From patchwork Tue Mar 3 16:06:51 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Richard Earnshaw \(lists\)" X-Patchwork-Id: 184141 Delivered-To: patch@linaro.org Received: by 2002:a92:1f12:0:0:0:0:0 with SMTP id i18csp3830776ile; Tue, 3 Mar 2020 08:07:14 -0800 (PST) X-Google-Smtp-Source: ADFU+vuGvlyQgbK7Z7ACmf4eGcsRP7HFvi1e1QbLzNS5JrSXkJ9kYszUcJ1hEHwV9MHdwoYtxQol X-Received: by 2002:a9d:12b4:: with SMTP id g49mr515673otg.50.1583251634647; Tue, 03 Mar 2020 08:07:14 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1583251634; cv=none; d=google.com; s=arc-20160816; b=miVH2IXMzztjpuGGLHpo6zUFNGcviwYz17pcZA+aGiPwpgX3nPpQJYMc3w3IsXrlgx eoPPkq4Z7KuzzDstMmagwM1fECPW22vBnckl8/dRAV3tVVNZuO840vNKI6WqRhk6v9dF OVIf3HBefKj45oQMWur4dhZBcuV8EqSYNL6AM7TAgZSbUKXZpnQ7EvPUIHjQNO5UX/c/ 2b+Z40iiPK6Fa9vPVLAteQByN6TZ9OTKYzyIK5WQmhGi2du6UcwJm3TNohuaQdb64fun in5WsCUyam/KiK7JisxNtZ8JA0HwgDuOee0EFKfqHzMudnur06QzqYTFHC9V23NiwAYZ aQkw== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=mime-version:user-agent:date:message-id:subject:from:to :delivered-to:sender:list-help:list-post:list-archive :list-unsubscribe:list-id:precedence:mailing-list:dkim-signature :domainkey-signature; bh=KjJt2hKIG2X+Xa/+Ov8xQWJdBcva9VZhLcxyMrfa8OE=; b=EZP4wFPAPZK8IFzJpCx7IYWcq2oYhE7w4O381zr6FSV5NyuoCl0LvDoXqXtslXy6Us wFWT+ifS4m5m3wnyazsv0IWQWe3XYTpJxdHUeUluQ7hmo8yprwQ+gf493nThY0pdB0+w IJF1Ln+5kCgbNuPVdYXycKT1FQ5gSdqrTEqbJz9mSUilM39Vn3c98tkywcN82CtiF9EM /HkYnf2HEhx3Dh6Xiw5SlHqTgBNY1/Li/gVC21od4NG7MFESaHYtCBdZvlLVFCC2Txmb aXCWHMjb3vvZvXJ3aXxWm8K36pKQiqBR1mykTZO1/mbSZnLRm1sEiwbru5cgxslCYhJm PNew== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@gcc.gnu.org header.s=default header.b=wv5ijbDX; spf=pass (google.com: domain of gcc-patches-return-520540-patch=linaro.org@gcc.gnu.org designates 209.132.180.131 as permitted sender) smtp.mailfrom="gcc-patches-return-520540-patch=linaro.org@gcc.gnu.org" Return-Path: Received: from sourceware.org (server1.sourceware.org. [209.132.180.131]) by mx.google.com with ESMTPS id j74si9939346otj.246.2020.03.03.08.07.14 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Tue, 03 Mar 2020 08:07:14 -0800 (PST) Received-SPF: pass (google.com: domain of gcc-patches-return-520540-patch=linaro.org@gcc.gnu.org designates 209.132.180.131 as permitted sender) client-ip=209.132.180.131; Authentication-Results: mx.google.com; dkim=pass header.i=@gcc.gnu.org header.s=default header.b=wv5ijbDX; spf=pass (google.com: domain of gcc-patches-return-520540-patch=linaro.org@gcc.gnu.org designates 209.132.180.131 as permitted sender) smtp.mailfrom="gcc-patches-return-520540-patch=linaro.org@gcc.gnu.org" DomainKey-Signature: a=rsa-sha1; c=nofws; d=gcc.gnu.org; h=list-id :list-unsubscribe:list-archive:list-post:list-help:sender:to :from:subject:message-id:date:mime-version:content-type; q=dns; s=default; b=UFhFTJ65/31ep32GzGJu7EEWDq+dDoWdPnym5v65uAKWVdYW36 jMBinyTVCREaJ/IgB1AHPSq+I3ip/8sbnd2mOIjYYjRAo+ys6PjWxPDZhTPK0wdq 4hvFG4czQY0iKx+uuZbM+nVHCRGr1ZW2fLoOhLjsQIBa+aGvRo0X1eP7E= DKIM-Signature: v=1; a=rsa-sha1; c=relaxed; d=gcc.gnu.org; h=list-id :list-unsubscribe:list-archive:list-post:list-help:sender:to :from:subject:message-id:date:mime-version:content-type; s= default; bh=TQcadadIDTq/7OoAg+4dgckBZVY=; b=wv5ijbDX3TTb8zgnRulb 3KqqPB+ivAwDdRYifY8AloK1LevdyoRlwft8h+UA+VIuNmWnjNVlyg7nyrSF4E8f rBrhFcir4AR/ytxiBU8sTQRrgneSlRzdkOiZn3HJa5FU9nUx3DyFm1dhA+uBQR0s VuTQewWsrf3Qb16rSQ5r2bk= Received: (qmail 121247 invoked by alias); 3 Mar 2020 16:06:59 -0000 Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org Delivered-To: mailing list