[mirror_ubuntu-eoan-kernel.git] / samples / bpf / bpf_load.c

// SPDX-License-Identifier: GPL-2.0
#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <libelf.h>
#include <gelf.h>
#include <errno.h>
#include <unistd.h>
#include <string.h>
#include <stdbool.h>
#include <stdlib.h>
#include <linux/bpf.h>
#include <linux/filter.h>
#include <linux/perf_event.h>
#include <linux/netlink.h>
#include <linux/rtnetlink.h>
#include <linux/types.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <sys/syscall.h>
#include <sys/ioctl.h>
#include <sys/mman.h>
#include <poll.h>
#include <ctype.h>
#include <assert.h>
#include "libbpf.h"
#include "bpf_load.h"
#include "perf-sys.h"

#define DEBUGFS "/sys/kernel/debug/tracing/"

static char license[128];
static int kern_version;
static bool processed_sec[128];
char bpf_log_buf[BPF_LOG_BUF_SIZE];
int map_fd[MAX_MAPS];
int prog_fd[MAX_PROGS];
int event_fd[MAX_PROGS];
int prog_cnt;
int prog_array_fd = -1;

struct bpf_map_data map_data[MAX_MAPS];
int map_data_count = 0;

static int populate_prog_array(const char *event, int prog_fd)
{
	int ind = atoi(event), err;

	err = bpf_map_update_elem(prog_array_fd, &ind, &prog_fd, BPF_ANY);
	if (err < 0) {
		printf("failed to store prog_fd in prog_array\n");
		return -1;
	}
	return 0;
}

static int load_and_attach(const char *event, struct bpf_insn *prog, int size)
{
	bool is_socket = strncmp(event, "socket", 6) == 0;
	bool is_kprobe = strncmp(event, "kprobe/", 7) == 0;
	bool is_kretprobe = strncmp(event, "kretprobe/", 10) == 0;
	bool is_tracepoint = strncmp(event, "tracepoint/", 11) == 0;
	bool is_xdp = strncmp(event, "xdp", 3) == 0;
	bool is_perf_event = strncmp(event, "perf_event", 10) == 0;
	bool is_cgroup_skb = strncmp(event, "cgroup/skb", 10) == 0;
	bool is_cgroup_sk = strncmp(event, "cgroup/sock", 11) == 0;
	bool is_sockops = strncmp(event, "sockops", 7) == 0;
	bool is_sk_skb = strncmp(event, "sk_skb", 6) == 0;
	bool is_sk_msg = strncmp(event, "sk_msg", 6) == 0;
	size_t insns_cnt = size / sizeof(struct bpf_insn);
	enum bpf_prog_type prog_type;
	char buf[256];
	int fd, efd, err, id;
	struct perf_event_attr attr = {};

	attr.type = PERF_TYPE_TRACEPOINT;
	attr.sample_type = PERF_SAMPLE_RAW;
	attr.sample_period = 1;
	attr.wakeup_events = 1;

	if (is_socket) {
		prog_type = BPF_PROG_TYPE_SOCKET_FILTER;
	} else if (is_kprobe || is_kretprobe) {
		prog_type = BPF_PROG_TYPE_KPROBE;
	} else if (is_tracepoint) {
		prog_type = BPF_PROG_TYPE_TRACEPOINT;
	} else if (is_xdp) {
		prog_type = BPF_PROG_TYPE_XDP;
	} else if (is_perf_event) {
		prog_type = BPF_PROG_TYPE_PERF_EVENT;
	} else if (is_cgroup_skb) {
		prog_type = BPF_PROG_TYPE_CGROUP_SKB;
	} else if (is_cgroup_sk) {
		prog_type = BPF_PROG_TYPE_CGROUP_SOCK;
	} else if (is_sockops) {
		prog_type = BPF_PROG_TYPE_SOCK_OPS;
	} else if (is_sk_skb) {
		prog_type = BPF_PROG_TYPE_SK_SKB;
	} else if (is_sk_msg) {
		prog_type = BPF_PROG_TYPE_SK_MSG;
	} else {
		printf("Unknown event '%s'\n", event);
		return -1;
	}

	fd = bpf_load_program(prog_type, prog, insns_cnt, license, kern_version,
			      bpf_log_buf, BPF_LOG_BUF_SIZE);
	if (fd < 0) {
		printf("bpf_load_program() err=%d\n%s", errno, bpf_log_buf);
		return -1;
	}

	prog_fd[prog_cnt++] = fd;

	if (is_xdp || is_perf_event || is_cgroup_skb || is_cgroup_sk)
		return 0;

	if (is_socket || is_sockops || is_sk_skb || is_sk_msg) {
		if (is_socket)
			event += 6;
		else
			event += 7;
		if (*event != '/')
			return 0;
		event++;
		if (!isdigit(*event)) {
			printf("invalid prog number\n");
			return -1;
		}
		return populate_prog_array(event, fd);
	}

	if (is_kprobe || is_kretprobe) {
		if (is_kprobe)
			event += 7;
		else
			event += 10;

		if (*event == 0) {
			printf("event name cannot be empty\n");
			return -1;
		}

		if (isdigit(*event))
			return populate_prog_array(event, fd);

		snprintf(buf, sizeof(buf),
			 "echo '%c:%s %s' >> /sys/kernel/debug/tracing/kprobe_events",
			 is_kprobe ? 'p' : 'r', event, event);
		err = system(buf);
		if (err < 0) {
			printf("failed to create kprobe '%s' error '%s'\n",
			       event, strerror(errno));
			return -1;
		}

		strcpy(buf, DEBUGFS);
		strcat(buf, "events/kprobes/");
		strcat(buf, event);
		strcat(buf, "/id");
	} else if (is_tracepoint) {
		event += 11;

		if (*event == 0) {
			printf("event name cannot be empty\n");
			return -1;
		}
		strcpy(buf, DEBUGFS);
		strcat(buf, "events/");
		strcat(buf, event);
		strcat(buf, "/id");
	}

	efd = open(buf, O_RDONLY, 0);
	if (efd < 0) {
		printf("failed to open event %s\n", event);
		return -1;
	}

	err = read(efd, buf, sizeof(buf));
	if (err < 0 || err >= sizeof(buf)) {
		printf("read from '%s' failed '%s'\n", event, strerror(errno));
		return -1;
	}

	close(efd);

	buf[err] = 0;
	id = atoi(buf);
	attr.config = id;

