<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head><meta http-equiv="content-type" content="text/html; charset=utf-8" />
<title>[197446] trunk/Source/JavaScriptCore</title>
</head>
<body>
<style type="text/css"><!--
#msg dl.meta { border: 1px #006 solid; background: #369; padding: 6px; color: #fff; }
#msg dl.meta dt { float: left; width: 6em; font-weight: bold; }
#msg dt:after { content:':';}
#msg dl, #msg dt, #msg ul, #msg li, #header, #footer, #logmsg { font-family: verdana,arial,helvetica,sans-serif; font-size: 10pt; }
#msg dl a { font-weight: bold}
#msg dl a:link { color:#fc3; }
#msg dl a:active { color:#ff0; }
#msg dl a:visited { color:#cc6; }
h3 { font-family: verdana,arial,helvetica,sans-serif; font-size: 10pt; font-weight: bold; }
#msg pre { overflow: auto; background: #ffc; border: 1px #fa0 solid; padding: 6px; }
#logmsg { background: #ffc; border: 1px #fa0 solid; padding: 1em 1em 0 1em; }
#logmsg p, #logmsg pre, #logmsg blockquote { margin: 0 0 1em 0; }
#logmsg p, #logmsg li, #logmsg dt, #logmsg dd { line-height: 14pt; }
#logmsg h1, #logmsg h2, #logmsg h3, #logmsg h4, #logmsg h5, #logmsg h6 { margin: .5em 0; }
#logmsg h1:first-child, #logmsg h2:first-child, #logmsg h3:first-child, #logmsg h4:first-child, #logmsg h5:first-child, #logmsg h6:first-child { margin-top: 0; }
#logmsg ul, #logmsg ol { padding: 0; list-style-position: inside; margin: 0 0 0 1em; }
#logmsg ul { text-indent: -1em; padding-left: 1em; }#logmsg ol { text-indent: -1.5em; padding-left: 1.5em; }
#logmsg > ul, #logmsg > ol { margin: 0 0 1em 0; }
#logmsg pre { background: #eee; padding: 1em; }
#logmsg blockquote { border: 1px solid #fa0; border-left-width: 10px; padding: 1em 1em 0 1em; background: white;}
#logmsg dl { margin: 0; }
#logmsg dt { font-weight: bold; }
#logmsg dd { margin: 0; padding: 0 0 0.5em 0; }
#logmsg dd:before { content:'\00bb';}
#logmsg table { border-spacing: 0px; border-collapse: collapse; border-top: 4px solid #fa0; border-bottom: 1px solid #fa0; background: #fff; }
#logmsg table th { text-align: left; font-weight: normal; padding: 0.2em 0.5em; border-top: 1px dotted #fa0; }
#logmsg table td { text-align: right; border-top: 1px dotted #fa0; padding: 0.2em 0.5em; }
#logmsg table thead th { text-align: center; border-bottom: 1px solid #fa0; }
#logmsg table th.Corner { text-align: left; }
#logmsg hr { border: none 0; border-top: 2px dashed #fa0; height: 1px; }
#header, #footer { color: #fff; background: #636; border: 1px #300 solid; padding: 6px; }
#patch { width: 100%; }
#patch h4 {font-family: verdana,arial,helvetica,sans-serif;font-size:10pt;padding:8px;background:#369;color:#fff;margin:0;}
#patch .propset h4, #patch .binary h4 {margin:0;}
#patch pre {padding:0;line-height:1.2em;margin:0;}
#patch .diff {width:100%;background:#eee;padding: 0 0 10px 0;overflow:auto;}
#patch .propset .diff, #patch .binary .diff {padding:10px 0;}
#patch span {display:block;padding:0 10px;}
#patch .modfile, #patch .addfile, #patch .delfile, #patch .propset, #patch .binary, #patch .copfile {border:1px solid #ccc;margin:10px 0;}
#patch ins {background:#dfd;text-decoration:none;display:block;padding:0 10px;}
#patch del {background:#fdd;text-decoration:none;display:block;padding:0 10px;}
#patch .lines, .info {color:#888;background:#fff;}
--></style>
<div id="msg">
<dl class="meta">
<dt>Revision</dt> <dd><a href="http://trac.webkit.org/projects/webkit/changeset/197446">197446</a></dd>
<dt>Author</dt> <dd>commit-queue@webkit.org</dd>
<dt>Date</dt> <dd>2016-03-02 02:22:27 -0800 (Wed, 02 Mar 2016)</dd>
</dl>
<h3>Log Message</h3>
<pre>[JSC] Use a Move without REX byte when possible
https://bugs.webkit.org/show_bug.cgi?id=154801
Patch by Benjamin Poulain <bpoulain@apple.com> on 2016-03-02
Reviewed by Alex Christensen.
