From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on dcvr.yhbt.net X-Spam-Level: X-Spam-ASN: AS3215 2.6.0.0/16 X-Spam-Status: No, score=-3.9 required=3.0 tests=AWL,BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,MAILING_LIST_MULTI, RCVD_IN_DNSWL_MED,SPF_HELO_PASS,SPF_PASS shortcircuit=no autolearn=ham autolearn_force=no version=3.4.2 Received: from sourceware.org (server2.sourceware.org [IPv6:2620:52:3:1:0:246e:9693:128c]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) by dcvr.yhbt.net (Postfix) with ESMTPS id 11EE31F910 for ; Fri, 18 Nov 2022 19:09:13 +0000 (UTC) Authentication-Results: dcvr.yhbt.net; dkim=pass (1024-bit key; secure) header.d=sourceware.org header.i=@sourceware.org header.b="yG8G3FR9"; dkim-atps=neutral Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id B53FC384F6FB for ; Fri, 18 Nov 2022 19:09:08 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B53FC384F6FB DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1668798548; bh=vQCXdl1sJVmPknv5WkMIacfiyDWZB/7ZwobTxUXULt4=; h=To:Cc:Subject:Date:List-Id:List-Unsubscribe:List-Archive: List-Post:List-Help:List-Subscribe:From:Reply-To:From; b=yG8G3FR9DLg91LK/CJCzv2AfPJUMAWhCr6XU1TT4aaC7JK1ioov+SNX00BiFTs8vy +bSGD+Ggo5TxC2ZgEwAMvAFl7Kqp4rmNCy6dqriu8rd2GzgCMF2k9jh9ftmlE+HvPB jSRo1oYDTzhHez3oNkivtH4gnjcDtoLvN2mSJTK4= Received: from mail-pf1-x431.google.com (mail-pf1-x431.google.com [IPv6:2607:f8b0:4864:20::431]) by sourceware.org (Postfix) with ESMTPS id 972BA3854551 for ; Fri, 18 Nov 2022 19:08:43 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 972BA3854551 Received: by mail-pf1-x431.google.com with SMTP id g62so5738545pfb.10 for ; Fri, 18 Nov 2022 11:08:43 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:message-id:date:subject:cc :to:from:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=vQCXdl1sJVmPknv5WkMIacfiyDWZB/7ZwobTxUXULt4=; b=kMPXgmiL8kckET9OryYLBGSUO8zJCN3CqqoaARXhSQVrKcOnWTWP7l6/NFt938ub0l /T2jNOOWwNL0XXmKa60oi2IQIKkUBr7K/nLng5bsJIaDCBAeH6sSw1Y1mDo/aIcVPyAh qhvlBwYA9Es4xl12xtHaKsWPQ5hIlsAOjJd6+O5u47AAQBfj176XeW5pxddWaFwbyu14 pj2VTUODPOduSR2k6jk3t0YLPzCrdlWMMZY58FB8ZBOTPZDNkgtwkVlwKnQ5ems+LRcX v+f4c21uSnX/TROq7SK1dw37BT+f4kGpQQ9OS5D61LzjOiCC0A62VhDkDGy11dz7FgAk q6Iw== X-Gm-Message-State: ANoB5pmoCbjAF6z+lSCszXelbzgbWC5SK1Nw4/b0XBwYaTW0BGoVNNiD fJyT0I5FR4D+15Op/lhZBJRIqP2Vov4= X-Google-Smtp-Source: AA0mqf6Skk0KZG6qIZBOrmm7feUZGZO1lYDptVFs0pjethJ8JCOAbnZ12a8f2DHwP27kAfZ0dNK2GQ== X-Received: by 2002:a63:1345:0:b0:476:f92f:69f0 with SMTP id 5-20020a631345000000b00476f92f69f0mr7763863pgt.463.1668798521633; Fri, 18 Nov 2022 11:08:41 -0800 (PST) Received: from noahgold-desk.. ([192.55.60.47]) by smtp.gmail.com with ESMTPSA id q16-20020a17090311d000b00182a9c27acfsm4120680plh.227.2022.11.18.11.08.40 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 18 Nov 2022 11:08:41 -0800 (PST) To: libc-alpha@sourceware.org Cc: goldstein.w.n@gmail.com, hjl.tools@gmail.com, andrey.kolesov@intel.com, carlos@systemhalted.org Subject: [PATCH v2 1/2] x86/fpu: Reformat svml_{s|d}_wrapper_impl.h.S Date: Fri, 18 Nov 2022 11:08:34 -0800 Message-Id: <20221118190835.1033248-1-goldstein.w.n@gmail.com> X-Mailer: git-send-email 2.34.1 MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 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: , From: Noah Goldstein via Libc-alpha Reply-To: Noah Goldstein Errors-To: libc-alpha-bounces+e=80x24.org@sourceware.org Sender: "Libc-alpha" Just reformat with the style convention used in other x86 assembler files. This doesn't change libm.so or libmvec.so. --- sysdeps/x86_64/fpu/svml_d_wrapper_impl.h | 474 ++++++++++---------- sysdeps/x86_64/fpu/svml_s_wrapper_impl.h | 546 +++++++++++------------ 2 files changed, 510 insertions(+), 510 deletions(-) diff --git a/sysdeps/x86_64/fpu/svml_d_wrapper_impl.h b/sysdeps/x86_64/fpu/svml_d_wrapper_impl.h index 2334713015..b03a2122b9 100644 --- a/sysdeps/x86_64/fpu/svml_d_wrapper_impl.h +++ b/sysdeps/x86_64/fpu/svml_d_wrapper_impl.h @@ -18,273 +18,273 @@ /* SSE2 ISA version as wrapper to scalar. */ .macro WRAPPER_IMPL_SSE2 callee - subq $40, %rsp - cfi_adjust_cfa_offset(40) - movaps %xmm0, (%rsp) - call JUMPTARGET(\callee) - movsd %xmm0, 16(%rsp) - movsd 8(%rsp), %xmm0 - call JUMPTARGET(\callee) - movsd 16(%rsp), %xmm1 - movsd %xmm0, 24(%rsp) - unpcklpd %xmm0, %xmm1 - movaps %xmm1, %xmm0 - addq $40, %rsp - cfi_adjust_cfa_offset(-40) - ret + subq $40, %rsp + cfi_adjust_cfa_offset (40) + movaps %xmm0, (%rsp) + call JUMPTARGET(\callee) + movsd %xmm0, 16(%rsp) + movsd 8(%rsp), %xmm0 + call JUMPTARGET(\callee) + movsd 16(%rsp), %xmm1 + movsd %xmm0, 24(%rsp) + unpcklpd %xmm0, %xmm1 + movaps %xmm1, %xmm0 + addq $40, %rsp + cfi_adjust_cfa_offset (-40) + ret .endm /* 2 argument SSE2 ISA version as wrapper to scalar. */ .macro WRAPPER_IMPL_SSE2_ff callee - subq $56, %rsp - cfi_adjust_cfa_offset(56) - movaps %xmm0, (%rsp) - movaps %xmm1, 16(%rsp) - call JUMPTARGET(\callee) - movsd %xmm0, 32(%rsp) - movsd 8(%rsp), %xmm0 - movsd 24(%rsp), %xmm1 - call JUMPTARGET(\callee) - movsd 32(%rsp), %xmm1 - movsd %xmm0, 40(%rsp) - unpcklpd %xmm0, %xmm1 - movaps %xmm1, %xmm0 - addq $56, %rsp - cfi_adjust_cfa_offset(-56) - ret + subq $56, %rsp + cfi_adjust_cfa_offset (56) + movaps %xmm0, (%rsp) + movaps %xmm1, 16(%rsp) + call JUMPTARGET(\callee) + movsd %xmm0, 32(%rsp) + movsd 8(%rsp), %xmm0 + movsd 24(%rsp), %xmm1 + call JUMPTARGET(\callee) + movsd 32(%rsp), %xmm1 + movsd %xmm0, 40(%rsp) + unpcklpd %xmm0, %xmm1 + movaps %xmm1, %xmm0 + addq $56, %rsp + cfi_adjust_cfa_offset (-56) + ret .endm /* 3 argument SSE2 ISA version as wrapper to scalar. */ .macro WRAPPER_IMPL_SSE2_fFF callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - pushq %rbx - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbx, 0) - movq %rdi, %rbp - movq %rsi, %rbx - subq $40, %rsp - cfi_adjust_cfa_offset(40) - leaq 16(%rsp), %rsi - leaq 24(%rsp), %rdi - movaps %xmm0, (%rsp) - call JUMPTARGET(\callee) - leaq 16(%rsp), %rsi - leaq 24(%rsp), %rdi - movsd 24(%rsp), %xmm0 - movapd (%rsp), %xmm1 - movsd %xmm0, 0(%rbp) - unpckhpd %xmm1, %xmm1 - movsd 16(%rsp), %xmm0 - movsd %xmm0, (%rbx) - movapd %xmm1, %xmm0 - call JUMPTARGET(\callee) - movsd 24(%rsp), %xmm0 - movsd %xmm0, 8(%rbp) - movsd 16(%rsp), %xmm0 - movsd %xmm0, 8(%rbx) - addq $40, %rsp - cfi_adjust_cfa_offset(-40) - popq %rbx - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbx) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + