Update the minimum external LLVM to 14

.github/workflows/ci.yml

@@ -54,7 +54,7 @@ jobs:
             tidy: true
             os: ubuntu-20.04-xl
             env: {}
-          - name: x86_64-gnu-llvm-13
+          - name: x86_64-gnu-llvm-14
             tidy: false
             os: ubuntu-20.04-xl
             env: {}
@@ -300,11 +300,7 @@ jobs:
             env:
               RUST_BACKTRACE: 1
             os: ubuntu-20.04-xl
-          - name: x86_64-gnu-llvm-13
-            env:
-              RUST_BACKTRACE: 1
-            os: ubuntu-20.04-xl
-          - name: x86_64-gnu-llvm-13-stage1
+          - name: x86_64-gnu-llvm-14-stage1
             env:
               RUST_BACKTRACE: 1
             os: ubuntu-20.04-xl

compiler/rustc_codegen_llvm/src/consts.rs

@@ -3,7 +3,6 @@ use crate::common::{self, CodegenCx};
 use crate::debuginfo;
 use crate::errors::{InvalidMinimumAlignment, SymbolAlreadyDefined};
 use crate::llvm::{self, True};
-use crate::llvm_util;
 use crate::type_::Type;
 use crate::type_of::LayoutLlvmExt;
 use crate::value::Value;
@@ -56,13 +55,7 @@ pub fn const_alloc_to_llvm<'ll>(cx: &CodegenCx<'ll, '_>, alloc: ConstAllocation<
         // to avoid the cost of generating large complex const expressions.
         // For example, `[(u32, u8); 1024 * 1024]` contains uninit padding in each element,
         // and would result in `{ [5 x i8] zeroinitializer, [3 x i8] undef, ...repeat 1M times... }`.
-        let max = if llvm_util::get_version() < (14, 0, 0) {
-            // Generating partially-uninit consts inhibits optimizations in LLVM < 14.
-            // See https://github.com/rust-lang/rust/issues/84565.
-            1
-        } else {
-            cx.sess().opts.unstable_opts.uninit_const_chunk_threshold
-        };
+        let max = cx.sess().opts.unstable_opts.uninit_const_chunk_threshold;
         let allow_uninit_chunks = chunks.clone().take(max.saturating_add(1)).count() <= max;
 
         if allow_uninit_chunks {

compiler/rustc_codegen_llvm/src/context.rs

@@ -143,17 +143,6 @@ pub unsafe fn create_module<'ll>(
 
     let mut target_data_layout = sess.target.data_layout.to_string();
     let llvm_version = llvm_util::get_version();
-    if llvm_version < (14, 0, 0) {
-        if sess.target.llvm_target == "i686-pc-windows-msvc"
-            || sess.target.llvm_target == "i586-pc-windows-msvc"
-        {
-            target_data_layout =
-                "e-m:x-p:32:32-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:32-n8:16:32-a:0:32-S32"
-                    .to_string();
-        } else if sess.target.arch == "wasm32" {
-            target_data_layout = target_data_layout.replace("-p10:8:8-p20:8:8", "");
-        }
-    }
     if llvm_version < (16, 0, 0) {
         if sess.target.arch == "s390x" {
             target_data_layout = target_data_layout.replace("-v128:64", "");