Filip wrote an optimization in the register allocator
to use 32bit "Move" when we don't care about the top bytes.
When I moved the commutative ops to the fake 3 operands instruction
I largely destroyed this since all the "Moves" became full register.
In this patch, I switch back to 32bit "Moves" for 32bit operations.
* assembler/MacroAssemblerX86Common.h:
(JSC::MacroAssemblerX86Common::and32):
(JSC::MacroAssemblerX86Common::lshift32):
(JSC::MacroAssemblerX86Common::mul32):
(JSC::MacroAssemblerX86Common::or32):
(JSC::MacroAssemblerX86Common::rshift32):
(JSC::MacroAssemblerX86Common::urshift32):
(JSC::MacroAssemblerX86Common::xor32):
(JSC::MacroAssemblerX86Common::branchAdd32):
(JSC::MacroAssemblerX86Common::branchMul32):
(JSC::MacroAssemblerX86Common::branchSub32):
(JSC::MacroAssemblerX86Common::move32IfNeeded):</pre>
<h3>Modified Paths</h3>
<ul>
<li><a href="#trunkSourceJavaScriptCoreChangeLog">trunk/Source/JavaScriptCore/ChangeLog</a></li>
<li><a href="#trunkSourceJavaScriptCoreassemblerMacroAssemblerX86Commonh">trunk/Source/JavaScriptCore/assembler/MacroAssemblerX86Common.h</a></li>
</ul>
</div>
<div id="patch">
<h3>Diff</h3>
<a id="trunkSourceJavaScriptCoreChangeLog"></a>
<div class="modfile"><h4>Modified: trunk/Source/JavaScriptCore/ChangeLog (197445 => 197446)</h4>
<pre class="diff"><span>
<span class="info">--- trunk/Source/JavaScriptCore/ChangeLog 2016-03-02 07:53:20 UTC (rev 197445)
+++ trunk/Source/JavaScriptCore/ChangeLog 2016-03-02 10:22:27 UTC (rev 197446)
</span><span class="lines">@@ -1,3 +1,31 @@
</span><ins>+2016-03-02 Benjamin Poulain <bpoulain@apple.com>
+
+ [JSC] Use a Move without REX byte when possible
+ https://bugs.webkit.org/show_bug.cgi?id=154801
+
+ Reviewed by Alex Christensen.
+
+ Filip wrote an optimization in the register allocator
+ to use 32bit "Move" when we don't care about the top bytes.
+
+ When I moved the commutative ops to the fake 3 operands instruction
+ I largely destroyed this since all the "Moves" became full register.
+
+ In this patch, I switch back to 32bit "Moves" for 32bit operations.
