From patchwork Wed Mar 23 21:57:46 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Noah Goldstein X-Patchwork-Id: 52290 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 4D8AF386EC0E for ; Wed, 23 Mar 2022 22:14:45 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 4D8AF386EC0E DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1648073685; bh=LDlnysSYbLACEjnax9zE6qJVSWOUq7mhcZ41uKPFyRs=; h=To:Subject:Date:In-Reply-To:References:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To: From; b=ih4YnHYn1uZQo7llquo4EnmDtbb4qpct94WzuZUXtlCMxOZJ++v0oXNYD4mryD2H/ i5MQyUu11fdxDs7iP95/pXNABYiNCnLnFvvPr7e30lH4K9Zj5vppCwxkD5a2f2EKIq x5Iw6rwjuKwfmHcSjRij+fROK8G3a5YL6jdiXhLo= X-Original-To: libc-alpha@sourceware.org Delivered-To: libc-alpha@sourceware.org Received: from mail-io1-xd34.google.com (mail-io1-xd34.google.com [IPv6:2607:f8b0:4864:20::d34]) by sourceware.org (Postfix) with ESMTPS id 65523388883C for ; Wed, 23 Mar 2022 22:03:06 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 65523388883C Received: by mail-io1-xd34.google.com with SMTP id k25so3396953iok.8 for ; Wed, 23 Mar 2022 15:03:06 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=LDlnysSYbLACEjnax9zE6qJVSWOUq7mhcZ41uKPFyRs=; b=NlftNX2ojfMUhQC3pJVreajtiCgefMffOV8pu6NkF6V+YekzWZyN/F/qc2dP0W9BIh fESSeoBIMEcXV6y4tnEGclWXAJT4eI/Atqy/QCbUsyjNlG2c1MvFqA3gL7y0YNWayTDA kiQi63pIpvMENvuAZl4FGT/PcjKM0Pc3KTvUgaaicuk7VSu0bUzleLwwl9B1DugK007k yDexhVMKc6VSOkOTbna/2Hww+sjMgPAzLlvkFr35BwIpaWsW68/yJWhM3k+9j/PeFT9x yc/fmFTc9dOC9BGVQZv0VTdaO8c8+50o00PNp7i/Qi5lJMLZ+DU5FBHD1nEmZO2Li0Bn HRrA== X-Gm-Message-State: AOAM532jA2p99WVIIEOrwYnNmX8OYBGU1LCBTYUX7McXgBqm3oEcX+pv t4N0C6EXF0q6tZNP12jAJ8y9/p+PuwQ= X-Google-Smtp-Source: ABdhPJwvaS+gaLcp1m5067S9ZPT2MXfNsrwJFYTFuSTJvyRUs1Af7vgi9g5D/F+jBRx9u3cjlyb09w== X-Received: by 2002:a02:bb8c:0:b0:317:b015:1998 with SMTP id g12-20020a02bb8c000000b00317b0151998mr1094902jan.134.1648072985063; Wed, 23 Mar 2022 15:03:05 -0700 (PDT) Received: from localhost.localdomain (node-17-161.flex.volo.net. [76.191.17.161]) by smtp.googlemail.com with ESMTPSA id c22-20020a5ea816000000b00649d360663asm529227ioa.40.2022.03.23.15.03.03 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 23 Mar 2022 15:03:04 -0700 (PDT) To: libc-alpha@sourceware.org Subject: [PATCH v1 23/23] x86: Remove AVX str{n}casecmp Date: Wed, 23 Mar 2022 16:57:46 -0500 Message-Id: <20220323215734.3927131-23-goldstein.w.n@gmail.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20220323215734.3927131-1-goldstein.w.n@gmail.com> References: <20220323215734.3927131-1-goldstein.w.n@gmail.com> MIME-Version: 1.0 X-Spam-Status: No, score=-12.5 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, FREEMAIL_FROM, GIT_PATCH_0, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: libc-alpha@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Libc-alpha mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Noah Goldstein via Libc-alpha From: Noah Goldstein Reply-To: Noah Goldstein Errors-To: libc-alpha-bounces+patchwork=sourceware.org@sourceware.org Sender: "Libc-alpha" The rational is: 1. SSE42 has nearly identical logic so any benefit is minimal (3.4% regression on Tigerlake using SSE42 versus AVX across the benchtest suite). 