	efd = sys_perf_event_open(&attr, -1/*pid*/, 0/*cpu*/, -1/*group_fd*/, 0);
	if (efd < 0) {
		printf("event %d fd %d err %s\n", id, efd, strerror(errno));
		return -1;
	}
	event_fd[prog_cnt - 1] = efd;
	err = ioctl(efd, PERF_EVENT_IOC_ENABLE, 0);
	if (err < 0) {
		printf("ioctl PERF_EVENT_IOC_ENABLE failed err %s\n",
		       strerror(errno));
		return -1;
	}
	err = ioctl(efd, PERF_EVENT_IOC_SET_BPF, fd);
	if (err < 0) {
		printf("ioctl PERF_EVENT_IOC_SET_BPF failed err %s\n",
		       strerror(errno));
		return -1;
	}

	return 0;
}

static int load_maps(struct bpf_map_data *maps, int nr_maps,
		     fixup_map_cb fixup_map)
{
	int i, numa_node;

	for (i = 0; i < nr_maps; i++) {
		if (fixup_map) {
			fixup_map(&maps[i], i);
			/* Allow userspace to assign map FD prior to creation */
			if (maps[i].fd != -1) {
				map_fd[i] = maps[i].fd;
				continue;
			}
		}

		numa_node = maps[i].def.map_flags & BPF_F_NUMA_NODE ?
			maps[i].def.numa_node : -1;

		if (maps[i].def.type == BPF_MAP_TYPE_ARRAY_OF_MAPS ||
		    maps[i].def.type == BPF_MAP_TYPE_HASH_OF_MAPS) {
			int inner_map_fd = map_fd[maps[i].def.inner_map_idx];

			map_fd[i] = bpf_create_map_in_map_node(maps[i].def.type,
							maps[i].name,
							maps[i].def.key_size,
							inner_map_fd,
							maps[i].def.max_entries,
							maps[i].def.map_flags,
							numa_node);
		} else {
			map_fd[i] = bpf_create_map_node(maps[i].def.type,
							maps[i].name,
							maps[i].def.key_size,
							maps[i].def.value_size,
							maps[i].def.max_entries,
							maps[i].def.map_flags,
							numa_node);
		}
		if (map_fd[i] < 0) {
			printf("failed to create a map: %d %s\n",
			       errno, strerror(errno));
			return 1;
		}
		maps[i].fd = map_fd[i];

		if (maps[i].def.type == BPF_MAP_TYPE_PROG_ARRAY)
			prog_array_fd = map_fd[i];
	}
	return 0;
}

static int get_sec(Elf *elf, int i, GElf_Ehdr *ehdr, char **shname,
		   GElf_Shdr *shdr, Elf_Data **data)
{
	Elf_Scn *scn;

	scn = elf_getscn(elf, i);
	if (!scn)
		return 1;

	if (gelf_getshdr(scn, shdr) != shdr)
		return 2;

	*shname = elf_strptr(elf, ehdr->e_shstrndx, shdr->sh_name);
	if (!*shname || !shdr->sh_size)
		return 3;

	*data = elf_getdata(scn, 0);
	if (!*data || elf_getdata(scn, *data) != NULL)
		return 4;

	return 0;
}

static int parse_relo_and_apply(Elf_Data *data, Elf_Data *symbols,
				GElf_Shdr *shdr, struct bpf_insn *insn,
				struct bpf_map_data *maps, int nr_maps)
{
	int i, nrels;

	nrels = shdr->sh_size / shdr->sh_entsize;

	for (i = 0; i < nrels; i++) {
		GElf_Sym sym;
		GElf_Rel rel;
		unsigned int insn_idx;
		bool match = false;
		int j, map_idx;

		gelf_getrel(data, i, &rel);

		insn_idx = rel.r_offset / sizeof(struct bpf_insn);

		gelf_getsym(symbols, GELF_R_SYM(rel.r_info), &sym);

		if (insn[insn_idx].code != (BPF_LD | BPF_IMM | BPF_DW)) {
			printf("invalid relo for insn[%d].code 0x%x\n",
			       insn_idx, insn[insn_idx].code);
			return 1;
		}
		insn[insn_idx].src_reg = BPF_PSEUDO_MAP_FD;

		/* Match FD relocation against recorded map_data[] offset */
		for (map_idx = 0; map_idx < nr_maps; map_idx++) {
			if (maps[map_idx].elf_offset == sym.st_value) {
				match = true;
				break;
			}
		}
		if (match) {
			insn[insn_idx].imm = maps[map_idx].fd;
		} else {
			printf("invalid relo for insn[%d] no map_data match\n",
			       insn_idx);
			return 1;
		}
	}

	return 0;
}

static int cmp_symbols(const void *l, const void *r)
{
	const GElf_Sym *lsym = (const GElf_Sym *)l;
	const GElf_Sym *rsym = (const GElf_Sym *)r;

	if (lsym->st_value < rsym->st_value)
		return -1;
	else if (lsym->st_value > rsym->st_value)
		return 1;
	else
		return 0;
}

static int load_elf_maps_section(struct bpf_map_data *maps, int maps_shndx,
				 Elf *elf, Elf_Data *symbols, int strtabidx)
{
	int map_sz_elf, map_sz_copy;
	bool validate_zero = false;
	Elf_Data *data_maps;
	int i, nr_maps;
	GElf_Sym *sym;
	Elf_Scn *scn;
	int copy_sz;

	if (maps_shndx < 0)
		return -EINVAL;
	if (!symbols)
		return -EINVAL;

	/* Get data for maps section via elf index */
	scn = elf_getscn(elf, maps_shndx);
	if (scn)
		data_maps = elf_getdata(scn, NULL);
	if (!scn || !data_maps) {
		printf("Failed to get Elf_Data from maps section %d\n",
		       maps_shndx);
		return -EINVAL;
	}

	/* For each map get corrosponding symbol table entry */
	sym = calloc(MAX_MAPS+1, sizeof(GElf_Sym));
	for (i = 0, nr_maps = 0; i < symbols->d_size / sizeof(GElf_Sym); i++) {
		assert(nr_maps < MAX_MAPS+1);
		if (!gelf_getsym(symbols, i, &sym[nr_maps]))
			continue;
		if (sym[nr_maps].st_shndx != maps_shndx)
			continue;
		/* Only increment iif maps section */
		nr_maps++;
	}

	/* Align to map_fd[] order, via sort on offset in sym.st_value */
	qsort(sym, nr_maps, sizeof(GElf_Sym), cmp_symbols);