+
+ * assembler/MacroAssemblerX86Common.h:
+ (JSC::MacroAssemblerX86Common::and32):
+ (JSC::MacroAssemblerX86Common::lshift32):
+ (JSC::MacroAssemblerX86Common::mul32):
+ (JSC::MacroAssemblerX86Common::or32):
+ (JSC::MacroAssemblerX86Common::rshift32):
+ (JSC::MacroAssemblerX86Common::urshift32):
+ (JSC::MacroAssemblerX86Common::xor32):
+ (JSC::MacroAssemblerX86Common::branchAdd32):
+ (JSC::MacroAssemblerX86Common::branchMul32):
+ (JSC::MacroAssemblerX86Common::branchSub32):
+ (JSC::MacroAssemblerX86Common::move32IfNeeded):
+
</ins><span class="cx"> 2016-03-01 Benjamin Poulain <benjamin@webkit.org>
</span><span class="cx">
</span><span class="cx"> [JSC] Simplify ArithMod(ArithMod(x, const1), const2) if const2 >= const1
</span></span></pre></div>
<a id="trunkSourceJavaScriptCoreassemblerMacroAssemblerX86Commonh"></a>
<div class="modfile"><h4>Modified: trunk/Source/JavaScriptCore/assembler/MacroAssemblerX86Common.h (197445 => 197446)</h4>
<pre class="diff"><span>
<span class="info">--- trunk/Source/JavaScriptCore/assembler/MacroAssemblerX86Common.h 2016-03-02 07:53:20 UTC (rev 197445)
+++ trunk/Source/JavaScriptCore/assembler/MacroAssemblerX86Common.h 2016-03-02 10:22:27 UTC (rev 197446)
</span><span class="lines">@@ -262,26 +262,26 @@
</span><span class="cx"> else if (op1 == dest)
</span><span class="cx"> and32(op2, dest);
</span><span class="cx"> else {
</span><del>- move(op2, dest);
</del><ins>+ move32IfNeeded(op2, dest);
</ins><span class="cx"> and32(op1, dest);
</span><span class="cx"> }
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void and32(Address op1, RegisterID op2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(op2, dest);
</del><ins>+ move32IfNeeded(op2, dest);
</ins><span class="cx"> and32(op1, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void and32(RegisterID op1, Address op2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(op1, dest);
</del><ins>+ move32IfNeeded(op1, dest);
</ins><span class="cx"> and32(op2, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void and32(TrustedImm32 imm, RegisterID src, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> and32(imm, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -324,8 +324,7 @@
</span><span class="cx"> {
</span><span class="cx"> ASSERT(shift_amount != dest);
</span><span class="cx">
</span><del>- if (src != dest)
- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> lshift32(shift_amount, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -336,8 +335,7 @@
</span><span class="cx">
</span><span class="cx"> void lshift32(RegisterID src, TrustedImm32 imm, RegisterID dest)
</span><span class="cx"> {
</span><del>- if (src != dest)
- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> lshift32(imm, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -352,7 +350,7 @@
</span><span class="cx"> m_assembler.imull_rr(src1, dest);
</span><span class="cx"> return;
</span><span class="cx"> }
</span><del>- move(src1, dest);
</del><ins>+ move32IfNeeded(src1, dest);
</ins><span class="cx"> m_assembler.imull_rr(src2, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -363,13 +361,13 @@
</span><span class="cx">
</span><span class="cx"> void mul32(Address src1, RegisterID src2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src2, dest);
</del><ins>+ move32IfNeeded(src2, dest);
</ins><span class="cx"> mul32(src1, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void mul32(RegisterID src1, Address src2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src1, dest);
</del><ins>+ move32IfNeeded(src1, dest);
</ins><span class="cx"> mul32(src2, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -444,26 +442,26 @@
</span><span class="cx"> else if (op1 == dest)
</span><span class="cx"> or32(op2, dest);
</span><span class="cx"> else {
</span><del>- move(op2, dest);
</del><ins>+ move32IfNeeded(op2, dest);
</ins><span class="cx"> or32(op1, dest);
</span><span class="cx"> }
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void or32(Address op1, RegisterID op2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(op2, dest);
</del><ins>+ move32IfNeeded(op2, dest);
</ins><span class="cx"> or32(op1, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void or32(RegisterID op1, Address op2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(op1, dest);
</del><ins>+ move32IfNeeded(op1, dest);
</ins><span class="cx"> or32(op2, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void or32(TrustedImm32 imm, RegisterID src, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> or32(imm, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -487,8 +485,7 @@
</span><span class="cx"> {
</span><span class="cx"> ASSERT(shift_amount != dest);
</span><span class="cx">
</span><del>- if (src != dest)
- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> rshift32(shift_amount, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -499,8 +496,7 @@
