Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752836AbaKCK6v (ORCPT ); Mon, 3 Nov 2014 05:58:51 -0500 Received: from mail-bl2on0086.outbound.protection.outlook.com ([65.55.169.86]:63567 "EHLO na01-bl2-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752486AbaKCKyo (ORCPT ); Mon, 3 Nov 2014 05:54:44 -0500 From: Ley Foon Tan To: , , CC: Arnd Bergmann , Ley Foon Tan , , Subject: [PATCH v6 19/29] nios2: Library functions Date: Mon, 3 Nov 2014 18:52:15 +0800 Message-ID: <1415011945-3297-20-git-send-email-lftan@altera.com> X-Mailer: git-send-email 1.8.3.2 In-Reply-To: <1415011945-3297-1-git-send-email-lftan@altera.com> References: <1415011945-3297-1-git-send-email-lftan@altera.com> MIME-Version: 1.0 Content-Type: text/plain X-EOPAttributedMessage: 0 X-Forefront-Antispam-Report: CIP:66.35.236.227;CTRY:US;IPV:NLI;EFV:NLI;SFV:NSPM;SFS:(10009020)(6009001)(199003)(22564002)(189002)(99396003)(120916001)(19580395003)(88136002)(50466002)(84676001)(86362001)(2201001)(15975445006)(89996001)(92566001)(575784001)(92726001)(93916002)(48376002)(46102003)(44976005)(6806004)(19580405001)(15202345003)(21056001)(68736004)(87936001)(47776003)(87286001)(64706001)(20776003)(104166001)(50986999)(50226001)(4396001)(42186005)(102836001)(107046002)(36756003)(106466001)(31966008)(76176999)(229853001)(33646002)(97736003)(77156002)(77096003)(62966003)(95666004)(105596002);DIR:OUT;SFP:1101;SCL:1;SRVR:BY2PR03MB224;H:sj-itexedge03.altera.priv.altera.com;FPR:;MLV:sfv;PTR:InfoDomainNonexistent;A:1;MX:1;LANG:en; X-Microsoft-Antispam: UriScan:; X-Microsoft-Antispam: BCL:0;PCL:0;RULEID:;SRVR:BY2PR03MB224; X-Forefront-PRVS: 0384275935 Authentication-Results: spf=softfail (sender IP is 66.35.236.227) smtp.mailfrom=lftan@altera.com; X-OriginatorOrg: altera.com Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Add optimised library functions for nios2. Signed-off-by: Ley Foon Tan --- arch/nios2/include/asm/checksum.h | 78 +++++++++++++++ arch/nios2/include/asm/string.h | 24 +++++ arch/nios2/lib/memcpy.c | 202 ++++++++++++++++++++++++++++++++++++++ arch/nios2/lib/memmove.c | 82 ++++++++++++++++ arch/nios2/lib/memset.c | 81 +++++++++++++++ 5 files changed, 467 insertions(+) create mode 100644 arch/nios2/include/asm/checksum.h create mode 100644 arch/nios2/include/asm/string.h create mode 100644 arch/nios2/lib/memcpy.c create mode 100644 arch/nios2/lib/memmove.c create mode 100644 arch/nios2/lib/memset.c diff --git a/arch/nios2/include/asm/checksum.h b/arch/nios2/include/asm/checksum.h new file mode 100644 index 0000000..6bc1f0d --- /dev/null +++ b/arch/nios2/include/asm/checksum.h @@ -0,0 +1,78 @@ +/* + * Copyright (C) 2010 Tobias Klauser + * Copyright (C) 2004 Microtronix Datacom Ltd. + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ + +#ifndef _ASM_NIOS_CHECKSUM_H +#define _ASM_NIOS_CHECKSUM_H + +/* Take these from lib/checksum.c */ +extern __wsum csum_partial(const void *buff, int len, __wsum sum); +extern __wsum csum_partial_copy(const void *src, void *dst, int len, + __wsum sum); +extern __wsum csum_partial_copy_from_user(const void __user *src, void *dst, + int len, __wsum sum, int *csum_err); +#define csum_partial_copy_nocheck(src, dst, len, sum) \ + csum_partial_copy((src), (dst), (len), (sum)) + +extern __sum16 ip_fast_csum(const void *iph, unsigned int ihl); +extern __sum16 ip_compute_csum(const void *buff, int len); + +/* + * Fold a partial checksum + */ +static inline __sum16 csum_fold(__wsum sum) +{ + __asm__ __volatile__( + "add %0, %1, %0\n" + "cmpltu r8, %0, %1\n" + "srli %0, %0, 16\n" + "add %0, %0, r8\n" + "nor %0, %0, %0\n" + : "=r" (sum) + : "r" (sum << 16), "0" (sum) + : "r8"); + return (__force __sum16) sum; +} + +/* + * computes the checksum of the TCP/UDP pseudo-header + * returns a 16-bit checksum, already complemented + */ +#define csum_tcpudp_nofold csum_tcpudp_nofold +static inline __wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr, + unsigned short len, + unsigned short proto, + __wsum sum) +{ + __asm__ __volatile__( + "add %0, %1, %0\n" + "cmpltu r8, %0, %1\n" + "add %0, %0, r8\n" /* add carry */ + "add %0, %2, %0\n" + "cmpltu r8, %0, %2\n" + "add %0, %0, r8\n" /* add carry */ + "add %0, %3, %0\n" + "cmpltu r8, %0, %3\n" + "add %0, %0, r8\n" /* add carry */ + : "=r" (sum), "=r" (saddr) + : "r" (daddr), "r" ((ntohs(len) << 16) + (proto * 256)), + "0" (sum), + "1" (saddr) + : "r8"); + + return sum; +} + +static inline __sum16 csum_tcpudp_magic(__be32 saddr, __be32 daddr, + unsigned short len, + unsigned short proto, __wsum sum) +{ + return csum_fold(csum_tcpudp_nofold(saddr, daddr, len, proto, sum)); +} + +#endif /* _ASM_NIOS_CHECKSUM_H */ diff --git a/arch/nios2/include/asm/string.h b/arch/nios2/include/asm/string.h new file mode 100644 index 0000000..14dd570 --- /dev/null +++ b/arch/nios2/include/asm/string.h @@ -0,0 +1,24 @@ +/* + * Copyright (C) 2004 Microtronix Datacom Ltd + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ + +#ifndef _ASM_NIOS2_STRING_H +#define _ASM_NIOS2_STRING_H + +#ifdef __KERNEL__ + +#define __HAVE_ARCH_MEMSET +#define __HAVE_ARCH_MEMCPY +#define __HAVE_ARCH_MEMMOVE + +extern void *memset(void *s, int c, size_t count); +extern void *memcpy(void *d, const void *s, size_t count); +extern void *memmove(void *d, const void *s, size_t count); + +#endif /* __KERNEL__ */ + +#endif /* _ASM_NIOS2_STRING_H */ diff --git a/arch/nios2/lib/memcpy.c b/arch/nios2/lib/memcpy.c new file mode 100644 index 0000000..1715f5d --- /dev/null +++ b/arch/nios2/lib/memcpy.c @@ -0,0 +1,202 @@ +/* Extracted from GLIBC memcpy.c and memcopy.h, which is: + Copyright (C) 1991, 1992, 1993, 1997, 2004 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Torbjorn Granlund (tege@sics.se). + + 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 + . */ + +#include + +/* Type to use for aligned memory operations. + This should normally be the biggest type supported by a single load + and store. */ +#define op_t unsigned long int +#define OPSIZ (sizeof(op_t)) + +/* Optimal type for storing bytes in registers. */ +#define reg_char char + +#define MERGE(w0, sh_1, w1, sh_2) (((w0) >> (sh_1)) | ((w1) << (sh_2))) + +/* Copy exactly NBYTES bytes from SRC_BP to DST_BP, + without any assumptions about alignment of the pointers. */ +#define BYTE_COPY_FWD(dst_bp, src_bp, nbytes) \ +do { \ + size_t __nbytes = (nbytes); \ + while (__nbytes > 0) { \ + unsigned char __x = ((unsigned char *) src_bp)[0]; \ + src_bp += 1; \ + __nbytes -= 1; \ + ((unsigned char *) dst_bp)[0] = __x; \ + dst_bp += 1; \ + } \ +} while (0) + +/* Copy *up to* NBYTES bytes from SRC_BP to DST_BP, with + the assumption that DST_BP is aligned on an OPSIZ multiple. If + not all bytes could be easily copied, store remaining number of bytes + in NBYTES_LEFT, otherwise store 0. */ +/* extern void _wordcopy_fwd_aligned __P ((long int, long int, size_t)); */ +/* extern void _wordcopy_fwd_dest_aligned __P ((long int, long int, size_t)); */ +#define WORD_COPY_FWD(dst_bp, src_bp, nbytes_left, nbytes) \ +do { \ + if (src_bp % OPSIZ == 0) \ + _wordcopy_fwd_aligned(dst_bp, src_bp, (nbytes) / OPSIZ);\ + else \ + _wordcopy_fwd_dest_aligned(dst_bp, src_bp, (nbytes) / OPSIZ);\ + src_bp += (nbytes) & -OPSIZ; \ + dst_bp += (nbytes) & -OPSIZ; \ + (nbytes_left) = (nbytes) % OPSIZ; \ +} while (0) + + +/* Threshold value for when to enter the unrolled loops. */ +#define OP_T_THRES 16 + +/* _wordcopy_fwd_aligned -- Copy block beginning at SRCP to + block beginning at DSTP with LEN `op_t' words (not LEN bytes!). + Both SRCP and DSTP should be aligned for memory operations on `op_t's. */ +/* stream-lined (read x8 + write x8) */ +static void _wordcopy_fwd_aligned(long int dstp, long int srcp, size_t len) +{ + while (len > 7) { + register op_t a0, a1, a2, a3, a4, a5, a6, a7; + + a0 = ((op_t *) srcp)[0]; + a1 = ((op_t *) srcp)[1]; + a2 = ((op_t *) srcp)[2]; + a3 = ((op_t *) srcp)[3]; + a4 = ((op_t *) srcp)[4]; + a5 = ((op_t *) srcp)[5]; + a6 = ((op_t *) srcp)[6]; + a7 = ((op_t *) srcp)[7]; + ((op_t *) dstp)[0] = a0; + ((op_t *) dstp)[1] = a1; + ((op_t *) dstp)[2] = a2; + ((op_t *) dstp)[3] = a3; + ((op_t *) dstp)[4] = a4; + ((op_t *) dstp)[5] = a5; + ((op_t *) dstp)[6] = a6; + ((op_t *) dstp)[7] = a7; + + srcp += 8 * OPSIZ; + dstp += 8 * OPSIZ; + len -= 8; + } + while (len > 0) { + *(op_t *)dstp = *(op_t *)srcp; + + srcp += OPSIZ; + dstp += OPSIZ; + len -= 1; + } +} + +/* _wordcopy_fwd_dest_aligned -- Copy block beginning at SRCP to + block beginning at DSTP with LEN `op_t' words (not LEN bytes!). + DSTP should be aligned for memory operations on `op_t's, but SRCP must + *not* be aligned. */ +/* stream-lined (read x4 + write x4) */ +static void _wordcopy_fwd_dest_aligned(long int dstp, long int srcp, + size_t len) +{ + op_t ap; + int sh_1, sh_2; + + /* Calculate how to shift a word read at the memory operation + aligned srcp to make it aligned for copy. */ + + sh_1 = 8 * (srcp % OPSIZ); + sh_2 = 8 * OPSIZ - sh_1; + + /* Make SRCP aligned by rounding it down to the beginning of the `op_t' + it points in the middle of. */ + srcp &= -OPSIZ; + ap = ((op_t *) srcp)[0]; + srcp += OPSIZ; + + while (len > 3) { + op_t a0, a1, a2, a3; + + a0 = ((op_t *) srcp)[0]; + a1 = ((op_t *) srcp)[1]; + a2 = ((op_t *) srcp)[2]; + a3 = ((op_t *) srcp)[3]; + ((op_t *) dstp)[0] = MERGE(ap, sh_1, a0, sh_2); + ((op_t *) dstp)[1] = MERGE(a0, sh_1, a1, sh_2); + ((op_t *) dstp)[2] = MERGE(a1, sh_1, a2, sh_2); + ((op_t *) dstp)[3] = MERGE(a2, sh_1, a3, sh_2); + + ap = a3; + srcp += 4 * OPSIZ; + dstp += 4 * OPSIZ; + len -= 4; + } + while (len > 0) { + register op_t a0; + + a0 = ((op_t *) srcp)[0]; + ((op_t *) dstp)[0] = MERGE(ap, sh_1, a0, sh_2); + + ap = a0; + srcp += OPSIZ; + dstp += OPSIZ; + len -= 1; + } +} + +void *memcpy(void *dstpp, const void *srcpp, size_t len) +{ + unsigned long int dstp = (long int) dstpp; + unsigned long int srcp = (long int) srcpp; + + /* Copy from the beginning to the end. */ + + /* If there not too few bytes to copy, use word copy. */ + if (len >= OP_T_THRES) { + /* Copy just a few bytes to make DSTP aligned. */ + len -= (-dstp) % OPSIZ; + BYTE_COPY_FWD(dstp, srcp, (-dstp) % OPSIZ); + + /* Copy whole pages from SRCP to DSTP by virtual address + manipulation, as much as possible. */ + + /* PAGE_COPY_FWD_MAYBE (dstp, srcp, len, len); */ + + /* Copy from SRCP to DSTP taking advantage of the known + alignment of DSTP. Number of bytes remaining is put in the + third argument, i.e. in LEN. This number may vary from + machine to machine. */ + + WORD_COPY_FWD(dstp, srcp, len, len); + + /* Fall out and copy the tail. */ + } + + /* There are just a few bytes to copy. Use byte memory operations. */ + BYTE_COPY_FWD(dstp, srcp, len); + + return dstpp; +} + +void *memcpyb(void *dstpp, const void *srcpp, unsigned len) +{ + unsigned long