	/* Keeping compatible with ELF maps section changes
	 * ------------------------------------------------
	 * The program size of struct bpf_map_def is known by loader
	 * code, but struct stored in ELF file can be different.
	 *
	 * Unfortunately sym[i].st_size is zero.  To calculate the
	 * struct size stored in the ELF file, assume all struct have
	 * the same size, and simply divide with number of map
	 * symbols.
	 */
	map_sz_elf = data_maps->d_size / nr_maps;
	map_sz_copy = sizeof(struct bpf_map_def);
	if (map_sz_elf < map_sz_copy) {
		/*
		 * Backward compat, loading older ELF file with
		 * smaller struct, keeping remaining bytes zero.
		 */
		map_sz_copy = map_sz_elf;
	} else if (map_sz_elf > map_sz_copy) {
		/*
		 * Forward compat, loading newer ELF file with larger
		 * struct with unknown features. Assume zero means
		 * feature not used.  Thus, validate rest of struct
		 * data is zero.
		 */
		validate_zero = true;
	}

	/* Memcpy relevant part of ELF maps data to loader maps */
	for (i = 0; i < nr_maps; i++) {
		unsigned char *addr, *end;
		struct bpf_map_def *def;
		const char *map_name;
		size_t offset;

		map_name = elf_strptr(elf, strtabidx, sym[i].st_name);
		maps[i].name = strdup(map_name);
		if (!maps[i].name) {
			printf("strdup(%s): %s(%d)\n", map_name,
			       strerror(errno), errno);
			free(sym);
			return -errno;
		}

		/* Symbol value is offset into ELF maps section data area */
		offset = sym[i].st_value;
		def = (struct bpf_map_def *)(data_maps->d_buf + offset);
		maps[i].elf_offset = offset;
		memset(&maps[i].def, 0, sizeof(struct bpf_map_def));
		memcpy(&maps[i].def, def, map_sz_copy);

		/* Verify no newer features were requested */
		if (validate_zero) {
			addr = (unsigned char*) def + map_sz_copy;
			end  = (unsigned char*) def + map_sz_elf;
			for (; addr < end; addr++) {
				if (*addr != 0) {
					free(sym);
					return -EFBIG;
				}
			}
		}
	}

	free(sym);
	return nr_maps;
}

static int do_load_bpf_file(const char *path, fixup_map_cb fixup_map)
{
	int fd, i, ret, maps_shndx = -1, strtabidx = -1;
	Elf *elf;
	GElf_Ehdr ehdr;
	GElf_Shdr shdr, shdr_prog;
	Elf_Data *data, *data_prog, *data_maps = NULL, *symbols = NULL;
	char *shname, *shname_prog;
	int nr_maps = 0;

	/* reset global variables */
	kern_version = 0;
	memset(license, 0, sizeof(license));
	memset(processed_sec, 0, sizeof(processed_sec));

	if (elf_version(EV_CURRENT) == EV_NONE)
		return 1;

	fd = open(path, O_RDONLY, 0);
	if (fd < 0)
		return 1;

	elf = elf_begin(fd, ELF_C_READ, NULL);

	if (!elf)
		return 1;

	if (gelf_getehdr(elf, &ehdr) != &ehdr)
		return 1;

	/* clear all kprobes */
	i = system("echo \"\" > /sys/kernel/debug/tracing/kprobe_events");

	/* scan over all elf sections to get license and map info */
	for (i = 1; i < ehdr.e_shnum; i++) {

		if (get_sec(elf, i, &ehdr, &shname, &shdr, &data))
			continue;

		if (0) /* helpful for llvm debugging */
			printf("section %d:%s data %p size %zd link %d flags %d\n",
			       i, shname, data->d_buf, data->d_size,
			       shdr.sh_link, (int) shdr.sh_flags);

		if (strcmp(shname, "license") == 0) {
			processed_sec[i] = true;
			memcpy(license, data->d_buf, data->d_size);
		} else if (strcmp(shname, "version") == 0) {
			processed_sec[i] = true;
			if (data->d_size != sizeof(int)) {
				printf("invalid size of version section %zd\n",
				       data->d_size);
				return 1;
			}
			memcpy(&kern_version, data->d_buf, sizeof(int));
		} else if (strcmp(shname, "maps") == 0) {
			int j;

			maps_shndx = i;
			data_maps = data;
			for (j = 0; j < MAX_MAPS; j++)
				map_data[j].fd = -1;
		} else if (shdr.sh_type == SHT_SYMTAB) {
			strtabidx = shdr.sh_link;
			symbols = data;
		}
	}

	ret = 1;

	if (!symbols) {
		printf("missing SHT_SYMTAB section\n");
		goto done;
	}

	if (data_maps) {
		nr_maps = load_elf_maps_section(map_data, maps_shndx,
						elf, symbols, strtabidx);
		if (nr_maps < 0) {
			printf("Error: Failed loading ELF maps (errno:%d):%s\n",
			       nr_maps, strerror(-nr_maps));
			ret = 1;
			goto done;
		}
		if (load_maps(map_data, nr_maps, fixup_map))
			goto done;
		map_data_count = nr_maps;

		processed_sec[maps_shndx] = true;
	}

	/* process all relo sections, and rewrite bpf insns for maps */
	for (i = 1; i < ehdr.e_shnum; i++) {
		if (processed_sec[i])
			continue;

		if (get_sec(elf, i, &ehdr, &shname, &shdr, &data))
			continue;

		if (shdr.sh_type == SHT_REL) {
			struct bpf_insn *insns;

			/* locate prog sec that need map fixup (relocations) */
			if (get_sec(elf, shdr.sh_info, &ehdr, &shname_prog,
				    &shdr_prog, &data_prog))
				continue;

			if (shdr_prog.sh_type != SHT_PROGBITS ||
			    !(shdr_prog.sh_flags & SHF_EXECINSTR))
				continue;

			insns = (struct bpf_insn *) data_prog->d_buf;
			processed_sec[i] = true; /* relo section */

			if (parse_relo_and_apply(data, symbols, &shdr, insns,
						 map_data, nr_maps))
				continue;
		}
	}

