[X86] Add FeatureINVLPGB and CPUID handling for INVLPGB/TLBSYNC instructions #121570
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…uctions Both instructions are available under the same CPUID bit (Fn8000_0008_EBX[INVLPGB]). NOTE: I could use some advice on the X86TargetParser.def ordering wrt matching gcc
|
@llvm/pr-subscribers-backend-x86 Author: Simon Pilgrim (RKSimon) ChangesIntroduced on znver3: https://reviews.llvm.org/D94134 Both instructions are available under the same CPUID bit (Fn8000_0008_EBX[INVLPGB]). NOTE: I could use some advice on the X86TargetParser.def ordering wrt matching gcc Full diff: https://github.com/llvm/llvm-project/pull/121570.diff 6 Files Affected:
diff --git a/llvm/include/llvm/TargetParser/X86TargetParser.def b/llvm/include/llvm/TargetParser/X86TargetParser.def
index 5b719a2b0eb399..4eb4877f4c8b0d 100644
--- a/llvm/include/llvm/TargetParser/X86TargetParser.def
+++ b/llvm/include/llvm/TargetParser/X86TargetParser.def
@@ -197,6 +197,7 @@ X86_FEATURE_COMPAT(ENQCMD, "enqcmd", 0)
X86_FEATURE_COMPAT(F16C, "f16c", 0)
X86_FEATURE_COMPAT(FSGSBASE, "fsgsbase", 0)
X86_FEATURE (CRC32, "crc32")
+X86_FEATURE (INVLPGB, "invlpgb")
X86_FEATURE (INVPCID, "invpcid")
X86_FEATURE (RDPRU, "rdpru")
X86_FEATURE (SAHF, "sahf")
diff --git a/llvm/lib/Target/X86/X86.td b/llvm/lib/Target/X86/X86.td
index 38761e1fd7eecc..f0e5de25bb9d6b 100644
--- a/llvm/lib/Target/X86/X86.td
+++ b/llvm/lib/Target/X86/X86.td
@@ -252,6 +252,8 @@ def FeatureMWAITX : SubtargetFeature<"mwaitx", "HasMWAITX", "true",
"Enable MONITORX/MWAITX timer functionality">;
def FeatureCLZERO : SubtargetFeature<"clzero", "HasCLZERO", "true",
"Enable Cache Line Zero">;
+def FeatureINVLPGB : SubtargetFeature<"invlpgb", "HasINVLPGB", "true",
+ "Support invlpgb/tlbsync instructions">;
def FeatureCLDEMOTE : SubtargetFeature<"cldemote", "HasCLDEMOTE", "true",
"Enable Cache Line Demote">;
def FeaturePTWRITE : SubtargetFeature<"ptwrite", "HasPTWRITE", "true",
@@ -1565,6 +1567,7 @@ def ProcessorFeatures {
!listconcat(ZNFeatures, ZN2AdditionalFeatures);
list<SubtargetFeature> ZN3AdditionalFeatures = [FeatureFSRM,
FeatureINVPCID,
+ FeatureINVLPGB,
FeaturePKU,
FeatureVAES,
FeatureVPCLMULQDQ];
diff --git a/llvm/lib/Target/X86/X86InstrMisc.td b/llvm/lib/Target/X86/X86InstrMisc.td
index 9fabe2acf00194..0af60980820b51 100644
--- a/llvm/lib/Target/X86/X86InstrMisc.td
+++ b/llvm/lib/Target/X86/X86InstrMisc.td
@@ -1603,11 +1603,11 @@ let SchedRW = [WriteSystem] in {
let Uses = [EAX, EDX] in
def INVLPGB32 : I<0x01, MRM_FE, (outs), (ins),
"invlpgb", []>,
- TB, Requires<[Not64BitMode]>;
+ TB, Requires<[HasINVLPGB, Not64BitMode]>;
let Uses = [RAX, EDX] in
def INVLPGB64 : I<0x01, MRM_FE, (outs), (ins),
"invlpgb", []>,
- TB, Requires<[In64BitMode]>;
+ TB, Requires<[HasINVLPGB, In64BitMode]>;
} // SchedRW
//===----------------------------------------------------------------------===//
@@ -1617,7 +1617,7 @@ let SchedRW = [WriteSystem] in {
let SchedRW = [WriteSystem] in {
def TLBSYNC : I<0x01, MRM_FF, (outs), (ins),
"tlbsync", []>,
- TB, Requires<[]>;
+ TB, Requires<[HasINVLPGB]>;
} // SchedRW
//===----------------------------------------------------------------------===//
diff --git a/llvm/lib/Target/X86/X86InstrPredicates.td b/llvm/lib/Target/X86/X86InstrPredicates.td
index 5bdcf51be9dd84..7a105d5a9346d7 100644
--- a/llvm/lib/Target/X86/X86InstrPredicates.td
+++ b/llvm/lib/Target/X86/X86InstrPredicates.td
@@ -166,6 +166,7 @@ def HasWBNOINVD : Predicate<"Subtarget->hasWBNOINVD()">;
def HasRDPID : Predicate<"Subtarget->hasRDPID()">;
def HasRDPRU : Predicate<"Subtarget->hasRDPRU()">;
