2005-10-06 Frank Ch. Eigler <fche@elastic.org>

	PR 1332.
	* translate.cxx (emit_symbol_data): New function to transcribe
	a processed address->symbol lookup table, based upon /proc/kallsyms.

2005-10-06  Frank Ch. Eigler  <fche@elastic.org>

	PR 1332.
	* sym.h: New file to declare explicit symbol table struct.
	* runtime.h (_stp_kallsyms_lookup_tabled): Use it if available.
	* sym.c (_stp_symbol_sprint): HAS_LOOKUP mooted.

2005-10-06  Frank Ch. Eigler  <fche@elastic.org>

	PR 1132.
	* systemtap.samples/symbols.*: New test.

6 files changed, 167 insertions, 15 deletions

diff --git a/ChangeLog b/ChangeLog
index f806e3fe..6f13cbbb 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+2005-10-06  Frank Ch. Eigler  <fche@elastic.org>
+
+	PR 1332.
+	* translate.cxx (emit_symbol_data): New function to transcribe
+	a processed address->symbol lookup table, based upon /proc/kallsyms.
+
 2005-10-05  Tom Zanussi  <zanussi@us.ibm.com>
 
 	* buildrun.cxx (run_pass): Add bulk/buffer_size flags to flags
diff --git a/runtime/ChangeLog b/runtime/ChangeLog
index 548f908e..196d9486 100644
--- a/runtime/ChangeLog
+++ b/runtime/ChangeLog
@@ -1,3 +1,10 @@
+2005-10-06  Frank Ch. Eigler  <fche@elastic.org>
+
+	PR 1332.
+	* sym.h: New file to declare explicit symbol table struct.
+	* runtime.h (_stp_kallsyms_lookup_tabled): Use it if available.
+	* sym.c (_stp_symbol_sprint): HAS_LOOKUP mooted.
+
 2005-09-30  Graydon Hoare  <graydon@redhat.com>
 
 	* loc2c-runtime.h (_put_user_asm): Fix bracket-matching.
diff --git a/runtime/runtime.h b/runtime/runtime.h
index 94a210a6..bcee85d5 100644
--- a/runtime/runtime.h
+++ b/runtime/runtime.h
@@ -60,29 +60,79 @@ static struct
 #include "string.c"
 #include "arith.c"
 #include "copy.c"
+#include "sym.h"
+
 
 /************* Module Stuff ********************/
-#if defined (__x86_64__) || defined (__i386__)
-#define HAS_LOOKUP 1
 static int (*_stp_kta)(unsigned long addr);
 static const char * (*_stp_kallsyms_lookup)(unsigned long addr,
 					    unsigned long *symbolsize,
 					    unsigned long *offset,
 					    char **modname, char *namebuf);
 
-int init_module (void)
+
+/* This implementation is used if stap_[num_]symbols are available. */
+static const char * _stp_kallsyms_lookup_tabled (unsigned long addr,
+						 unsigned long *symbolsize,
+						 unsigned long *offset,
+						 char **modname,
+						 char *namebuf)
 {
-  _stp_kta = (int (*)(unsigned long))kallsyms_lookup_name("__kernel_text_address");
-  _stp_kallsyms_lookup = (const char * (*)(unsigned long,unsigned long *,unsigned long *,char **,char *))
-    kallsyms_lookup_name("kallsyms_lookup");
-  return _stp_transport_init();
+  unsigned begin = 0;
+  unsigned end = stap_num_symbols;
+  /*const*/ struct stap_symbol* s;
+
+  /* binary search on index [begin,end) */
+  do
+    {
+      unsigned mid = (begin + end) / 2;
+      if (addr < stap_symbols[mid].addr)
+	end = mid;
+      else
+	begin = mid;
+    } while (begin + 1 < end);
+  /* result index in $begin, guaranteed between [0,stap_num_symbols) */
+
+  s = & stap_symbols [begin];
+  if (addr < s->addr)
+    return NULL;
+  else
+    {
+      if (offset) *offset = addr - s->addr;
+      if (modname) *modname = (char *) s->modname;
+      if (symbolsize)
+	{
+	  if ((begin + 1) < stap_num_symbols)
+	    *symbolsize = stap_symbols[begin+1].addr - s->addr;
+	  else
+	    *symbolsize = 0;
+	  // NB: This is only a heuristic.  Sometimes there are large
+	  // gaps between text areas of modules.
+	}
+      if (namebuf)
+	{
+	  strlcpy (namebuf, s->symbol, KSYM_NAME_LEN+1);
+	  return namebuf;
+	}
+      else
+	return s->symbol;
+    }
 }
-#else
+
+
+
 int init_module (void)
 {
+  _stp_kta = (int (*)(unsigned long))kallsyms_lookup_name("__kernel_text_address");
+
+  if (stap_num_symbols > 0)
+    _stp_kallsyms_lookup = & _stp_kallsyms_lookup_tabled;
+  else
+    _stp_kallsyms_lookup = (const char * (*)(unsigned long,unsigned long *,unsigned long *,char **,char *))
+      kallsyms_lookup_name("kallsyms_lookup");
+
   return _stp_transport_init();
 }
-#endif
 
 int probe_start(void);
 
diff --git a/runtime/sym.c b/runtime/sym.c
index 344df373..6e686051 100644
--- a/runtime/sym.c
+++ b/runtime/sym.c
@@ -18,7 +18,6 @@
 
