* dll_init.cc (dll_list::find_by_modname): New function to search the dll list

for a module name only (no path).
(dll_list::alloc): Initialize newly-added members of struct dll.
(dll_list::append): New function to factor out the append operation
(used by dll_list::topsort).
(dll_list::populate_deps): New function to identify dll dependencies.
(dll_list::topsort): New function to sort the dll list topologically by
dependencies.
(dll_list::topsort_visit): New helper function for the above.
* dll_init.h (dll::ndeps): New class member.
(dll::deps): Ditto.
(dll::modname): Ditto.
(dll_list::find_by_modname): New function related to topsort.
(dll_list::populate_all_deps): Ditto.
(dll_list::populate_deps): Ditto.
(dll_list::topsort): Ditto.
(dll_list::topsort_visit): Ditto.
(dll_list::append): Ditto.
(pefile): New struct allowing simple introspection of dll images.
* fork.cc (fork): Topologically sort the dll list before forking.

4 files changed, 191 insertions, 9 deletions

diff --git a/winsup/cygwin/ChangeLog b/winsup/cygwin/ChangeLog
index 1670f7f8f..aba346b9e 100644
--- a/winsup/cygwin/ChangeLog
+++ b/winsup/cygwin/ChangeLog
@@ -1,5 +1,28 @@
+2011-05-28  Ryan Johnson  <ryan.johnson@cs.utoronto.ca>
+
+	* dll_init.cc (dll_list::find_by_modname): New function to search the
+	dll list for a module name only (no path).
+	(dll_list::alloc): Initialize newly-added members of struct dll.
+	(dll_list::append): New function to factor out the append operation
+	(used by dll_list::topsort).
+	(dll_list::populate_deps): New function to identify dll dependencies.
+	(dll_list::topsort): New function to sort the dll list topologically by
+	dependencies.
+	(dll_list::topsort_visit): New helper function for the above.
+	* dll_init.h (dll::ndeps): New class member.
+	(dll::deps): Ditto.
+	(dll::modname): Ditto.
+	(dll_list::find_by_modname): New function related to topsort.
+	(dll_list::populate_all_deps): Ditto.
+	(dll_list::populate_deps): Ditto.
+	(dll_list::topsort): Ditto.
+	(dll_list::topsort_visit): Ditto.
+	(dll_list::append): Ditto.
+	(pefile): New struct allowing simple introspection of dll images.
+	* fork.cc (fork): Topologically sort the dll list before forking.
+
 2011-05-28  Christopher Faylor  <me.cygwin2011@cgf.cx>
-	    Ryan Johnson  <ryan.johnson@cs.utoronto.ca>
+	    
 
 	* dll_init.c (dll_list::load_after_fork): Don't clear in_forkee here.
 	* fork.cc (frok::errmsg): Rename from 'error'.
diff --git a/winsup/cygwin/dll_init.cc b/winsup/cygwin/dll_init.cc
index 68a974b7b..6fc117c0a 100644
--- a/winsup/cygwin/dll_init.cc
+++ b/winsup/cygwin/dll_init.cc
@@ -116,6 +116,18 @@ dll_list::operator[] (const PWCHAR name)
   return NULL;
 }
 
+/* Look for a dll based on is short name only (no path) */
+dll *
+dll_list::find_by_modname (const PWCHAR name)
+{
+  dll *d = &start;
+  while ((d = d->next) != NULL)
+    if (!wcscasecmp (name, d->modname))
+      return d;
+
+  return NULL;
+}
+
 #define RETRIES 1000
 
 /* Allocate space for a dll struct. */
@@ -161,14 +173,13 @@ dll_list::alloc (HINSTANCE h, per_process *p, dll_type type)
       d->handle = h;
       d->has_dtors = true;
       d->p = p;
+      d->ndeps = 0;
+      d->deps = NULL;
+      d->modname = wcsrchr (d->name, L'\\');
+      if (d->modname)
+	d->modname++;
       d->type = type;
-      if (end == NULL)
-	end = &start;	/* Point to "end" of dll chain. */
-      end->next = d;	/* Standard linked list stuff. */
-      d->next = NULL;
-      d->prev = end;
-      end = d;
-      tot++;
+      append (d);
       if (type == DLL_LOAD)
 	loaded_dlls++;
     }
@@ -177,6 +188,119 @@ dll_list::alloc (HINSTANCE h, per_process *p, dll_type type)
   return d;
 }
 