compiler/rustc_codegen_llvm/src/llvm_util.rs

@@ -152,13 +152,7 @@ pub fn time_trace_profiler_finish(file_name: &Path) {
 pub fn to_llvm_features<'a>(sess: &Session, s: &'a str) -> SmallVec<[&'a str; 2]> {
     let arch = if sess.target.arch == "x86_64" { "x86" } else { &*sess.target.arch };
     match (arch, s) {
-        ("x86", "sse4.2") => {
-            if get_version() >= (14, 0, 0) {
-                smallvec!["sse4.2", "crc32"]
-            } else {
-                smallvec!["sse4.2"]
-            }
-        }
+        ("x86", "sse4.2") => smallvec!["sse4.2", "crc32"],
         ("x86", "pclmulqdq") => smallvec!["pclmul"],
         ("x86", "rdrand") => smallvec!["rdrnd"],
         ("x86", "bmi1") => smallvec!["bmi"],
@@ -243,7 +237,7 @@ pub fn target_features(sess: &Session, allow_unstable: bool) -> Vec<Symbol> {
     // (see https://reviews.llvm.org/D110413). This unstable target feature is intended for use
     // by compiler-builtins, to export the builtins with the expected, LLVM-version-dependent ABI.
     // The target feature can be dropped once we no longer support older LLVM versions.
-    if sess.is_nightly_build() && get_version() >= (14, 0, 0) {
+    if sess.is_nightly_build() {
         features.push(Symbol::intern("llvm14-builtins-abi"));
     }
     features
@@ -494,11 +488,6 @@ pub(crate) fn global_llvm_features(sess: &Session, diagnostics: bool) -> Vec<Str
         .flatten();
     features.extend(feats);
 
-    // FIXME: Move v8a to target definition list when earliest supported LLVM is 14.
-    if get_version() >= (14, 0, 0) && sess.target.arch == "aarch64" {
-        features.push("+v8a".into());
-    }
-
     if diagnostics && let Some(f) = check_tied_features(sess, &featsmap) {
         sess.emit_err(TargetFeatureDisableOrEnable {
             features: f,

compiler/rustc_codegen_llvm/src/type_of.rs

@@ -1,6 +1,5 @@
 use crate::common::*;
 use crate::context::TypeLowering;
-use crate::llvm_util::get_version;
 use crate::type_::Type;
 use rustc_codegen_ssa::traits::*;
 use rustc_middle::bug;
@@ -43,10 +42,8 @@ fn uncached_llvm_type<'a, 'tcx>(
         // in problematically distinct types due to HRTB and subtyping (see #47638).
         // ty::Dynamic(..) |
         ty::Adt(..) | ty::Closure(..) | ty::Foreign(..) | ty::Generator(..) | ty::Str
-            // For performance reasons we use names only when emitting LLVM IR. Unless we are on
-            // LLVM < 14, where the use of unnamed types resulted in various issues, e.g., #76213,
-            // #79564, and #79246.
-            if get_version() < (14, 0, 0) || !cx.sess().fewer_names() =>
+            // For performance reasons we use names only when emitting LLVM IR.
+            if !cx.sess().fewer_names() =>
         {
             let mut name = with_no_visible_paths!(with_no_trimmed_paths!(layout.ty.to_string()));
             if let (&ty::Adt(def, _), &Variants::Single { index }) =

compiler/rustc_llvm/llvm-wrapper/PassWrapper.cpp

@@ -14,6 +14,7 @@
 #include "llvm/IR/AssemblyAnnotationWriter.h"
 #include "llvm/IR/IntrinsicInst.h"
 #include "llvm/IR/Verifier.h"
+#include "llvm/MC/TargetRegistry.h"
 #include "llvm/Object/ObjectFile.h"
 #include "llvm/Object/IRObjectFile.h"
 #include "llvm/Passes/PassBuilder.h"
@@ -25,11 +26,6 @@
 #include "llvm/Support/VirtualFileSystem.h"
 #endif
 #include "llvm/Support/Host.h"
-#if LLVM_VERSION_LT(14, 0)
-#include "llvm/Support/TargetRegistry.h"
-#else
-#include "llvm/MC/TargetRegistry.h"
-#endif
 #include "llvm/Target/TargetMachine.h"
 #include "llvm/Transforms/IPO/PassManagerBuilder.h"
 #include "llvm/Transforms/IPO/AlwaysInliner.h"
@@ -267,10 +263,6 @@ enum class LLVMRustPassBuilderOptLevel {
   Oz,
 };
 