gcc-patches@gcc.gnu.org Received: (qmail 121238 invoked by uid 89); 3 Mar 2020 16:06:59 -0000 Authentication-Results: sourceware.org; auth=none X-Spam-SWARE-Status: No, score=-16.4 required=5.0 tests=AWL, BAYES_00, GIT_PATCH_0, GIT_PATCH_1, GIT_PATCH_2, GIT_PATCH_3, KAM_SHORT, SCC_10_SHORT_WORD_LINES, SCC_20_SHORT_WORD_LINES, SCC_35_SHORT_WORD_LINES, SCC_5_SHORT_WORD_LINES, SPF_PASS autolearn=ham version=3.3.1 spammy=1968, Body, Everything, danger X-HELO: foss.arm.com Received: from foss.arm.com (HELO foss.arm.com) (217.140.110.172) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Tue, 03 Mar 2020 16:06:55 +0000 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 4DD332F; Tue, 3 Mar 2020 08:06:53 -0800 (PST) Received: from e120077-lin.cambridge.arm.com (e120077-lin.cambridge.arm.com [10.2.78.81]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id D2D0D3F6C4; Tue, 3 Mar 2020 08:06:52 -0800 (PST) To: "gcc-patches@gcc.gnu.org" From: "Richard Earnshaw (lists)" Subject: [committed] libgcc: arm: convert thumb1 code to unified syntax Message-ID: <36d699ca-c555-8b20-08d0-a9328ec25771@arm.com> Date: Tue, 3 Mar 2020 16:06:51 +0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.9.0 MIME-Version: 1.0 Unified syntax has been the official syntax for thumb1 assembly for over 10 years now. It's time we made preparations for that becoming the default in the assembler. But before we can start doing that we really need to clean up some laggards from the olden days. Libgcc support for thumb1 is one such example. This patch converts all of the legacy (disjoint) syntax that I could find over to unified code. The identification was done by using a trick version of gas that defaulted to unified mode which then faults if legacy syntax is encountered. The code produced was then compared against the old code to check for differences. One such difference does exist, but that is because in unified syntax 'movs rd, rn' is encoded as 'lsls rd, rn, #0', rather than 'adds rd, rn, #0'; but that is a deliberate change that was introduced because the lsls encoding more closely reflects the behaviour of 'movs' in arm state (where only some of the condition flags are modified). libgcc: * config/arm/bpabi-v6m.S (aeabi_lcmp): Convert thumb1 code to unified syntax. (aeabi_ulcmp, aeabi_ldivmod, aeabi_uldivmod): Likewise. (aeabi_frsub, aeabi_cfcmpeq, aeabi_fcmpeq): Likewise. (aeabi_fcmp, aeabi_drsub, aeabi_cdrcmple): Likewise. (aeabi_cdcmpeq, aeabi_dcmpeq, aeabi_dcmp): Likewise. * config/arm/lib1funcs.S (Lend_fde): Convert thumb1 code to unified syntax. (divsi3, modsi3): Likewise. (clzdi2, ctzsi2): Likewise. * config/arm/libunwind.S (restore_core_regs): Convert thumb1 code to unified syntax. (UNWIND_WRAPPER): Likewise. Committed to trunk. R. diff --git a/libgcc/config/arm/bpabi-v6m.S b/libgcc/config/arm/bpabi-v6m.S index 29fe8faa6e6..1a403efc872 100644 --- a/libgcc/config/arm/bpabi-v6m.S +++ b/libgcc/config/arm/bpabi-v6m.S @@ -39,21 +39,21 @@ FUNC_START aeabi_lcmp cmp xxh, yyh beq 1f bgt 2f - mov r0, #1 - neg r0, r0 + movs r0, #1 + negs r0, r0 RET 2: - mov r0, #1 + movs r0, #1 RET 1: - sub r0, xxl, yyl + subs r0, xxl, yyl beq 1f bhi 2f - mov r0, #1 - neg r0, r0 + movs r0, #1 + negs r0, r0 RET 2: - mov r0, #1 + movs r0, #1 1: RET FUNC_END aeabi_lcmp @@ -65,15 +65,15 @@ FUNC_START aeabi_lcmp FUNC_START aeabi_ulcmp cmp xxh, yyh bne 1f - sub r0, xxl, yyl + subs r0, xxl, yyl beq 2f 1: bcs 1f - mov r0, #1 - neg r0, r0 + movs r0, #1 + negs r0, r0 RET 1: - mov r0, #1 + movs r0, #1 2: RET FUNC_END aeabi_ulcmp @@ -91,29 +91,29 @@ FUNC_START