pushq %rbx + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbx, 0) + movq %rdi, %rbp + movq %rsi, %rbx + subq $40, %rsp + cfi_adjust_cfa_offset (40) + leaq 16(%rsp), %rsi + leaq 24(%rsp), %rdi + movaps %xmm0, (%rsp) + call JUMPTARGET(\callee) + leaq 16(%rsp), %rsi + leaq 24(%rsp), %rdi + movsd 24(%rsp), %xmm0 + movapd (%rsp), %xmm1 + movsd %xmm0, 0(%rbp) + unpckhpd %xmm1, %xmm1 + movsd 16(%rsp), %xmm0 + movsd %xmm0, (%rbx) + movapd %xmm1, %xmm0 + call JUMPTARGET(\callee) + movsd 24(%rsp), %xmm0 + movsd %xmm0, 8(%rbp) + movsd 16(%rsp), %xmm0 + movsd %xmm0, 8(%rbx) + addq $40, %rsp + cfi_adjust_cfa_offset (-40) + popq %rbx + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbx) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* AVX/AVX2 ISA version as wrapper to SSE ISA version. */ .macro WRAPPER_IMPL_AVX callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-32, %rsp - subq $32, %rsp - vextractf128 $1, %ymm0, (%rsp) - vzeroupper - call HIDDEN_JUMPTARGET(\callee) - vmovapd %xmm0, 16(%rsp) - vmovaps (%rsp), %xmm0 - call HIDDEN_JUMPTARGET(\callee) - vmovapd %xmm0, %xmm1 - vmovapd 16(%rsp), %xmm0 - vinsertf128 $1, %xmm1, %ymm0, %ymm0 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-32, %rsp + subq $32, %rsp + vextractf128 $1, %ymm0, (%rsp) + vzeroupper + call HIDDEN_JUMPTARGET(\callee) + vmovapd %xmm0, 16(%rsp) + vmovaps (%rsp), %xmm0 + call HIDDEN_JUMPTARGET(\callee) + vmovapd %xmm0, %xmm1 + vmovapd 16(%rsp), %xmm0 + vinsertf128 $1, %xmm1, %ymm0, %ymm0 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* 2 argument AVX/AVX2 ISA version as wrapper to SSE ISA version. */ .macro WRAPPER_IMPL_AVX_ff callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-32, %rsp - subq $64, %rsp - vextractf128 $1, %ymm0, 16(%rsp) - vextractf128 $1, %ymm1, (%rsp) - vzeroupper - call HIDDEN_JUMPTARGET(\callee) - vmovaps %xmm0, 32(%rsp) - vmovaps 16(%rsp), %xmm0 - vmovaps (%rsp), %xmm1 - call HIDDEN_JUMPTARGET(\callee) - vmovaps %xmm0, %xmm1 - vmovaps 32(%rsp), %xmm0 - vinsertf128 $1, %xmm1, %ymm0, %ymm0 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-32, %rsp + subq $64, %rsp + vextractf128 $1, %ymm0, 16(%rsp) + vextractf128 $1, %ymm1, (%rsp) + vzeroupper + call HIDDEN_JUMPTARGET(\callee) + vmovaps %xmm0, 32(%rsp) + vmovaps 16(%rsp), %xmm0 + vmovaps (%rsp), %xmm1 + call HIDDEN_JUMPTARGET(\callee) + vmovaps %xmm0, %xmm1 + vmovaps 32(%rsp), %xmm0 + vinsertf128 $1, %xmm1, %ymm0, %ymm0 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* 3 argument AVX/AVX2 ISA version as wrapper to SSE ISA version. */ .macro WRAPPER_IMPL_AVX_fFF callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-32, %rsp - pushq %r13 - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%r13, 0) - pushq %r14 - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%r14, 0) - subq $48, %rsp - movq %rsi, %r14 - movq %rdi, %r13 - vextractf128 $1, %ymm0, 32(%rsp) - vzeroupper - call HIDDEN_JUMPTARGET(\callee) - vmovaps 32(%rsp), %xmm0 - lea (%rsp), %rdi - lea 16(%rsp), %rsi - call HIDDEN_JUMPTARGET(\callee) - vmovapd (%rsp), %xmm0 - vmovapd 16(%rsp), %xmm1 - vmovapd %xmm0, 16(%r13) - vmovapd %xmm1, 16(%r14) - addq $48, %rsp - popq %r14 - cfi_adjust_cfa_offset (-8) - cfi_restore (%r14) - popq %r13 - cfi_adjust_cfa_offset (-8) - cfi_restore (%r13) - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-32, %rsp + pushq %r13 + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%r13, 0) + pushq %r14 + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%r14, 0) + subq $48, %rsp + movq %rsi, %r14 + movq %rdi, %r13 + vextractf128 $1, %ymm0, 32(%rsp) + vzeroupper + call HIDDEN_JUMPTARGET(\callee) + vmovaps 32(%rsp), %xmm0 + lea (%rsp), %rdi + lea 16(%rsp), %rsi + call HIDDEN_JUMPTARGET(\callee) + vmovapd (%rsp), %xmm0 + vmovapd 16(%rsp), %xmm1 + vmovapd %xmm0, 16(%r13) + vmovapd %xmm1, 16(%r14) + addq $48, %rsp + popq %r14 + cfi_adjust_cfa_offset (-8) + cfi_restore (%r14) + popq %r13 + cfi_adjust_cfa_offset (-8) + cfi_restore (%r13) + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* AVX512 ISA version as wrapper to AVX2 ISA version. */ .macro WRAPPER_IMPL_AVX512 callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-64, %rsp - subq $128, %rsp - vmovups %zmm0, (%rsp) - vmovupd (%rsp), %ymm0 - call HIDDEN_JUMPTARGET(\callee) - vmovupd %ymm0, 64(%rsp) - vmovupd 32(%rsp), %ymm0 - call HIDDEN_JUMPTARGET(\callee) - vmovupd %ymm0, 96(%rsp) - vmovups 64(%rsp), %zmm0 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-64, %rsp + subq $128, %rsp + vmovups %zmm0, (%rsp) + vmovupd (%rsp), %ymm0 + call HIDDEN_JUMPTARGET(\callee) + vmovupd %ymm0, 64(%rsp) + vmovupd 32(%rsp), %ymm0 + call HIDDEN_JUMPTARGET(\callee) + vmovupd %ymm0, 96(%rsp) + vmovups 64(%rsp), %zmm0 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* 2 argument AVX512 ISA version as wrapper to AVX2 ISA version. */ .macro WRAPPER_IMPL_AVX512_ff callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-64, %rsp - subq $192, %rsp - vmovups %zmm0, (%rsp) - vmovups %zmm1, 64(%rsp) - vmovupd (%rsp), %ymm0 - vmovupd 64(%rsp), %ymm1 - call HIDDEN_JUMPTARGET(\callee) - vmovupd %ymm0, 128(%rsp) - vmovupd 32(%rsp), %ymm0 - vmovupd 96(%rsp), %ymm1 - call HIDDEN_JUMPTARGET(\callee) - vmovupd %ymm0, 160(%rsp) - vmovups 128(%rsp), %zmm0 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-64, %rsp + subq $192, %rsp + vmovups %zmm0, (%rsp) + vmovups %zmm1, 64(%rsp) + vmovupd (%rsp), %ymm0 + vmovupd 64(%rsp), %ymm1 + call HIDDEN_JUMPTARGET(\callee) + vmovupd %ymm0, 128(%rsp) + vmovupd 32(%rsp), %ymm0 + vmovupd 96(%rsp), %ymm1 + call HIDDEN_JUMPTARGET(\callee) + vmovupd %ymm0, 160(%rsp) + vmovups 128(%rsp), %zmm0 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* 3 argument AVX512 ISA version as wrapper to AVX2 ISA version. */ .macro WRAPPER_IMPL_AVX512_fFF callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-64, %rsp - pushq %r12 - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%r12, 0) - pushq %r13 - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%r13, 0) - subq $176, %rsp - movq %rsi, %r13 - vmovups %zmm0, (%rsp) - movq %rdi, %r12 - vmovupd (%rsp), %ymm0 - call HIDDEN_JUMPTARGET(\callee) - vmovupd 32(%rsp), %ymm0 - lea 64(%rsp), %rdi - lea 96(%rsp), %rsi - call HIDDEN_JUMPTARGET(\callee) - vmovupd 64(%rsp), %ymm0 - vmovupd 96(%rsp), %ymm1 - vmovupd %ymm0, 32(%r12) - vmovupd %ymm1, 