powerpc: Only enable TLE with PPC_FEATURE2_HTM_NOSC

Linux from 3.9 through 4.2 does not abort HTM transaction on syscalls,
instead it suspend and resume it when leaving the kernel.  The
side-effects of the syscall will always remain visible, even if the
transaction is aborted.  This is an issue when transaction is used along
with futex syscall, on pthread_cond_wait for instance, where the futex
call might succeed but the transaction is rolled back leading the
pthread_cond object in an inconsistent state.

Glibc used to prevent it by always aborting a transaction before issuing
a syscall.  Linux 4.2 also decided to abort active transaction in
syscalls which makes the glibc workaround superfluous.  Worse, glibc
transaction abortion leads to a performance issue on recent kernels
where the HTM state is saved/restore lazily (v4.9).  By aborting a
transaction on every syscalls, regardless whether a transaction has being
initiated before, GLIBS makes the kernel always save/restore HTM state
(it can not even lazily disable it after a certain number of syscall
iterations).

Because of this shortcoming, Transactional Lock Elision is just enabled
when it has been explicitly set (either by tunables of by a configure
switch) and if kernel aborts HTM transactions on syscalls
(PPC_FEATURE2_HTM_NOSC).  It is reported that using simple benchmark [1],
the context-switch is about 5% faster by not issuing a tabort in every
syscall in newer kernels.

Checked on powerpc64le-linux-gnu with 4.4.0 kernel (Ubuntu 16.04).

	* NEWS: Add note about new TLE support on powerpc64le.
	* sysdeps/powerpc/nptl/tcb-offsets.sym (TM_CAPABLE): Remove.
	* sysdeps/powerpc/nptl/tls.h (tcbhead_t): Rename tm_capable to
	__ununsed1.
	(TLS_INIT_TP, TLS_DEFINE_INIT_TP): Remove tm_capable setup.
	(THREAD_GET_TM_CAPABLE, THREAD_SET_TM_CAPABLE): Remove macros.
	* sysdeps/powerpc/powerpc32/sysdep.h,
	sysdeps/powerpc/powerpc64/sysdep.h (ABORT_TRANSACTION_IMPL,
	ABORT_TRANSACTION): Remove macros.
	* sysdeps/powerpc/sysdep.h (ABORT_TRANSACTION): Likewise.
	* sysdeps/unix/sysv/linux/powerpc/elision-conf.c (elision_init): Set
	__pthread_force_elision iff PPC_FEATURE2_HTM_NOSC is set.
	* sysdeps/unix/sysv/linux/powerpc/powerpc32/sysdep.h,
	sysdeps/unix/sysv/linux/powerpc/powerpc64/sysdep.h
	sysdeps/unix/sysv/linux/powerpc/syscall.S (ABORT_TRANSACTION): Remove
	usage.
	* sysdeps/unix/sysv/linux/powerpc/not-errno.h: Remove file.

Reported-by: Breno Leitão <leitao@debian.org>

1 files changed, 20 insertions, 0 deletions

diff --git a/sysdeps/unix/sysv/linux/powerpc/elision-conf.c b/sysdeps/unix/sysv/linux/powerpc/elision-conf.c
index 906882a65e..fc82bd1ad8 100644
--- a/sysdeps/unix/sysv/linux/powerpc/elision-conf.c
+++ b/sysdeps/unix/sysv/linux/powerpc/elision-conf.c
@@ -127,6 +127,26 @@ elision_init (int argc __attribute__ ((unused)),
 	       TUNABLE_CALLBACK (set_elision_skip_trylock_internal_abort));
 #endif
 
+  /* Linux from 3.9 through 4.2 do not abort HTM transaction on syscalls,
+     instead it suspends the transaction and resumes it when returning to
+     usercode.  The side-effects of the syscall will always remain visible,
+     even if the transaction is aborted.  This is an issue when a transaction
+     is used along with futex syscall, on pthread_cond_wait for instance,
+     where futex might succeed but the transaction is rolled back leading
+     the condition variable object in an inconsistent state.
+
+     Glibc used to prevent it by always aborting a transaction before issuing
+     a syscall.  Linux 4.2 also decided to abort active transaction in
+     syscalls which makes the glibc workaround superflours.  Worse, glibc
+     transaction abortions leads to a performance issues on recent kernels.
+
+     So Lock Elision is just enabled when it has been explict set (either
+     by tunables of by a configure switch) and if kernel aborts HTM
+     transactions on syscalls (PPC_FEATURE2_HTM_NOSC)  */
+
+  __pthread_force_elision = (__pthread_force_elision
+			     && GLRO (dl_hwcap2) & PPC_FEATURE2_HTM_NOSC);
+
   if (!__pthread_force_elision)
     __elision_aconf.try_tbegin = 0; /* Disable elision on rwlocks.  */
 }