powerpc: Fix float128 IFUNC relocations [BZ #21707]

The patch proposed by Peter Bergner [1] to libgcc in order to fix
[BZ #21707] adds a dependency on a symbol provided by the loader,
forcing the loader to be linked to tests after libgcc was linked.

It also requires to read the thread pointer during IRELA relocations.

Tested on powerpc, powerpc64, powerpc64le, s390x and x86_64.

[1] https://sourceware.org/ml/libc-alpha/2017-06/msg01383.html

	[BZ #21707]
	* csu/libc-start.c (LIBC_START_MAIN): Perform IREL{,A}
	relocations before or after initializing the TCB on statically
	linked executables.  That's a per-architecture definition.
	* elf/rtld.c (dl_main): Add a comment about thread-local
	variables initialization.
	* sysdeps/generic/libc-start.h: New file.  Define
	ARCH_APPLY_IREL and ARCH_SETUP_IREL.
	* sysdeps/powerpc/Makefile:
	[$(subdir) = elf && $(multi-arch) != no] (tests-static-internal): Add tst-tlsifunc-static.
	[$(subdir) = elf && $(multi-arch) != no && $(build-shared) == yes]
	(tests-internal): Add tst-tlsifunc.
	* sysdeps/powerpc/tst-tlsifunc.c: New file.
	* sysdeps/powerpc/tst-tlsifunc-static.c: Likewise.
	* sysdeps/powerpc/powerpc64le/Makefile (f128-loader-link): New
	variable.
	[$(subdir) = math] (test-float128% test-ifloat128%): Force
	linking to the loader after linking to libgcc.
	[$(subdir) = wcsmbs || $(subdir) = stdlib] (bug-strtod bug-strtod2)
	(bug-strtod2 tst-strtod-round tst-wcstod-round tst-strtod6 tst-strrom)
	(tst-strfrom-locale strfrom-skeleton): Likewise.
	* sysdeps/unix/sysv/linux/powerpc/libc-start.h: New file.  Define
	ARCH_APPLY_IREL and ARCH_SETUP_IREL.

1 files changed, 3 insertions, 1 deletions

diff --git a/elf/rtld.c b/elf/rtld.c
index 65647fb..1772f89 100644
--- a/elf/rtld.c
+++ b/elf/rtld.c
@@ -2208,7 +2208,9 @@ ERROR: ld.so: object '%s' cannot be loaded as audit interface: %s; ignored.\n",
 
   /* Now that we have completed relocation, the initializer data
      for the TLS blocks has its final values and we can copy them
-     into the main thread's TLS area, which we allocated above.  */
+     into the main thread's TLS area, which we allocated above.
+     Note: thread-local variables must only be accessed after completing
+     the next step.  */
   _dl_allocate_tls_init (tcbp);
 
   /* And finally install it for the main thread.  */