	/* load programs */
	for (i = 1; i < ehdr.e_shnum; i++) {

		if (processed_sec[i])
			continue;

		if (get_sec(elf, i, &ehdr, &shname, &shdr, &data))
			continue;

		if (memcmp(shname, "kprobe/", 7) == 0 ||
		    memcmp(shname, "kretprobe/", 10) == 0 ||
		    memcmp(shname, "tracepoint/", 11) == 0 ||
		    memcmp(shname, "xdp", 3) == 0 ||
		    memcmp(shname, "perf_event", 10) == 0 ||
		    memcmp(shname, "socket", 6) == 0 ||
		    memcmp(shname, "cgroup/", 7) == 0 ||
		    memcmp(shname, "sockops", 7) == 0 ||
		    memcmp(shname, "sk_skb", 6) == 0 ||
		    memcmp(shname, "sk_msg", 6) == 0) {
			ret = load_and_attach(shname, data->d_buf,
					      data->d_size);
			if (ret != 0)
				goto done;
		}
	}

	ret = 0;
done:
	close(fd);
	return ret;
}

int load_bpf_file(char *path)
{
	return do_load_bpf_file(path, NULL);
}

int load_bpf_file_fixup_map(const char *path, fixup_map_cb fixup_map)
{
	return do_load_bpf_file(path, fixup_map);
}

void read_trace_pipe(void)
{
	int trace_fd;

	trace_fd = open(DEBUGFS "trace_pipe", O_RDONLY, 0);
	if (trace_fd < 0)
		return;

	while (1) {
		static char buf[4096];
		ssize_t sz;

		sz = read(trace_fd, buf, sizeof(buf));
		if (sz > 0) {
			buf[sz] = 0;
			puts(buf);
		}
	}
}

#define MAX_SYMS 300000
static struct ksym syms[MAX_SYMS];
static int sym_cnt;

static int ksym_cmp(const void *p1, const void *p2)
{
	return ((struct ksym *)p1)->addr - ((struct ksym *)p2)->addr;
}

int load_kallsyms(void)
{
	FILE *f = fopen("/proc/kallsyms", "r");
	char func[256], buf[256];
	char symbol;
	void *addr;
	int i = 0;

	if (!f)
		return -ENOENT;

	while (!feof(f)) {
		if (!fgets(buf, sizeof(buf), f))
			break;
		if (sscanf(buf, "%p %c %s", &addr, &symbol, func) != 3)
			break;
		if (!addr)
			continue;
		syms[i].addr = (long) addr;
		syms[i].name = strdup(func);
		i++;
	}
	sym_cnt = i;
	qsort(syms, sym_cnt, sizeof(struct ksym), ksym_cmp);
	return 0;
}

struct ksym *ksym_search(long key)
{
	int start = 0, end = sym_cnt;
	int result;

	while (start < end) {
		size_t mid = start + (end - start) / 2;

		result = key - syms[mid].addr;
		if (result < 0)
			end = mid;
		else if (result > 0)
			start = mid + 1;
		else
			return &syms[mid];
	}

	if (start >= 1 && syms[start - 1].addr < key &&
	    key < syms[start].addr)
		/* valid ksym */
		return &syms[start - 1];