32(%r13) - vzeroupper - addq $176, %rsp - popq %r13 - cfi_adjust_cfa_offset (-8) - cfi_restore (%r13) - popq %r12 - cfi_adjust_cfa_offset (-8) - cfi_restore (%r12) - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-64, %rsp + pushq %r12 + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%r12, 0) + pushq %r13 + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%r13, 0) + subq $176, %rsp + movq %rsi, %r13 + vmovups %zmm0, (%rsp) + movq %rdi, %r12 + vmovupd (%rsp), %ymm0 + call HIDDEN_JUMPTARGET(\callee) + vmovupd 32(%rsp), %ymm0 + lea 64(%rsp), %rdi + lea 96(%rsp), %rsi + call HIDDEN_JUMPTARGET(\callee) + vmovupd 64(%rsp), %ymm0 + vmovupd 96(%rsp), %ymm1 + vmovupd %ymm0, 32(%r12) + vmovupd %ymm1, 32(%r13) + vzeroupper + addq $176, %rsp + popq %r13 + cfi_adjust_cfa_offset (-8) + cfi_restore (%r13) + popq %r12 + cfi_adjust_cfa_offset (-8) + cfi_restore (%r12) + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm diff --git a/sysdeps/x86_64/fpu/svml_s_wrapper_impl.h b/sysdeps/x86_64/fpu/svml_s_wrapper_impl.h index c23da7ec83..cecf6c8384 100644 --- a/sysdeps/x86_64/fpu/svml_s_wrapper_impl.h +++ b/sysdeps/x86_64/fpu/svml_s_wrapper_impl.h @@ -18,309 +18,309 @@ /* SSE2 ISA version as wrapper to scalar. */ .macro WRAPPER_IMPL_SSE2 callee - subq $40, %rsp - cfi_adjust_cfa_offset(40) - movaps %xmm0, (%rsp) - call JUMPTARGET(\callee) - movss %xmm0, 16(%rsp) - movss 4(%rsp), %xmm0 - call JUMPTARGET(\callee) - movss %xmm0, 20(%rsp) - movss 8(%rsp), %xmm0 - call JUMPTARGET(\callee) - movss %xmm0, 24(%rsp) - movss 12(%rsp), %xmm0 - call JUMPTARGET(\callee) - movss 16(%rsp), %xmm3 - movss 20(%rsp), %xmm2 - movss 24(%rsp), %xmm1 - movss %xmm0, 28(%rsp) - unpcklps %xmm1, %xmm3 - unpcklps %xmm0, %xmm2 - unpcklps %xmm2, %xmm3 - movaps %xmm3, %xmm0 - addq $40, %rsp - cfi_adjust_cfa_offset(-40) - ret + subq $40, %rsp + cfi_adjust_cfa_offset (40) + movaps %xmm0, (%rsp) + call JUMPTARGET(\callee) + movss %xmm0, 16(%rsp) + movss 4(%rsp), %xmm0 + call JUMPTARGET(\callee) + movss %xmm0, 20(%rsp) + movss 8(%rsp), %xmm0 + call JUMPTARGET(\callee) + movss %xmm0, 24(%rsp) + movss 12(%rsp), %xmm0 + call JUMPTARGET(\callee) + movss 16(%rsp), %xmm3 + movss 20(%rsp), %xmm2 + movss 24(%rsp), %xmm1 + movss %xmm0, 28(%rsp) + unpcklps %xmm1, %xmm3 + unpcklps %xmm0, %xmm2 + unpcklps %xmm2, %xmm3 + movaps %xmm3, %xmm0 + addq $40, %rsp + cfi_adjust_cfa_offset (-40) + ret .endm /* 2 argument SSE2 ISA version as wrapper to scalar. */ .macro WRAPPER_IMPL_SSE2_ff callee - subq $56, %rsp - cfi_adjust_cfa_offset(56) - movaps %xmm0, (%rsp) - movaps %xmm1, 16(%rsp) - call JUMPTARGET(\callee) - movss %xmm0, 32(%rsp) - movss 4(%rsp), %xmm0 - movss 20(%rsp), %xmm1 - call JUMPTARGET(\callee) - movss %xmm0, 36(%rsp) - movss 8(%rsp), %xmm0 - movss 24(%rsp), %xmm1 - call JUMPTARGET(\callee) - movss %xmm0, 40(%rsp) - movss 12(%rsp), %xmm0 - movss 28(%rsp), %xmm1 - call JUMPTARGET(\callee) - movss 32(%rsp), %xmm3 - movss 36(%rsp), %xmm2 - movss 40(%rsp), %xmm1 - movss %xmm0, 44(%rsp) - unpcklps %xmm1, %xmm3 - unpcklps %xmm0, %xmm2 - unpcklps %xmm2, %xmm3 - movaps %xmm3, %xmm0 - addq $56, %rsp - cfi_adjust_cfa_offset(-56) - ret + subq $56, %rsp + cfi_adjust_cfa_offset (56) + movaps %xmm0, (%rsp) + movaps %xmm1, 16(%rsp) + call JUMPTARGET(\callee) + movss %xmm0, 32(%rsp) + movss 4(%rsp), %xmm0 + movss 20(%rsp), %xmm1 + call JUMPTARGET(\callee) + movss %xmm0, 36(%rsp) + movss 8(%rsp), %xmm0 + movss 24(%rsp), %xmm1 + call JUMPTARGET(\callee) + movss %xmm0, 40(%rsp) + movss 12(%rsp), %xmm0 + movss 28(%rsp), %xmm1 + call JUMPTARGET(\callee) + movss 32(%rsp), %xmm3 + movss 36(%rsp), %xmm2 + movss 40(%rsp), %xmm1 + movss %xmm0, 44(%rsp) + unpcklps %xmm1, %xmm3 + unpcklps %xmm0, %xmm2 + unpcklps %xmm2, %xmm3 + movaps %xmm3, %xmm0 + addq $56, %rsp + cfi_adjust_cfa_offset (-56) + ret .endm /* 3 argument SSE2 ISA version as wrapper to scalar. */ .macro WRAPPER_IMPL_SSE2_fFF callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - pushq %rbx - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbx, 0) - movq %rdi, %rbp - movq %rsi, %rbx - subq $40, %rsp - cfi_adjust_cfa_offset(40) - leaq 24(%rsp), %rsi - leaq 28(%rsp), %rdi - movaps %xmm0, (%rsp) - call JUMPTARGET(\callee) - leaq 24(%rsp), %rsi - leaq 28(%rsp), %rdi - movss 28(%rsp), %xmm0 - movss %xmm0, 0(%rbp) - movaps (%rsp), %xmm1 - movss 24(%rsp), %xmm0 - movss %xmm0, (%rbx) - movaps %xmm1, %xmm0 - shufps $85, %xmm1, %xmm0 - call JUMPTARGET(\callee) - movss 28(%rsp), %xmm0 - leaq 24(%rsp), %rsi - movss %xmm0, 4(%rbp) - leaq 28(%rsp), %rdi - movaps (%rsp), %xmm1 - movss 24(%rsp), %xmm0 - movss %xmm0, 4(%rbx) - movaps %xmm1, %xmm0 - unpckhps %xmm1, %xmm0 - call JUMPTARGET(\callee) - movaps (%rsp), %xmm1 - leaq 24(%rsp), %rsi - leaq 28(%rsp), %rdi - movss 28(%rsp), %xmm0 - shufps $255, %xmm1, %xmm1 - movss %xmm0, 8(%rbp) - movss 24(%rsp), %xmm0 - movss %xmm0, 8(%rbx) - movaps %xmm1, %xmm0 - call JUMPTARGET(\callee) - movss 28(%rsp), %xmm0 - movss %xmm0, 12(%rbp) - movss 24(%rsp), %xmm0 - movss %xmm0, 12(%rbx) - addq $40, %rsp - cfi_adjust_cfa_offset(-40) - popq %rbx - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbx) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + pushq %rbx + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbx, 0) + movq %rdi, %rbp + movq %rsi, %rbx + subq $40, %rsp + cfi_adjust_cfa_offset (40) + leaq 24(%rsp), %rsi + leaq 28(%rsp), %rdi + movaps %xmm0, (%rsp) + call JUMPTARGET(\callee) + leaq 24(%rsp), %rsi + leaq 28(%rsp), %rdi + movss 28(%rsp), %xmm0 + movss %xmm0, 0(%rbp) + movaps (%rsp), %xmm1 + movss 24(%rsp), %xmm0 + movss %xmm0, (%rbx) + movaps %xmm1, %xmm0 + shufps $85, %xmm1, %xmm0 + call JUMPTARGET(\callee) + movss 28(%rsp), %xmm0 + leaq 24(%rsp), %rsi + movss %xmm0, 4(%rbp) + leaq 28(%rsp), %rdi + movaps (%rsp), %xmm1 + movss 24(%rsp), %xmm0 + movss %xmm0, 4(%rbx) + movaps %xmm1, %xmm0 + unpckhps %xmm1, %xmm0 + call JUMPTARGET(\callee) + movaps (%rsp), %xmm1 + leaq 24(%rsp), %rsi + leaq 28(%rsp), %rdi + movss 28(%rsp), %xmm0 + shufps $255, %xmm1, %xmm1 + movss %xmm0, 8(%rbp) + movss 24(%rsp), %xmm0 + movss %xmm0, 8(%rbx) + movaps %xmm1, %xmm0 + call JUMPTARGET(\callee) + movss 28(%rsp), %xmm0 + movss %xmm0, 12(%rbp) + movss 24(%rsp), %xmm0 + movss %xmm0, 12(%rbx) + addq $40, %rsp + cfi_adjust_cfa_offset (-40) + popq %rbx + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbx) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* AVX/AVX2 