/checksum.c + */ +static unsigned int do_csum(const unsigned char *buff, int len) +{ + unsigned int result = 0; + int odd; + + if (len <= 0) + goto out; + odd = 1 & (unsigned long)buff; + if (odd) { +#ifdef __LITTLE_ENDIAN + result += (*buff << 8); +#else + result = *buff; +#endif + len--; + buff++; + } + if (len >= 2) { + if (2 & (unsigned long)buff) { + result += *(unsigned short *)buff; + len -= 2; + buff += 2; + } + if (len >= 4) { + const unsigned char *end = buff + + ((unsigned int)len & ~3); + unsigned int carry = 0; + + do { + unsigned int w = *(unsigned int *)buff; + + buff += 4; + result += carry; + result += w; + carry = (w > result); + } while (buff < end); + result += carry; + result = (result & 0xffff) + (result >> 16); + } + if (len & 2) { + result += *(unsigned short *)buff; + buff += 2; + } + } + if (len & 1) +#ifdef __LITTLE_ENDIAN + result += *buff; +#else + result += (*buff << 8); +#endif + result = from32to16(result); + if (odd) + result = ((result >> 8) & 0xff) | ((result & 0xff) << 8); +out: + return result; +} + +__sum16 ip_fast_csum(const void *iph, unsigned int ihl); + +/* + * This is a version of ip_compute_csum() optimized for IP headers, + * which always checksum on 4 octet boundaries. + * This function code has been taken from + * Linux kernel lib/checksum.c + */ +__sum16 ip_fast_csum(const void *iph, unsigned int ihl) +{ + return (__force __sum16)~do_csum(iph, ihl * 4); +} + +/* + * Fold a partial checksum + * This function code has been taken from + * Linux kernel include/asm-generic/checksum.h + */ +static inline __sum16 csum_fold(__wsum csum) +{ + u32 sum = (__force u32)csum; + + sum = (sum & 0xffff) + (sum >> 16); + sum = (sum & 0xffff) + (sum >> 16); + return (__force __sum16)~sum; +} + +/* + * This function code has been taken from + * Linux kernel lib/checksum.c + */ +static inline u32 from64to32(u64 x) +{ + /* add up 32-bit and 32-bit for 32+c bit */ + x = (x & 0xffffffff) + (x >> 32); + /* add up carry.. */ + x = (x & 0xffffffff) + (x >> 32); + return (u32)x; +} + +__wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr, + __u32 len, __u8 proto, __wsum sum); + +/* + * This function code has been taken from + * Linux kernel lib/checksum.c + */ +__wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr, + __u32 len, __u8 proto, __wsum sum) +{ + unsigned long long s = (__force u32)sum; + + s += (__force u32)saddr; + s += (__force u32)daddr; +#ifdef __BIG_ENDIAN__ + s += proto + len; +#else + s += (proto + len) << 8; +#endif + return (__force __wsum)from64to32(s); +} + +/* + * This function has been taken from + * Linux kernel include/asm-generic/checksum.h + */ +static inline __sum16 +csum_tcpudp_magic(__be32 saddr, __be32 daddr, __u32 len, + __u8 proto, __wsum sum) +{ + return csum_fold(csum_tcpudp_nofold(saddr, daddr, len, proto, sum)); +} + +static inline u16 udp_csum(u32 saddr, u32 daddr, u32 len, + u8 proto, u16 *udp_pkt) +{ + u32 csum = 0; + u32 cnt = 0; + + /* udp hdr and data */ + for (; cnt < len; cnt += 2) + csum += udp_pkt[cnt >> 1]; + + return csum_tcpudp_magic(saddr, daddr, len, proto, csum); +} + +#define ETH_FCS_SIZE 4 + +#define PKT_HDR_SIZE (sizeof(struct ethhdr) + sizeof(struct iphdr) + \ + sizeof(struct