Provide source code info for Symbol

For function symbols, provide the source code file, optional
line number, and optional column.

Signed-off-by: Paul Zuchowski <pzuchowski@datto.com>

_drgn.pyi

@@ -1529,6 +1529,20 @@ class Symbol:
     kind: SymbolKind
     """Kind of entity represented by this symbol."""
 
+    def source(self, offset: int, line_number: bool, line_column: bool) -> str:
+        """
+        Return a string with the source code file, optional line number and
+        optional column for the function symbol plus offset.
+
+        >>> s = prog.symbol('mutex_lock')
+        >>> s.source(offset=0, line_number=True, line_column=False)
+        '/build/linux-hwe-5.8-Ox7OXf/linux-hwe-5.8-5.8.0/kernel/locking/mutex.c: 280'
+
+        param: offset: Offset of instruction from function base address.
+        param: line_number: Include source code line number.
+        param: line_column: Include source code column.
+        """
+
 class SymbolBinding(enum.Enum):
     """
     A ``SymbolBinding`` describes the linkage behavior and visibility of a

libdrgn/debug_info.c

@@ -303,6 +303,25 @@ static int drgn_dwfl_module_removed(Dwfl_Module *dwfl_module, void *userdatap,
 	return DWARF_CB_OK;
 }
 
+struct drgn_debug_info_module *
+drgn_debug_info_module_byaddress(struct drgn_debug_info *dbinfo, uint64_t addr)
+{
+
+	for (struct drgn_debug_info_module_table_iterator it =
+             drgn_debug_info_module_table_first(&dbinfo->modules); it.entry; ) {
+                struct drgn_debug_info_module *module = *it.entry;
+                do {
+			struct drgn_debug_info_module *next = module->next;
+			if (!next)
+				it = drgn_debug_info_module_table_next(it);
+			if (addr >= module->start && addr <= module->end)
+				return module;
+			module = next;
+                } while (module);
+	}
+	return NULL;
+}
+
 static void drgn_debug_info_free_modules(struct drgn_debug_info *dbinfo,
 					 bool finish_indexing, bool free_all)
 {

libdrgn/debug_info.h

@@ -171,6 +171,10 @@ drgn_debug_info_buffer_init(struct drgn_debug_info_buffer *buffer,
 	buffer->scn = scn;
 }
 
+struct drgn_debug_info_module *
+drgn_debug_info_module_byaddress(struct drgn_debug_info *dbinfo, uint64_t addr);
+
+
 DEFINE_HASH_TABLE_TYPE(drgn_debug_info_module_table,
 		       struct drgn_debug_info_module *)
 

libdrgn/drgn.h.in

@@ -2675,6 +2675,15 @@ void drgn_symbol_destroy(struct drgn_symbol *sym);
  */
 void drgn_symbols_destroy(struct drgn_symbol **syms, size_t count);
 
+/** Get the source code file, line, and column of a @ref drgn_symbol */
+void drgn_symbol_source(struct drgn_symbol *sym,
+			struct drgn_program *prog,
+			unsigned long offset,
+			size_t size,
+			char *src_file_ret,
+			int *line_num_ret,
+			int *line_col_ret);
+
 /**
  * Get the name of a @ref drgn_symbol.
  *

libdrgn/python/symbol.c

@@ -36,6 +36,43 @@ static PyObject *Symbol_richcompare(Symbol *self, PyObject *other, int op)
 	Py_RETURN_BOOL(ret);
 }
 
