PR6866: First pass at translating addresses to symbol names through vma.

13 files changed, 238 insertions, 27 deletions

diff --git a/ChangeLog b/ChangeLog
index 6b31fb76..c46c0ef2 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,10 @@
+2008-12-10  Mark Wielaard  <mjw@redhat.com>
+
+	* tapsets.cxx
+	(utrace_derived_probe_group::emit_vm_callback_probe_decl):
+	Test for STP_NEED_TASK_FINDER_VMA to enable
+	emit_vm_callback_probe_decl.
+
 2008-12-09  Frank Ch. Eigler  <fche@elastic.org>
 
 	PR6961
diff --git a/runtime/ChangeLog b/runtime/ChangeLog
index 3f88bae2..d83ac12c 100644
--- a/runtime/ChangeLog
+++ b/runtime/ChangeLog
@@ -1,3 +1,14 @@
+2008-12-10  Mark Wielaard  <mjw@redhat.com>
+
+	* runtime.h: Define __stp_tf_vma_entry here.
+	* sym.c (_stp_mod_sec_lookup): Try vma matching first if task given.
+	(_stp_kallsyms_lookup): Take an optional task.
+	(_stp_symbol_snprint): Likewise.
+	task_finder.c (__stp_tf_vm_cb): Use vm_path to match and attach
+	module to vma entry.
+	task_finder_vma.c (stap_add_vma_map_info): Take an optional module.
+	(__stp_tf_get_vma_entry_addr): New lookup function.
+
 2008-12-09  Frank Ch. Eigler  <fche@elastic.org>
 
 	* time.c (_stp_gettimeofday_ns): Protect some more against freq=0.
diff --git a/runtime/runtime.h b/runtime/runtime.h
index 3b3e117d..055d3f27 100644
--- a/runtime/runtime.h
+++ b/runtime/runtime.h
@@ -86,6 +86,24 @@ static struct
 #include "io.c"
 #include "arith.c"
 #include "copy.c"
+
+/* Lifted task_finder, internal details used in sym.c - XXX */
+struct __stp_tf_vma_entry {
+  struct hlist_node hlist;
+
+  pid_t pid;
+  unsigned long addr;
+  unsigned long vm_start;
+  unsigned long vm_end;
+  unsigned long vm_pgoff;
+  // Is that enough?  Should we store a dcookie for vm_file?              
+
+  // Module that this vma entry is mapped from, if any.                   
+  struct _stp_module *module;
+};
+static struct __stp_tf_vma_entry *
+__stp_tf_get_vma_entry_addr(struct task_struct *, unsigned long);
+
 #include "sym.c"
 #ifdef STP_PERFMON
 #include "perf.c"
diff --git a/runtime/sym.c b/runtime/sym.c
index 06ac14a5..9b295f58 100644
--- a/runtime/sym.c
+++ b/runtime/sym.c
@@ -77,11 +77,28 @@ unsigned long _stp_module_relocate(const char *module, const char *section, unsi
    if found, return NULL otherwise.
    XXX: needs to be address-space-specific. */
 static struct _stp_module *_stp_mod_sec_lookup(unsigned long addr,
+					       struct task_struct *task,
 					       struct _stp_section **sec)
 {
   struct _stp_module *m = NULL;
   unsigned midx = 0;
   unsigned long closest_section_offset = ~0;
+
+  // Try vma matching first if task given.
+  struct __stp_tf_vma_entry *entry;
+  if (task)
+    {
+      entry = __stp_tf_get_vma_entry_addr(task, addr);
+      if (entry != NULL && entry->module != NULL)
+	{
+	  m = entry->module;
+	  *sec = & m->sections[0]; // XXX check actual section and relocate
+	  if (strcmp(".dynamic", m->sections[0].name) == 0)
+	    m->sections[0].addr = entry->vm_start; // cheat...
+	  return m;
+	}
+    }
+
   for (midx = 0; midx < _stp_num_modules; midx++)
     {
       unsigned secidx;
@@ -109,7 +126,8 @@ static const char *_stp_kallsyms_lookup(unsigned long addr, unsigned long *symbo
                                         unsigned long *offset, 
                                         const char **modname, 
                                         /* char ** secname? */
-                                        char *namebuf)
+                                        char *namebuf,
+					struct task_struct *task)
 {
 	struct _stp_module *m = NULL;
 	struct _stp_section *sec = NULL;
@@ -117,7 +135,7 @@ static const char *_stp_kallsyms_lookup(unsigned long addr, unsigned long *symbo
 	unsigned long flags;
 	unsigned end, begin = 0;
 
