Yeah I intend to update the AutoDetectSubtarget. I didn't do it yet because it requires capturing the max leaf supported from leaf 0 so that we know if leaf 7 is legal. Then we also need to pass 0 in ecx into cpuid to get the right subleaf of leaf 7. That mucks with the cpuid routine that is branched for different compilers so I need to be careful there.<br>
<br><div class="gmail_quote">On Thu, Oct 13, 2011 at 9:45 AM, Joe Abbey <span dir="ltr"><<a href="mailto:jabbey@arxan.com">jabbey@arxan.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div style="word-wrap:break-word">I see. Are you going to updating AutoDetectSubtarget features to check bit 8 of the Structured Extended Feature Leaf? Looks like there's more features.<div><br></div><div><div>Bits 00: FSGS BASE </div>
<div>Bits 02-01: Reserved.</div><div>Bits 03: BMI1
</div><div>Bits 04: Reserved. </div><div>Bits 05: AVX2
</div><div>Bits 07-06: Reserved. </div><div>Bits 08: BMI2</div><div>
Bits 09: ERMS</div><div>
Bits 10: INVPCID
</div>
<div>Bits 31-11: Reserved.</div><div><br></div><div>And add an initializer to the list for the constructor?</div><div><br></div><div><div class="im"><div>
Joe Abbey<br>Software Architect<br>Arxan Technologies, Inc.<br>1305 Cumberland Ave, Ste 215<br>West Lafayette, IN 47906<br>W: <a href="tel:765-889-4756%20x2" value="+17658894756" target="_blank">765-889-4756 x2</a><br>C: <a href="tel:765-464-9893" value="+17654649893" target="_blank">765-464-9893</a><br>
<a href="mailto:jabbey@arxan.com" target="_blank">jabbey@arxan.com</a><br><a href="http://www.arxan.com" target="_blank">www.arxan.com</a><br><br>
</div>
<br></div><div><div></div><div class="h5"><div><div>On Oct 13, 2011, at 12:18 PM, Craig Topper wrote:</div><br><blockquote type="cite">The test failed because the X86Subtarget constructor didn't initialize HasBMI to false. So the feature check in X86ISelLowering ends up reading from uninitialized memory and sometimes thinks its enabled. This causes ISelLowering to not run the custom lowering code, but the pattern matching code for instruction selection gets the feature enable a different way that correctly knows that BMI is disabled. So it was unable to find an instruction to use. I have a fix that I'll probably submit tonight.<br>
<br><div class="gmail_quote">On Thu, Oct 13, 2011 at 9:04 AM, Joe Abbey <span dir="ltr"><<a href="mailto:jabbey@arxan.com" target="_blank">jabbey@arxan.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div>
<div>
<font size="2"><div>This patch enhances r141855 and fixes the failing test by adding the new -mattr=+bmi target attribute.<br>
<br>
Craig, is that correct?<br>
<br>
</div></font>
</div>
<div>
<font size="2"><div><br>
<br>
Joe Abbey<br>
Software Architect<br>
Arxan Technologies, Inc.<br>
1305 Cumberland Ave, Ste 215<br>
West Lafayette, IN 47906<br>
W: <a href="tel:765-889-4756%20x2" value="+17658894756" target="_blank">765-889-4756 x2</a><br>
C: <a href="tel:765-464-9893" value="+17654649893" target="_blank">765-464-9893</a><br>