+static PyObject *Symbol_source(PyObject *self, PyObject *args, PyObject* kwargs)
+{
+	char source[256];
+	char formatted[296];
+	Program *prgm = ((Symbol *)self)->prog;
+	struct drgn_program *p = &prgm->prog;
+	int ln_num = 0;
+	int ln_col = 0;
+	int offset, line_number, line_column;
+	static char* keywords[] = {"offset", "line_number", "line_column", NULL};
+
+	if (!PyArg_ParseTupleAndKeywords(args,
+					 kwargs,
+					 "iii",
+					 keywords,
+					 &offset,
+					 &line_number,
+					 &line_column))
+		return NULL;
+
+	(void) drgn_symbol_source(((Symbol *)self)->sym, p,
+				  (unsigned long)offset,
+				  sizeof(source), source,
+				  &ln_num, &ln_col);
+	/* we do not print the column without the line number */
+	if (line_number && line_column)
+		snprintf(formatted, sizeof(formatted), "%s: %d:%d",
+			 source, ln_num, ln_col);
+	else if (line_number)
+		snprintf(formatted, sizeof(formatted), "%s: %d",
+			 source, ln_num);
+	else
+		snprintf(formatted, sizeof(formatted), "%s", source);
+
+	return PyUnicode_FromString(formatted);
+}
+
 static PyObject *Symbol_get_name(Symbol *self, void *arg)
 {
 	return PyUnicode_FromString(drgn_symbol_name(self->sym));
@@ -93,6 +130,12 @@ static PyObject *Symbol_repr(Symbol *self)
 
 }
 
+static PyMethodDef Symbol_methods[] = {
+	{"source", (PyCFunction)Symbol_source, METH_VARARGS | METH_KEYWORDS,
+	 "get source code reference for symbol"},
+	{},
+};
+
 static PyGetSetDef Symbol_getset[] = {
 	{"name", (getter)Symbol_get_name, NULL, drgn_Symbol_name_DOC},
 	{"address", (getter)Symbol_get_address, NULL, drgn_Symbol_address_DOC},
@@ -111,5 +154,6 @@ PyTypeObject Symbol_type = {
 	.tp_flags = Py_TPFLAGS_DEFAULT,
 	.tp_doc = drgn_Symbol_DOC,
 	.tp_richcompare = (richcmpfunc)Symbol_richcompare,
+	.tp_methods = Symbol_methods,
 	.tp_getset = Symbol_getset,
 };

libdrgn/symbol.c

@@ -4,8 +4,13 @@
 #include <elf.h>
 #include <stdlib.h>
 #include <string.h>
+#include <dwarf.h>
+#include <elfutils/libdw.h>
+#include <elfutils/libdwfl.h>
 
+#include "debug_info.h"
 #include "drgn.h"
+#include "program.h"
 #include "symbol.h"
 #include "util.h"
 
@@ -40,6 +45,39 @@ void drgn_symbol_from_elf(const char *name, uint64_t address,
 		ret->kind = DRGN_SYMBOL_KIND_UNKNOWN;
 }
 
+LIBDRGN_PUBLIC void drgn_symbol_source(struct drgn_symbol *sym,
+				       struct drgn_program *prog,
+				       unsigned long offset,
+				       size_t size,
+				       char *src_file_ret,
+				       int *line_num_ret,
+				       int *line_col_ret)
+{
+	if (sym->kind != DRGN_SYMBOL_KIND_FUNC)
+		return;
+
+	unsigned long sym_address = sym->address + offset;
+	struct drgn_debug_info_module *dim;
+	dim = drgn_debug_info_module_byaddress(prog->dbinfo, sym_address);
+	if (dim == NULL)
+		return;
+
+	Dwfl_Line *line = dwfl_module_getsrc(dim->dwfl_module, sym_address);
+	if (line == NULL)
+		return;
+
+	const char *src;
+	int lineno, linecol;
+	if ((src = dwfl_lineinfo(line, &sym_address, &lineno, &linecol,
+				 NULL, NULL)) != NULL) {
+		strncpy(src_file_ret, src, size);
+		*line_num_ret = lineno;
+		*line_col_ret = linecol;
+	}
+
+	return;
+}
+
 LIBDRGN_PUBLIC const char *drgn_symbol_name(struct drgn_symbol *sym)
 {
 	return sym->name;