2. AVX2 version covers the majority of targets that previously prefered it. 3. The targets where AVX would still be best (SnB and IVB) are becoming outdated. All in all the saving the code size is worth it. All string/memory tests pass. Reviewed-by: H.J. Lu --- Geomtric Mean N=40 runs; All functions page aligned length, align1, align2, max_char, AVX Time / SSE42 Time 1, 1, 1, 127, 0.928 2, 2, 2, 127, 0.934 3, 3, 3, 127, 0.975 4, 4, 4, 127, 0.96 5, 5, 5, 127, 0.935 6, 6, 6, 127, 0.929 7, 7, 7, 127, 0.959 8, 0, 0, 127, 0.955 9, 1, 1, 127, 0.944 10, 2, 2, 127, 0.975 11, 3, 3, 127, 0.935 12, 4, 4, 127, 0.931 13, 5, 5, 127, 0.926 14, 6, 6, 127, 0.901 15, 7, 7, 127, 0.951 4, 0, 0, 127, 0.958 4, 0, 0, 254, 0.956 8, 0, 0, 254, 0.977 16, 0, 0, 127, 0.955 16, 0, 0, 254, 0.953 32, 0, 0, 127, 0.943 32, 0, 0, 254, 0.941 64, 0, 0, 127, 0.941 64, 0, 0, 254, 0.955 128, 0, 0, 127, 0.972 128, 0, 0, 254, 0.975 256, 0, 0, 127, 0.996 256, 0, 0, 254, 0.993 512, 0, 0, 127, 0.992 512, 0, 0, 254, 0.986 1024, 0, 0, 127, 0.994 1024, 0, 0, 254, 0.993 16, 1, 2, 127, 0.933 16, 2, 1, 254, 0.953 32, 2, 4, 127, 0.927 32, 4, 2, 254, 0.986 64, 3, 6, 127, 0.991 64, 6, 3, 254, 1.014 128, 4, 0, 127, 1.001 128, 0, 4, 254, 0.991 256, 5, 2, 127, 1.011 256, 2, 5, 254, 1.013 512, 6, 4, 127, 1.056 512, 4, 6, 254, 0.916 1024, 7, 6, 127, 1.059 1024, 6, 7, 254, 1.043 sysdeps/x86_64/multiarch/Makefile | 2 - sysdeps/x86_64/multiarch/ifunc-impl-list.c | 12 - sysdeps/x86_64/multiarch/ifunc-strcasecmp.h | 4 - sysdeps/x86_64/multiarch/strcasecmp_l-avx.S | 22 -- sysdeps/x86_64/multiarch/strcmp-sse42.S | 240 +++++++++----------- sysdeps/x86_64/multiarch/strncase_l-avx.S | 22 -- 6 files changed, 105 insertions(+), 197 deletions(-) delete mode 100644 sysdeps/x86_64/multiarch/strcasecmp_l-avx.S delete mode 100644 sysdeps/x86_64/multiarch/strncase_l-avx.S diff --git a/sysdeps/x86_64/multiarch/Makefile b/sysdeps/x86_64/multiarch/Makefile index 35d80dc2ff..6507d1b7fa 100644 --- a/sysdeps/x86_64/multiarch/Makefile +++ b/sysdeps/x86_64/multiarch/Makefile @@ -54,7 +54,6 @@ sysdep_routines += \ stpncpy-evex \ stpncpy-sse2-unaligned \ stpncpy-ssse3 \ - strcasecmp_l-avx \ strcasecmp_l-avx2 \ strcasecmp_l-avx2-rtm \ strcasecmp_l-evex \ @@ -95,7 +94,6 @@ sysdep_routines += \ strlen-avx2-rtm \ strlen-evex \ strlen-sse2 \ - strncase_l-avx \ strncase_l-avx2 \ strncase_l-avx2-rtm \ strncase_l-evex \ diff --git a/sysdeps/x86_64/multiarch/ifunc-impl-list.c b/sysdeps/x86_64/multiarch/ifunc-impl-list.c index f1a4d3dac2..40cc6cc49e 100644 --- a/sysdeps/x86_64/multiarch/ifunc-impl-list.c +++ b/sysdeps/x86_64/multiarch/ifunc-impl-list.c @@ -447,9 +447,6 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array, (CPU_FEATURE_USABLE (AVX2) && CPU_FEATURE_USABLE (RTM)), __strcasecmp_avx2_rtm) - IFUNC_IMPL_ADD (array, i, strcasecmp, - CPU_FEATURE_USABLE (AVX), - __strcasecmp_avx) IFUNC_IMPL_ADD (array, i, strcasecmp, CPU_FEATURE_USABLE (SSE4_2), __strcasecmp_sse42) @@ -471,9 +468,6 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array, (CPU_FEATURE_USABLE (AVX2) && CPU_FEATURE_USABLE (RTM)), __strcasecmp_l_avx2_rtm) - IFUNC_IMPL_ADD (array, i, strcasecmp_l, - CPU_FEATURE_USABLE (AVX), - __strcasecmp_l_avx) IFUNC_IMPL_ADD (array, i, strcasecmp_l, CPU_FEATURE_USABLE (SSE4_2), __strcasecmp_l_sse42) @@ -609,9 +603,6 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array, (CPU_FEATURE_USABLE (AVX2) && CPU_FEATURE_USABLE (RTM)), __strncasecmp_avx2_rtm) - IFUNC_IMPL_ADD (array, i, strncasecmp, - CPU_FEATURE_USABLE (AVX), - __strncasecmp_avx) IFUNC_IMPL_ADD (array, i, strncasecmp, CPU_FEATURE_USABLE (SSE4_2), __strncasecmp_sse42) @@ -634,9 +625,6 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array, (CPU_FEATURE_USABLE (AVX2) && CPU_FEATURE_USABLE (RTM)), __strncasecmp_l_avx2_rtm) - IFUNC_IMPL_ADD (array, i, strncasecmp_l, - CPU_FEATURE_USABLE (AVX), - __strncasecmp_l_avx) IFUNC_IMPL_ADD (array, i, strncasecmp_l, CPU_FEATURE_USABLE (SSE4_2), __strncasecmp_l_sse42) diff --git a/sysdeps/x86_64/multiarch/ifunc-strcasecmp.h b/sysdeps/x86_64/multiarch/ifunc-strcasecmp.h index bf0d146e7f..766539c241 100644 --- a/sysdeps/x86_64/multiarch/ifunc-strcasecmp.h +++ b/sysdeps/x86_64/multiarch/ifunc-strcasecmp.h @@ -22,7 +22,6 @@ extern __typeof (REDIRECT_NAME) OPTIMIZE (sse2) attribute_hidden; extern __typeof (REDIRECT_NAME) OPTIMIZE (ssse3) attribute_hidden; extern __typeof (REDIRECT_NAME) OPTIMIZE (sse42) attribute_hidden; -extern __typeof (REDIRECT_NAME) OPTIMIZE (avx) attribute_hidden; extern __typeof (REDIRECT_NAME) OPTIMIZE (avx2) attribute_hidden; extern __typeof (REDIRECT_NAME) OPTIMIZE (avx2_rtm) attribute_hidden; extern __typeof (REDIRECT_NAME) OPTIMIZE (evex) attribute_hidden; @@ -46,9 +45,6 @@ IFUNC_SELECTOR (void) return OPTIMIZE (avx2); } - if (CPU_FEATURE_USABLE_P (cpu_features, AVX)) - return OPTIMIZE (avx); - if (CPU_FEATURE_USABLE_P (cpu_features, SSE4_2) && !CPU_FEATURES_ARCH_P (cpu_features, Slow_SSE4_2)) return OPTIMIZE (sse42); diff --git a/sysdeps/x86_64/multiarch/strcasecmp_l-avx.S b/sysdeps/x86_64/multiarch/strcasecmp_l-avx.S deleted file mode 100644 index 7ec7c21b5a..0000000000 --- a/sysdeps/x86_64/multiarch/strcasecmp_l-avx.S +++ /dev/null @@ -1,22 +0,0 @@ -/* strcasecmp_l optimized with AVX. - Copyright (C) 2017-2022 Free Software Foundation, Inc. - 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 - . */ - -#define STRCMP_SSE42 __strcasecmp_l_avx -#define USE_AVX 1 -#define USE_AS_STRCASECMP_L -#include "strcmp-sse42.S" diff --git a/sysdeps/x86_64/multiarch/strcmp-sse42.S b/sysdeps/x86_64/multiarch/strcmp-sse42.S index 7805ae9d41..a9178ad25c 100644 --- a/sysdeps/x86_64/multiarch/strcmp-sse42.S +++ b/sysdeps/x86_64/multiarch/strcmp-sse42.S @@ -41,13 +41,8 @@ # define UPDATE_STRNCMP_COUNTER #endif -#ifdef USE_AVX -# define SECTION avx -# define GLABEL(l) l##_avx -#else -# define SECTION sse4.2 -# define GLABEL(l) l##_sse42 -#endif +#define SECTION sse4.2 +#define GLABEL(l) l##_sse42 #define LABEL(l) .L##l @@ -105,21 +100,7 @@ END (GLABEL(__strncasecmp)) #endif -#ifdef USE_AVX -# define movdqa vmovdqa -# define movdqu vmovdqu -# define pmovmskb vpmovmskb -# define pcmpistri vpcmpistri -# define psubb vpsubb -# define pcmpeqb vpcmpeqb -# define psrldq vpsrldq -# define pslldq vpslldq -# define palignr vpalignr -# define pxor vpxor -# define D(arg) arg, arg -#else -# define D(arg) arg -#endif +#define arg arg