bfd/

	* elf64-ppc.c: Define more insns used in plt call stubs.
	(ppc64_elf_brtaken_reloc): Assume isa version 2 or above.
	(ppc64_elf_relocate_section): Likewise.
	(enum ppc_stub_type): Add ppc_stub_plt_call_r2save.
	(struct ppc_link_hash_table): Increase size of stub_count array.
	Add plt_stub_align and plt_thread_safe.
	(ALWAYS_USE_FAKE_DEP, ALWAYS_EMIT_R2SAVE): Define.
	(plt_stub_size, plt_stub_pad): New functions.
	(build_plt_stub): Emit barriers for power7 thread safety.  Don't
	emit needless save of r2.
	(build_tls_get_addr_stub): Adjust params.
	(ppc_build_one_stub): Handle ppc_stub_plt_call_r2save and aligning
	plt stubs.  Adjust build_*plt_stub calls.
	(ppc_size_one_stub): Similarly.
	(ppc64_elf_size_stubs): Accept plt_thread_safe and plt_stub_align
	params.  Choose default for plt_thread_safe based on existence of
	calls to thread creation functions.  Modify plt_call to
	plt_call_r2save when no tocsave reloc found.  Align tail of stub
	sections.
	(ppc64_elf_build_stubs): Align tail of stub sections.  Adjust
	output of stub statistics.
	(ppc64_elf_relocate_section): Handle ppc_stub_plt_call_r2save.
	* elf64-ppc.h (ppc64_elf_size_stubs): Update prototype.
ld/
	* emultempl/ppc64elf.em (PARSE_AND_LIST_PROLOGUE,
	PARSE_AND_LIST_LONGOPTS, PARSE_AND_LIST_OPTIONS,
	PARSE_AND_LIST_ARGS_CASES): Handle --{no-,}plt-thread-safe and
	--{no-,}plt-align.
	(plt_thread_safe, plt_stub_align): New vars.
	(gld${EMULATION_NAME}_after_allocation): Pass them to
	ppc64_elf_size_stubs.  Align stub sections according to plt_stub_align.
	* ld.texinfo: Document new command line options, and an old
	undocumented option.

3 files changed, 122 insertions, 8 deletions

diff --git a/ld/ChangeLog b/ld/ChangeLog
index d779176..8d4027f 100644
--- a/ld/ChangeLog
+++ b/ld/ChangeLog
@@ -1,3 +1,15 @@
+2012-01-11  Alan Modra  <amodra@gmail.com>
+
+	* emultempl/ppc64elf.em (PARSE_AND_LIST_PROLOGUE,
+	PARSE_AND_LIST_LONGOPTS, PARSE_AND_LIST_OPTIONS,
+	PARSE_AND_LIST_ARGS_CASES): Handle --{no-,}plt-thread-safe and
+	--{no-,}plt-align.
+	(plt_thread_safe, plt_stub_align): New vars.
+	(gld${EMULATION_NAME}_after_allocation): Pass them to
+	ppc64_elf_size_stubs.  Align stub sections according to plt_stub_align.
+	* ld.texinfo: Document new command line options, and an old
+	undocumented option.
+
 2012-01-09  Roland McGrath  <mcgrathr@google.com>
 
 	* configure.in: Use AM_ZLIB.
diff --git a/ld/emultempl/ppc64elf.em b/ld/emultempl/ppc64elf.em
index 9c352ee..5b637e1 100644
--- a/ld/emultempl/ppc64elf.em
+++ b/ld/emultempl/ppc64elf.em
@@ -1,5 +1,5 @@
 # This shell script emits a C file. -*- C -*-
-# Copyright 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
+# Copyright 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
 # Free Software Foundation, Inc.
 #
 # This file is part of the GNU Binutils.
@@ -64,6 +64,12 @@ static int no_toc_sort = 0;
 /* Set if PLT call stubs should load r11.  */
 static int plt_static_chain = ${DEFAULT_PLT_STATIC_CHAIN-0};
 
+/* Set if PLT call stubs need to be thread safe on power7+.  */
+static int plt_thread_safe = -1;
+
+/* Set if individual PLT call stubs should be aligned.  */
+static int plt_stub_align = 0;
+
 /* Whether to emit symbols for stubs.  */
 static int emit_stub_syms = -1;
 