<a href="mailto:jabbey@arxan.com" target="_blank">jabbey@arxan.com</a><br>
<a href="http://www.arxan.com/" target="_blank">www.arxan.com</a><br>
<br>
<br>
On Oct 13, 2011, at 3:27 AM, Bill Wendling wrote:<br>
<br>
> Hi Craig,<br>
> <br>
> This is causing a failure on one of the public buildbots:<br>
> <br>
> <a href="http://lab.llvm.org:8011/builders/llvm-x86_64-linux/builds/101" target="_blank">http://lab.llvm.org:8011/builders/llvm-x86_64-linux/builds/101</a><br>
> <br>
> Could you take a look please?<br>
> <br>
> -bw<br>
> <br>
> On Oct 13, 2011, at 12:09 AM, Craig Topper wrote:<br>
> <br>
>> Author: ctopper<br>
>> Date: Thu Oct 13 02:09:14 2011<br>
>> New Revision: 141854<br>
>> <br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project?rev=141854&view=rev" target="_blank">http://llvm.org/viewvc/llvm-project?rev=141854&view=rev</a><br>
>> Log:<br>
>> Add X86 TZCNT instruction and patterns to select it. Also added core-avx2 processor which is gcc's name for Haswell.<br>
>> <br>
>> Added:<br>
>> llvm/trunk/test/CodeGen/X86/bmi.ll<br>
>> Modified:<br>
>> llvm/trunk/lib/Target/X86/X86.td<br>
>> llvm/trunk/lib/Target/X86/X86ISelLowering.cpp<br>
>> llvm/trunk/lib/Target/X86/X86InstrInfo.td<br>
>> llvm/trunk/lib/Target/X86/X86Subtarget.h<br>
>> llvm/trunk/test/MC/Disassembler/X86/simple-tests.txt<br>
>> llvm/trunk/test/MC/Disassembler/X86/x86-32.txt<br>
>> <br>
>> Modified: llvm/trunk/lib/Target/X86/X86.td<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86.td?rev=141854&r1=141853&r2=141854&view=diff" target="_blank">http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86.td?rev=141854&r1=141853&r2=141854&view=diff</a><br>
>> ==============================================================================<br>
>> --- llvm/trunk/lib/Target/X86/X86.td (original)<br>
>> +++ llvm/trunk/lib/Target/X86/X86.td Thu Oct 13 02:09:14 2011<br>
>> @@ -104,6 +104,8 @@<br>
>> "Support 16-bit floating point conversion instructions">;<br>
>> def FeatureLZCNT : SubtargetFeature<"lzcnt", "HasLZCNT", "true",<br>
>> "Support LZCNT instruction">;<br>
>> +def FeatureBMI : SubtargetFeature<"bmi", "HasBMI", "true",<br>
>> + "Support BMI instructions">;<br>
>> <br>
>> //===----------------------------------------------------------------------===//<br>
>> // X86 processors supported.<br>
>> @@ -157,6 +159,11 @@<br>
>> def : Proc<"core-avx-i", [FeatureSSE42, FeatureCMPXCHG16B,<br>
>> FeatureAES, FeatureCLMUL,<br>
>> FeatureRDRAND, FeatureF16C]>;<br>
>> +// Haswell<br>
>> +def : Proc<"core-avx2", [FeatureSSE42, FeatureCMPXCHG16B, FeatureAES,<br>
>> + FeatureCLMUL, FeatureRDRAND, FeatureF16C,<br>
>> + FeatureFMA3, FeatureMOVBE, FeatureLZCNT,<br>
>> + FeatureBMI]>;<br>
>> <br>
>> def : Proc<"k6", [FeatureMMX]>;<br>
>> def : Proc<"k6-2", [Feature3DNow]>;<br>
>> <br>
>> Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=141854&r1=141853&r2=141854&view=diff" target="_blank">http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=141854&r1=141853&r2=141854&view=diff</a><br>
>> ==============================================================================<br>