	/* out of range. return _stext */
	return &syms[0];
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
249b812d AS	2	#include <stdio.h>
	3	#include <sys/types.h>
	4	#include <sys/stat.h>
	5	#include <fcntl.h>
	6	#include <libelf.h>
	7	#include <gelf.h>
	8	#include <errno.h>
	9	#include <unistd.h>
	10	#include <string.h>
	11	#include <stdbool.h>
b896c4f9	12	#include <stdlib.h>
249b812d AS	13	#include <linux/bpf.h>
249b812d AS	14	#include <linux/filter.h>
b896c4f9	15	#include <linux/perf_event.h>
12d8bb64 MKL	16	#include <linux/netlink.h>
12d8bb64 MKL	17	#include <linux/rtnetlink.h>
6387d011	18	#include <linux/types.h>
12d8bb64 MKL	19	#include <sys/types.h>
12d8bb64 MKL	20	#include <sys/socket.h>
b896c4f9 AS	21	#include <sys/syscall.h>
	22	#include <sys/ioctl.h>
	23	#include <sys/mman.h>
	24	#include <poll.h>
5bacd780	25	#include <ctype.h>
9fd63d05	26	#include <assert.h>
249b812d	27	#include "libbpf.h"
249b812d	28	#include "bpf_load.h"
205c8ada	29	#include "perf-sys.h"
249b812d	30
b896c4f9 AS	31	#define DEBUGFS "/sys/kernel/debug/tracing/"
b896c4f9 AS	32
249b812d	33	static char license[128];
b896c4f9	34	static int kern_version;
249b812d	35	static bool processed_sec[128];
d40fc181	36	char bpf_log_buf[BPF_LOG_BUF_SIZE];
249b812d AS	37	int map_fd[MAX_MAPS];
249b812d AS	38	int prog_fd[MAX_PROGS];
b896c4f9	39	int event_fd[MAX_PROGS];
249b812d	40	int prog_cnt;
5bacd780 AS	41	int prog_array_fd = -1;
5bacd780 AS	42
156450d9 JDB	43	struct bpf_map_data map_data[MAX_MAPS];
	44	int map_data_count = 0;
	45
5bacd780 AS	46	static int populate_prog_array(const char *event, int prog_fd)
	47	{
	48	int ind = atoi(event), err;
	49
d40fc181	50	err = bpf_map_update_elem(prog_array_fd, &ind, &prog_fd, BPF_ANY);
5bacd780 AS	51	if (err < 0) {
	52	printf("failed to store prog_fd in prog_array\n");
	53	return -1;
	54	}
	55	return 0;
	56	}
249b812d AS	57
	58	static int load_and_attach(const char event, struct bpf_insn prog, int size)
	59	{
249b812d	60	bool is_socket = strncmp(event, "socket", 6) == 0;
b896c4f9 AS	61	bool is_kprobe = strncmp(event, "kprobe/", 7) == 0;
b896c4f9 AS	62	bool is_kretprobe = strncmp(event, "kretprobe/", 10) == 0;
c0766040	63	bool is_tracepoint = strncmp(event, "tracepoint/", 11) == 0;
86af8b41	64	bool is_xdp = strncmp(event, "xdp", 3) == 0;
1c47910e	65	bool is_perf_event = strncmp(event, "perf_event", 10) == 0;
4f2e7ae5 DA	66	bool is_cgroup_skb = strncmp(event, "cgroup/skb", 10) == 0;
4f2e7ae5 DA	67	bool is_cgroup_sk = strncmp(event, "cgroup/sock", 11) == 0;
40304b2a	68	bool is_sockops = strncmp(event, "sockops", 7) == 0;
69e8cc13	69	bool is_sk_skb = strncmp(event, "sk_skb", 6) == 0;
4c4c3c27	70	bool is_sk_msg = strncmp(event, "sk_msg", 6) == 0;
43371c83	71	size_t insns_cnt = size / sizeof(struct bpf_insn);
b896c4f9 AS	72	enum bpf_prog_type prog_type;
	73	char buf[256];
	74	int fd, efd, err, id;
	75	struct perf_event_attr attr = {};
	76
	77	attr.type = PERF_TYPE_TRACEPOINT;
	78	attr.sample_type = PERF_SAMPLE_RAW;
	79	attr.sample_period = 1;
	80	attr.wakeup_events = 1;
	81
	82	if (is_socket) {
	83	prog_type = BPF_PROG_TYPE_SOCKET_FILTER;
	84	} else if (is_kprobe \|\| is_kretprobe) {
	85	prog_type = BPF_PROG_TYPE_KPROBE;
c0766040 AS	86	} else if (is_tracepoint) {
c0766040 AS	87	prog_type = BPF_PROG_TYPE_TRACEPOINT;
86af8b41 BB	88	} else if (is_xdp) {
86af8b41 BB	89	prog_type = BPF_PROG_TYPE_XDP;
1c47910e AS	90	} else if (is_perf_event) {
1c47910e AS	91	prog_type = BPF_PROG_TYPE_PERF_EVENT;
4f2e7ae5 DA	92	} else if (is_cgroup_skb) {
	93	prog_type = BPF_PROG_TYPE_CGROUP_SKB;
	94	} else if (is_cgroup_sk) {
	95	prog_type = BPF_PROG_TYPE_CGROUP_SOCK;
40304b2a LB	96	} else if (is_sockops) {
40304b2a LB	97	prog_type = BPF_PROG_TYPE_SOCK_OPS;
69e8cc13 JF	98	} else if (is_sk_skb) {
69e8cc13 JF	99	prog_type = BPF_PROG_TYPE_SK_SKB;
4c4c3c27 JF	100	} else if (is_sk_msg) {
4c4c3c27 JF	101	prog_type = BPF_PROG_TYPE_SK_MSG;
b896c4f9 AS	102	} else {
b896c4f9 AS	103	printf("Unknown event '%s'\n", event);
249b812d	104	return -1;
b896c4f9 AS	105	}
b896c4f9 AS	106
43371c83	107	fd = bpf_load_program(prog_type, prog, insns_cnt, license, kern_version,
d40fc181	108	bpf_log_buf, BPF_LOG_BUF_SIZE);
5bacd780	109	if (fd < 0) {
d40fc181	110	printf("bpf_load_program() err=%d\n%s", errno, bpf_log_buf);
5bacd780 AS	111	return -1;
	112	}
	113
	114	prog_fd[prog_cnt++] = fd;
	115
4f2e7ae5	116	if (is_xdp \|\| is_perf_event \|\| is_cgroup_skb \|\| is_cgroup_sk)
86af8b41 BB	117	return 0;
86af8b41 BB	118
4c4c3c27	119	if (is_socket \|\| is_sockops \|\| is_sk_skb \|\| is_sk_msg) {
40304b2a LB	120	if (is_socket)
	121	event += 6;
	122	else
	123	event += 7;
5bacd780 AS	124	if (*event != '/')
	125	return 0;
	126	event++;
	127	if (!isdigit(*event)) {
	128	printf("invalid prog number\n");
	129	return -1;
	130	}
	131	return populate_prog_array(event, fd);
	132	}
	133
b896c4f9 AS	134	if (is_kprobe \|\| is_kretprobe) {
	135	if (is_kprobe)
	136	event += 7;
	137	else
	138	event += 10;
	139
5bacd780 AS	140	if (*event == 0) {
	141	printf("event name cannot be empty\n");
	142	return -1;
	143	}
	144
	145	if (isdigit(*event))
	146	return populate_prog_array(event, fd);
	147
b896c4f9 AS	148	snprintf(buf, sizeof(buf),
	149	"echo '%c:%s %s' >> /sys/kernel/debug/tracing/kprobe_events",
	150	is_kprobe ? 'p' : 'r', event, event);