</span><span class="cx">
</span><span class="cx"> void rshift32(RegisterID src, TrustedImm32 imm, RegisterID dest)
</span><span class="cx"> {
</span><del>- if (src != dest)
- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> rshift32(imm, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -524,8 +520,7 @@
</span><span class="cx"> {
</span><span class="cx"> ASSERT(shift_amount != dest);
</span><span class="cx">
</span><del>- if (src != dest)
- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> urshift32(shift_amount, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -536,8 +531,7 @@
</span><span class="cx">
</span><span class="cx"> void urshift32(RegisterID src, TrustedImm32 imm, RegisterID dest)
</span><span class="cx"> {
</span><del>- if (src != dest)
- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> urshift32(imm, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -607,26 +601,26 @@
</span><span class="cx"> else if (op1 == dest)
</span><span class="cx"> xor32(op2, dest);
</span><span class="cx"> else {
</span><del>- move(op2, dest);
</del><ins>+ move32IfNeeded(op2, dest);
</ins><span class="cx"> xor32(op1, dest);
</span><span class="cx"> }
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void xor32(Address op1, RegisterID op2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(op2, dest);
</del><ins>+ move32IfNeeded(op2, dest);
</ins><span class="cx"> xor32(op1, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void xor32(RegisterID op1, Address op2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(op1, dest);
</del><ins>+ move32IfNeeded(op1, dest);
</ins><span class="cx"> xor32(op2, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> void xor32(TrustedImm32 imm, RegisterID src, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> xor32(imm, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -2036,25 +2030,25 @@
</span><span class="cx"> {
</span><span class="cx"> if (src1 == dest)
</span><span class="cx"> return branchAdd32(cond, src2, dest);
</span><del>- move(src2, dest);
</del><ins>+ move32IfNeeded(src2, dest);
</ins><span class="cx"> return branchAdd32(cond, src1, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> Jump branchAdd32(ResultCondition cond, Address src1, RegisterID src2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src2, dest);
</del><ins>+ move32IfNeeded(src2, dest);
</ins><span class="cx"> return branchAdd32(cond, src1, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> Jump branchAdd32(ResultCondition cond, RegisterID src1, Address src2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src1, dest);
</del><ins>+ move32IfNeeded(src1, dest);
</ins><span class="cx"> return branchAdd32(cond, src2, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> Jump branchAdd32(ResultCondition cond, RegisterID src, TrustedImm32 imm, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src, dest);
</del><ins>+ move32IfNeeded(src, dest);
</ins><span class="cx"> return branchAdd32(cond, imm, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -2086,7 +2080,7 @@
</span><span class="cx"> {
</span><span class="cx"> if (src1 == dest)
</span><span class="cx"> return branchMul32(cond, src2, dest);
</span><del>- move(src2, dest);
</del><ins>+ move32IfNeeded(src2, dest);
</ins><span class="cx"> return branchMul32(cond, src1, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -2125,13 +2119,13 @@
</span><span class="cx"> // B := A - B is invalid.
</span><span class="cx"> ASSERT(src1 == dest || src2 != dest);
</span><span class="cx">
</span><del>- move(src1, dest);
</del><ins>+ move32IfNeeded(src1, dest);
</ins><span class="cx"> return branchSub32(cond, src2, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="cx"> Jump branchSub32(ResultCondition cond, RegisterID src1, TrustedImm32 src2, RegisterID dest)
</span><span class="cx"> {
</span><del>- move(src1, dest);
</del><ins>+ move32IfNeeded(src1, dest);
</ins><span class="cx"> return branchSub32(cond, src2, dest);
</span><span class="cx"> }
</span><span class="cx">
</span><span class="lines">@@ -2492,6 +2486,15 @@
</span><span class="cx"> return Jump(m_assembler.jCC(static_cast<X86Assembler::Condition>(cond & ~DoubleConditionBits)));
</span><span class="cx"> }
</span><span class="cx">
</span><ins>+ // The 32bit Move does not need the REX byte for low registers, making it shorter.
+ // Use this if the top bits are irrelevant because they will be reset by the next instruction.
+ void move32IfNeeded(RegisterID src, RegisterID dest)
+ {
+ if (src == dest)
+ return;
+ m_assembler.movl_rr(src, dest);
+ }
+
</ins><span class="cx"> #if CPU(X86_64)
</span><span class="cx"> void moveConditionallyAfterFloatingPointCompare(DoubleCondition cond, FPRegisterID left, FPRegisterID right, RegisterID src, RegisterID dest)
</span><span class="cx"> {
</span></span></pre>
</div>
</div>
</body>
</html>