[PATCH v5 bpf-next 2/5] libbpf: support function name-based attach uprobes

[Alan Maguire <alan.maguire@oracle.com>]

kprobe attach is name-based, using lookups of kallsyms to translate
a function name to an address.  Currently uprobe attach is done
via an offset value as described in [1].  Extend uprobe opts
for attach to include a function name which can then be converted
into a uprobe-friendly offset.  The calcualation is done in
several steps:

1. First, determine the symbol address using libelf; this gives us
   the offset as reported by objdump
2. If the function is a shared library function - and the binary
   provided is a shared library - no further work is required;
   the address found is the required address
3. Finally, if the function is local, subtract the base address
   associated with the object, retrieved from ELF program headers.

The resultant value is then added to the func_offset value passed
in to specify the uprobe attach address.  So specifying a func_offset
of 0 along with a function name "printf" will attach to printf entry.

The modes of operation supported are then

1. to attach to a local function in a binary; function "foo1" in
   "/usr/bin/foo"
2. to attach to a shared library function in a shared library -
   function "malloc" in libc.

[1] https://www.kernel.org/doc/html/latest/trace/uprobetracer.html

Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
---
 tools/lib/bpf/libbpf.c | 203 +++++++++++++++++++++++++++++++++++++++++++++++++
 tools/lib/bpf/libbpf.h |  10 ++-
 2 files changed, 212 insertions(+), 1 deletion(-)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index a7d5954..eda724c 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -10517,6 +10517,195 @@ static int perf_event_uprobe_open_legacy(const char *probe_name, bool retprobe,
 	return pfd;
 }
 
+/* uprobes deal in relative offsets; subtract the base address associated with
+ * the mapped binary.  See Documentation/trace/uprobetracer.rst for more
+ * details.
+ */
+static long elf_find_relative_offset(const char *filename, Elf *elf, long addr)
+{
+	size_t n;
+	int i;
+
+	if (elf_getphdrnum(elf, &n)) {
+		pr_warn("elf: failed to find program headers for '%s': %s\n", filename,
+			elf_errmsg(-1));
+		return -ENOENT;
+	}
+
+	for (i = 0; i < n; i++) {
+		int seg_start, seg_end, seg_offset;
+		GElf_Phdr phdr;
+
+		if (!gelf_getphdr(elf, i, &phdr)) {
+			pr_warn("elf: failed to get program header %d from '%s': %s\n", i, filename,
+				elf_errmsg(-1));
+			return -ENOENT;
+		}
+		if (phdr.p_type != PT_LOAD || !(phdr.p_flags & PF_X))
+			continue;
+
+		seg_start = phdr.p_vaddr;
+		seg_end = seg_start + phdr.p_memsz;
+		seg_offset = phdr.p_offset;
+		if (addr >= seg_start && addr < seg_end)
+			return addr - seg_start + seg_offset;
+	}
+	pr_warn("elf: failed to find prog header containing 0x%lx in '%s'\n", addr, filename);
+	return -ENOENT;
+}
+
+/* Return next ELF section of sh_type after scn, or first of that type if scn is NULL. */
+static Elf_Scn *elf_find_next_scn_by_type(Elf *elf, int sh_type, Elf_Scn *scn)
+{
+	while ((scn = elf_nextscn(elf, scn)) != NULL) {
+		GElf_Shdr sh;
+
+		if (!gelf_getshdr(scn, &sh))
+			continue;
+		if (sh.sh_type == sh_type)
+			return scn;
+	}
+	return NULL;
+}
+
+/* Find offset of function name in object specified by path.  "name" matches
+ * symbol name or name@@LIB for library functions.
+ */
+static long elf_find_func_offset(const char *binary_path, const char *name)
+{
+	int fd, i, sh_types[2] = { SHT_DYNSYM, SHT_SYMTAB };
+	bool is_shared_lib, is_name_qualified;
+	char errmsg[STRERR_BUFSIZE];
+	long ret = -ENOENT;
+	size_t name_len;
+	GElf_Ehdr ehdr;
+	Elf *elf;
+
+	fd = open(binary_path, O_RDONLY | O_CLOEXEC);
+	if (fd < 0) {
+		ret = -errno;
+		pr_warn("failed to open %s: %s\n", binary_path,
+			libbpf_strerror_r(ret, errmsg, sizeof(errmsg)));
+		return ret;
+	}
+	elf = elf_begin(fd, ELF_C_READ_MMAP, NULL);
+	if (!elf) {
+		pr_warn("elf: could not read elf from %s: %s\n", binary_path, elf_errmsg(-1));
+		close(fd);
+		return -LIBBPF_ERRNO__FORMAT;
+	}
+	if (!gelf_getehdr(elf, &ehdr)) {
+		pr_warn("elf: failed to get ehdr from %s: %s\n", binary_path, elf_errmsg(-1));
+		ret = -LIBBPF_ERRNO__FORMAT;
+		goto out;
+	}
+	/* for shared lib case, we do not need to calculate relative offset */
+	is_shared_lib = ehdr.e_type == ET_DYN;
+
+	name_len = strlen(name);
+	/* Does name specify "@@LIB"? */
+	is_name_qualified = strstr(name, "@@") != NULL;
+
+	/* Search SHT_DYNSYM, SHT_SYMTAB for symbol.  This search order is used because if
+	 * a binary is stripped, it may only have SHT_DYNSYM, and a fully-statically
+	 * linked binary may not have SHT_DYMSYM, so absence of a section should not be
+	 * reported as a warning/error.
+	 */
+	for (i = 0; i < ARRAY_SIZE(sh_types); i++) {
+		size_t nr_syms, strtabidx, idx;
+		Elf_Data *symbols = NULL;
+		Elf_Scn *scn = NULL;
+		int last_bind = -1;
+		const char *sname;
+		GElf_Shdr sh;
+
+		scn = elf_find_next_scn_by_type(elf, sh_types[i], NULL);
+		if (!scn) {
+			pr_debug("elf: failed to find symbol table ELF sections in '%s'\n",
+				 binary_path);
+			continue;
+		}
+		if (!gelf_getshdr(scn, &sh))
+			continue;
+		strtabidx = sh.sh_link;
+		symbols = elf_getdata(scn, 0);
+		if (!symbols) {
+			pr_warn("elf: failed to get symbols for symtab section in '%s': %s\n",
+				binary_path, elf_errmsg(-1));
+			ret = -LIBBPF_ERRNO__FORMAT;
+			goto out;
+		}
+		nr_syms = symbols->d_size / sh.sh_entsize;
+
+		for (idx = 0; idx < nr_syms; idx++) {
+			int curr_bind;
+			GElf_Sym sym;
+
+			if (!gelf_getsym(symbols, idx, &sym))
+				continue;
+
+			if (GELF_ST_TYPE(sym.st_info) != STT_FUNC)
+				continue;
+
+			sname = elf_strptr(elf, strtabidx, sym.st_name);
+			if (!sname)
+				continue;
+
+			curr_bind = GELF_ST_BIND(sym.st_info);
+
+			/* User can specify func, func@@LIB or func@@LIB_VERSION. */
+			if (strncmp(sname, name, name_len) != 0)
+				continue;
+			/* ...but we don't want a search for "foo" to match 'foo2" also, so any
+			 * additional characters in sname should be of the form "@@LIB".
+			 */
+			if (!is_name_qualified && sname[name_len] != '\0' && sname[name_len] != '@')
+				continue;
+
+			if (ret >= 0) {
+				/* handle multiple matches */
+				if (last_bind != STB_WEAK && curr_bind != STB_WEAK) {
+					/* Only accept one non-weak bind. */
+					pr_warn("elf: ambiguous match for '%s', '%s' in '%s'\n",
+						sname, name, binary_path);
+					ret = -LIBBPF_ERRNO__FORMAT;
+					goto out;
+				} else if (curr_bind == STB_WEAK) {
+					/* already have a non-weak bind, and
+					 * this is a weak bind, so ignore.
+					 */
+					continue;
+				}
+			}
+			ret = sym.st_value;
+			last_bind = curr_bind;
+		}
+		/* For binaries that are not shared libraries, we need relative offset */
+		if (ret > 0 && !is_shared_lib)
+			ret = elf_find_relative_offset(binary_path, elf, ret);
+		if (ret > 0)
+			break;
+	}
+
+	if (ret > 0) {
+		pr_debug("elf: symbol address match for '%s' in '%s': 0x%lx\n", name, binary_path,
+			 ret);
+	} else {
+		if (ret == 0) {
+			pr_warn("elf: '%s' is 0 in symtab for '%s': %s\n", name, binary_path,
+				is_shared_lib ? "should not be 0 in a shared library" :
+						"try using shared library path instead");
+			ret = -ENOENT;
+		} else {
+			pr_warn("elf: failed to find symbol '%s' in '%s'\n", name, binary_path);
+		}
+	}
+out:
+	elf_end(elf);
+	close(fd);
+	return ret;
+}
+
 /* Get full path to program/shared library. */
 static int resolve_full_path(const char *file, char *result, size_t result_sz)
 {
@@ -10569,6 +10758,7 @@ static int resolve_full_path(const char *file, char *result, size_t result_sz)
 	size_t ref_ctr_off;
 	int pfd, err;
 	bool retprobe, legacy;
+	const char *func_name;
 