-	m = _stp_mod_sec_lookup(addr, &sec);
+	m = _stp_mod_sec_lookup(addr, task, &sec);
         if (unlikely (m == NULL || sec == NULL))
           return NULL;
         
@@ -242,7 +260,7 @@ void _stp_symbol_print(unsigned long address)
 	const char *name;
 	unsigned long offset, size;
 
-	name = _stp_kallsyms_lookup(address, &size, &offset, &modname, NULL);
+	name = _stp_kallsyms_lookup(address, &size, &offset, &modname, NULL, NULL);
 
 	_stp_printf("%p", (int64_t) address);
 
@@ -267,7 +285,7 @@ int _stp_func_print(unsigned long address, int verbose, int exact)
 	else
 		exstr = " (inexact)";
 
-	name = _stp_kallsyms_lookup(address, &size, &offset, &modname, NULL);
+	name = _stp_kallsyms_lookup(address, &size, &offset, &modname, NULL, NULL);
 
 	if (name) {
 		if (verbose) {
@@ -283,13 +301,15 @@ int _stp_func_print(unsigned long address, int verbose, int exact)
 	return 0;
 }
 
-void _stp_symbol_snprint(char *str, size_t len, unsigned long address)
+void _stp_symbol_snprint(char *str, size_t len, unsigned long address,
+			 struct task_struct *task)
 {
 	const char *modname;
 	const char *name;
 	unsigned long offset, size;
 
-	name = _stp_kallsyms_lookup(address, &size, &offset, &modname, NULL);
+	name = _stp_kallsyms_lookup(address, &size, &offset, &modname, NULL,
+				    task);
 	if (name)
 		strlcpy(str, name, len);
 	else
diff --git a/runtime/task_finder.c b/runtime/task_finder.c
index 25fad1a4..1e0a8474 100644
--- a/runtime/task_finder.c
+++ b/runtime/task_finder.c
@@ -55,7 +55,6 @@ typedef int (*stap_task_finder_vm_callback)(struct stap_task_finder_target *tgt,
 					    unsigned long vm_end,
 					    unsigned long vm_pgoff);
 
-#ifdef DEBUG_TASK_FINDER_VMA
 int __stp_tf_vm_cb(struct stap_task_finder_target *tgt,
 		   struct task_struct *tsk,
 		   int map_p, char *vm_path,
@@ -63,21 +62,32 @@ int __stp_tf_vm_cb(struct stap_task_finder_target *tgt,
 		   unsigned long vm_end,
 		   unsigned long vm_pgoff)
 {
+  int i;
+#ifdef DEBUG_TASK_FINDER_VMA
 	_stp_dbug(__FUNCTION__, __LINE__,
 		  "vm_cb: tsk %d:%d path %s, start 0x%08lx, end 0x%08lx, offset 0x%lx\n",
 		  tsk->pid, map_p, vm_path, vm_start, vm_end, vm_pgoff);
+#endif
 	if (map_p) {
-		// FIXME: What should we do with vm_path?  We can't save
-		// the vm_path pointer itself, but we don't have any
-		// storage space allocated to save it in...
-		stap_add_vma_map_info(tsk, vm_start, vm_end, vm_pgoff);
+	  struct _stp_module *module = NULL;
+	  if (vm_path != NULL)
+	    for (i = 0; i < _stp_num_modules; i++)
+	      if (strcmp(vm_path, _stp_modules[i]->name) == 0)
+		{
+#ifdef DEBUG_TASK_FINDER_VMA
+		  _stp_dbug(__FUNCTION__, __LINE__,
+			    "vm_cb: matched path %s to module\n", vm_path);
+#endif
+		  module = _stp_modules[i];
+		  break;
+		}
+	  stap_add_vma_map_info(tsk, vm_start, vm_end, vm_pgoff, module);
 	}
 	else {
 		stap_remove_vma_map_info(tsk, vm_start, vm_end, vm_pgoff);
 	}
 	return 0;
 }
-#endif
 