aeabi_ulcmp cmp xxl, #0 2: beq 3f - mov xxh, #0 - mvn xxh, xxh @ 0xffffffff - mov xxl, xxh + movs xxh, #0 + mvns xxh, xxh @ 0xffffffff + movs xxl, xxh 3: .else blt 6f bgt 4f cmp xxl, #0 beq 5f -4: mov xxl, #0 - mvn xxl, xxl @ 0xffffffff - lsr xxh, xxl, #1 @ 0x7fffffff +4: movs xxl, #0 + mvns xxl, xxl @ 0xffffffff + lsrs xxh, xxl, #1 @ 0x7fffffff b 5f -6: mov xxh, #0x80 - lsl xxh, xxh, #24 @ 0x80000000 - mov xxl, #0 +6: movs xxh, #0x80 + lsls xxh, xxh, #24 @ 0x80000000 + movs xxl, #0 5: .endif @ tailcalls are tricky on v6-m. push {r0, r1, r2} ldr r0, 1f adr r1, 1f - add r0, r1 + adds r0, r1 str r0, [sp, #8] @ We know we are not on armv4t, so pop pc is safe. pop {r0, r1, pc} @@ -128,15 +128,15 @@ FUNC_START aeabi_ulcmp FUNC_START aeabi_ldivmod test_div_by_zero signed - push {r0, r1} - mov r0, sp - push {r0, lr} - ldr r0, [sp, #8] - bl SYM(__gnu_ldivmod_helper) - ldr r3, [sp, #4] - mov lr, r3 - add sp, sp, #8 - pop {r2, r3} + push {r0, r1} + mov r0, sp + push {r0, lr} + ldr r0, [sp, #8] + bl SYM(__gnu_ldivmod_helper) + ldr r3, [sp, #4] + mov lr, r3 + add sp, sp, #8 + pop {r2, r3} RET FUNC_END aeabi_ldivmod @@ -147,15 +147,15 @@ FUNC_START aeabi_ldivmod FUNC_START aeabi_uldivmod test_div_by_zero unsigned - push {r0, r1} - mov r0, sp - push {r0, lr} - ldr r0, [sp, #8] - bl SYM(__udivmoddi4) - ldr r3, [sp, #4] - mov lr, r3 - add sp, sp, #8 - pop {r2, r3} + push {r0, r1} + mov r0, sp + push {r0, lr} + ldr r0, [sp, #8] + bl SYM(__udivmoddi4) + ldr r3, [sp, #4] + mov lr, r3 + add sp, sp, #8 + pop {r2, r3} RET FUNC_END aeabi_uldivmod @@ -166,9 +166,9 @@ FUNC_START aeabi_uldivmod FUNC_START aeabi_frsub push {r4, lr} - mov r4, #1 - lsl r4, #31 - eor r0, r0, r4 + movs r4, #1 + lsls r4, #31 + eors r0, r0, r4 bl __aeabi_fadd pop {r4, pc} @@ -181,7 +181,7 @@ FUNC_START aeabi_frsub FUNC_START aeabi_cfrcmple mov ip, r0 - mov r0, r1 + movs r0, r1 mov r1, ip b 6f @@ -196,8 +196,8 @@ FUNC_ALIAS aeabi_cfcmple aeabi_cfcmpeq cmp r0, #0 @ Clear the C flag if the return value was -1, indicating @ that the first operand was smaller than the second. - bmi 1f - mov r1, #0 + bmi 1f + movs r1, #0 cmn r0, r1 1: pop {r0, r1, r2, r3, r4, pc} @@ -210,8 +210,8 @@ FUNC_START aeabi_fcmpeq push {r4, lr} bl __eqsf2 - neg r0, r0 - add r0, r0, #1 + negs r0, r0 + adds r0, r0, #1 pop {r4, pc} FUNC_END aeabi_fcmpeq @@ -223,10 +223,10 @@ FUNC_START aeabi_fcmp\cond bl __\helper\mode cmp r0, #0 b\cond 1f - mov r0, #0 + movs r0, #0 pop {r4, pc} 1: - mov r0, #1 + movs r0, #1 pop {r4, pc} FUNC_END aeabi_fcmp\cond @@ -244,9 +244,9 @@ COMPARISON ge, ge FUNC_START aeabi_drsub push {r4, lr} - mov r4, #1 - lsl r4, #31 - eor xxh, xxh, r4 + movs r4, #1 + lsls r4, #31 + eors xxh, xxh, r4 bl __aeabi_dadd pop {r4, pc} @@ -259,10 +259,10 @@ FUNC_START aeabi_drsub FUNC_START aeabi_cdrcmple mov ip, r0 - mov r0, r2 + movs r0, r2 mov r2, ip mov ip, r1 - mov r1, r3 + movs r1, r3 mov r3, ip b 6f @@ -277,8 +277,8 @@ FUNC_ALIAS aeabi_cdcmple aeabi_cdcmpeq cmp r0, #0 @ Clear the C flag if the return value was -1, indicating @ that the first operand was smaller than the second. - bmi 1f - mov r1, #0 + bmi 1f + movs r1, #0 cmn r0, r1 1: pop {r0, r1, r2, r3, r4, pc} @@ -291,8 +291,8 @@ FUNC_START aeabi_dcmpeq push {r4, lr} bl __eqdf2 - neg r0, r0 - add r0, r0, #1 + negs r0, r0 + adds r0, r0, #1 pop {r4, pc} FUNC_END aeabi_dcmpeq @@ -304,10 +304,10 @@ FUNC_START aeabi_dcmp\cond bl __\helper\mode cmp r0, #0 b\cond 1f - mov r0, #0 + movs r0, #0 pop {r4, pc} 1: - mov r0, #1 + movs r0, #1 pop {r4, pc} FUNC_END aeabi_dcmp\cond diff --git a/libgcc/config/arm/lib1funcs.S b/libgcc/config/arm/lib1funcs.S