Relocate alpha from ports to libc

Also fixed the following whitespace nits to satisfy the push:

sysdeps/alpha/alphaev6/memset.S:142: space before tab in indent.
sysdeps/alpha/configure:1: new blank line at EOF.
sysdeps/alpha/fpu/e_sqrt.c:126: space before tab in indent.
sysdeps/alpha/preconfigure:1: new blank line at EOF.
sysdeps/unix/sysv/linux/alpha/syscalls.list:1: new blank line at EOF.

1 files changed, 110 insertions, 0 deletions

diff --git a/sysdeps/alpha/bzero.S b/sysdeps/alpha/bzero.S
new file mode 100644
index 0000000000..36e891ca85
--- /dev/null
+++ b/sysdeps/alpha/bzero.S
@@ -0,0 +1,110 @@
+/* Copyright (C) 1996-2014 Free Software Foundation, Inc.
+   Contributed by Richard Henderson (rth@tamu.edu)
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library.  If not, see
+   <http://www.gnu.org/licenses/>.  */
+
+/* Fill a block of memory with zeros.  Optimized for the Alpha architecture:
+
+   - memory accessed as aligned quadwords only
+   - destination memory not read unless needed for good cache behaviour
+   - basic blocks arranged to optimize branch prediction for full-quadword
+     aligned memory blocks.
+   - partial head and tail quadwords constructed with byte-mask instructions
+
+   This is generally scheduled for the EV5 (got to look out for my own
+   interests :-), but with EV4 needs in mind.  There *should* be no more
+   stalls for the EV4 than there are for the EV5.
+*/
+
+
+#include <sysdep.h>
+
+	.set noat
+	.set noreorder
+
+	.text
+	.type	__bzero, @function
+	.globl	__bzero
+	.usepv	__bzero, USEPV_PROF
+
+	cfi_startproc
+
+	/* On entry to this basic block:
+	   t3 == loop counter
+	   t4 == bytes in partial final word
+	   a0 == possibly misaligned destination pointer  */
+
+	.align 3
+bzero_loop:
+	beq	t3, $tail	#
+	blbc	t3, 0f		# skip single store if count even
+
+	stq_u	zero, 0(a0)	# e0    : store one word
+	subq	t3, 1, t3	# .. e1 :
+	addq	a0, 8, a0	# e0    :
+	beq	t3, $tail	# .. e1 :
+
+0:	stq_u	zero, 0(a0)	# e0    : store two words
+	subq	t3, 2, t3	# .. e1 :
+	stq_u	zero, 8(a0)	# e0    :
+	addq	a0, 16, a0	# .. e1 :
+	bne	t3, 0b		# e1    :
+
+$tail:	bne	t4, 1f		# is there a tail to do?
+	ret			# no
+
+1:	ldq_u	t0, 0(a0)	# yes, load original data
+	mskqh	t0, t4, t0	#
+	stq_u	t0, 0(a0)	#
+	ret			#
+
+__bzero:
+#ifdef PROF
+	ldgp	gp, 0(pv)
+	lda	AT, _mcount
+	jsr	AT, (AT), _mcount
+#endif
+
+	mov	a0, v0		# e0    : move return value in place
+	beq	a1, $done	# .. e1 : early exit for zero-length store
+	and	a0, 7, t1	# e0    :
+	addq	a1, t1, a1	# e1    : add dest misalignment to count
+	srl	a1, 3, t3	# e0    : loop = count >> 3
+	and	a1, 7, t4	# .. e1 : find number of bytes in tail
+	unop			#       :
+	beq	t1, bzero_loop	# e1    : aligned head, jump right in
+
+	ldq_u	t0, 0(a0)	# e0    : load original data to mask into
+	cmpult	a1, 8, t2	# .. e1 : is this a sub-word set?
+	bne	t2, $oneq	# e1    :
+
+	mskql	t0, a0, t0	# e0    : we span words.  finish this partial
+	subq	t3, 1, t3	# e0    :
+	addq	a0, 8, a0	# .. e1 :
+	stq_u	t0, -8(a0)	# e0    :
+	br 	bzero_loop	# .. e1 :
+
+	.align 3
+$oneq:
+	mskql	t0, a0, t2	# e0    :
+	mskqh	t0, a1, t3	# e0    :
+	or	t2, t3, t0	# e1    :
+	stq_u	t0, 0(a0)	# e0    :
+
+$done:	ret
+
+	cfi_endproc
+weak_alias (__bzero, bzero)