 String _stp_symbol_sprint (String str, unsigned long address)
 { 
-#ifdef HAS_LOOKUP
 	char *modname;
         const char *name;
         unsigned long offset, size;
@@ -34,9 +33,6 @@ String _stp_symbol_sprint (String str, unsigned long address)
 		else
 			_stp_sprintf (str, " : %s+%#lx/%#lx", name, offset, size);
 	}
-#else
-	_stp_sprintf (str, "0x%lx", address);
-#endif
 	return str;
 }
 
diff --git a/runtime/sym.h b/runtime/sym.h
new file mode 100644
index 00000000..6ed08b06
--- /dev/null
+++ b/runtime/sym.h
@@ -0,0 +1,25 @@
+/*
+ * Copyright (C) 2005 Red Hat Inc.
+ *
+ * This file is part of systemtap, and is free software.  You can
+ * redistribute it and/or modify it under the terms of the GNU General
+ * Public License (GPL); either version 2, or (at your option) any
+ * later version.
+ */
+
+#ifndef _STAP_SYMBOLS_H_
+#define _STAP_SYMBOLS_H_
+
+
+/* A symbol table defined by the translator. */
+struct stap_symbol {
+  unsigned long addr;
+  const char *symbol;
+  const char *modname; 
+};
+
+extern struct stap_symbol stap_symbols [];
+extern unsigned stap_num_symbols;
+
+
+#endif /* _RUNTIME_H_ */
diff --git a/translate.cxx b/translate.cxx
index a7caf0bd..981d71df 100644
--- a/translate.cxx
+++ b/translate.cxx
@@ -2362,6 +2362,72 @@ c_unparser::visit_functioncall (functioncall* e)
 }
 
 
+
+int
+emit_symbol_data (systemtap_session& s)
+{
+  int rc = 0;
+
+  // Instead of processing elf symbol tables, for now we just snatch
+  // /proc/kallsyms and convert it to our use.  We need it sorted by
+  // address (so we can binary search) , and filtered (to show text
+  // symbols only), a task that we defer to grep(1) and sort(1).  It
+  // may be useful to cache the symbols.sorted file, perhaps indexed
+  // by md5sum(/proc/modules), but let's not until this simple method
+  // proves too costly.  All that LC_ALL=C stuff is there to avoid the
+  // excessive penalty of i18n code in some glibc/coreutils versions.
+
+  string sorted_kallsyms = s.tmpdir + "/symbols.sorted";
+  string sortcmd = "/bin/env LC_ALL=C /bin/grep \" [tT] \" /proc/kallsyms | ";
+ 
+  sortcmd += "/bin/env LC_ALL=C /bin/sort ";
+#if __LP64__
+  sortcmd += "-k 1,16 ";
+#else
+  sortcmd += "-k 1,8 ";
+#endif
+  sortcmd += "-s -o " + sorted_kallsyms;
+
+  if (s.verbose) clog << "Running " << sortcmd << endl;
+  rc = system(sortcmd.c_str());
+  if (rc == 0)
+    {
+      ifstream kallsyms (sorted_kallsyms.c_str());
+
+      unsigned i=0;
+      s.op->newline() << "struct stap_symbol stap_symbols [] = {";
+      s.op->indent(1);
+      string lastaddr;
+      while (! kallsyms.eof())
+	{
+	  string addr, type, sym, module;
+	  kallsyms >> addr >> type >> sym;
+	  kallsyms >> ws;
+	  if (kallsyms.peek() == '[')
+	    {
+	      string bracketed;
+	      kallsyms >> bracketed;
+	      module = bracketed.substr (1, bracketed.length()-2);
+	    }
+	  
+	  // NB: kallsyms includes some duplicate addresses
+	  if ((type == "t" || type == "T") && lastaddr != addr)
+	    {
+	      s.op->newline() << "{ 0x" << addr << ", "
+			      << "\"" << sym << "\", "
+			      << "\"" << module << "\" },";
+	      lastaddr = addr;
+	      i ++;
+	    }
+	}
+      s.op->newline(-1) << "};";
+      s.op->newline() << "unsigned stap_num_symbols = " << i << ";" << endl;
+    }
+
+  return rc;
+}
+
+
 int
 translate_pass (systemtap_session& s)
 {
@@ -2468,16 +2534,18 @@ translate_pass (systemtap_session& s)
       s.op->newline() << "MODULE_DESCRIPTION(\"systemtap probe\");";
       s.op->newline() << "MODULE_LICENSE(\"GPL\");"; // XXX
       s.op->newline() << "#endif";
-
-      s.op->line() << endl;
     }
   catch (const semantic_error& e)
     {
       s.print_error (e);
     }
 
+  rc |= emit_symbol_data (s);
+  s.op->line() << endl;
+
   delete s.op;
   s.op = 0;
   s.up = 0;
+
   return rc + s.num_errors;
 }