>> --- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)<br>
>> +++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Thu Oct 13 02:09:14 2011<br>
>> @@ -379,11 +379,15 @@<br>
>> setOperationAction(ISD::FREM , MVT::f80 , Expand);<br>
>> setOperationAction(ISD::FLT_ROUNDS_ , MVT::i32 , Custom);<br>
>> <br>
>> - setOperationAction(ISD::CTTZ , MVT::i8 , Custom);<br>
>> - setOperationAction(ISD::CTTZ , MVT::i16 , Custom);<br>
>> - setOperationAction(ISD::CTTZ , MVT::i32 , Custom);<br>
>> - if (Subtarget->is64Bit())<br>
>> - setOperationAction(ISD::CTTZ , MVT::i64 , Custom);<br>
>> + if (Subtarget->hasBMI()) {<br>
>> + setOperationAction(ISD::CTTZ , MVT::i8 , Promote);<br>
>> + } else {<br>
>> + setOperationAction(ISD::CTTZ , MVT::i8 , Custom);<br>
>> + setOperationAction(ISD::CTTZ , MVT::i16 , Custom);<br>
>> + setOperationAction(ISD::CTTZ , MVT::i32 , Custom);<br>
>> + if (Subtarget->is64Bit())<br>
>> + setOperationAction(ISD::CTTZ , MVT::i64 , Custom);<br>
>> + }<br>
>> <br>
>> if (Subtarget->hasLZCNT()) {<br>
>> setOperationAction(ISD::CTLZ , MVT::i8 , Promote);<br>
>> <br>
>> Modified: llvm/trunk/lib/Target/X86/X86InstrInfo.td<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrInfo.td?rev=141854&r1=141853&r2=141854&view=diff" target="_blank">http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrInfo.td?rev=141854&r1=141853&r2=141854&view=diff</a><br>
>> ==============================================================================<br>
>> --- llvm/trunk/lib/Target/X86/X86InstrInfo.td (original)<br>
>> +++ llvm/trunk/lib/Target/X86/X86InstrInfo.td Thu Oct 13 02:09:14 2011<br>
>> @@ -478,6 +478,7 @@<br>
>> def HasRDRAND : Predicate<"Subtarget->hasRDRAND()">;<br>
>> def HasF16C : Predicate<"Subtarget->hasF16C()">;<br>
>> def HasLZCNT : Predicate<"Subtarget->hasLZCNT()">;<br>
>> +def HasBMI : Predicate<"Subtarget->hasBMI()">;<br>
>> def FPStackf32 : Predicate<"!Subtarget->hasXMM()">;<br>
>> def FPStackf64 : Predicate<"!Subtarget->hasXMMInt()">;<br>
>> def HasCmpxchg16b: Predicate<"Subtarget->hasCmpxchg16b()">;<br>
>> @@ -1373,6 +1374,37 @@<br>
>> }<br>
>> <br>
>> //===----------------------------------------------------------------------===//<br>
>> +// BMI Instructions<br>
>> +//<br>
>> +let Predicates = [HasBMI], Defs = [EFLAGS] in {<br>
>> + def TZCNT16rr : I<0xBC, MRMSrcReg, (outs GR16:$dst), (ins GR16:$src),<br>
>> + "tzcnt{w}\t{$src, $dst|$dst, $src}",<br>
>> + [(set GR16:$dst, (cttz GR16:$src)), (implicit EFLAGS)]>, XS,<br>
>> + OpSize;<br>
>> + def TZCNT16rm : I<0xBC, MRMSrcMem, (outs GR16:$dst), (ins i16mem:$src),<br>
>> + "tzcnt{w}\t{$src, $dst|$dst, $src}",<br>
>> + [(set GR16:$dst, (cttz (loadi16 addr:$src))),<br>
>> + (implicit EFLAGS)]>, XS, OpSize;<br>
>> +<br>
>> + def TZCNT32rr : I<0xBC, MRMSrcReg, (outs GR32:$dst), (ins GR32:$src),<br>
>> + "tzcnt{l}\t{$src, $dst|$dst, $src}",<br>
>> + [(set GR32:$dst, (cttz GR32:$src)), (implicit EFLAGS)]>, XS;<br>
>> + def TZCNT32rm : I<0xBC, MRMSrcMem, (outs GR32:$dst), (ins i32mem:$src),<br>