@@ -379,7 +385,8 @@ ppc_add_stub_section (const char *stub_sec_name, asection *input_section)
   stub_sec = bfd_make_section_anyway_with_flags (stub_file->the_bfd,
 						 stub_sec_name, flags);
   if (stub_sec == NULL
-      || !bfd_set_section_alignment (stub_file->the_bfd, stub_sec, 5))
+      || !bfd_set_section_alignment (stub_file->the_bfd, stub_sec,
+				     plt_stub_align > 5 ? plt_stub_align : 5))
     goto err_ret;
 
   output_section = input_section->output_section;
@@ -504,7 +511,9 @@ gld${EMULATION_NAME}_after_allocation (void)
 	    einfo ("%P: .init/.fini fragments use differing TOC pointers\n");
 
 	  /* Call into the BFD backend to do the real work.  */
-	  if (!ppc64_elf_size_stubs (&link_info, group_size, plt_static_chain))
+	  if (!ppc64_elf_size_stubs (&link_info, group_size,
+				     plt_static_chain, plt_thread_safe,
+				     plt_stub_align))
 	    einfo ("%X%P: can not size stub section: %E\n");
 	}
     }
@@ -649,7 +658,11 @@ PARSE_AND_LIST_PROLOGUE=${PARSE_AND_LIST_PROLOGUE}'
 #define OPTION_STUBGROUP_SIZE		321
 #define OPTION_PLT_STATIC_CHAIN		(OPTION_STUBGROUP_SIZE + 1)
 #define OPTION_NO_PLT_STATIC_CHAIN	(OPTION_PLT_STATIC_CHAIN + 1)
-#define OPTION_STUBSYMS			(OPTION_NO_PLT_STATIC_CHAIN + 1)
+#define OPTION_PLT_THREAD_SAFE		(OPTION_NO_PLT_STATIC_CHAIN + 1)
+#define OPTION_NO_PLT_THREAD_SAFE	(OPTION_PLT_THREAD_SAFE + 1)
+#define OPTION_PLT_ALIGN		(OPTION_NO_PLT_THREAD_SAFE + 1)
+#define OPTION_NO_PLT_ALIGN		(OPTION_PLT_ALIGN + 1)
+#define OPTION_STUBSYMS			(OPTION_NO_PLT_ALIGN + 1)
 #define OPTION_NO_STUBSYMS		(OPTION_STUBSYMS + 1)
 #define OPTION_DOTSYMS			(OPTION_NO_STUBSYMS + 1)
 #define OPTION_NO_DOTSYMS		(OPTION_DOTSYMS + 1)
@@ -666,6 +679,10 @@ PARSE_AND_LIST_LONGOPTS=${PARSE_AND_LIST_LONGOPTS}'
   { "stub-group-size", required_argument, NULL, OPTION_STUBGROUP_SIZE },
   { "plt-static-chain", no_argument, NULL, OPTION_PLT_STATIC_CHAIN },
   { "no-plt-static-chain", no_argument, NULL, OPTION_NO_PLT_STATIC_CHAIN },
+  { "plt-thread-safe", no_argument, NULL, OPTION_PLT_THREAD_SAFE },
+  { "no-plt-thread-safe", no_argument, NULL, OPTION_NO_PLT_THREAD_SAFE },
+  { "plt-align", optional_argument, NULL, OPTION_PLT_ALIGN },
+  { "no-plt-align", no_argument, NULL, OPTION_NO_PLT_ALIGN },
   { "emit-stub-syms", no_argument, NULL, OPTION_STUBSYMS },
   { "no-emit-stub-syms", no_argument, NULL, OPTION_NO_STUBSYMS },
   { "dotsyms", no_argument, NULL, OPTION_DOTSYMS },
@@ -691,10 +708,22 @@ PARSE_AND_LIST_OPTIONS=${PARSE_AND_LIST_OPTIONS}'
                                 choose suitable defaults.\n"
 		   ));
   fprintf (file, _("\
-  --plt-static-chain          PLT call stubs should load r11.\n"
+  --plt-static-chain          PLT call stubs should load r11.${DEFAULT_PLT_STATIC_CHAIN- (default)}\n"
+		   ));
+  fprintf (file, _("\
+  --no-plt-static-chain       PLT call stubs should not load r11.${DEFAULT_PLT_STATIC_CHAIN+ (default)}\n"
+		   ));
+  fprintf (file, _("\
+  --plt-thread-safe           PLT call stubs with load-load barrier.\n"
+		   ));
+  fprintf (file, _("\
+  --no-plt-thread-safe        PLT call stubs without barrier.\n"
+		   ));
+  fprintf (file, _("\
+  --plt-align [=<align>]      Align PLT call stubs to fit cache lines.\n"
 		   ));
   fprintf (file, _("\
-  --no-plt-static-chain       PLT call stubs should not load r11. (default)\n"
+  --no-plt-align              Dont'\''t align individual PLT call stubs.\n"
 		   ));
   fprintf (file, _("\
   --emit-stub-syms            Label linker stubs with a symbol.\n"
@@ -753,6 +782,31 @@ PARSE_AND_LIST_ARGS_CASES=${PARSE_AND_LIST_ARGS_CASES}'
       plt_static_chain = 0;
       break;
 