 struct stap_task_finder_target {
 /* private: */
diff --git a/runtime/task_finder_vma.c b/runtime/task_finder_vma.c
index 4dce4be8..b65d9ea4 100644
--- a/runtime/task_finder_vma.c
+++ b/runtime/task_finder_vma.c
@@ -16,17 +16,6 @@ static DEFINE_MUTEX(__stp_tf_vma_mutex);
 #define TASK_FINDER_VMA_ENTRY_ITEMS 100
 #endif
 
-struct __stp_tf_vma_entry {
-	struct hlist_node hlist;
-
-	pid_t pid;
-	unsigned long addr;
-	unsigned long vm_start;
-	unsigned long vm_end;
-	unsigned long vm_pgoff;
-	// Is that enough?  Should we store a dcookie for vm_file?
-};
-
 static struct __stp_tf_vma_entry
 __stp_tf_vma_free_list_items[TASK_FINDER_VMA_ENTRY_ITEMS];
 
@@ -211,7 +200,8 @@ __stp_tf_get_vma_map_entry_internal(struct task_struct *tsk,
 // Add the vma info to the vma map hash table.
 static int
 stap_add_vma_map_info(struct task_struct *tsk, unsigned long vm_start,
-			  unsigned long vm_end, unsigned long vm_pgoff)
+		      unsigned long vm_end, unsigned long vm_pgoff,
+		      struct _stp_module *module)
 {
 	struct hlist_head *head;
 	struct hlist_node *node;
@@ -242,6 +232,7 @@ stap_add_vma_map_info(struct task_struct *tsk, unsigned long vm_start,
 	entry->vm_start = vm_start;
 	entry->vm_end = vm_end;
 	entry->vm_pgoff = vm_pgoff;
+	entry->module = module;
 
 	head = &__stp_tf_vma_map[__stp_tf_vma_map_hash(tsk)];
 	hlist_add_head(&entry->hlist, head);
@@ -305,3 +296,26 @@ stap_find_vma_map_info(struct task_struct *tsk, unsigned long vm_addr,
 	mutex_unlock(&__stp_tf_vma_mutex);
 	return rc;
 }
+
+// Get vma_entry of the address (vm_start/vm_end) if the vma is
+// present in the vma hash table containing.
+// Returns NULL if not present.
+static struct __stp_tf_vma_entry *
+__stp_tf_get_vma_entry_addr(struct task_struct *tsk, unsigned long addr)
+{
+  struct hlist_head *head;
+  struct hlist_node *node;
+  struct __stp_tf_vma_entry *entry;
+  
+  mutex_lock(&__stp_tf_vma_mutex);
+  head = &__stp_tf_vma_map[__stp_tf_vma_map_hash(tsk)];
+  hlist_for_each_entry(entry, node, head, hlist) {
+    if (tsk->pid == entry->pid
+	&& addr >= entry->vm_start && addr < entry->vm_end) {
+      mutex_unlock(&__stp_tf_vma_mutex);
+      return entry;
+    }
+  }
+  mutex_unlock(&__stp_tf_vma_mutex);
+  return NULL;
+}
diff --git a/tapset/ChangeLog b/tapset/ChangeLog
index 626ad67b..30634bcc 100644
--- a/tapset/ChangeLog
+++ b/tapset/ChangeLog
@@ -1,3 +1,10 @@
+2008-12-10  Mark Wielaard  <mjw@redhat.com>
+
+	* context-symbols.stp: Define STP_NEED_TASK_FINDER_VMA.
+	(symbolname): New function.
+	* context-unwind.stp (caller): Pass current task to
+	_stp_symbol_snprint.
+
 2008-12-09  Frank Ch. Eigler  <fche@elastic.org>
 