>> + "tzcnt{l}\t{$src, $dst|$dst, $src}",<br>
>> + [(set GR32:$dst, (cttz (loadi32 addr:$src))),<br>
>> + (implicit EFLAGS)]>, XS;<br>
>> +<br>
>> + def TZCNT64rr : RI<0xBC, MRMSrcReg, (outs GR64:$dst), (ins GR64:$src),<br>
>> + "tzcnt{q}\t{$src, $dst|$dst, $src}",<br>
>> + [(set GR64:$dst, (cttz GR64:$src)), (implicit EFLAGS)]>,<br>
>> + XS;<br>
>> + def TZCNT64rm : RI<0xBC, MRMSrcMem, (outs GR64:$dst), (ins i64mem:$src),<br>
>> + "tzcnt{q}\t{$src, $dst|$dst, $src}",<br>
>> + [(set GR64:$dst, (cttz (loadi64 addr:$src))),<br>
>> + (implicit EFLAGS)]>, XS;<br>
>> +}<br>
>> +<br>
>> +//===----------------------------------------------------------------------===//<br>
>> // Subsystems.<br>
>> //===----------------------------------------------------------------------===//<br>
>> <br>
>> <br>
>> Modified: llvm/trunk/lib/Target/X86/X86Subtarget.h<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86Subtarget.h?rev=141854&r1=141853&r2=141854&view=diff" target="_blank">http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86Subtarget.h?rev=141854&r1=141853&r2=141854&view=diff</a><br>
>> ==============================================================================<br>
>> --- llvm/trunk/lib/Target/X86/X86Subtarget.h (original)<br>
>> +++ llvm/trunk/lib/Target/X86/X86Subtarget.h Thu Oct 13 02:09:14 2011<br>
>> @@ -102,6 +102,9 @@<br>
>> /// HasLZCNT - Processor has LZCNT instruction.<br>
>> bool HasLZCNT;<br>
>> <br>
>> + /// HasBMI - Processor has BMI1 instructions.<br>
>> + bool HasBMI;<br>
>> +<br>
>> /// IsBTMemSlow - True if BT (bit test) of memory instructions are slow.<br>
>> bool IsBTMemSlow;<br>
>> <br>
>> @@ -188,6 +191,7 @@<br>
>> bool hasRDRAND() const { return HasRDRAND; }<br>
>> bool hasF16C() const { return HasF16C; }<br>
>> bool hasLZCNT() const { return HasLZCNT; }<br>
>> + bool hasBMI() const { return HasBMI; }<br>
>> bool isBTMemSlow() const { return IsBTMemSlow; }<br>
>> bool isUnalignedMemAccessFast() const { return IsUAMemFast; }<br>
>> bool hasVectorUAMem() const { return HasVectorUAMem; }<br>
>> <br>
>> Added: llvm/trunk/test/CodeGen/X86/bmi.ll<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/bmi.ll?rev=141854&view=auto" target="_blank">http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/bmi.ll?rev=141854&view=auto</a><br>
>> ==============================================================================<br>
>> --- llvm/trunk/test/CodeGen/X86/bmi.ll (added)<br>
>> +++ llvm/trunk/test/CodeGen/X86/bmi.ll Thu Oct 13 02:09:14 2011<br>
>> @@ -0,0 +1,38 @@<br>
>> +; RUN: llc < %s -march=x86-64 -mattr=+bmi | FileCheck %s<br>
>> +<br>
>> +define i32 @t1(i32 %x) nounwind {<br>
>> + %tmp = tail call i32 @llvm.cttz.i32( i32 %x )<br>
>> + ret i32 %tmp<br>
>> +; CHECK: t1:<br>
>> +; CHECK: tzcntl<br>
>> +}<br>
>> +<br>
>> +declare i32 @llvm.cttz.i32(i32) nounwind readnone<br>
>> +<br>
>> +define i16 @t2(i16 %x) nounwind {<br>
>> + %tmp = tail call i16 @llvm.cttz.i16( i16 %x )<br>
>> + ret i16 %tmp<br>
>> +; CHECK: t2:<br>
>> +; CHECK: tzcntw<br>
>> +}<br>
>> +<br>