STRCMP_SSE42: cfi_startproc @@ -191,18 +172,7 @@ LABEL(case_add): movdqu (%rdi), %xmm1 movdqu (%rsi), %xmm2 #if defined USE_AS_STRCASECMP_L || defined USE_AS_STRNCASECMP_L -# ifdef USE_AVX -# define TOLOWER(reg1, reg2) \ - vpaddb LCASE_MIN_reg, reg1, %xmm7; \ - vpaddb LCASE_MIN_reg, reg2, %xmm8; \ - vpcmpgtb LCASE_MAX_reg, %xmm7, %xmm7; \ - vpcmpgtb LCASE_MAX_reg, %xmm8, %xmm8; \ - vpandn CASE_ADD_reg, %xmm7, %xmm7; \ - vpandn CASE_ADD_reg, %xmm8, %xmm8; \ - vpaddb %xmm7, reg1, reg1; \ - vpaddb %xmm8, reg2, reg2 -# else -# define TOLOWER(reg1, reg2) \ +# define TOLOWER(reg1, reg2) \ movdqa LCASE_MIN_reg, %xmm7; \ movdqa LCASE_MIN_reg, %xmm8; \ paddb reg1, %xmm7; \ @@ -213,15 +183,15 @@ LABEL(case_add): pandn CASE_ADD_reg, %xmm8; \ paddb %xmm7, reg1; \ paddb %xmm8, reg2 -# endif + TOLOWER (%xmm1, %xmm2) #else # define TOLOWER(reg1, reg2) #endif - pxor %xmm0, D(%xmm0) /* clear %xmm0 for null char checks */ - pcmpeqb %xmm1, D(%xmm0) /* Any null chars? */ - pcmpeqb %xmm2, D(%xmm1) /* compare first 16 bytes for equality */ - psubb %xmm0, D(%xmm1) /* packed sub of comparison results*/ + pxor %xmm0, %xmm0 /* clear %xmm0 for null char checks */ + pcmpeqb %xmm1, %xmm0 /* Any null chars? */ + pcmpeqb %xmm2, %xmm1 /* compare first 16 bytes for equality */ + psubb %xmm0, %xmm1 /* packed sub of comparison results*/ pmovmskb %xmm1, %edx sub $0xffff, %edx /* if first 16 bytes are same, edx == 0xffff */ jnz LABEL(less16bytes)/* If not, find different value or null char */ @@ -245,7 +215,7 @@ LABEL(crosscache): xor %r8d, %r8d and $0xf, %ecx /* offset of rsi */ and $0xf, %eax /* offset of rdi */ - pxor %xmm0, D(%xmm0) /* clear %xmm0 for null char check */ + pxor %xmm0, %xmm0 /* clear %xmm0 for null char check */ cmp %eax, %ecx je LABEL(ashr_0) /* rsi and rdi relative offset same */ ja LABEL(bigger) @@ -259,7 +229,7 @@ LABEL(bigger): sub %rcx, %r9 lea LABEL(unaligned_table)(%rip), %r10 movslq (%r10, %r9,4), %r9 - pcmpeqb %xmm1, D(%xmm0) /* Any null chars? */ + pcmpeqb %xmm1, %xmm0 /* Any null chars? */ lea (%r10, %r9), %r10 _CET_NOTRACK jmp *%r10 /* jump to corresponding case */ @@ -272,15 +242,15 @@ LABEL(bigger): LABEL(ashr_0): movdqa (%rsi), %xmm1 - pcmpeqb %xmm1, D(%xmm0) /* Any null chars? */ + pcmpeqb %xmm1, %xmm0 /* Any null chars? */ #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L - pcmpeqb (%rdi), D(%xmm1) /* compare 16 bytes for equality */ + pcmpeqb (%rdi), %xmm1 /* compare 16 bytes for equality */ #else movdqa (%rdi), %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm2, D(%xmm1) /* compare 16 bytes for equality */ + pcmpeqb %xmm2, %xmm1 /* compare 16 bytes for equality */ #endif - psubb %xmm0, D(%xmm1) /* packed sub of comparison results*/ + psubb %xmm0, %xmm1 /* packed sub of comparison results*/ pmovmskb %xmm1, %r9d shr %cl, %edx /* adjust 0xffff for offset */ shr %cl, %r9d /* adjust for 16-byte offset */ @@ -360,10 +330,10 @@ LABEL(ashr_0_exit_use): */ .p2align 4 LABEL(ashr_1): - pslldq $15, D(%xmm2) /* shift first string to align with second */ + pslldq $15, %xmm2 /* shift first string to align with second */ TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) /* compare 16 bytes for equality */ - psubb %xmm0, D(%xmm2) /* packed sub