udphdr)) + +#define PKT_SIZE (opt_pkt_size - ETH_FCS_SIZE) +#define IP_PKT_SIZE (PKT_SIZE - sizeof(struct ethhdr)) +#define UDP_PKT_SIZE (IP_PKT_SIZE - sizeof(struct iphdr)) +#define UDP_PKT_DATA_SIZE (UDP_PKT_SIZE - sizeof(struct udphdr)) + +static u8 pkt_data[XSK_UMEM__DEFAULT_FRAME_SIZE]; + +static void gen_eth_hdr_data(void) +{ + struct udphdr *udp_hdr = (struct udphdr *)(pkt_data + + sizeof(struct ethhdr) + + sizeof(struct iphdr)); + struct iphdr *ip_hdr = (struct iphdr *)(pkt_data + + sizeof(struct ethhdr)); + struct ethhdr *eth_hdr = (struct ethhdr *)pkt_data; + + /* ethernet header */ + memcpy(eth_hdr->h_dest, "\x3c\xfd\xfe\x9e\x7f\x71", ETH_ALEN); + memcpy(eth_hdr->h_source, "\xec\xb1\xd7\x98\x3a\xc0", ETH_ALEN); + eth_hdr->h_proto = htons(ETH_P_IP); + + /* IP header */ + ip_hdr->version = IPVERSION; + ip_hdr->ihl = 0x5; /* 20 byte header */ + ip_hdr->tos = 0x0; + ip_hdr->tot_len = htons(IP_PKT_SIZE); + ip_hdr->id = 0; + ip_hdr->frag_off = 0; + ip_hdr->ttl = IPDEFTTL; + ip_hdr->protocol = IPPROTO_UDP; + ip_hdr->saddr = htonl(0x0a0a0a10); + ip_hdr->daddr = htonl(0x0a0a0a20); + + /* IP header checksum */ + ip_hdr->check = 0; + ip_hdr->check = ip_fast_csum((const void *)ip_hdr, ip_hdr->ihl); + + /* UDP header */ + udp_hdr->source = htons(0x1000); + udp_hdr->dest = htons(0x1000); + udp_hdr->len = htons(UDP_PKT_SIZE); + + /* UDP data */ + memset32_htonl(pkt_data + PKT_HDR_SIZE, opt_pkt_fill_pattern, + UDP_PKT_DATA_SIZE); + + /* UDP header checksum */ + udp_hdr->check = 0; + udp_hdr->check = udp_csum(ip_hdr->saddr, ip_hdr->daddr, UDP_PKT_SIZE, + IPPROTO_UDP, (u16 *)udp_hdr); +} + +static void gen_eth_frame(struct xsk_umem_info *umem, u64 addr) { memcpy(xsk_umem__get_data(umem->buffer, addr), pkt_data, - sizeof(pkt_data) - 1); - return sizeof(pkt_data) - 1; + PKT_SIZE); } static struct xsk_umem_info *xsk_configure_umem(void *buffer, u64 size) @@ -375,6 +626,11 @@ static struct option long_options[] = { {"unaligned", no_argument, 0, 'u'}, {"shared-umem", no_argument, 0, 'M'}, {"force", no_argument, 0, 'F'}, + {"duration", required_argument, 0, 'd'}, + {"batch-size", required_argument, 0, 'b'}, + {"tx-pkt-count", required_argument, 0, 'C'}, + {"tx-pkt-size", required_argument, 0, 's'}, + {"tx-pkt-pattern", required_argument, 0, 'P'}, {0, 0, 0, 0} }; @@ -399,8 +655,21 @@ static void usage(const char *prog) " -u, --unaligned Enable unaligned chunk placement\n" " -M, --shared-umem Enable XDP_SHARED_UMEM\n" " -F, --force Force loading the XDP prog\n" + " -d, --duration=n Duration in secs to run command.\n" + " Default: forever.\n" + " -b, --batch-size=n Batch size for sending or receiving\n" + " packets. Default: %d\n" + " -C, --tx-pkt-count=n Number of packets to send.\n" + " Default: Continuous packets.\n" + " -s, --tx-pkt-size=n Transmit packet size.\n" + " (Default: %d bytes)\n" + " Min size: %d, Max size %d.