 	if (!OPTS_VALID(opts, bpf_uprobe_opts))
 		return libbpf_err_ptr(-EINVAL);
@@ -10586,6 +10776,19 @@ static int resolve_full_path(const char *file, char *result, size_t result_sz)
 		}
 		binary_path = full_binary_path;
 	}
+	func_name = OPTS_GET(opts, func_name, NULL);
+	if (func_name) {
+		long sym_off;
+
+		if (!binary_path) {
+			pr_warn("name-based attach requires binary_path\n");
+			return libbpf_err_ptr(-EINVAL);
+		}
+		sym_off = elf_find_func_offset(binary_path, func_name);
+		if (sym_off < 0)
+			return libbpf_err_ptr(sym_off);
+		func_offset += sym_off;
+	}
 
 	legacy = determine_uprobe_perf_type() < 0;
 	if (!legacy) {
diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h
index 05dde85..28cd206 100644
--- a/tools/lib/bpf/libbpf.h
+++ b/tools/lib/bpf/libbpf.h
@@ -459,9 +459,17 @@ struct bpf_uprobe_opts {
 	__u64 bpf_cookie;
 	/* uprobe is return probe, invoked at function return time */
 	bool retprobe;
+	/* Function name to attach to.  Could be an unqualified ("abc") or library-qualified
+	 * "abc@LIBXYZ" name.  To specify function entry, func_name should be set while
+	 * func_offset argument to bpf_prog__attach_uprobe_opts() should be 0.  To trace an
+	 * offset within a function, specify func_name and use func_offset argument to specify
+	 * offset within the function.  Shared library functions must specify the shared library
+	 * binary_path.
+	 */
+	const char *func_name;
 	size_t :0;
 };
-#define bpf_uprobe_opts__last_field retprobe
+#define bpf_uprobe_opts__last_field func_name
 
 /**
  * @brief **bpf_program__attach_uprobe()** attaches a BPF program
-- 
1.8.3.1