>> +declare i16 @llvm.cttz.i16(i16) nounwind readnone<br>
>> +<br>
>> +define i64 @t3(i64 %x) nounwind {<br>
>> + %tmp = tail call i64 @llvm.cttz.i64( i64 %x )<br>
>> + ret i64 %tmp<br>
>> +; CHECK: t3:<br>
>> +; CHECK: tzcntq<br>
>> +}<br>
>> +<br>
>> +declare i64 @llvm.cttz.i64(i64) nounwind readnone<br>
>> +<br>
>> +define i8 @t4(i8 %x) nounwind {<br>
>> + %tmp = tail call i8 @llvm.cttz.i8( i8 %x )<br>
>> + ret i8 %tmp<br>
>> +; CHECK: t4:<br>
>> +; CHECK: tzcntw<br>
>> +}<br>
>> +<br>
>> +declare i8 @llvm.cttz.i8(i8) nounwind readnone<br>
>> +<br>
>> <br>
>> Modified: llvm/trunk/test/MC/Disassembler/X86/simple-tests.txt<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/MC/Disassembler/X86/simple-tests.txt?rev=141854&r1=141853&r2=141854&view=diff" target="_blank">http://llvm.org/viewvc/llvm-project/llvm/trunk/test/MC/Disassembler/X86/simple-tests.txt?rev=141854&r1=141853&r2=141854&view=diff</a><br>
>> ==============================================================================<br>
>> --- llvm/trunk/test/MC/Disassembler/X86/simple-tests.txt (original)<br>
>> +++ llvm/trunk/test/MC/Disassembler/X86/simple-tests.txt Thu Oct 13 02:09:14 2011<br>
>> @@ -497,3 +497,12 @@<br>
>> <br>
>> # CHECK: lzcntq %rax, %rax<br>
>> 0xf3 0x48 0x0f 0xbd 0xc0<br>
>> +<br>
>> +# CHECK: tzcntl %eax, %eax<br>
>> +0xf3 0x0f 0xbc 0xc0<br>
>> +<br>
>> +# CHECK: tzcntw %ax, %ax<br>
>> +0x66 0xf3 0x0f 0xbc 0xc0<br>
>> +<br>
>> +# CHECK: tzcntq %rax, %rax<br>
>> +0xf3 0x48 0x0f 0xbc 0xc0<br>
>> <br>
>> Modified: llvm/trunk/test/MC/Disassembler/X86/x86-32.txt<br>
>> URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/MC/Disassembler/X86/x86-32.txt?rev=141854&r1=141853&r2=141854&view=diff" target="_blank">http://llvm.org/viewvc/llvm-project/llvm/trunk/test/MC/Disassembler/X86/x86-32.txt?rev=141854&r1=141853&r2=141854&view=diff</a><br>
>> ==============================================================================<br>
>> --- llvm/trunk/test/MC/Disassembler/X86/x86-32.txt (original)<br>
>> +++ llvm/trunk/test/MC/Disassembler/X86/x86-32.txt Thu Oct 13 02:09:14 2011<br>
>> @@ -477,3 +477,9 @@<br>
>> <br>
>> # CHECK: lzcntw %ax, %ax<br>
>> 0x66 0xf3 0x0f 0xbd 0xc0<br>
>> +<br>
>> +# CHECK: tzcntl %eax, %eax<br>
>> +0xf3 0x0f 0xbc 0xc0<br>
>> +<br>
>> +# CHECK: tzcntw %ax, %ax<br>
>> +0x66 0xf3 0x0f 0xbc 0xc0<br>
>> <br>
>> <br>
>> _______________________________________________<br>
>> llvm-commits mailing list<br>
>> <a href="mailto:llvm-commits@cs.uiuc.edu" target="_blank">llvm-commits@cs.uiuc.edu</a><br>
>> <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a><br>
> <br>
> _______________________________________________<br>
> llvm-commits mailing list<br>
> <a href="mailto:llvm-commits@cs.uiuc.edu" target="_blank">llvm-commits@cs.uiuc.edu</a><br>
> <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a><br>
<br>
</div></font>
</div>
</div></blockquote></div><br><br clear="all"><br>-- <br>~Craig<br>
</blockquote></div><br></div></div></div></div></div></blockquote></div><br><br clear="all"><br>-- <br>~Craig<br>