Index: lib/Target/X86/X86.td =================================================================== --- lib/Target/X86/X86.td +++ lib/Target/X86/X86.td @@ -88,12 +88,12 @@ [Feature3DNow]>; // All x86-64 hardware has SSE2, but we don't mark SSE2 as an implied // feature, because SSE2 can be disabled (e.g. for compiling OS kernels) -// without disabling 64-bit mode. +// without disabling 64-bit mode. Nothing should imply this feature bit. It +// is used to enforce that only 64-bit capable CPUs are used in 64-bit mode. def Feature64Bit : SubtargetFeature<"64bit", "HasX86_64", "true", "Support 64-bit instructions">; def FeatureCMPXCHG16B : SubtargetFeature<"cx16", "HasCmpxchg16b", "true", - "64-bit with cmpxchg16b", - [Feature64Bit]>; + "64-bit with cmpxchg16b">; def FeatureSlowSHLD : SubtargetFeature<"slow-shld", "IsSHLDSlow", "true", "SHLD instruction is slow">; def FeatureSlowPMULLD : SubtargetFeature<"slow-pmulld", "IsPMULLDSlow", "true", @@ -520,6 +520,7 @@ FeatureSSE3, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B ]>; @@ -532,6 +533,7 @@ FeatureSSSE3, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeatureLAHFSAHF, FeatureMacroFusion @@ -544,6 +546,7 @@ FeatureSSE41, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeatureLAHFSAHF, FeatureMacroFusion @@ -559,6 +562,7 @@ FeatureSSSE3, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeatureMOVBE, FeatureLEAForSP, @@ -580,6 +584,7 @@ FeatureSSE42, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeatureMOVBE, FeaturePOPCNT, @@ -615,6 +620,7 @@ FeatureSSE42, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeatureMOVBE, FeaturePOPCNT, @@ -675,6 +681,7 @@ FeatureSSE42, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeaturePOPCNT, FeatureLAHFSAHF, @@ -692,6 +699,7 @@ FeatureSSE42, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeaturePOPCNT, FeatureAES, @@ -710,6 +718,7 @@ FeatureAVX, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeaturePOPCNT, FeatureAES, @@ -927,13 +936,13 @@ foreach P = ["k8-sse3", "opteron-sse3", "athlon64-sse3"] in { def : Proc; + FeatureCMOV, Feature64Bit]>; } foreach P = ["amdfam10", "barcelona"] in { def : Proc; + FeatureSlowSHLD, FeatureLAHFSAHF, FeatureCMOV, Feature64Bit]>; } // Bobcat @@ -945,6 +954,7 @@ FeatureSSE4A, FeatureFXSR, FeatureNOPL, + Feature64Bit, FeatureCMPXCHG16B, FeaturePRFCHW, FeatureLZCNT, @@ -963,6 +973,7 @@ FeatureFXSR, FeatureNOPL, FeatureSSE4A, + Feature64Bit, FeatureCMPXCHG16B, FeaturePRFCHW, FeatureAES, @@ -987,6 +998,7 @@ FeatureCMOV, FeatureXOP, FeatureFMA4, + Feature64Bit, FeatureCMPXCHG16B, FeatureAES, FeaturePRFCHW, @@ -1011,6 +1023,7 @@ FeatureCMOV, FeatureXOP, FeatureFMA4, + Feature64Bit, FeatureCMPXCHG16B, FeatureAES, FeaturePRFCHW, @@ -1040,6 +1053,7 @@ FeatureCMOV, FeatureXOP, FeatureFMA4, + Feature64Bit, FeatureCMPXCHG16B, FeatureAES, FeaturePRFCHW, @@ -1075,6 +1089,7 @@ FeatureNOPL, FeatureXOP, FeatureFMA4, + Feature64Bit, FeatureCMPXCHG16B, FeatureAES, FeaturePRFCHW, @@ -1107,6 +1122,7 @@ FeatureCLFLUSHOPT, FeatureCLZERO, FeatureCMOV, + Feature64Bit, FeatureCMPXCHG16B, FeatureF16C, FeatureFMA, Index: lib/Target/X86/X86Subtarget.cpp =================================================================== --- lib/Target/X86/X86Subtarget.cpp +++ lib/Target/X86/X86Subtarget.cpp @@ -222,14 +222,22 @@ if (CPUName.empty()) CPUName = "generic"; - // Make sure 64-bit features are available in 64-bit mode. (But make sure - // SSE2 can be turned off explicitly.) std::string FullFS = FS; if (In64BitMode) { + // SSE2 should default to enabled in 64-bit mode, but can be turned off + // explicitly. if (!FullFS.empty()) - FullFS = "+64bit,+sse2," + FullFS; + FullFS = "+sse2," + FullFS; else - FullFS = "+64bit,+sse2"; + FullFS = "+sse2"; + + // If no CPU was specified, enable 64bit feature to satisy later check. + if (CPUName == "generic") { + if (!FullFS.empty()) + FullFS = "+64bit," + FullFS; + else + FullFS = "+64bit"; + } } // LAHF/SAHF are always supported in non-64-bit mode. @@ -264,8 +272,9 @@ LLVM_DEBUG(dbgs() << "Subtarget features: SSELevel " << X86SSELevel << ", 3DNowLevel " << X863DNowLevel << ", 64bit " << HasX86_64 << "\n"); - assert((!In64BitMode || HasX86_64) && - "64-bit code requested on a subtarget that doesn't support it!"); + if (In64BitMode && !HasX86_64) + report_fatal_error("64-bit code requested on a subtarget that doesn't " + "support it!"); // Stack alignment is 16 bytes on Darwin, Linux, kFreeBSD and Solaris (both // 32 and 64 bit) and for all 64-bit targets. Index: test/CodeGen/X86/early-ifcvt-crash.ll =================================================================== --- test/CodeGen/X86/early-ifcvt-crash.ll +++ test/CodeGen/X86/early-ifcvt-crash.ll @@ -1,7 +1,7 @@ ; RUN: llc < %s -x86-early-ifcvt -verify-machineinstrs ; RUN: llc < %s -x86-early-ifcvt -stress-early-ifcvt -verify-machineinstrs ; CPU without a scheduling model: -; RUN: llc < %s -x86-early-ifcvt -mcpu=pentium3 -verify-machineinstrs +; RUN: llc < %s -x86-early-ifcvt -mcpu=k8 -verify-machineinstrs ; ; Run these tests with and without -stress-early-ifcvt to exercise heuristics. ; Index: test/CodeGen/X86/pr11985.ll =================================================================== --- test/CodeGen/X86/pr11985.ll +++ test/CodeGen/X86/pr11985.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -mtriple=x86_64-pc-linux -mcpu=prescott | FileCheck %s --check-prefix=PRESCOTT +; RUN: llc < %s -mtriple=x86_64-pc-linux -mcpu=nocona | FileCheck %s --check-prefix=PRESCOTT ; RUN: llc < %s -mtriple=x86_64-pc-linux -mcpu=nehalem | FileCheck %s --check-prefix=NEHALEM ;;; TODO: (1) Some of the loads and stores are certainly unaligned and (2) the first load and first Index: test/CodeGen/X86/pr34080.ll =================================================================== --- test/CodeGen/X86/pr34080.ll +++ test/CodeGen/X86/pr34080.ll @@ -2,7 +2,7 @@ ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse2 | FileCheck %s --check-prefix=SSE2 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse2 -mcpu=x86-64 | FileCheck %s --check-prefix=SSE2-SCHEDULE ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse3 | FileCheck %s --check-prefix=SSE3 -; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse3 -mcpu=prescott | FileCheck %s --check-prefix=SSE3 +; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse3 -mcpu=nocona | FileCheck %s --check-prefix=SSE3 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx | FileCheck %s --check-prefix=AVX ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx -mcpu=sandybridge | FileCheck %s --check-prefix=AVX Index: test/CodeGen/X86/x86-64-double-shifts-var.ll =================================================================== --- test/CodeGen/X86/x86-64-double-shifts-var.ll +++ test/CodeGen/X86/x86-64-double-shifts-var.ll @@ -1,8 +1,3 @@ -; RUN: llc < %s -mtriple=x86_64-- -mcpu=athlon | FileCheck %s -; RUN: llc < %s -mtriple=x86_64-- -mcpu=athlon-tbird | FileCheck %s -; RUN: llc < %s -mtriple=x86_64-- -mcpu=athlon-4 | FileCheck %s -; RUN: llc < %s -mtriple=x86_64-- -mcpu=athlon-xp | FileCheck %s -; RUN: llc < %s -mtriple=x86_64-- -mcpu=athlon-mp | FileCheck %s ; RUN: llc < %s -mtriple=x86_64-- -mcpu=k8 | FileCheck %s ; RUN: llc < %s -mtriple=x86_64-- -mcpu=opteron | FileCheck %s ; RUN: llc < %s -mtriple=x86_64-- -mcpu=athlon64 | FileCheck %s Index: test/ThinLTO/X86/cache-config.ll =================================================================== --- test/ThinLTO/X86/cache-config.ll +++ test/ThinLTO/X86/cache-config.ll @@ -2,7 +2,7 @@ ; RUN: opt -module-hash -module-summary %s -o %t.bc ; RUN: llvm-lto2 run -o %t.o %t.bc -cache-dir %t.cache -r=%t.bc,globalfunc,plx -; RUN: llvm-lto2 run -o %t.o %t.bc -cache-dir %t.cache -r=%t.bc,globalfunc,plx -mcpu=yonah +; RUN: llvm-lto2 run -o %t.o %t.bc -cache-dir %t.cache -r=%t.bc,globalfunc,plx -mcpu=core2 ; RUN: llvm-lto2 run -o %t.o %t.bc -cache-dir %t.cache -r=%t.bc,globalfunc,plx -relax-elf-relocations ; RUN: llvm-lto2 run -o %t.o %t.bc -cache-dir %t.cache -r=%t.bc,globalfunc,plx -function-sections ; RUN: llvm-lto2 run -o %t.o %t.bc -cache-dir %t.cache -r=%t.bc,globalfunc,plx -data-sections Index: test/Transforms/LoopUnroll/X86/partial.ll =================================================================== --- test/Transforms/LoopUnroll/X86/partial.ll +++ test/Transforms/LoopUnroll/X86/partial.ll @@ -1,5 +1,5 @@ ; RUN: opt < %s -S -loop-unroll -mcpu=nehalem | FileCheck %s -; RUN: opt < %s -S -loop-unroll -mcpu=core -unroll-runtime=0 | FileCheck -check-prefix=CHECK-NOUNRL %s +; RUN: opt < %s -S -loop-unroll -unroll-runtime=0 | FileCheck -check-prefix=CHECK-NOUNRL %s target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" target triple = "x86_64-unknown-linux-gnu"