int dstp = (long int) dstpp; + unsigned long int srcp = (long int) srcpp; + + BYTE_COPY_FWD(dstp, srcp, len); + + return dstpp; +} diff --git a/arch/nios2/lib/memmove.c b/arch/nios2/lib/memmove.c new file mode 100644 index 0000000..c65ef51 --- /dev/null +++ b/arch/nios2/lib/memmove.c @@ -0,0 +1,82 @@ +/* + * Copyright (C) 2011 Tobias Klauser + * Copyright (C) 2004 Microtronix Datacom Ltd + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ + +#include +#include + +#ifdef __HAVE_ARCH_MEMMOVE +void *memmove(void *d, const void *s, size_t count) +{ + unsigned long dst, src; + + if (!count) + return d; + + if (d < s) { + dst = (unsigned long) d; + src = (unsigned long) s; + + if ((count < 8) || ((dst ^ src) & 3)) + goto restup; + + if (dst & 1) { + *(char *)dst++ = *(char *)src++; + count--; + } + if (dst & 2) { + *(short *)dst = *(short *)src; + src += 2; + dst += 2; + count -= 2; + } + while (count > 3) { + *(long *)dst = *(long *)src; + src += 4; + dst += 4; + count -= 4; + } +restup: + while (count--) + *(char *)dst++ = *(char *)src++; + } else { + dst = (unsigned long) d + count; + src = (unsigned long) s + count; + + if ((count < 8) || ((dst ^ src) & 3)) + goto restdown; + + if (dst & 1) { + src--; + dst--; + count--; + *(char *)dst = *(char *)src; + } + if (dst & 2) { + src -= 2; + dst -= 2; + count -= 2; + *(short *)dst = *(short *)src; + } + while (count > 3) { + src -= 4; + dst -= 4; + count -= 4; + *(long *)dst = *(long *)src; + } +restdown: + while (count--) { + src--; + dst--; + *(char *)dst = *(char *)src; + } + } + + return d; +} +#endif /* __HAVE_ARCH_MEMMOVE */ diff --git a/arch/nios2/lib/memset.c b/arch/nios2/lib/memset.c new file mode 100644 index 0000000..65e9780 --- /dev/null +++ b/arch/nios2/lib/memset.c @@ -0,0 +1,81 @@ +/* + * Copyright (C) 2011 Tobias Klauser + * Copyright (C) 2004 Microtronix Datacom Ltd + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ + +#include +#include + +#ifdef __HAVE_ARCH_MEMSET +void *memset(void *s, int c, size_t count) +{ + int destptr, charcnt, dwordcnt, fill8reg, wrkrega; + + if (!count) + return s; + + c &= 0xFF; + + if (count <= 8) { + char *xs = (char *) s; + + while (count--) + *xs++ = c; + return s; + } + + __asm__ __volatile__ ( + /* fill8 %3, %5 (c & 0xff) */ + " slli %4, %5, 8\n" + " or %4, %4, %5\n" + " slli %3, %4, 16\n" + " or %3, %3, %4\n" + /* Word-align %0 (s) if necessary */ + " andi %4, %0, 0x01\n" + " beq %4, zero, 1f\n" + " addi %1, %1, -1\n" + " stb %3, 0(%0)\n" + " addi %0, %0, 1\n" + "1: mov %2, %1\n" + /* Dword-align %0 (s) if necessary */ + " andi %4, %0, 0x02\n" + " beq %4, zero, 2f\n" + " addi %1, %1, -2\n" + " sth %3, 0(%0)\n" + " addi %0, %0, 2\n" + " mov %2, %1\n" + /* %1 and %2 are how many more bytes to set */ + "2: srli %2, %2, 2\n" + /* %2 is how many dwords to set */ + "3: stw %3, 0(%0)\n" + " addi %0, %0, 4\n" + " addi %2, %2, -1\n" + " bne %2, zero, 3b\n" + /* store residual word and/or byte if necessary */ + " andi %4, %1, 0x02\n" + " beq %4, zero, 4f\n" + " sth %3, 0(%0)\n" + " addi %0, %0, 2\n" + /* store residual byte if necessary */ + "4: andi %4, %1, 0x01\n" + " beq %4, zero, 5f\n" + " stb %3, 0(%0)\n" + "5:\n" + : "=r" (destptr), /* %0 Output */ + "=r" (charcnt), /* %1 Output */ + "=r" (dwordcnt), /* %2 Output */ + "=r" (fill8reg), /* %3 Output */ + "=r" (wrkrega) /* %4 Output */ + : "r" (c), /* %5 Input */ + "0" (s), /* %0 Input/Output */ + "1" (count) /* %1 Input/Output */ + : "memory" /* clobbered */ + ); + + return s; +} +#endif /* __HAVE_ARCH_MEMSET */ -- 1.8.2.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/