def HasWAITPKG : Predicate<"Subtarget->hasWAITPKG()">;
+def HasINVLPGB : Predicate<"Subtarget->hasINVLPGB()">;
def HasINVPCID : Predicate<"Subtarget->hasINVPCID()">;
def HasCX8 : Predicate<"Subtarget->hasCX8()">;
def HasCX16 : Predicate<"Subtarget->hasCX16()">;
diff --git a/llvm/lib/TargetParser/Host.cpp b/llvm/lib/TargetParser/Host.cpp
index 45b4cafc995986..fd67285fbecb70 100644
--- a/llvm/lib/TargetParser/Host.cpp
+++ b/llvm/lib/TargetParser/Host.cpp
@@ -1825,6 +1825,7 @@ const StringMap<bool> sys::getHostCPUFeatures() {
bool HasExtLeaf8 = MaxExtLevel >= 0x80000008 &&
!getX86CpuIDAndInfo(0x80000008, &EAX, &EBX, &ECX, &EDX);
Features["clzero"] = HasExtLeaf8 && ((EBX >> 0) & 1);
+ Features["invlpgb"] = HasExtLeaf8 && ((EBX >> 3) & 1);
Features["rdpru"] = HasExtLeaf8 && ((EBX >> 4) & 1);
Features["wbnoinvd"] = HasExtLeaf8 && ((EBX >> 9) & 1);
diff --git a/llvm/lib/TargetParser/X86TargetParser.cpp b/llvm/lib/TargetParser/X86TargetParser.cpp
index e4b7ed7cf9b61f..a6ac6a305b0e28 100644
--- a/llvm/lib/TargetParser/X86TargetParser.cpp
+++ b/llvm/lib/TargetParser/X86TargetParser.cpp
@@ -236,9 +236,9 @@ constexpr FeatureBitset FeaturesZNVER1 =
constexpr FeatureBitset FeaturesZNVER2 = FeaturesZNVER1 | FeatureCLWB |
FeatureRDPID | FeatureRDPRU |
FeatureWBNOINVD;
-static constexpr FeatureBitset FeaturesZNVER3 = FeaturesZNVER2 |
- FeatureINVPCID | FeaturePKU |
- FeatureVAES | FeatureVPCLMULQDQ;
+static constexpr FeatureBitset FeaturesZNVER3 =
+ FeaturesZNVER2 | FeatureINVPCID | FeatureINVLPGB | FeaturePKU |
+ FeatureVAES | FeatureVPCLMULQDQ;
static constexpr FeatureBitset FeaturesZNVER4 =
FeaturesZNVER3 | FeatureAVX512F | FeatureEVEX512 | FeatureAVX512CD |
FeatureAVX512DQ | FeatureAVX512BW | FeatureAVX512VL | FeatureAVX512IFMA |
@@ -505,6 +505,7 @@ constexpr FeatureBitset ImpliedFeaturesCRC32 = {};
constexpr FeatureBitset ImpliedFeaturesENQCMD = {};
constexpr FeatureBitset ImpliedFeaturesFSGSBASE = {};
constexpr FeatureBitset ImpliedFeaturesFXSR = {};
+constexpr FeatureBitset ImpliedFeaturesINVLPGB = {};
constexpr FeatureBitset ImpliedFeaturesINVPCID = {};
constexpr FeatureBitset ImpliedFeaturesLWP = {};
constexpr FeatureBitset ImpliedFeaturesLZCNT = {};
|
You can test this locally with the following command:git-clang-format --diff f87a9db8322643ccbc324e317a75b55903129b55 f1524c81933c699bd7ec26a48d2991ff78556951 --extensions cpp -- llvm/lib/TargetParser/Host.cpp llvm/lib/TargetParser/X86TargetParser.cppView the diff from clang-format here.diff --git a/llvm/lib/TargetParser/Host.cpp b/llvm/lib/TargetParser/Host.cpp
index fd67285fbe..9471bf5d05 100644
--- a/llvm/lib/TargetParser/Host.cpp
+++ b/llvm/lib/TargetParser/Host.cpp
@@ -1825,7 +1825,7 @@ const StringMap<bool> sys::getHostCPUFeatures() {
bool HasExtLeaf8 = MaxExtLevel >= 0x80000008 &&
!getX86CpuIDAndInfo(0x80000008, &EAX, &EBX, &ECX, &EDX);
Features["clzero"] = HasExtLeaf8 && ((EBX >> 0) & 1);
- Features["invlpgb"] = HasExtLeaf8 && ((EBX >> 3) & 1);
+ Features["invlpgb"] = HasExtLeaf8 && ((EBX >> 3) & 1);
Features["rdpru"] = HasExtLeaf8 && ((EBX >> 4) & 1);
Features["wbnoinvd"] = HasExtLeaf8 && ((EBX >> 9) & 1);
|
|
I think compiler doesn't need aware of a feature if it never generates the instructions. |
OK - are we consistent with this approach though? |
I think so, here is a recent example. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Introduced on znver3: https://reviews.llvm.org/D94134
Both instructions are available under the same CPUID bit (Fn8000_0008_EBX[INVLPGB]).
NOTE: I could use some advice on the X86TargetParser.def ordering wrt matching gcc