	151	err = system(buf);
	152	if (err < 0) {
	153	printf("failed to create kprobe '%s' error '%s'\n",
	154	event, strerror(errno));
	155	return -1;
	156	}
249b812d	157
c0766040 AS	158	strcpy(buf, DEBUGFS);
	159	strcat(buf, "events/kprobes/");
	160	strcat(buf, event);
	161	strcat(buf, "/id");
	162	} else if (is_tracepoint) {
	163	event += 11;
	164
	165	if (*event == 0) {
	166	printf("event name cannot be empty\n");
	167	return -1;
	168	}
	169	strcpy(buf, DEBUGFS);
	170	strcat(buf, "events/");
	171	strcat(buf, event);
	172	strcat(buf, "/id");
	173	}
b896c4f9 AS	174
	175	efd = open(buf, O_RDONLY, 0);
	176	if (efd < 0) {
	177	printf("failed to open event %s\n", event);
	178	return -1;
	179	}
	180
	181	err = read(efd, buf, sizeof(buf));
	182	if (err < 0 \|\| err >= sizeof(buf)) {
	183	printf("read from '%s' failed '%s'\n", event, strerror(errno));
	184	return -1;
	185	}
	186
	187	close(efd);
	188
	189	buf[err] = 0;
	190	id = atoi(buf);
	191	attr.config = id;
	192
205c8ada	193	efd = sys_perf_event_open(&attr, -1/pid/, 0/cpu/, -1/group_fd/, 0);
b896c4f9 AS	194	if (efd < 0) {
	195	printf("event %d fd %d err %s\n", id, efd, strerror(errno));
	196	return -1;
	197	}
	198	event_fd[prog_cnt - 1] = efd;
0ec9552b YS	199	err = ioctl(efd, PERF_EVENT_IOC_ENABLE, 0);
	200	if (err < 0) {
	201	printf("ioctl PERF_EVENT_IOC_ENABLE failed err %s\n",
	202	strerror(errno));
	203	return -1;
	204	}
	205	err = ioctl(efd, PERF_EVENT_IOC_SET_BPF, fd);
	206	if (err < 0) {
	207	printf("ioctl PERF_EVENT_IOC_SET_BPF failed err %s\n",
	208	strerror(errno));
	209	return -1;
	210	}
b896c4f9	211
249b812d AS	212	return 0;
	213	}
	214
156450d9 JDB	215	static int load_maps(struct bpf_map_data *maps, int nr_maps,
156450d9 JDB	216	fixup_map_cb fixup_map)
249b812d	217	{
ad17d0e6	218	int i, numa_node;
156450d9	219
5010e948	220	for (i = 0; i < nr_maps; i++) {
6979bcc7 JDB	221	if (fixup_map) {
	222	fixup_map(&maps[i], i);
	223	/* Allow userspace to assign map FD prior to creation */
	224	if (maps[i].fd != -1) {
	225	map_fd[i] = maps[i].fd;
	226	continue;
	227	}
	228	}
249b812d	229
ad17d0e6 MKL	230	numa_node = maps[i].def.map_flags & BPF_F_NUMA_NODE ?
	231	maps[i].def.numa_node : -1;
	232
156450d9 JDB	233	if (maps[i].def.type == BPF_MAP_TYPE_ARRAY_OF_MAPS \|\|
	234	maps[i].def.type == BPF_MAP_TYPE_HASH_OF_MAPS) {
	235	int inner_map_fd = map_fd[maps[i].def.inner_map_idx];
fb30d4b7	236
ad17d0e6	237	map_fd[i] = bpf_create_map_in_map_node(maps[i].def.type,
88cda1c9	238	maps[i].name,
156450d9 JDB	239	maps[i].def.key_size,
	240	inner_map_fd,
	241	maps[i].def.max_entries,
ad17d0e6 MKL	242	maps[i].def.map_flags,
ad17d0e6 MKL	243	numa_node);
fb30d4b7	244	} else {
ad17d0e6	245	map_fd[i] = bpf_create_map_node(maps[i].def.type,
88cda1c9	246	maps[i].name,
ad17d0e6 MKL	247	maps[i].def.key_size,
	248	maps[i].def.value_size,
	249	maps[i].def.max_entries,
	250	maps[i].def.map_flags,
	251	numa_node);
fb30d4b7	252	}
618ec9a7 AS	253	if (map_fd[i] < 0) {
	254	printf("failed to create a map: %d %s\n",
	255	errno, strerror(errno));
249b812d	256	return 1;
618ec9a7	257	}
156450d9	258	maps[i].fd = map_fd[i];
5bacd780	259
156450d9	260	if (maps[i].def.type == BPF_MAP_TYPE_PROG_ARRAY)
5bacd780	261	prog_array_fd = map_fd[i];
249b812d AS	262	}
	263	return 0;
	264	}
	265
	266	static int get_sec(Elf elf, int i, GElf_Ehdr ehdr, char **shname,
	267	GElf_Shdr shdr, Elf_Data *data)
	268	{
	269	Elf_Scn *scn;
	270
	271	scn = elf_getscn(elf, i);
	272	if (!scn)
	273	return 1;
	274
	275	if (gelf_getshdr(scn, shdr) != shdr)
	276	return 2;
	277
	278	*shname = elf_strptr(elf, ehdr->e_shstrndx, shdr->sh_name);
	279	if (!*shname \|\| !shdr->sh_size)
	280	return 3;
	281
	282	*data = elf_getdata(scn, 0);
	283	if (!data \|\| elf_getdata(scn, data) != NULL)
	284	return 4;
	285
	286	return 0;
	287	}
	288
	289	static int parse_relo_and_apply(Elf_Data data, Elf_Data symbols,
156450d9 JDB	290	GElf_Shdr shdr, struct bpf_insn insn,
156450d9 JDB	291	struct bpf_map_data *maps, int nr_maps)
249b812d AS	292	{
	293	int i, nrels;
	294
	295	nrels = shdr->sh_size / shdr->sh_entsize;
	296
	297	for (i = 0; i < nrels; i++) {
	298	GElf_Sym sym;
	299	GElf_Rel rel;
	300	unsigned int insn_idx;
156450d9 JDB	301	bool match = false;
156450d9 JDB	302	int j, map_idx;
249b812d AS	303
	304	gelf_getrel(data, i, &rel);
	305
	306	insn_idx = rel.r_offset / sizeof(struct bpf_insn);
	307
	308	gelf_getsym(symbols, GELF_R_SYM(rel.r_info), &sym);
	309
	310	if (insn[insn_idx].code != (BPF_LD \| BPF_IMM \| BPF_DW)) {
	311	printf("invalid relo for insn[%d].code 0x%x\n",
	312	insn_idx, insn[insn_idx].code);
	313	return 1;
	314	}
	315	insn[insn_idx].src_reg = BPF_PSEUDO_MAP_FD;
156450d9 JDB	316
	317	/* Match FD relocation against recorded map_data[] offset */
	318	for (map_idx = 0; map_idx < nr_maps; map_idx++) {
	319	if (maps[map_idx].elf_offset == sym.st_value) {
	320	match = true;
	321	break;
	322	}
	323	}
	324	if (match) {
	325	insn[insn_idx].imm = maps[map_idx].fd;
	326	} else {
	327	printf("invalid relo for insn[%d] no map_data match\n",
	328	insn_idx);
	329	return 1;
	330	}
249b812d AS	331	}
	332
	333	return 0;
	334	}
	335
9fd63d05 MKL	336	static int cmp_symbols(const void l, const void r)
	337	{
	338	const GElf_Sym lsym = (const GElf_Sym )l;
	339	const GElf_Sym rsym = (const GElf_Sym )r;
	340