+void
+dll_list::append (dll* d)
+{
+  if (end == NULL)
+    end = &start;	/* Point to "end" of dll chain. */
+  end->next = d;	/* Standard linked list stuff. */
+  d->next = NULL;
+  d->prev = end;
+  end = d;
+  tot++;
+}
+
+void dll_list::populate_deps (dll* d)
+{
+  WCHAR wmodname[NT_MAX_PATH];
+  pefile* pef = (pefile*) d->handle;
+  PIMAGE_DATA_DIRECTORY dd = pef->idata_dir (IMAGE_DIRECTORY_ENTRY_IMPORT);
+  /* Annoyance: calling crealloc with a NULL pointer will use the
+     wrong heap and crash, so we have to replicate some code */
+  long maxdeps = 4;
+  d->deps = (dll**) cmalloc (HEAP_2_DLL, maxdeps*sizeof (dll*));
+  d->ndeps = 0;
+  for (PIMAGE_IMPORT_DESCRIPTOR id=
+	(PIMAGE_IMPORT_DESCRIPTOR) pef->rva (dd->VirtualAddress);
+      dd->Size && id->Name;
+      id++)
+    {
+      char* modname = pef->rva (id->Name);
+      sys_mbstowcs (wmodname, NT_MAX_PATH, modname);
+      if (dll* dep = find_by_modname (wmodname))
+	{
+	  if (d->ndeps >= maxdeps)
+	    {
+	      maxdeps = 2*(1+maxdeps);
+	      d->deps = (dll**) crealloc (d->deps, maxdeps*sizeof (dll*));
+	    }
+	  d->deps[d->ndeps++] = dep;
+	}
+    }
+  
+  /* add one to differentiate no deps from unknown */
+  d->ndeps++;
+}
+
+
+void
+dll_list::topsort ()
+{
+  /* Anything to do? */
+  if (!end)
+    return;
+  
+  /* make sure we have all the deps available */
+  dll* d = &start;
+  while ((d = d->next))
+    if (!d->ndeps)
+      populate_deps (d);
+  
+  /* unlink head and tail pointers so the sort can rebuild the list */
+  d = start.next;
+  start.next = end = NULL;
+  topsort_visit (d, true);
+
+  /* clear node markings made by the sort */
+  d = &start;
+  while ((d = d->next))
+    {
+      debug_printf ("%W", d->modname);
+      for (int i=1; i < -d->ndeps; i++)
+	debug_printf ("-> %W", d->deps[i-1]->modname);
+
+      /* It would be really nice to be able to keep this information
+	 around for next time, but we don't have an easy way to
+	 invalidate cached dependencies when a module unloads. */
+      d->ndeps = 0;
+      cfree (d->deps);
+      d->deps = NULL;
+    }
+}
+
+/* A recursive in-place topological sort. The result is ordered so that
+   dependencies of a dll appear before it in the list.
+
+   NOTE: this algorithm is guaranteed to terminate with a "partial
+   order" of dlls but does not do anything smart about cycles: an
+   arbitrary dependent dll will necessarily appear first. Perhaps not
+   surprisingly, Windows ships several dlls containing dependency
+   cycles, including SspiCli/RPCRT4.dll and a lovely tangle involving
+   USP10/LPK/GDI32/USER32.dll). Fortunately, we don't care about
+   Windows DLLs here, and cygwin dlls should behave better */
+void
+dll_list::topsort_visit (dll* d, bool seek_tail)
+{
+  /* Recurse to the end of the dll chain, then visit nodes as we
+     unwind. We do this because once we start visiting nodes we can no
+     longer trust any _next_ pointers.
+
+     We "mark" visited nodes (to avoid revisiting them) by negating
+     ndeps (undone once the sort completes). */
+  if (seek_tail && d->next)
+    topsort_visit (d->next, true);
+  
+  if (d->ndeps > 0)
+    {
+      d->ndeps = -d->ndeps;
+      for (long i=1; i < -d->ndeps; i++)
+	topsort_visit (d->deps[i-1], false);
+
+      append (d);
+    }
+}
+
+
 dll *
 dll_list::find (void *retaddr)
 {
diff --git a/winsup/cygwin/dll_init.h b/winsup/cygwin/dll_init.h
index d14cc6cb9..3a9a5c4f0 100644
--- a/winsup/cygwin/dll_init.h
+++ b/winsup/cygwin/dll_init.h
@@ -52,6 +52,9 @@ struct dll
   int count;
   bool has_dtors;
   dll_type type;
+  long ndeps;
+  dll** deps;
+  PWCHAR modname;
   WCHAR name[1];
   void detach ();
   int init ();
@@ -84,6 +87,13 @@ public:
   void detach (void *);
   void init ();
   void load_after_fork (HANDLE);
+  dll *find_by_modname (const PWCHAR name);
+  void populate_all_deps ();
+  void populate_deps (dll* d);
+  void topsort ();
+  void topsort_visit (dll* d, bool goto_tail);
+  void append (dll* d);
+  
   dll *inext ()
   {
     while ((hold = hold->next))
@@ -109,6 +119,25 @@ public:
   dll_list () { protect.init ("dll_list"); }
 };
 
+/* References:
+   http://msdn.microsoft.com/en-us/windows/hardware/gg463125
+   http://msdn.microsoft.com/en-us/library/ms809762.aspx
+*/
+/* FIXME: Integrate with other similar uses in source. */
+struct pefile
+{
+  IMAGE_DOS_HEADER dos_hdr;
+
+  char* rva (long offset) { return (char*) this + offset; }
+  PIMAGE_NT_HEADERS32 pe_hdr () { return (PIMAGE_NT_HEADERS32) rva (dos_hdr.e_lfanew); }
+  PIMAGE_OPTIONAL_HEADER32 optional_hdr () { return &pe_hdr ()->OptionalHeader; }
+  PIMAGE_DATA_DIRECTORY idata_dir (DWORD which)
+  {
+    PIMAGE_OPTIONAL_HEADER32 oh = optional_hdr ();
+    return (which < oh->NumberOfRvaAndSizes)? oh->DataDirectory + which : 0;
+  }
+};
+
 extern dll_list dlls;
 void dll_global_dtors ();
 
diff --git a/winsup/cygwin/fork.cc b/winsup/cygwin/fork.cc
index 3ec7fa8c5..1ac2b8a45 100644
--- a/winsup/cygwin/fork.cc
+++ b/winsup/cygwin/fork.cc
@@ -625,7 +625,7 @@ fork ()
        the thread-local storage.  A process forking too deeply will run into
        the problem to be out of temporary TLS path buffers. */
     tmp_pathbuf tp;
-
+  
     if (!held_everything)
       {
 	if (exit_state)
@@ -634,6 +634,12 @@ fork ()
 	return -1;
       }
 
+    /* Put the dll list in topological dependency ordering, in
+       hopes that the child will have a better shot at loading dlls
+       properly if it only has to deal with one at a time.
+    */
+    dlls.topsort ();
+
     ischild = !!setjmp (grouped.ch.jmp);
 
     volatile char * volatile esp;