index e4b73caabbd..e8d2158f8d6 100644 --- a/libgcc/config/arm/lib1funcs.S +++ b/libgcc/config/arm/lib1funcs.S @@ -22,6 +22,10 @@ a copy of the GCC Runtime Library Exception along with this program; see the files COPYING3 and COPYING.RUNTIME respectively. If not, see . */ +/* Everything in this file should now use unified syntax. */ + + .syntax unified + /* An executable stack is *not* required for these functions. */ #if defined(__ELF__) && defined(__linux__) .section .note.GNU-stack,"",%progbits @@ -270,7 +274,7 @@ LSYM(Lend_fde): #ifdef NOT_ISA_TARGET_32BIT push {r0, lr} - mov r0, #0 + movs r0, #0 bl SYM(__aeabi_idiv0) @ We know we are not on armv4t, so pop pc is safe. pop {r1, pc} @@ -310,7 +314,7 @@ LSYM(Lend_fde): push { r1, lr } 98: cfi_push 98b - __\name, 0xe, -0x4, 0x8 bl SYM (__div0) - mov r0, #0 @ About as wrong as it could be. + movs r0, #0 @ About as wrong as it could be. #if defined (__INTERWORKING__) pop { r1, r2 } bx r2 @@ -349,7 +353,7 @@ SYM (\name): #define THUMB_FUNC .thumb_func #define THUMB_CODE .force_thumb # if defined(__thumb2__) -#define THUMB_SYNTAX .syntax divided +#define THUMB_SYNTAX # else #define THUMB_SYNTAX # endif @@ -725,8 +729,8 @@ pc .req r15 /* ------------------------------------------------------------------------ */ .macro THUMB_DIV_MOD_BODY modulo @ Load the constant 0x10000000 into our work register. - mov work, #1 - lsl work, #28 + movs work, #1 + lsls work, #28 LSYM(Loop1): @ Unless the divisor is very big, shift it up in multiples of @ four bits, since this is the amount of unwinding in the main @@ -736,12 +740,12 @@ LSYM(Loop1): bhs LSYM(Lbignum) cmp divisor, dividend bhs LSYM(Lbignum) - lsl divisor, #4 - lsl curbit, #4 + lsls divisor, #4 + lsls curbit, #4 b LSYM(Loop1) LSYM(Lbignum): @ Set work to 0x80000000 - lsl work, #3 + lsls work, #3 LSYM(Loop2): @ For very big divisors, we must shift it a bit at a time, or @ we will be in danger of overflowing. @@ -749,8 +753,8 @@ LSYM(Loop2): bhs LSYM(Loop3) cmp divisor, dividend bhs LSYM(Loop3) - lsl divisor, #1 - lsl curbit, #1 + lsls divisor, #1 + lsls curbit, #1 b LSYM(Loop2) LSYM(Loop3): @ Test for possible subtractions ... @@ -758,39 +762,39 @@ LSYM(Loop3): @ ... On the final pass, this may subtract too much from the dividend, @ so keep track of which subtractions are done, we can fix them up @ afterwards. - mov overdone, #0 + movs overdone, #0 cmp dividend, divisor blo LSYM(Lover1) - sub dividend, dividend, divisor + subs dividend, dividend, divisor LSYM(Lover1): - lsr work, divisor, #1 + lsrs work, divisor, #1 cmp dividend, work blo LSYM(Lover2) - sub dividend, dividend, work + subs dividend, dividend, work mov ip, curbit - mov work, #1 - ror curbit, work - orr overdone, curbit + movs work, #1 + rors curbit, work + orrs overdone, curbit mov curbit, ip LSYM(Lover2): - lsr work, divisor, #2 + lsrs work, divisor, #2 cmp dividend, work blo LSYM(Lover3) - sub dividend, dividend, work + subs dividend, dividend, work mov ip, curbit - mov work, #2 - ror curbit, work - orr overdone, curbit + movs work, #2 + rors curbit, work + orrs overdone, curbit mov curbit, ip LSYM(Lover3): - lsr work, divisor, #3 + lsrs work, divisor, #3 cmp dividend, work blo LSYM(Lover4) - sub dividend, dividend, work + subs dividend, dividend, work mov ip, curbit - mov work, #3 - ror curbit, work - orr overdone, curbit + movs work, #3 + rors curbit, work + orrs overdone, curbit mov curbit, ip LSYM(Lover4): mov ip, curbit @@ -800,46 +804,46 @@ LSYM(Lover4): @ since the "bit" will have been shifted out at the bottom. cmp dividend, divisor