	341	if (lsym->st_value < rsym->st_value)
	342	return -1;
	343	else if (lsym->st_value > rsym->st_value)
	344	return 1;
	345	else
	346	return 0;
	347	}
	348
156450d9 JDB	349	static int load_elf_maps_section(struct bpf_map_data *maps, int maps_shndx,
156450d9 JDB	350	Elf elf, Elf_Data symbols, int strtabidx)
9fd63d05	351	{
156450d9 JDB	352	int map_sz_elf, map_sz_copy;
	353	bool validate_zero = false;
	354	Elf_Data *data_maps;
	355	int i, nr_maps;
	356	GElf_Sym *sym;
	357	Elf_Scn *scn;
	358	int copy_sz;
	359
	360	if (maps_shndx < 0)
	361	return -EINVAL;
	362	if (!symbols)
	363	return -EINVAL;
	364
	365	/* Get data for maps section via elf index */
	366	scn = elf_getscn(elf, maps_shndx);
	367	if (scn)
	368	data_maps = elf_getdata(scn, NULL);
	369	if (!scn \|\| !data_maps) {
	370	printf("Failed to get Elf_Data from maps section %d\n",
	371	maps_shndx);
	372	return -EINVAL;
	373	}
9fd63d05	374
156450d9 JDB	375	/* For each map get corrosponding symbol table entry */
	376	sym = calloc(MAX_MAPS+1, sizeof(GElf_Sym));
	377	for (i = 0, nr_maps = 0; i < symbols->d_size / sizeof(GElf_Sym); i++) {
	378	assert(nr_maps < MAX_MAPS+1);
	379	if (!gelf_getsym(symbols, i, &sym[nr_maps]))
9fd63d05	380	continue;
156450d9	381	if (sym[nr_maps].st_shndx != maps_shndx)
9fd63d05	382	continue;
156450d9	383	/* Only increment iif maps section */
9fd63d05 MKL	384	nr_maps++;
	385	}
	386
156450d9 JDB	387	/* Align to map_fd[] order, via sort on offset in sym.st_value */
	388	qsort(sym, nr_maps, sizeof(GElf_Sym), cmp_symbols);
	389
	390	/* Keeping compatible with ELF maps section changes
	391	* ------------------------------------------------
	392	* The program size of struct bpf_map_def is known by loader
	393	* code, but struct stored in ELF file can be different.
	394	*
	395	* Unfortunately sym[i].st_size is zero. To calculate the
	396	* struct size stored in the ELF file, assume all struct have
	397	* the same size, and simply divide with number of map
	398	* symbols.
	399	*/
	400	map_sz_elf = data_maps->d_size / nr_maps;
	401	map_sz_copy = sizeof(struct bpf_map_def);
	402	if (map_sz_elf < map_sz_copy) {
	403	/*
	404	* Backward compat, loading older ELF file with
	405	* smaller struct, keeping remaining bytes zero.
	406	*/
	407	map_sz_copy = map_sz_elf;
	408	} else if (map_sz_elf > map_sz_copy) {
	409	/*
	410	* Forward compat, loading newer ELF file with larger
	411	* struct with unknown features. Assume zero means
	412	* feature not used. Thus, validate rest of struct
	413	* data is zero.
	414	*/
	415	validate_zero = true;
	416	}
9fd63d05	417
156450d9	418	/* Memcpy relevant part of ELF maps data to loader maps */
9fd63d05	419	for (i = 0; i < nr_maps; i++) {
156450d9 JDB	420	unsigned char addr, end;
	421	struct bpf_map_def *def;
	422	const char *map_name;
	423	size_t offset;
	424
	425	map_name = elf_strptr(elf, strtabidx, sym[i].st_name);
	426	maps[i].name = strdup(map_name);
	427	if (!maps[i].name) {
9fd63d05 MKL	428	printf("strdup(%s): %s(%d)\n", map_name,
9fd63d05 MKL	429	strerror(errno), errno);
156450d9 JDB	430	free(sym);
	431	return -errno;
	432	}
	433
	434	/* Symbol value is offset into ELF maps section data area */
	435	offset = sym[i].st_value;
	436	def = (struct bpf_map_def *)(data_maps->d_buf + offset);
	437	maps[i].elf_offset = offset;
	438	memset(&maps[i].def, 0, sizeof(struct bpf_map_def));
	439	memcpy(&maps[i].def, def, map_sz_copy);
	440
	441	/* Verify no newer features were requested */
	442	if (validate_zero) {
	443	addr = (unsigned char*) def + map_sz_copy;
	444	end = (unsigned char*) def + map_sz_elf;
	445	for (; addr < end; addr++) {
	446	if (*addr != 0) {
	447	free(sym);
	448	return -EFBIG;
	449	}
	450	}
9fd63d05 MKL	451	}
	452	}
	453
156450d9	454	free(sym);
5010e948	455	return nr_maps;
9fd63d05 MKL	456	}
	457
	458	static int do_load_bpf_file(const char *path, fixup_map_cb fixup_map)
249b812d	459	{
9fd63d05	460	int fd, i, ret, maps_shndx = -1, strtabidx = -1;
249b812d AS	461	Elf *elf;
	462	GElf_Ehdr ehdr;
	463	GElf_Shdr shdr, shdr_prog;
9fd63d05	464	Elf_Data data, data_prog, data_maps = NULL, symbols = NULL;
156450d9 JDB	465	char shname, shname_prog;
156450d9 JDB	466	int nr_maps = 0;
249b812d	467
a734fb5d MS	468	/* reset global variables */
	469	kern_version = 0;
	470	memset(license, 0, sizeof(license));
	471	memset(processed_sec, 0, sizeof(processed_sec));
	472
249b812d AS	473	if (elf_version(EV_CURRENT) == EV_NONE)
	474	return 1;
	475
	476	fd = open(path, O_RDONLY, 0);
	477	if (fd < 0)
	478	return 1;
	479
	480	elf = elf_begin(fd, ELF_C_READ, NULL);
	481
	482	if (!elf)
	483	return 1;
	484
	485	if (gelf_getehdr(elf, &ehdr) != &ehdr)
	486	return 1;
	487
b896c4f9 AS	488	/* clear all kprobes */
	489	i = system("echo \"\" > /sys/kernel/debug/tracing/kprobe_events");
	490
249b812d AS	491	/* scan over all elf sections to get license and map info */
	492	for (i = 1; i < ehdr.e_shnum; i++) {
	493
	494	if (get_sec(elf, i, &ehdr, &shname, &shdr, &data))
	495	continue;
	496
	497	if (0) /* helpful for llvm debugging */
	498	printf("section %d:%s data %p size %zd link %d flags %d\n",
	499	i, shname, data->d_buf, data->d_size,
	500	shdr.sh_link, (int) shdr.sh_flags);
	501
	502	if (strcmp(shname, "license") == 0) {
	503	processed_sec[i] = true;
	504	memcpy(license, data->d_buf, data->d_size);