+    case OPTION_PLT_THREAD_SAFE:
+      plt_thread_safe = 1;
+      break;
+
+    case OPTION_NO_PLT_THREAD_SAFE:
+      plt_thread_safe = 0;
+      break;
+
+    case OPTION_PLT_ALIGN:
+      if (optarg != NULL)
+	{
+	  char *end;
+	  unsigned long val = strtoul (optarg, &end, 0);
+	  if (*end || val > 8)
+	    einfo (_("%P%F: invalid --plt-align `%s'\''\n"), optarg);
+	  plt_stub_align = val;
+	}
+      else
+	plt_stub_align = 5;
+      break;
+
+    case OPTION_NO_PLT_ALIGN:
+      plt_stub_align = 0;
+      break;
+
     case OPTION_STUBSYMS:
       emit_stub_syms = 1;
       break;
diff --git a/ld/ld.texinfo b/ld/ld.texinfo
index 592e38c..6c74ff5 100644
--- a/ld/ld.texinfo
+++ b/ld/ld.texinfo
@@ -1,7 +1,7 @@
 \input texinfo
 @setfilename ld.info
 @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
-@c 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
+@c 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
 @c Free Software Foundation, Inc.
 @syncodeindex ky cp
 @c man begin INCLUDE
@@ -6710,7 +6710,9 @@ optimization.
 @cindex PowerPC64 multi-TOC
 @kindex --no-multi-toc
 @item --no-multi-toc
-By default, PowerPC64 GCC generates code for a TOC model where TOC
+If given any toc option besides @code{-mcmodel=medium} or
+@code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
+where TOC
 entries are accessed with a 16-bit offset from r2.  This limits the
 total TOC size to 64K.  PowerPC64 @command{ld} extends this limit by
 grouping code sections such that each group uses less than 64K for its
@@ -6719,6 +6721,52 @@ calls.  @command{ld} does not split apart input sections, so cannot
 help if a single input file has a @code{.toc} section that exceeds
 64K, most likely from linking multiple files with @command{ld -r}.
 Use this option to turn off this feature.
+
+@cindex PowerPC64 TOC sorting
+@kindex --no-toc-sort
+@item --no-toc-sort
+By default, @command{ld} sorts TOC sections so that those whose file
+happens to have a section called @code{.init} or @code{.fini} are
+placed first, followed by TOC sections referenced by code generated
+with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
+referenced only by code generated with PowerPC64 gcc's
+@code{-mcmodel=medium} or @code{-mcmodel=large} options.  Doing this
+results in better TOC grouping for multi-TOC.  Use this option to turn
+off this feature.
+
+@cindex PowerPC64 PLT stub alignment
+@kindex --plt-align
+@kindex --no-plt-align
+@item --plt-align
+@itemx --no-plt-align
+Use these options to control whether individual PLT call stubs are
+aligned to a 32-byte boundary, or to the specified power of two
+boundary when using @code{--plt-align=}.  By default PLT call stubs
+are packed tightly.
+
+@cindex PowerPC64 PLT call stub static chain
+@kindex --plt-static-chain
+@kindex --no-plt-static-chain
+@item --plt-static-chain
+@itemx --no-plt-static-chain
+Use these options to control whether PLT call stubs load the static
+chain pointer (r11).  @code{ld} defaults to not loading the static
+chain since there is never any need to do so on a PLT call.
+
+@cindex PowerPC64 PLT call stub thread safety
+@kindex --plt-thread-safe
+@kindex --no-plt-thread-safe
+@item --plt-thread-safe
+@itemx --no-thread-safe
+With power7's weakly ordered memory model, it is possible when using
+lazy binding for ld.so to update a plt entry in one thread and have
+another thread see the individual plt entry words update in the wrong
+order, despite ld.so carefully writing in the correct order and using
+memory write barriers.  To avoid this we need some sort of read
+barrier in the call stub, or use LD_BIND_NOW=1.  By default, @code{ld}
+looks for calls to commonly used functions that create threads, and if
+seen, adds the necessary barriers.  Use these options to change the
+default behaviour.
 @end table
 
 @ifclear GENERIC