ISA version as wrapper to SSE ISA version. */ .macro WRAPPER_IMPL_AVX callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-32, %rsp - subq $32, %rsp - vextractf128 $1, %ymm0, (%rsp) - vzeroupper - call HIDDEN_JUMPTARGET(\callee) - vmovaps %xmm0, 16(%rsp) - vmovaps (%rsp), %xmm0 - call HIDDEN_JUMPTARGET(\callee) - vmovaps %xmm0, %xmm1 - vmovaps 16(%rsp), %xmm0 - vinsertf128 $1, %xmm1, %ymm0, %ymm0 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-32, %rsp + subq $32, %rsp + vextractf128 $1, %ymm0, (%rsp) + vzeroupper + call HIDDEN_JUMPTARGET(\callee) + vmovaps %xmm0, 16(%rsp) + vmovaps (%rsp), %xmm0 + call HIDDEN_JUMPTARGET(\callee) + vmovaps %xmm0, %xmm1 + vmovaps 16(%rsp), %xmm0 + vinsertf128 $1, %xmm1, %ymm0, %ymm0 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* 2 argument AVX/AVX2 ISA version as wrapper to SSE ISA version. */ .macro WRAPPER_IMPL_AVX_ff callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-32, %rsp - subq $64, %rsp - vextractf128 $1, %ymm0, 16(%rsp) - vextractf128 $1, %ymm1, (%rsp) - vzeroupper - call HIDDEN_JUMPTARGET(\callee) - vmovaps %xmm0, 32(%rsp) - vmovaps 16(%rsp), %xmm0 - vmovaps (%rsp), %xmm1 - call HIDDEN_JUMPTARGET(\callee) - vmovaps %xmm0, %xmm1 - vmovaps 32(%rsp), %xmm0 - vinsertf128 $1, %xmm1, %ymm0, %ymm0 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-32, %rsp + subq $64, %rsp + vextractf128 $1, %ymm0, 16(%rsp) + vextractf128 $1, %ymm1, (%rsp) + vzeroupper + call HIDDEN_JUMPTARGET(\callee) + vmovaps %xmm0, 32(%rsp) + vmovaps 16(%rsp), %xmm0 + vmovaps (%rsp), %xmm1 + call HIDDEN_JUMPTARGET(\callee) + vmovaps %xmm0, %xmm1 + vmovaps 32(%rsp), %xmm0 + vinsertf128 $1, %xmm1, %ymm0, %ymm0 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* 3 argument AVX/AVX2 ISA version as wrapper to SSE ISA version. */ .macro WRAPPER_IMPL_AVX_fFF callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-32, %rsp - pushq %r13 - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%r13, 0) - pushq %r14 - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%r14, 0) - subq $48, %rsp - movq %rsi, %r14 - vmovaps %ymm0, (%rsp) - movq %rdi, %r13 - vmovaps 16(%rsp), %xmm1 - vmovaps %xmm1, 32(%rsp) - vzeroupper - vmovaps (%rsp), %xmm0 - call HIDDEN_JUMPTARGET(\callee) - vmovaps 32(%rsp), %xmm0 - lea (%rsp), %rdi - lea 16(%rsp), %rsi - call HIDDEN_JUMPTARGET(\callee) - vmovaps (%rsp), %xmm0 - vmovaps 16(%rsp), %xmm1 - vmovaps %xmm0, 16(%r13) - vmovaps %xmm1, 16(%r14) - addq $48, %rsp - popq %r14 - cfi_adjust_cfa_offset (-8) - cfi_restore (%r14) - popq %r13 - cfi_adjust_cfa_offset (-8) - cfi_restore (%r13) - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-32, %rsp + pushq %r13 + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%r13, 0) + pushq %r14 + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%r14, 0) + subq $48, %rsp + movq %rsi, %r14 + vmovaps %ymm0, (%rsp) + movq %rdi, %r13 + vmovaps 16(%rsp), %xmm1 + vmovaps %xmm1, 32(%rsp) + vzeroupper + vmovaps (%rsp), %xmm0 + call HIDDEN_JUMPTARGET(\callee) + vmovaps 32(%rsp), %xmm0 + lea (%rsp), %rdi + lea 16(%rsp), %rsi + call HIDDEN_JUMPTARGET(\callee) + vmovaps (%rsp), %xmm0 + vmovaps 16(%rsp), %xmm1 + vmovaps %xmm0, 16(%r13) + vmovaps %xmm1, 16(%r14) + addq $48, %rsp + popq %r14 + cfi_adjust_cfa_offset (-8) + cfi_restore (%r14) + popq %r13 + cfi_adjust_cfa_offset (-8) + cfi_restore (%r13) + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* AVX512 ISA version as wrapper to AVX2 ISA version. */ .macro WRAPPER_IMPL_AVX512 callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-64, %rsp - subq $128, %rsp - vmovups %zmm0, (%rsp) - vmovupd (%rsp), %ymm0 - call HIDDEN_JUMPTARGET(\callee) - vmovupd %ymm0, 64(%rsp) - vmovupd 32(%rsp), %ymm0 - call HIDDEN_JUMPTARGET(\callee) - vmovupd %ymm0, 96(%rsp) - vmovups 64(%rsp), %zmm0 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-64, %rsp + subq $128, %rsp + vmovups %zmm0, (%rsp) + vmovupd (%rsp), %ymm0 + call HIDDEN_JUMPTARGET(\callee) + vmovupd %ymm0, 64(%rsp) + vmovupd 32(%rsp), %ymm0 + call HIDDEN_JUMPTARGET(\callee) + vmovupd %ymm0, 96(%rsp) + vmovups 64(%rsp), %zmm0 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* 2 argument AVX512 ISA version as wrapper to AVX2 ISA version. */ .macro WRAPPER_IMPL_AVX512_ff callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-64, %rsp - subq $192, %rsp - vmovups %zmm0, (%rsp) - vmovups %zmm1, 64(%rsp) - vmovups (%rsp), %ymm0 - vmovups 64(%rsp), %ymm1 - call HIDDEN_JUMPTARGET(\callee) - vmovups %ymm0, 128(%rsp) - vmovups 32(%rsp), %ymm0 - vmovups 96(%rsp), %ymm1 - call HIDDEN_JUMPTARGET(\callee) - vmovups %ymm0, 160(%rsp) - vmovups 128(%rsp), %zmm0 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-64, %rsp + subq $192, %rsp + vmovups %zmm0, (%rsp) + vmovups %zmm1, 64(%rsp) + vmovups (%rsp), %ymm0 + vmovups 64(%rsp), %ymm1 + call HIDDEN_JUMPTARGET(\callee) + vmovups %ymm0, 128(%rsp) + vmovups 32(%rsp), %ymm0 + vmovups 96(%rsp), %ymm1 + call HIDDEN_JUMPTARGET(\callee) + vmovups %ymm0, 160(%rsp) + vmovups 128(%rsp), %zmm0 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm /* 3 argument AVX512 ISA version as wrapper to AVX2 ISA version. */ .macro WRAPPER_IMPL_AVX512_fFF callee - pushq %rbp - cfi_adjust_cfa_offset (8) - cfi_rel_offset (%rbp, 0) - movq %rsp, %rbp - cfi_def_cfa_register (%rbp) - andq $-64, %rsp - pushq %r12 - pushq %r13 - subq $176, %rsp - movq %rsi, %r13 - vmovaps %zmm0, (%rsp) - movq %rdi, %r12 - vmovaps (%rsp), %ymm0 - call HIDDEN_JUMPTARGET(\callee) - vmovaps 32(%rsp), %ymm0 - lea 64(%rsp), %rdi - lea 96(%rsp), %rsi - call HIDDEN_JUMPTARGET(\callee) - vmovaps 64(%rsp), %ymm0 - vmovaps 96(%rsp), %ymm1 - vmovaps %ymm0, 32(%r12) - vmovaps %ymm1, 32(%r13) - addq $176, %rsp - popq %r13 - popq %r12 - movq %rbp, %rsp - cfi_def_cfa_register (%rsp) - popq %rbp - cfi_adjust_cfa_offset (-8) - cfi_restore (%rbp) - ret + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-64, %rsp + pushq %r12 + pushq %r13 + subq $176, %rsp + movq %rsi, %r13 + vmovaps %zmm0, (%rsp) + movq %rdi, %r12 + vmovaps (%rsp), %ymm0 + call HIDDEN_JUMPTARGET(\callee) + vmovaps 32(%rsp), %ymm0 + lea 64(%rsp), %rdi + lea 96(%rsp), %rsi + call HIDDEN_JUMPTARGET(\callee) + vmovaps 64(%rsp), %ymm0 + vmovaps 96(%rsp), %ymm1 + vmovaps %ymm0, 32(%r12) + vmovaps %ymm1, 32(%r13) + addq $176, %rsp + popq %r13 + popq %r12 + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret .endm -- 2.34.1