b896c4f9 AS	505	} else if (strcmp(shname, "version") == 0) {
	506	processed_sec[i] = true;
	507	if (data->d_size != sizeof(int)) {
	508	printf("invalid size of version section %zd\n",
	509	data->d_size);
	510	return 1;
	511	}
	512	memcpy(&kern_version, data->d_buf, sizeof(int));
249b812d	513	} else if (strcmp(shname, "maps") == 0) {
156450d9 JDB	514	int j;
156450d9 JDB	515
9fd63d05 MKL	516	maps_shndx = i;
9fd63d05 MKL	517	data_maps = data;
156450d9 JDB	518	for (j = 0; j < MAX_MAPS; j++)
156450d9 JDB	519	map_data[j].fd = -1;
249b812d	520	} else if (shdr.sh_type == SHT_SYMTAB) {
9fd63d05	521	strtabidx = shdr.sh_link;
249b812d AS	522	symbols = data;
	523	}
	524	}
	525
9fd63d05 MKL	526	ret = 1;
	527
	528	if (!symbols) {
	529	printf("missing SHT_SYMTAB section\n");
	530	goto done;
	531	}
	532
	533	if (data_maps) {
156450d9 JDB	534	nr_maps = load_elf_maps_section(map_data, maps_shndx,
	535	elf, symbols, strtabidx);
	536	if (nr_maps < 0) {
	537	printf("Error: Failed loading ELF maps (errno:%d):%s\n",
	538	nr_maps, strerror(-nr_maps));
5010e948 JDB	539	ret = 1;
	540	goto done;
	541	}
156450d9	542	if (load_maps(map_data, nr_maps, fixup_map))
9fd63d05	543	goto done;
156450d9	544	map_data_count = nr_maps;
9fd63d05 MKL	545
	546	processed_sec[maps_shndx] = true;
	547	}
	548
7bc57950	549	/* process all relo sections, and rewrite bpf insns for maps */
249b812d	550	for (i = 1; i < ehdr.e_shnum; i++) {
16ad1329 MS	551	if (processed_sec[i])
16ad1329 MS	552	continue;
249b812d AS	553
	554	if (get_sec(elf, i, &ehdr, &shname, &shdr, &data))
	555	continue;
7bc57950	556
249b812d AS	557	if (shdr.sh_type == SHT_REL) {
	558	struct bpf_insn *insns;
	559
7bc57950	560	/* locate prog sec that need map fixup (relocations) */
249b812d AS	561	if (get_sec(elf, shdr.sh_info, &ehdr, &shname_prog,
	562	&shdr_prog, &data_prog))
	563	continue;
	564
db6a71dd AS	565	if (shdr_prog.sh_type != SHT_PROGBITS \|\|
	566	!(shdr_prog.sh_flags & SHF_EXECINSTR))
	567	continue;
	568
249b812d	569	insns = (struct bpf_insn *) data_prog->d_buf;
7bc57950	570	processed_sec[i] = true; /* relo section */
249b812d	571
156450d9 JDB	572	if (parse_relo_and_apply(data, symbols, &shdr, insns,
156450d9 JDB	573	map_data, nr_maps))
249b812d	574	continue;
249b812d AS	575	}
	576	}
	577
7bc57950	578	/* load programs */
249b812d AS	579	for (i = 1; i < ehdr.e_shnum; i++) {
	580
	581	if (processed_sec[i])
	582	continue;
	583
	584	if (get_sec(elf, i, &ehdr, &shname, &shdr, &data))
	585	continue;
	586
b896c4f9 AS	587	if (memcmp(shname, "kprobe/", 7) == 0 \|\|
b896c4f9 AS	588	memcmp(shname, "kretprobe/", 10) == 0 \|\|
c0766040	589	memcmp(shname, "tracepoint/", 11) == 0 \|\|
86af8b41	590	memcmp(shname, "xdp", 3) == 0 \|\|
1c47910e	591	memcmp(shname, "perf_event", 10) == 0 \|\|
4f2e7ae5	592	memcmp(shname, "socket", 6) == 0 \|\|
40304b2a	593	memcmp(shname, "cgroup/", 7) == 0 \|\|
69e8cc13	594	memcmp(shname, "sockops", 7) == 0 \|\|
4c4c3c27 JF	595	memcmp(shname, "sk_skb", 6) == 0 \|\|
4c4c3c27 JF	596	memcmp(shname, "sk_msg", 6) == 0) {
f856e469 LB	597	ret = load_and_attach(shname, data->d_buf,
	598	data->d_size);
	599	if (ret != 0)
	600	goto done;
	601	}
249b812d AS	602	}
249b812d AS	603
9fd63d05 MKL	604	ret = 0;
9fd63d05 MKL	605	done:
249b812d	606	close(fd);
9fd63d05 MKL	607	return ret;
	608	}
	609
	610	int load_bpf_file(char *path)
	611	{
	612	return do_load_bpf_file(path, NULL);
	613	}
	614
	615	int load_bpf_file_fixup_map(const char *path, fixup_map_cb fixup_map)
	616	{
	617	return do_load_bpf_file(path, fixup_map);
249b812d	618	}
b896c4f9 AS	619
	620	void read_trace_pipe(void)
	621	{
	622	int trace_fd;
	623
	624	trace_fd = open(DEBUGFS "trace_pipe", O_RDONLY, 0);
	625	if (trace_fd < 0)
	626	return;
	627
	628	while (1) {
	629	static char buf[4096];
	630	ssize_t sz;
	631
	632	sz = read(trace_fd, buf, sizeof(buf));
	633	if (sz > 0) {
	634	buf[sz] = 0;
	635	puts(buf);
	636	}
	637	}
	638	}
3622e7e4 AS	639
	640	#define MAX_SYMS 300000
	641	static struct ksym syms[MAX_SYMS];
	642	static int sym_cnt;
	643
	644	static int ksym_cmp(const void p1, const void p2)
	645	{
	646	return ((struct ksym )p1)->addr - ((struct ksym )p2)->addr;
	647	}
	648
	649	int load_kallsyms(void)
	650	{
	651	FILE *f = fopen("/proc/kallsyms", "r");
	652	char func[256], buf[256];
	653	char symbol;
	654	void *addr;
	655	int i = 0;
	656
	657	if (!f)
	658	return -ENOENT;
	659
	660	while (!feof(f)) {
	661	if (!fgets(buf, sizeof(buf), f))
	662	break;
	663	if (sscanf(buf, "%p %c %s", &addr, &symbol, func) != 3)
	664	break;
	665	if (!addr)
	666	continue;
	667	syms[i].addr = (long) addr;
	668	syms[i].name = strdup(func);
	669	i++;
	670	}
	671	sym_cnt = i;
	672	qsort(syms, sym_cnt, sizeof(struct ksym), ksym_cmp);
	673	return 0;
	674	}
	675
	676	struct ksym *ksym_search(long key)
	677	{
	678	int start = 0, end = sym_cnt;
	679	int result;
	680
	681	while (start < end) {
	682	size_t mid = start + (end - start) / 2;
	683
	684	result = key - syms[mid].addr;
	685	if (result < 0)
	686	end = mid;
	687	else if (result > 0)
	688	start = mid + 1;
	689	else
	690	return &syms[mid];
	691	}
	692
	693	if (start >= 1 && syms[start - 1].addr < key &&
	694	key < syms[start].addr)
	695	/* valid ksym */
	696	return &syms[start - 1];
	697
	698	/* out of range. return _stext */
	699	return &syms[0];
	700	}
12d8bb64	701