 	PR 6961.
diff --git a/tapset/context-symbols.stp b/tapset/context-symbols.stp
index 79645f4f..fbb51767 100644
--- a/tapset/context-symbols.stp
+++ b/tapset/context-symbols.stp
@@ -11,6 +11,9 @@
 #ifndef STP_NEED_SYMBOL_DATA
 #define STP_NEED_SYMBOL_DATA 1
 #endif
+#ifndef STP_NEED_TASK_FINDER_VMA
+#define STP_NEED_TASK_FINDER_VMA 1
+#endif
 %}
 
 /**
@@ -93,3 +96,8 @@ function probemod:string () %{ /* pure */
 		THIS->__retvalue[0] = '\0';
 	}
 %}
+
+function symbolname:string (addr:long) %{ /* pure */
+	 _stp_symbol_snprint(THIS->__retvalue, MAXSTRINGLEN, THIS->addr,
+	                     current);
+%}
diff --git a/tapset/context-unwind.stp b/tapset/context-unwind.stp
index 4c5ed34b..59d111ee 100644
--- a/tapset/context-unwind.stp
+++ b/tapset/context-unwind.stp
@@ -51,7 +51,8 @@ function backtrace:string () %{ /* pure */
 function caller:string() %{ /* pure */
         if (CONTEXT->pi) 
 		_stp_symbol_snprint( THIS->__retvalue, MAXSTRINGLEN, 
-			(unsigned long)_stp_ret_addr_r(CONTEXT->pi));
+			(unsigned long)_stp_ret_addr_r(CONTEXT->pi),
+                        current);
         else
 		strlcpy(THIS->__retvalue,"unknown",MAXSTRINGLEN);
 %}
diff --git a/tapsets.cxx b/tapsets.cxx
index 4d9a021d..9fe1d236 100644
--- a/tapsets.cxx
+++ b/tapsets.cxx
@@ -6800,7 +6800,7 @@ utrace_derived_probe_group::emit_module_decls (systemtap_session& s)
 	  // Emit a "fake" probe decl that is really a hook for to get
 	  // our vm_callback called.
 	  string path = it->first;
-	  s.op->newline() << "#ifdef DEBUG_TASK_FINDER_VMA";
+	  s.op->newline() << "#ifdef STP_NEED_TASK_FINDER_VMA";
 	  emit_vm_callback_probe_decl (s, true, path, (int64_t)0,
 				       "__stp_tf_vm_cb");
 	  s.op->newline() << "#endif";
@@ -6821,7 +6821,7 @@ utrace_derived_probe_group::emit_module_decls (systemtap_session& s)
         {
 	  // Emit a "fake" probe decl that is really a hook for to get
 	  // our vm_callback called.
-	  s.op->newline() << "#ifdef DEBUG_TASK_FINDER_VMA";
+	  s.op->newline() << "#ifdef STP_NEED_TASK_FINDER_VMA";
 	  emit_vm_callback_probe_decl (s, false, "", it->first,
 				       "__stp_tf_vm_cb");
 	  s.op->newline() << "#endif";
diff --git a/testsuite/ChangeLog b/testsuite/ChangeLog
index 8e174efc..44261176 100644
--- a/testsuite/ChangeLog
+++ b/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2008-12-10  Mark Wielaard  <mjw@redhat.com>
+
+	* systemtap.context/usymbols.c: New test program.
+	* systemtap.context/usymbols.exp: New dejagnu test.
+
 2008-12-09  Frank Ch. Eigler  <fche@elastic.org>
 