of comparison results*/ + pcmpeqb %xmm1, %xmm2 /* compare 16 bytes for equality */ + psubb %xmm0, %xmm2 /* packed sub of comparison results*/ pmovmskb %xmm2, %r9d shr %cl, %edx /* adjust 0xffff for offset */ shr %cl, %r9d /* adjust for 16-byte offset */ @@ -391,7 +361,7 @@ LABEL(loop_ashr_1_use): LABEL(nibble_ashr_1_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $1, -16(%rdi, %rdx), D(%xmm0) + palignr $1, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -410,7 +380,7 @@ LABEL(nibble_ashr_1_restart_use): jg LABEL(nibble_ashr_1_use) movdqa (%rdi, %rdx), %xmm0 - palignr $1, -16(%rdi, %rdx), D(%xmm0) + palignr $1, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -430,7 +400,7 @@ LABEL(nibble_ashr_1_restart_use): LABEL(nibble_ashr_1_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $1, D(%xmm0) + psrldq $1, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -448,10 +418,10 @@ LABEL(nibble_ashr_1_use): */ .p2align 4 LABEL(ashr_2): - pslldq $14, D(%xmm2) + pslldq $14, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -479,7 +449,7 @@ LABEL(loop_ashr_2_use): LABEL(nibble_ashr_2_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $2, -16(%rdi, %rdx), D(%xmm0) + palignr $2, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -498,7 +468,7 @@ LABEL(nibble_ashr_2_restart_use): jg LABEL(nibble_ashr_2_use) movdqa (%rdi, %rdx), %xmm0 - palignr $2, -16(%rdi, %rdx), D(%xmm0) + palignr $2, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -518,7 +488,7 @@ LABEL(nibble_ashr_2_restart_use): LABEL(nibble_ashr_2_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $2, D(%xmm0) + psrldq $2, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -536,10 +506,10 @@ LABEL(nibble_ashr_2_use): */ .p2align 4 LABEL(ashr_3): - pslldq $13, D(%xmm2) + pslldq $13, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -567,7 +537,7 @@ LABEL(loop_ashr_3_use): LABEL(nibble_ashr_3_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $3, -16(%rdi, %rdx), D(%xmm0) + palignr $3, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -586,7 +556,7 @@ LABEL(nibble_ashr_3_restart_use): jg LABEL(nibble_ashr_3_use) movdqa (%rdi, %rdx), %xmm0 - palignr $3, -16(%rdi, %rdx), D(%xmm0) + palignr $3, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -606,7 +576,7 @@ LABEL(nibble_ashr_3_restart_use): LABEL(nibble_ashr_3_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $3, D(%xmm0) + psrldq $3, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -624,10 +594,10 @@ LABEL(nibble_ashr_3_use): */ .p2align 4 LABEL(ashr_4): - pslldq $12, D(%xmm2) + pslldq $12, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -656,7 +626,7 @@ LABEL(loop_ashr_4_use): LABEL(nibble_ashr_4_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $4, -16(%rdi, %rdx), D(%xmm0) + palignr $4, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -675,7 +645,7 @@ LABEL(nibble_ashr_4_restart_use): jg LABEL(nibble_ashr_4_use) movdqa (%rdi, %rdx), %xmm0 - palignr $4, -16(%rdi, %rdx), D(%xmm0) + palignr $4, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -695,7 +665,7 @@ LABEL(nibble_ashr_4_restart_use): LABEL(nibble_ashr_4_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $4, D(%xmm0) + psrldq $4, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -713,10 +683,10 @@ LABEL(nibble_ashr_4_use): */ .p2align 4 LABEL(ashr_5): - pslldq $11, D(%xmm2) + pslldq $11, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -745,7 +715,7 @@ LABEL(loop_ashr_5_use): LABEL(nibble_ashr_5_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $5, -16(%rdi, %rdx), D(%xmm0) + palignr $5, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -765,7 +735,7 @@ LABEL(nibble_ashr_5_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $5, -16(%rdi, %rdx), D(%xmm0) + palignr $5, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -785,7 +755,7 @@ LABEL(nibble_ashr_5_restart_use): LABEL(nibble_ashr_5_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $5, D(%xmm0) + psrldq $5, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -803,10 +773,10 @@ LABEL(nibble_ashr_5_use): */ .p2align 4 LABEL(ashr_6): - pslldq $10, D(%xmm2) + pslldq $10, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -835,7 +805,7 @@ LABEL(loop_ashr_6_use): LABEL(nibble_ashr_6_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $6, -16(%rdi, %rdx), D(%xmm0) + palignr $6, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -854,7 +824,7 @@ LABEL(nibble_ashr_6_restart_use): jg LABEL(nibble_ashr_6_use) movdqa (%rdi, %rdx), %xmm0 - palignr $6, -16(%rdi, %rdx), D(%xmm0) + palignr $6, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -874,7 +844,7 @@ LABEL(nibble_ashr_6_restart_use): LABEL(nibble_ashr_6_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $6, D(%xmm0) + psrldq $6, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -892,10 +862,10 @@ LABEL(nibble_ashr_6_use): */ .p2align 4 LABEL(ashr_7): - pslldq $9, D(%xmm2) + pslldq $9, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -924,7 +894,7 @@ LABEL(loop_ashr_7_use): LABEL(nibble_ashr_7_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $7, -16(%rdi, %rdx), D(%xmm0) + palignr $7, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -943,7 +913,7 @@ LABEL(nibble_ashr_7_restart_use): jg LABEL(nibble_ashr_7_use) movdqa (%rdi, %rdx), %xmm0 - palignr $7, -16(%rdi, %rdx), D(%xmm0) + palignr $7, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a,(%rsi,%rdx), %xmm0 #else @@ -963,7 +933,7 @@ LABEL(nibble_ashr_7_restart_use): LABEL(nibble_ashr_7_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $7, D(%xmm0) + psrldq $7, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -981,10 +951,10 @@ LABEL(nibble_ashr_7_use): */ .p2align 4 LABEL(ashr_8): - pslldq $8, D(%xmm2) + pslldq $8, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -1013,7 +983,7 @@ LABEL(loop_ashr_8_use): LABEL(nibble_ashr_8_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $8, -16(%rdi, %rdx), D(%xmm0) + palignr $8, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1032,7 +1002,7 @@ LABEL(nibble_ashr_8_restart_use): jg LABEL(nibble_ashr_8_use) movdqa (%rdi, %rdx), %xmm0 - palignr $8, -16(%rdi, %rdx), D(%xmm0) + palignr $8, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1052,7 +1022,7 @@ LABEL(nibble_ashr_8_restart_use): LABEL(nibble_ashr_8_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $8, D(%xmm0) + psrldq $8, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -1070,10 +1040,10 @@ LABEL(nibble_ashr_8_use): */ .p2align 4 LABEL(ashr_9): - pslldq $7, D(%xmm2) + pslldq $7, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -1103,7 +1073,7 @@ LABEL(loop_ashr_9_use): LABEL(nibble_ashr_9_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $9, -16(%rdi, %rdx), D(%xmm0) + palignr $9, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1122,7 +1092,7 @@ LABEL(nibble_ashr_9_restart_use): jg LABEL(nibble_ashr_9_use) movdqa (%rdi, %rdx), %xmm0 - palignr $9, -16(%rdi, %rdx), D(%xmm0) + palignr $9, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1142,7 +1112,7 @@ LABEL(nibble_ashr_9_restart_use): LABEL(nibble_ashr_9_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $9, D(%xmm0) + psrldq $9, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -1160,10 +1130,10 @@ LABEL(nibble_ashr_9_use): */ .p2align 4 LABEL(ashr_10): - pslldq $6, D(%xmm2) + pslldq $6, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -1192,7 +1162,7 @@ LABEL(loop_ashr_10_use): LABEL(nibble_ashr_10_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $10, -16(%rdi, %rdx), D(%xmm0) + palignr $10, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1211,7 +1181,7 @@ LABEL(nibble_ashr_10_restart_use): jg LABEL(nibble_ashr_10_use) movdqa (%rdi, %rdx), %xmm0 - palignr $10, -16(%rdi, %rdx), D(%xmm0) + palignr $10, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1231,7 +1201,7 @@ LABEL(nibble_ashr_10_restart_use): LABEL(nibble_ashr_10_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $10, D(%xmm0) + psrldq $10, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -1249,10 +1219,10 @@ LABEL(nibble_ashr_10_use): */ .p2align 4 LABEL(ashr_11): - pslldq $5, D(%xmm2) + pslldq $5, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -1281,7 +1251,7 @@ LABEL(loop_ashr_11_use): LABEL(nibble_ashr_11_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $11, -16(%rdi, %rdx), D(%xmm0) + palignr $11, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1300,7 +1270,7 @@ LABEL(nibble_ashr_11_restart_use): jg LABEL(nibble_ashr_11_use) movdqa (%rdi, %rdx), %xmm0 - palignr $11, -16(%rdi, %rdx), D(%xmm0) + palignr $11, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1320,7 +1290,7 @@ LABEL(nibble_ashr_11_restart_use): LABEL(nibble_ashr_11_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $11, D(%xmm0) + psrldq $11, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -1338,10 +1308,10 @@ LABEL(nibble_ashr_11_use): */ .p2align 4 LABEL(ashr_12): - pslldq $4, D(%xmm2) + pslldq $4, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -1370,7 +1340,7 @@ LABEL(loop_ashr_12_use): LABEL(nibble_ashr_12_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $12, -16(%rdi, %rdx), D(%xmm0) + palignr $12, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1389,7 +1359,7 @@ LABEL(nibble_ashr_12_restart_use): jg LABEL(nibble_ashr_12_use) movdqa (%rdi, %rdx), %xmm0 - palignr $12, -16(%rdi, %rdx), D(%xmm0) + palignr $12, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1409,7 +1379,7 @@ LABEL(nibble_ashr_12_restart_use): LABEL(nibble_ashr_12_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $12, D(%xmm0) + psrldq $12, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -1427,10 +1397,10 @@ LABEL(nibble_ashr_12_use): */ .p2align 4 LABEL(ashr_13): - pslldq $3, D(%xmm2) + pslldq $3, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -1460,7 +1430,7 @@ LABEL(loop_ashr_13_use): LABEL(nibble_ashr_13_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $13, -16(%rdi, %rdx), D(%xmm0) + palignr $13, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1479,7 +1449,7 @@ LABEL(nibble_ashr_13_restart_use): jg LABEL(nibble_ashr_13_use) movdqa (%rdi, %rdx), %xmm0 - palignr $13, -16(%rdi, %rdx), D(%xmm0) + palignr $13, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1499,7 +1469,7 @@ LABEL(nibble_ashr_13_restart_use): LABEL(nibble_ashr_13_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $13, D(%xmm0) + psrldq $13, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -1517,10 +1487,10 @@ LABEL(nibble_ashr_13_use): */ .p2align 4 LABEL(ashr_14): - pslldq $2, D(%xmm2) + pslldq $2, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -1550,7 +1520,7 @@ LABEL(loop_ashr_14_use): LABEL(nibble_ashr_14_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $14, -16(%rdi, %rdx), D(%xmm0) + palignr $14, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1569,7 +1539,7 @@ LABEL(nibble_ashr_14_restart_use): jg LABEL(nibble_ashr_14_use) movdqa (%rdi, %rdx), %xmm0 - palignr $14, -16(%rdi, %rdx), D(%xmm0) + palignr $14, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1589,7 +1559,7 @@ LABEL(nibble_ashr_14_restart_use): LABEL(nibble_ashr_14_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $14, D(%xmm0) + psrldq $14, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx @@ -1607,10 +1577,10 @@ LABEL(nibble_ashr_14_use): */ .p2align 4 LABEL(ashr_15): - pslldq $1, D(%xmm2) + pslldq $1, %xmm2 TOLOWER (%xmm1, %xmm2) - pcmpeqb %xmm1, D(%xmm2) - psubb %xmm0, D(%xmm2) + pcmpeqb %xmm1, %xmm2 + psubb %xmm0, %xmm2 pmovmskb %xmm2, %r9d shr %cl, %edx shr %cl, %r9d @@ -1642,7 +1612,7 @@ LABEL(loop_ashr_15_use): LABEL(nibble_ashr_15_restart_use): movdqa (%rdi, %rdx), %xmm0 - palignr $15, -16(%rdi, %rdx), D(%xmm0) + palignr $15, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1661,7 +1631,7 @@ LABEL(nibble_ashr_15_restart_use): jg LABEL(nibble_ashr_15_use) movdqa (%rdi, %rdx), %xmm0 - palignr $15, -16(%rdi, %rdx), D(%xmm0) + palignr $15, -16(%rdi, %rdx), %xmm0 #if !defined USE_AS_STRCASECMP_L && !defined USE_AS_STRNCASECMP_L pcmpistri $0x1a, (%rsi,%rdx), %xmm0 #else @@ -1681,7 +1651,7 @@ LABEL(nibble_ashr_15_restart_use): LABEL(nibble_ashr_15_use): sub $0x1000, %r10 movdqa -16(%rdi, %rdx), %xmm0 - psrldq $15, D(%xmm0) + psrldq $15, %xmm0 pcmpistri $0x3a,%xmm0, %xmm0 #if defined USE_AS_STRNCMP || defined USE_AS_STRNCASECMP_L cmp %r11, %rcx diff --git a/sysdeps/x86_64/multiarch/strncase_l-avx.S b/sysdeps/x86_64/multiarch/strncase_l-avx.S deleted file mode 100644 index b51b86d223..0000000000 --- a/sysdeps/x86_64/multiarch/strncase_l-avx.S +++ /dev/null @@ -1,22 +0,0 @@ -/* strncasecmp_l optimized with AVX. - Copyright (C) 2017-2022 Free Software Foundation, Inc. - 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 - . */ - -#define STRCMP_SSE42 __strncasecmp_l_avx -#define USE_AVX 1 -#define USE_AS_STRNCASECMP_L -#include "strcmp-sse42.S"