blo LSYM(Lover1) - sub dividend, dividend, divisor - orr result, result, curbit + subs dividend, dividend, divisor + orrs result, result, curbit LSYM(Lover1): - lsr work, divisor, #1 + lsrs work, divisor, #1 cmp dividend, work blo LSYM(Lover2) - sub dividend, dividend, work - lsr work, curbit, #1 - orr result, work + subs dividend, dividend, work + lsrs work, curbit, #1 + orrs result, work LSYM(Lover2): - lsr work, divisor, #2 + lsrs work, divisor, #2 cmp dividend, work blo LSYM(Lover3) - sub dividend, dividend, work - lsr work, curbit, #2 - orr result, work + subs dividend, dividend, work + lsrs work, curbit, #2 + orrs result, work LSYM(Lover3): - lsr work, divisor, #3 + lsrs work, divisor, #3 cmp dividend, work blo LSYM(Lover4) - sub dividend, dividend, work - lsr work, curbit, #3 - orr result, work + subs dividend, dividend, work + lsrs work, curbit, #3 + orrs result, work LSYM(Lover4): .endif cmp dividend, #0 @ Early termination? beq LSYM(Lover5) - lsr curbit, #4 @ No, any more bits to do? + lsrs curbit, #4 @ No, any more bits to do? beq LSYM(Lover5) - lsr divisor, #4 + lsrs divisor, #4 b LSYM(Loop3) LSYM(Lover5): .if \modulo @ Any subtractions that we should not have done will be recorded in @ the top three bits of "overdone". Exactly which were not needed @ are governed by the position of the bit, stored in ip. - mov work, #0xe - lsl work, #28 - and overdone, work + movs work, #0xe + lsls work, #28 + ands overdone, work beq LSYM(Lgot_result) @ If we terminated early, because dividend became zero, then the @@ -849,33 +853,33 @@ LSYM(Lover5): @ the bit in ip could be in the top two bits which might then match @ with one of the smaller RORs. mov curbit, ip - mov work, #0x7 + movs work, #0x7 tst curbit, work beq LSYM(Lgot_result) mov curbit, ip - mov work, #3 - ror curbit, work + movs work, #3 + rors curbit, work tst overdone, curbit beq LSYM(Lover6) - lsr work, divisor, #3 - add dividend, work + lsrs work, divisor, #3 + adds dividend, work LSYM(Lover6): mov curbit, ip - mov work, #2 - ror curbit, work + movs work, #2 + rors curbit, work tst overdone, curbit beq LSYM(Lover7) - lsr work, divisor, #2 - add dividend, work + lsrs work, divisor, #2 + adds dividend, work LSYM(Lover7): mov curbit, ip - mov work, #1 - ror curbit, work + movs work, #1 + rors curbit, work tst overdone, curbit beq LSYM(Lgot_result) - lsr work, divisor, #1 - add dividend, work + lsrs work, divisor, #1 + adds dividend, work .endif LSYM(Lgot_result): .endm @@ -885,7 +889,7 @@ LSYM(Lgot_result): /* Branch to div(n), and jump to label if curbit is lo than divisior. */ .macro BranchToDiv n, label - lsr curbit, dividend, \n + lsrs curbit, dividend, \n cmp curbit, divisor blo \label .endm @@ -893,13 +897,13 @@ LSYM(Lgot_result): /* Body of div(n). Shift the divisor in n bits and compare the divisor and dividend. Update the dividend as the substruction result. */ .macro DoDiv n - lsr curbit, dividend, \n + lsrs curbit, dividend, \n cmp curbit, divisor bcc 1f - lsl curbit, divisor, \n - sub dividend, dividend, curbit + lsls curbit, divisor, \n + subs dividend, dividend, curbit -1: adc result, result +1: adcs result, result .endm /* The body of division with positive divisor. Unless the divisor is very @@ -907,29 +911,29 @@ LSYM(Lgot_result): unwinding in the main division loop. Continue shifting until the divisor is larger than the dividend. */ .macro THUMB1_Div_Positive - mov result, #0 + movs result, #0 BranchToDiv #1, LSYM(Lthumb1_div1) BranchToDiv #4, LSYM(Lthumb1_div4) BranchToDiv #8, LSYM(Lthumb1_div8) BranchToDiv #12, LSYM(Lthumb1_div12) BranchToDiv #16, LSYM(Lthumb1_div16) LSYM(Lthumb1_div_large_positive): - mov result, #0xff - lsl divisor, divisor, #8 + movs result, #0xff + lsls divisor, divisor, #8 rev result, result - lsr curbit, dividend, #16 + lsrs curbit, dividend, #16 cmp curbit, divisor blo 1f - asr result, #8 - lsl divisor, divisor, #8 + asrs result, #8 + lsls divisor, divisor, #8 beq LSYM(Ldivbyzero_waypoint) -1: lsr curbit, dividend, #12 +1: lsrs curbit, dividend, #12 cmp curbit, divisor blo LSYM(Lthumb1_div12) b LSYM(Lthumb1_div16) LSYM(Lthumb1_div_loop): - lsr divisor, divisor, #8 + lsrs divisor, divisor, #8 LSYM(Lthumb1_div16): Dodiv #15 Dodiv #14 @@ -954,11 +958,11 @@ LSYM(Lthumb1_div3): LSYM(Lthumb1_div2): Dodiv #1 LSYM(Lthumb1_div1): - sub divisor, dividend, divisor + subs divisor, dividend, divisor bcs 1f cpy divisor, dividend -1: adc result, result +1: adcs result, result cpy dividend, result RET @@ -970,43 +974,43 @@ LSYM(Ldivbyzero_waypoint): THUMB1_Div_Positive except that the shift steps are in multiples of six bits. */ .macro THUMB1_Div_Negative - lsr result, divisor, #31 + lsrs result, divisor, #31 beq 1f - neg divisor, divisor + negs divisor, divisor -1: asr curbit, dividend, #32 +1: asrs curbit, dividend, #32 bcc 2f - neg dividend, dividend + negs dividend, dividend -2: eor curbit, result - mov result, #0 +2: eors curbit, result + movs result, #0 cpy ip, curbit BranchToDiv #4, LSYM(Lthumb1_div_negative4) BranchToDiv #8, LSYM(Lthumb1_div_negative8) LSYM(Lthumb1_div_large): - mov result, #0xfc - lsl divisor, divisor, #6 + movs result, #0xfc + lsls divisor, divisor, #6 rev result, result - lsr curbit, dividend, #8 + lsrs curbit, dividend, #8 cmp curbit, divisor blo LSYM(Lthumb1_div_negative8) - lsl divisor, divisor, #6 - asr result, result, #6 + lsls divisor, divisor, #6 + asrs result, result, #6 cmp curbit, divisor blo LSYM(Lthumb1_div_negative8) - lsl divisor, divisor, #6 - asr result, result, #6 + lsls divisor, divisor, #6 + asrs result, result, #6 cmp curbit, divisor blo LSYM(Lthumb1_div_negative8) - lsl divisor, divisor, #6 + lsls divisor, divisor, #6 beq LSYM(Ldivbyzero_negative) - asr result, result, #6 + asrs result, result, #6 b LSYM(Lthumb1_div_negative8) LSYM(Lthumb1_div_negative_loop): - lsr divisor, divisor, #6 + lsrs divisor, divisor, #6 LSYM(Lthumb1_div_negative8): DoDiv #7 DoDiv #6 @@ -1017,28 +1021,28 @@ LSYM(Lthumb1_div_negative4): DoDiv #2 bcs LSYM(Lthumb1_div_negative_loop) DoDiv #1 - sub divisor, dividend, divisor + subs divisor, dividend, divisor bcs 1f cpy divisor, dividend 1: cpy curbit, ip - adc result, result - asr curbit, curbit, #1 + adcs result, result + asrs curbit, curbit, #1 cpy dividend, result bcc 2f - neg dividend, dividend + negs dividend, dividend cmp curbit, #0 2: bpl 3f - neg divisor, divisor + negs divisor, divisor 3: RET LSYM(Ldivbyzero_negative): cpy curbit, ip - asr curbit, curbit, #1 + asrs curbit, curbit, #1 bcc LSYM(Ldiv0) - neg dividend, dividend + negs dividend, dividend .endm #endif /* ARM Thumb version. */ @@ -1056,8 +1060,8 @@ LSYM(Ldivbyzero_negative): cmp divisor, #0 beq LSYM(Ldiv0) LSYM(udivsi3_skip_div0_test): - mov curbit, #1 - mov result, #0 + movs curbit, #1 + movs result, #0 push { work } cmp dividend, divisor @@ -1065,7 +1069,7 @@ LSYM(udivsi3_skip_div0_test): THUMB_DIV_MOD_BODY 0 - mov r0, result + movs r0, result pop { work } RET @@ -1184,7 +1188,7 @@ ARM_FUNC_START aeabi_uidivmod cmp divisor, #0 beq LSYM(Ldiv0) - mov curbit, #1 + movs curbit, #1 cmp dividend, divisor bhs LSYM(Lover10) RET @@ -1263,7 +1267,7 @@ LSYM(Lover12): #else