\n" + " -P, --tx-pkt-pattern=nPacket fill pattern. Default: 0x%x\n" "\n"; - fprintf(stderr, str, prog, XSK_UMEM__DEFAULT_FRAME_SIZE); + fprintf(stderr, str, prog, XSK_UMEM__DEFAULT_FRAME_SIZE, + opt_batch_size, MIN_PKT_SIZE, MIN_PKT_SIZE, + XSK_UMEM__DEFAULT_FRAME_SIZE, opt_pkt_fill_pattern); + exit(EXIT_FAILURE); } @@ -411,7 +680,7 @@ static void parse_command_line(int argc, char **argv) opterr = 0; for (;;) { - c = getopt_long(argc, argv, "Frtli:q:psSNn:czf:muM", + c = getopt_long(argc, argv, "Frtli:q:pSNn:czf:muMd:b:C:s:P:", long_options, &option_index); if (c == -1) break; @@ -440,7 +709,7 @@ static void parse_command_line(int argc, char **argv) opt_xdp_bind_flags |= XDP_COPY; break; case 'N': - opt_xdp_flags |= XDP_FLAGS_DRV_MODE; + /* default, set below */ break; case 'n': opt_interval = atoi(optarg); @@ -469,11 +738,37 @@ static void parse_command_line(int argc, char **argv) case 'M': opt_num_xsks = MAX_SOCKS; break; + case 'd': + opt_duration = atoi(optarg); + opt_duration *= 1000000000; + break; + case 'b': + opt_batch_size = atoi(optarg); + break; + case 'C': + opt_pkt_count = atoi(optarg); + break; + case 's': + opt_pkt_size = atoi(optarg); + if (opt_pkt_size > (XSK_UMEM__DEFAULT_FRAME_SIZE) || + opt_pkt_size < MIN_PKT_SIZE) { + fprintf(stderr, + "ERROR: Invalid frame size %d\n", + opt_pkt_size); + usage(basename(argv[0])); + } + break; + case 'P': + opt_pkt_fill_pattern = strtol(optarg, NULL, 16); + break; default: usage(basename(argv[0])); } } + if (!(opt_xdp_flags & XDP_FLAGS_SKB_MODE)) + opt_xdp_flags |= XDP_FLAGS_DRV_MODE; + opt_ifindex = if_nametoindex(opt_if); if (!opt_ifindex) { fprintf(stderr, "ERROR: interface \"%s\" does not exist\n", @@ -513,7 +808,7 @@ static inline void complete_tx_l2fwd(struct xsk_socket_info *xsk, if (!opt_need_wakeup || xsk_ring_prod__needs_wakeup(&xsk->tx)) kick_tx(xsk); - ndescs = (xsk->outstanding_tx > BATCH_SIZE) ? BATCH_SIZE : + ndescs = (xsk->outstanding_tx > opt_batch_size) ? opt_batch_size : xsk->outstanding_tx; /* re-add completed Tx buffers */ @@ -542,7 +837,8 @@ static inline void complete_tx_l2fwd(struct xsk_socket_info *xsk, } } -static inline void complete_tx_only(struct xsk_socket_info *xsk) +static inline void complete_tx_only(struct xsk_socket_info *xsk, + int batch_size) { unsigned int rcvd; u32 idx; @@ -553,7 +849,7 @@ static inline void complete_tx_only(struct xsk_socket_info *xsk) if (!opt_need_wakeup || xsk_ring_prod__needs_wakeup(&xsk->tx)) kick_tx(xsk); - rcvd = xsk_ring_cons__peek(&xsk->umem->cq, BATCH_SIZE, &idx); + rcvd = xsk_ring_cons__peek(&xsk->umem->cq, batch_size, &idx); if (rcvd > 0) { xsk_ring_cons__release(&xsk->umem->cq, rcvd); xsk->outstanding_tx -= rcvd; @@ -567,7 +863,7 @@ static void rx_drop(struct xsk_socket_info *xsk, struct pollfd *fds) u32 idx_rx = 0, idx_fq = 0; int ret; - rcvd = xsk_ring_cons__peek(&xsk->rx, BATCH_SIZE, &idx_rx); + rcvd = xsk_ring_cons__peek(&xsk->rx, opt_batch_size, &idx_rx); if (!rcvd) { if (xsk_ring_prod__needs_wakeup(&xsk->umem->fq)) ret = poll(fds, num_socks, opt_timeout); @@ -619,36 +915,68 @@ static void rx_drop_all(void) for (i = 0; i < num_socks; i++) rx_drop(xsks[i], fds); + + if (benchmark_done) + break; } } -static void tx_only(struct xsk_socket_info *xsk, u32 frame_nb) +static void tx_only(struct xsk_socket_info *xsk, u32 frame_nb, int batch_size) { u32 idx; + unsigned int i; - if (xsk_ring_prod__reserve(&xsk->tx, BATCH_SIZE, &idx) == BATCH_SIZE) { - unsigned int i; - - for (i = 0; i < BATCH_SIZE; i++) { - xsk_ring_prod__tx_desc(&xsk->tx, idx + i)->addr = - (frame_nb + i) << XSK_UMEM__DEFAULT_FRAME_SHIFT; - xsk_ring_prod__tx_desc(&xsk->tx, idx + i)->len = - sizeof(pkt_data) - 1; - } + while (xsk_ring_prod__reserve(&xsk->tx, batch_size, &idx) < + batch_size) { + complete_tx_only(xsk, batch_size); + } - xsk_ring_prod__submit(&xsk->tx, BATCH_SIZE); - xsk->outstanding_tx += BATCH_SIZE; - frame_nb += BATCH_SIZE; - frame_nb %= NUM_FRAMES; + for (i = 0; i < batch_size; i++) { + struct xdp_desc *tx_desc = xsk_ring_prod__tx_desc(&xsk->tx, + idx + i); + tx_desc->addr = (frame_nb + i) << XSK_UMEM__DEFAULT_FRAME_SHIFT; + tx_desc->len = PKT_SIZE; } - complete_tx_only(xsk); + xsk_ring_prod__submit(&xsk->tx, batch_size); + xsk->outstanding_tx += batch_size; + frame_nb += batch_size; + frame_nb %= NUM_FRAMES; + complete_tx_only(xsk, batch_size); +} + +static inline int get_batch_size(int pkt_cnt) +{ + if (!opt_pkt_count) + return opt_batch_size; + + if (pkt_cnt + opt_batch_size <= opt_pkt_count) + return opt_batch_size; + + return opt_pkt_count - pkt_cnt; +} + +static void complete_tx_only_all(void) +{ + bool pending; + int i; + + do { + pending = false; + for (i = 0; i < num_socks; i++) { + if (xsks[i]->outstanding_tx) { + complete_tx_only(xsks[i], opt_batch_size); + pending = !!xsks[i]->outstanding_tx; + } + } + } while (pending); } static void tx_only_all(void) { struct pollfd fds[MAX_SOCKS] = {}; u32 frame_nb[MAX_SOCKS] = {}; + int pkt_cnt = 0; int i, ret; for (i = 0; i < num_socks; i++) { @@ -656,7 +984,9 @@ static void tx_only_all(void) fds[0].events = POLLOUT; } - for (;;) { + while ((opt_pkt_count && pkt_cnt < opt_pkt_count) || !opt_pkt_count) { + int batch_size = get_batch_size(pkt_cnt); + if (opt_poll) { ret = poll(fds, num_socks, opt_timeout); if (ret <= 0) @@ -667,8 +997,16 @@ static void tx_only_all(void) } for (i = 0; i < num_socks; i++) - tx_only(xsks[i], frame_nb[i]); + tx_only(xsks[i], frame_nb[i], batch_size); + + pkt_cnt += batch_size; + + if (benchmark_done) + break; } + + if (opt_pkt_count) + complete_tx_only_all(); } static void l2fwd(struct xsk_socket_info *xsk, struct pollfd *fds) @@ -679,7 +1017,7 @@ static void l2fwd(struct xsk_socket_info *xsk, struct pollfd *fds) complete_tx_l2fwd(xsk, fds); - rcvd = xsk_ring_cons__peek(&xsk->rx, BATCH_SIZE, &idx_rx); + rcvd = xsk_ring_cons__peek(&xsk->rx, opt_batch_size, &idx_rx); if (!rcvd) { if (xsk_ring_prod__needs_wakeup(&xsk->umem->fq)) ret = poll(fds, num_socks, opt_timeout); @@ -736,6 +1074,9 @@ static void l2fwd_all(void) for (i = 0; i < num_socks; i++) l2fwd(xsks[i], fds); + + if (benchmark_done) + break; } } @@ -831,9 +1172,12 @@ int main(int argc, char **argv) for (i = 0; i < opt_num_xsks; i++) xsks[num_socks++] = xsk_configure_socket(umem, rx, tx); - if (opt_bench == BENCH_TXONLY) + if (opt_bench == BENCH_TXONLY) { + gen_eth_hdr_data(); + for (i = 0; i < NUM_FRAMES; i++) gen_eth_frame(umem, i * opt_xsk_frame_size); + } if (opt_num_xsks > 1 && opt_bench != BENCH_TXONLY) enter_xsks_into_map(obj); @@ -849,6 +1193,7 @@ int main(int argc, char **argv) exit_with_error(ret); prev_time = get_nsecs(); + start_time = prev_time; if (opt_bench == BENCH_RXDROP) rx_drop_all(); @@ -857,5 +1202,11 @@ int main(int argc, char **argv) else l2fwd_all(); + benchmark_done = true; + + pthread_join(pt, NULL); + + xdpsock_cleanup(); + return 0; } |