 b6cbf7203b
			
		
	
	
		b6cbf7203b
		
	
	
	
	
		
			
			This patch imports the unmodified current version of NetBSD libc. The NetBSD includes are in /nbsd_include, while the libc code itself is split between lib/nbsd_libc and common/lib/libc.
		
			
				
	
	
		
			197 lines
		
	
	
		
			5.4 KiB
		
	
	
	
		
			ArmAsm
		
	
	
	
	
	
			
		
		
	
	
			197 lines
		
	
	
		
			5.4 KiB
		
	
	
	
		
			ArmAsm
		
	
	
	
	
	
| /*	$NetBSD: memset.S,v 1.4 2001/08/02 01:17:28 eeh Exp $	*/
 | |
| 
 | |
| /*
 | |
|  * Copyright (c) 2001, Eduardo E. Horvath
 | |
|  *
 | |
|  * This software was developed by the Computer Systems Engineering group
 | |
|  * at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
 | |
|  * contributed to Berkeley.
 | |
|  *
 | |
|  * Redistribution and use in source and binary forms, with or without
 | |
|  * modification, are permitted provided that the following conditions
 | |
|  * are met:
 | |
|  * 1. Redistributions of source code must retain the above copyright
 | |
|  *    notice, this list of conditions and the following disclaimer.
 | |
|  * 2. Redistributions in binary form must reproduce the above copyright
 | |
|  *    notice, this list of conditions and the following disclaimer in the
 | |
|  *    documentation and/or other materials provided with the distribution.
 | |
|  * 3. All advertising materials mentioning features or use of this software
 | |
|  *    must display the following acknowledgement:
 | |
|  *	This product includes software developed by the University of
 | |
|  *	California, Berkeley and its contributors.
 | |
|  * 4. Neither the name of the University nor the names of its contributors
 | |
|  *    may be used to endorse or promote products derived from this software
 | |
|  *    without specific prior written permission.
 | |
|  *
 | |
|  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 | |
|  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 | |
|  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 | |
|  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 | |
|  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 | |
|  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 | |
|  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 | |
|  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 | |
|  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 | |
|  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 | |
|  * SUCH DAMAGE.
 | |
|  *
 | |
|  * from: Header: bzero.s,v 1.1 92/06/25 12:52:46 torek Exp
 | |
|  */
 | |
| 
 | |
| #include <machine/asm.h>
 | |
| #ifndef _LOCORE
 | |
| #define _LOCORE
 | |
| #endif
 | |
| #include <machine/ctlreg.h>
 | |
| #include <machine/frame.h>
 | |
| #include <machine/psl.h>
 | |
| 
 | |
| #if defined(LIBC_SCCS) && !defined(lint)
 | |
| 	RCSID("$NetBSD: memset.S,v 1.4 2001/08/02 01:17:28 eeh Exp $")
 | |
| #endif  /* LIBC_SCCS and not lint */
 | |
| 
 | |
| /*
 | |
|  * bzero(addr, len)
 | |
|  *
 | |
|  * We want to use VIS instructions if we're clearing out more than
 | |
|  * 256 bytes, but to do that we need to properly save and restore the
 | |
|  * FP registers.  Unfortunately the code to do that in the kernel needs
 | |
|  * to keep track of the current owner of the FPU, hence the different
 | |
|  * code.
 | |
|  *
 | |
|  * XXXXX To produce more efficient code, we do not allow lengths
 | |
|  * greater than 0x80000000000000000, which are negative numbers.
 | |
|  * This should not really be an issue since the VA hole should
 | |
|  * cause any such ranges to fail anyway.