LSYM(divsi3_skip_div0_test): cpy curbit, dividend - orr curbit, divisor + orrs curbit, divisor bmi LSYM(Lthumb1_div_negative) LSYM(Lthumb1_div_positive): @@ -1395,11 +1399,11 @@ ARM_FUNC_START aeabi_idivmod FUNC_START modsi3 - mov curbit, #1 + movs curbit, #1 cmp divisor, #0 beq LSYM(Ldiv0) bpl LSYM(Lover10) - neg divisor, divisor @ Loops below use unsigned. + negs divisor, divisor @ Loops below use unsigned. LSYM(Lover10): push { work } @ Need to save the sign of the dividend, unfortunately, we need @@ -1408,7 +1412,7 @@ LSYM(Lover10): push { dividend } cmp dividend, #0 bpl LSYM(Lover11) - neg dividend, dividend + negs dividend, dividend LSYM(Lover11): cmp dividend, divisor blo LSYM(Lgot_result) @@ -1418,7 +1422,7 @@ LSYM(Lover11): pop { work } cmp work, #0 bpl LSYM(Lover12) - neg dividend, dividend + negs dividend, dividend LSYM(Lover12): pop { work } RET @@ -1540,12 +1544,12 @@ LSYM(Lover12): address, so just clear pc..pc+1. */ #if defined __thumb__ && !defined __thumb2__ push {r7} - mov r7, #0xf - lsl r7, #16 - add r7, #2 + movs r7, #0xf + lsls r7, #16 + adds r7, #2 adr r0, . + 4 - add r1, r0, #1 - mov r2, #0 + adds r1, r0, #1 + movs r2, #0 svc 0 pop {r7} #else @@ -1595,17 +1599,17 @@ LSYM(Lover12): FUNC_ALIAS aeabi_llsr lshrdi3 #ifdef __thumb__ - lsr al, r2 - mov r3, ah - lsr ah, r2 + lsrs al, r2 + movs r3, ah + lsrs ah, r2 mov ip, r3 - sub r2, #32 - lsr r3, r2 - orr al, r3 - neg r2, r2 + subs r2, #32 + lsrs r3, r2 + orrs al, r3 + negs r2, r2 mov r3, ip - lsl r3, r2 - orr al, r3 + lsls r3, r2 + orrs al, r3 RET #else subs r3, r2, #32 @@ -1627,21 +1631,21 @@ LSYM(Lover12): FUNC_ALIAS aeabi_lasr ashrdi3 #ifdef __thumb__ - lsr al, r2 - mov r3, ah - asr ah, r2 - sub r2, #32 + lsrs al, r2 + movs r3, ah + asrs ah, r2 + subs r2, #32 @ If r2 is negative at this point the following step would OR @ the sign bit into all of AL. That's not what we want... bmi 1f mov ip, r3 - asr r3, r2 - orr al, r3 + asrs r3, r2 + orrs al, r3 mov r3, ip 1: - neg r2, r2 - lsl r3, r2 - orr al, r3 + negs r2, r2 + lsls r3, r2 + orrs al, r3 RET #else subs r3, r2, #32 @@ -1664,17 +1668,17 @@ LSYM(Lover12): FUNC_ALIAS aeabi_llsl ashldi3 #ifdef __thumb__ - lsl ah, r2 - mov r3, al - lsl al, r2 + lsls ah, r2 + movs r3, al + lsls al, r2 mov ip, r3 - sub r2, #32 - lsl r3, r2 - orr ah, r3 - neg r2, r2 + subs r2, #32 + lsls r3, r2 + orrs ah, r3 + negs r2, r2 mov r3, ip - lsr r3, r2 - orr ah, r3 + lsrs r3, r2 + orrs ah, r3 RET #else subs r3, r2, #32 @@ -1695,26 +1699,26 @@ LSYM(Lover12): #ifdef L_clzsi2 #ifdef NOT_ISA_TARGET_32BIT FUNC_START clzsi2 - mov r1, #28 - mov r3, #1 - lsl r3, r3, #16 + movs r1, #28 + movs r3, #1 + lsls r3, r3, #16 cmp r0, r3 /* 0x10000 */ bcc 2f - lsr r0, r0, #16 - sub r1, r1, #16 -2: lsr r3, r3, #8 + lsrs r0, r0, #16 + subs r1, r1, #16 +2: lsrs r3, r3, #8 cmp r0, r3 /* #0x100 */ bcc 2f - lsr r0, r0, #8 - sub r1, r1, #8 -2: lsr r3, r3, #4 + lsrs r0, r0, #8 + subs r1, r1, #8 +2: lsrs r3, r3, #4 cmp r0, r3 /* #0x10 */ bcc 2f - lsr r0, r0, #4 - sub r1, r1, #4 + lsrs r0, r0, #4 + subs r1, r1, #4 2: adr r2, 1f ldrb r0, [r2, r0] - add r0, r0, r1 + adds r0, r0, r1 bx lr .align 2 1: @@ -1757,34 +1761,49 @@ ARM_FUNC_START clzsi2 # ifdef NOT_ISA_TARGET_32BIT FUNC_START clzdi2 push {r4, lr} -# else + cmp xxh, #0 + bne 1f +# ifdef __ARMEB__ + movs r0, xxl + bl __clzsi2 + adds r0, r0, #32 + b 2f +1: + bl __clzsi2 +# else + bl __clzsi2 + adds r0, r0, #32 + b 2f +1: + movs r0, xxh + bl __clzsi2 +# endif +2: + pop {r4, pc} +# else /* NOT_ISA_TARGET_32BIT */ ARM_FUNC_START clzdi2 do_push {r4, lr} -# endif cmp xxh, #0 bne 1f -# ifdef __ARMEB__ +# ifdef __ARMEB__ mov r0, xxl bl __clzsi2 add r0, r0, #32 b 2f 1: bl __clzsi2 -# else +# else bl __clzsi2 add r0, r0, #32 b 2f 1: mov r0, xxh bl __clzsi2 -# endif +# endif 2: -# ifdef NOT_ISA_TARGET_32BIT - pop {r4, pc} -# else RETLDM r4 -# endif FUNC_END clzdi2 +# endif /* NOT_ISA_TARGET_32BIT */ #else /* defined (__ARM_FEATURE_CLZ) */ @@ -1803,28 +1822,28 @@ ARM_FUNC_START clzdi2 #ifdef L_ctzsi2 #ifdef NOT_ISA_TARGET_32BIT FUNC_START ctzsi2 - neg r1, r0 - and r0, r0, r1 - mov r1, #28 - mov r3, #1 - lsl r3, r3, #16 + negs r1, r0 + ands r0, r0, r1 + movs r1, #28 + movs r3, #1 + lsls r3, r3, #16 cmp r0, r3 /* 0x10000 */ bcc 2f - lsr r0, r0, #16 - sub r1, r1, #16 -2: lsr r3, r3, #8 + lsrs r0, r0, #16 + subs r1, r1, #16 +2: lsrs r3, r3, #8 cmp r0, r3 /* #0x100 */ bcc 2f - lsr r0, r0, #8 - sub r1, r1, #8 -2: lsr r3, r3, #4 + lsrs r0, r0, #8 + subs r1, r1, #8 +2: lsrs r3, r3, #4 cmp r0, r3 /* #0x10 */ bcc 2f - lsr r0, r0, #4 - sub r1, r1, #4 + lsrs r0, r0, #4 + subs r1, r1, #4 2: adr r2, 1f ldrb r0, [r2, r0] - sub r0, r0, r1 + subs r0, r0, r1 bx lr .align 2 1: diff --git a/libgcc/config/arm/libunwind.S b/libgcc/config/arm/libunwind.S index 176ba5e73b7..08e0fcc0c58 100644 --- a/libgcc/config/arm/libunwind.S +++ b/libgcc/config/arm/libunwind.S @@ -63,28 +63,28 @@ /* r0 points to a 16-word block. Upload these values to the actual core state. */ FUNC_START restore_core_regs - mov r1, r0 - add r1, r1, #52 - ldmia r1!, {r3, r4, r5} - sub r3, r3, #4 - mov ip, r3 - str r5, [r3] - mov lr, r4 + movs r1, r0 + adds r1, r1, #52 + ldmia r1!, {r3, r4, r5} + subs r3, r3, #4 + mov ip, r3 + str r5, [r3] + mov lr, r4 /* Restore r8-r11. */ - mov r1, r0 - add r1, r1, #32 - ldmia r1!, {r2, r3, r4, r5} - mov r8, r2 - mov r9, r3 - mov sl, r4 - mov fp, r5 - mov r1, r0 - add r1, r1, #8 - ldmia r1!, {r2, r3, r4, r5, r6, r7} - ldr r1, [r0, #4] - ldr r0, [r0] - mov sp, ip - pop {pc} + movs r1, r0 + adds r1, r1, #32 + ldmia r1!, {r2, r3, r4, r5} + mov r8, r2 + mov r9, r3 + mov sl, r4 + mov fp, r5 + movs r1, r0 + adds r1, r1, #8 + ldmia r1!, {r2, r3, r4, r5, r6, r7} + ldr r1, [r0, #4] + ldr r0, [r0] + mov sp, ip + pop {pc} FUNC_END restore_core_regs UNPREFIX restore_core_regs @@ -132,38 +132,38 @@ FUNC_START gnu_Unwind_Save_WMMXC FUNC_START \name /* Create a phase2_vrs structure. */ /* Save r0 in the PC slot so we can use it as a scratch register. */ - push {r0} - add r0, sp, #4 - push {r0, lr} /* Push original SP and LR. */ + push {r0} + add r0, sp, #4 + push {r0, lr} /* Push original SP and LR. */ /* Make space for r8-r12. */ - sub sp, sp, #20 + sub sp, sp, #20 /* Save low registers. */ - push {r0, r1, r2, r3, r4, r5, r6, r7} + push {r0, r1, r2, r3, r4, r5, r6, r7} /* Save high registers. */ - add r0, sp, #32 - mov r1, r8 - mov r2, r9 - mov r3, sl - mov r4, fp - mov r5, ip - stmia r0!, {r1, r2, r3, r4, r5} + add r0, sp, #32 + mov r1, r8 + mov r2, r9 + mov r3, sl + mov r4, fp + mov r5, ip + stmia r0!, {r1, r2, r3, r4, r5} /* Restore original low register values. */ - add r0, sp, #4 - ldmia r0!, {r1, r2, r3, r4, r5} + add r0, sp, #4 + ldmia r0!, {r1, r2, r3, r4, r5} /* Restore orginial r0. */ - ldr r0, [sp, #60] - str r0, [sp] + ldr r0, [sp, #60] + str r0, [sp] /* Demand-save flags, plus an extra word for alignment. */ - mov r3, #0 - push {r2, r3} + movs r3, #0 + push {r2, r3} /* Point r1 at the block. Pass r[0..nargs) unchanged. */ - add r\nargs, sp, #4 + add r\nargs, sp, #4 - bl SYM (__gnu\name) + bl SYM (__gnu\name) - ldr r3, [sp, #64] - add sp, sp, #72 - bx r3 + ldr r3, [sp, #64] + add sp, sp, #72 + bx r3 FUNC_END \name UNPREFIX \name