 	PR6961.
diff --git a/testsuite/systemtap.context/usymbols.c b/testsuite/systemtap.context/usymbols.c
new file mode 100644
index 00000000..f8ee05b5
--- /dev/null
+++ b/testsuite/systemtap.context/usymbols.c
@@ -0,0 +1,41 @@
+/* usymbol test case
+ * Copyright (C) 2008, 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.
+ *
+ * Uses signal to tranfer user space addresses into the kernel where a
+ * probe on sigaction will extract them and produce the symbols.  To
+ * poke into the executable we get the sa_handler, to poke into glibc
+ * we get the sa_restorer fields in the stap script.
+ *
+ * XXX - Seems sa_restorer isn't set on all architectures.  should use
+ * our own shared library and set signal handler from there.  Also
+ * need to handle @plt symbols (setting a handler in the main
+ * executable that is in a shared library will have the @plt address,
+ * not the address inside the shared library).
+ */
+
+#include <signal.h>
+typedef void (*sighandler_t)(int);
+
+void
+handler (int signum)
+{
+  /* dummy handler, just used for the address... */
+}
+
+sighandler_t
+libc_handler (void *func)
+{
+  return (sighandler_t) func;
+}
+
+int
+main (int argc, char *argv[], char *envp[])
+{
+  // Use SIGFPE since we never expect that to be triggered.
+  signal(SIGFPE, handler);
+}
diff --git a/testsuite/systemtap.context/usymbols.exp b/testsuite/systemtap.context/usymbols.exp
new file mode 100644
index 00000000..ebaa058e
--- /dev/null
+++ b/testsuite/systemtap.context/usymbols.exp
@@ -0,0 +1,69 @@
+set test "./usymbols"
+set testpath "$srcdir/$subdir"
+set testsrc  "$testpath/usymbols.c"
+set testexe  "[pwd]/usymbols"
+set testflags "additional_flags=-g additional_flags=-O"
+
+# Only run on make installcheck
+if {! [installtest_p]} { untested "$test -p5"; return }
+
+# Compile out test program
+set res [target_compile $testsrc $testexe executable $testflags]
+if { $res != "" } {
+    verbose "target_compile failed: $res" 2
+    fail "unable to compile $testsrc"
+    return
+}
+
+# We need the execname() trick to work around (the workaround of) PR6964
+# otherwise we get also the rt_sigactions of stapio. Get the handler
+# (comes from the executable) and the restorer (comes from glibc).
+set testscript {
+    probe syscall.rt_sigaction {
+      if (pid() == target() && execname() == "%s") {
+	  handler = $act->sa_handler;
+	  printf("handler: %%s\n", symbolname(handler));
+	  restorer = $act->sa_restorer;
+	  printf("restorer: %%s\n", symbolname(restorer));   
+      }
+    }
+    probe process("%s").syscall { printf(""); /* XXX trigger tracker */ }
+}
+
+set output {handler: handler
+restorer: __restore_rt}
+
+# Got to run stap with both the exe and the libraries used as -d args.
+# XXX Note how we need the fully resolved (absolute) path...
+set script [format $testscript usymbols $testexe]
+catch {eval exec [concat ldd $testexe | grep libc.so]} libc
+set libc [lindex [split $libc " "] 2]
+send_log "libc: $libc\n"
+if {[string equal "link" [file type $libc]]} {
+    set libc [file join [file dirname $libc] [file readlink $libc]]
+}
+send_log "libc: $libc\n"
+set cmd [concat stap -d $libc -d $testexe -c $testexe -e {$script}]
+send_log "cmd: $cmd\n"
+catch {eval exec $cmd} res
+send_log "cmd output: $res\n"
+
+set n 0
+set m [llength [split $output "\n"]]
+set expected [split $output "\n"]
+foreach line [split $res "\n"] {
+    if {![string equal $line [lindex $expected $n]]} {
+      fail usymbols
+      send_log "line [expr $n + 1]: expected \"[lindex $expected $n]\", "
+      send_log "Got \"$line\"\n"
+      return
+    }
+    incr n
+}
+if { $n != $m } {
+  fail usymbols
+  send_log "Got \"$n\" lines, expected \"$m\" lines\n"
+} else {
+  pass usymbols
+}
+exec rm -f $testexe