 | |
|  */
 | |
| ENTRY(bzero)
 | |
| 	! %o0 = addr, %o1 = len
 | |
| 	mov	%o1, %o2
 | |
| 	clr	%o1			! Initialize our pattern
 | |
| /*
 | |
|  * memset(addr, c, len)
 | |
|  *
 | |
|  */
 | |
| ENTRY(memset)
 | |
| 	! %o0 = addr, %o1 = pattern, %o2 = len
 | |
| 	mov	%o0, %o4		! Save original pointer
 | |
| 
 | |
| Lbzero_internal:
 | |
| 	btst	7, %o0			! Word aligned?
 | |
| 	bz,pn	%xcc, 0f
 | |
| 	 nop
 | |
| 	inc	%o0
 | |
| 	deccc	%o2			! Store up to 7 bytes
 | |
| 	bge,a,pt	%xcc, Lbzero_internal
 | |
| 	 stb	%o1, [%o0 - 1]
 | |
| 
 | |
| 	retl				! Duplicate Lbzero_done
 | |
| 	 mov	%o4, %o0
 | |
| 0:
 | |
| 	/*
 | |
| 	 * Duplicate the pattern so it fills 64-bits.
 | |
| 	 */
 | |
| 	andcc	%o1, 0x0ff, %o1		! No need to extend zero
 | |
| 	bz,pt	%icc, 1f
 | |
| 	 sllx	%o1, 8, %o3		! sigh.  all dependent insns.
 | |
| 	or	%o1, %o3, %o1
 | |
| 	sllx	%o1, 16, %o3
 | |
| 	or	%o1, %o3, %o1
 | |
| 	sllx	%o1, 32, %o3
 | |
| 	 or	%o1, %o3, %o1
 | |
| 1:	
 | |
| #if 1
 | |
| 	!! Now we are 64-bit aligned
 | |
| 	cmp	%o2, 256		! Use block clear if len > 256
 | |
| 	bge,pt	%xcc, Lbzero_block	! use block store insns
 | |
| #endif	
 | |
| 	 deccc	8, %o2
 | |
| Lbzero_longs:
 | |
| 	bl,pn	%xcc, Lbzero_cleanup	! Less than 8 bytes left
 | |
| 	 nop
 | |
| 3:	
 | |
| 	inc	8, %o0
 | |
| 	deccc	8, %o2
 | |
| 	bge,pt	%xcc, 3b
 | |
| 	 stx	%o1, [%o0 - 8]		! Do 1 longword at a time
 | |
| 
 | |
| 	/*
 | |
| 	 * Len is in [-8..-1] where -8 => done, -7 => 1 byte to zero,
 | |
| 	 * -6 => two bytes, etc.  Mop up this remainder, if any.
 | |
| 	 */
 | |
| Lbzero_cleanup:	
 | |
| 	btst	4, %o2
 | |
| 	bz,pt	%xcc, 5f		! if (len & 4) {
 | |
| 	 nop
 | |
| 	stw	%o1, [%o0]		!	*(int *)addr = 0;
 | |
| 	inc	4, %o0			!	addr += 4;
 | |
| 5:	
 | |
| 	btst	2, %o2
 | |
| 	bz,pt	%xcc, 7f		! if (len & 2) {
 | |
| 	 nop
 | |
| 	sth	%o1, [%o0]		!	*(short *)addr = 0;
 | |
| 	inc	2, %o0			!	addr += 2;
 | |
| 7:	
 | |
| 	btst	1, %o2
 | |
| 	bnz,a	%icc, Lbzero_done	! if (len & 1)
 | |
| 	 stb	%o1, [%o0]		!	*addr = 0;
 | |
| Lbzero_done:
 | |
| 	retl
 | |
| 	 mov	%o4, %o0		! Restore ponter for memset (ugh)
 | |
| 
 | |
| #if 1	
 | |
| Lbzero_block:
 | |
| /*
 | |
|  * Userland:
 | |
|  *
 | |
|  * Floating point registers are volatile.  What luck.
 | |
|  *
 | |
|  * See locore.s for the kernel version.
 | |
|  *
 | |
|  */	
 | |
| !	wr	%g0, FPRS_FEF, %fprs			! Enable FPU
 | |
| 
 | |
| 	!! We are now 8-byte aligned.  We need to become 64-byte aligned.
 | |
| 	btst	63, %o0
 | |
| 	bz,pt	%xcc, 2f
 | |
| 	 nop
 | |
| 1:	
 | |
| 	stx	%o1, [%o0]
 | |
| 	inc	8, %o0
 | |
| 	btst	63, %o0
 | |
| 	bnz,pt	%xcc, 1b
 | |
| 	 dec	8, %o2
 | |
| 
 | |
| 2:
 | |
| 	brz	%o1, 3f					! Skip the memory op
 | |
| 	 fzero	%f0					! for bzero
 | |
| 	
 | |
| 	stx	%o1, [%o0]				! Flush this puppy to RAM
 | |
| 	membar	#StoreLoad
 | |
| 	ldd	[%o0], %f0
 | |
| 3:	
 | |
| 	fmovd	%f0, %f2				! Duplicate the pattern
 | |
| 	fmovd	%f0, %f4
 | |
| 	fmovd	%f0, %f6
 | |
| 	fmovd	%f0, %f8
 | |
| 	fmovd	%f0, %f10
 | |
| 	fmovd	%f0, %f12
 | |
| 	fmovd	%f0, %f14
 | |
| 	
 | |
| 	!! Remember: we were 8 bytes too far
 | |
| 	dec	56, %o2					! Go one iteration too far
 | |
| 5:
 | |
| 	stda	%f0, [%o0] ASI_BLK_P			! Store 64 bytes
 | |
| 	deccc	64, %o2
 | |
| 	bg,pn	%xcc, 5b
 | |
| 	 inc	64, %o0
 | |
| 	
 | |
| 	membar	#Sync
 | |
| /*
 | |
|  * Now we're done we need to load the FPU state from where
 | |
|  * we put it.
 | |
|  */
 | |
| 	ba,pt	%xcc, Lbzero_longs	! Finish up the remainder
 | |
| 	 inccc	56, %o2		! Restore the count
 | |
| #endif
 |