-#if LLVM_VERSION_LT(14,0)
-using OptimizationLevel = PassBuilder::OptimizationLevel;
-#endif
-
 static OptimizationLevel fromRust(LLVMRustPassBuilderOptLevel Level) {
   switch (Level) {
   case LLVMRustPassBuilderOptLevel::O0:
@@ -747,27 +739,18 @@ LLVMRustOptimize(
 
   if (SanitizerOptions) {
     if (SanitizerOptions->SanitizeMemory) {
-#if LLVM_VERSION_GE(14, 0)
       MemorySanitizerOptions Options(
           SanitizerOptions->SanitizeMemoryTrackOrigins,
           SanitizerOptions->SanitizeMemoryRecover,
           /*CompileKernel=*/false,
           /*EagerChecks=*/true);
-#else
-      MemorySanitizerOptions Options(
-          SanitizerOptions->SanitizeMemoryTrackOrigins,
-          SanitizerOptions->SanitizeMemoryRecover,
-          /*CompileKernel=*/false);
-#endif
       OptimizerLastEPCallbacks.push_back(
         [Options](ModulePassManager &MPM, OptimizationLevel Level) {
-#if LLVM_VERSION_GE(14, 0) && LLVM_VERSION_LT(16, 0)
+#if LLVM_VERSION_LT(16, 0)
           MPM.addPass(ModuleMemorySanitizerPass(Options));
+          MPM.addPass(createModuleToFunctionPassAdaptor(MemorySanitizerPass(Options)));
 #else
           MPM.addPass(MemorySanitizerPass(Options));
-#endif
-#if LLVM_VERSION_LT(16, 0)
-          MPM.addPass(createModuleToFunctionPassAdaptor(MemorySanitizerPass(Options)));
 #endif
         }
       );
@@ -776,11 +759,7 @@ LLVMRustOptimize(
     if (SanitizerOptions->SanitizeThread) {
       OptimizerLastEPCallbacks.push_back(
         [](ModulePassManager &MPM, OptimizationLevel Level) {
-#if LLVM_VERSION_GE(14, 0)
           MPM.addPass(ModuleThreadSanitizerPass());
-#else
-          MPM.addPass(ThreadSanitizerPass());
-#endif
           MPM.addPass(createModuleToFunctionPassAdaptor(ThreadSanitizerPass()));
         }
       );
@@ -792,7 +771,6 @@ LLVMRustOptimize(
 #if LLVM_VERSION_LT(15, 0)
           MPM.addPass(RequireAnalysisPass<ASanGlobalsMetadataAnalysis, Module>());
 #endif
-#if LLVM_VERSION_GE(14, 0)
           AddressSanitizerOptions opts = AddressSanitizerOptions{
             /*CompileKernel=*/false,
             SanitizerOptions->SanitizeAddressRecover,
@@ -803,29 +781,17 @@ LLVMRustOptimize(
           MPM.addPass(ModuleAddressSanitizerPass(opts));
 #else
           MPM.addPass(AddressSanitizerPass(opts));
-#endif
-#else
-          MPM.addPass(ModuleAddressSanitizerPass(
-              /*CompileKernel=*/false, SanitizerOptions->SanitizeAddressRecover));
-          MPM.addPass(createModuleToFunctionPassAdaptor(AddressSanitizerPass(
-              /*CompileKernel=*/false, SanitizerOptions->SanitizeAddressRecover,
-              /*UseAfterScope=*/true)));
 #endif
         }
       );
     }
     if (SanitizerOptions->SanitizeHWAddress) {
       OptimizerLastEPCallbacks.push_back(
         [SanitizerOptions](ModulePassManager &MPM, OptimizationLevel Level) {
-#if LLVM_VERSION_GE(14, 0)
           HWAddressSanitizerOptions opts(
               /*CompileKernel=*/false, SanitizerOptions->SanitizeHWAddressRecover,
               /*DisableOptimization=*/false);
           MPM.addPass(HWAddressSanitizerPass(opts));
-#else
-          MPM.addPass(HWAddressSanitizerPass(
-              /*CompileKernel=*/false, SanitizerOptions->SanitizeHWAddressRecover));
-#endif
         }
       );
     }
@@ -1328,11 +1294,7 @@ extern "C" bool
 LLVMRustPrepareThinLTOResolveWeak(const LLVMRustThinLTOData *Data, LLVMModuleRef M) {
   Module &Mod = *unwrap(M);
   const auto &DefinedGlobals = Data->ModuleToDefinedGVSummaries.lookup(Mod.getModuleIdentifier());
-#if LLVM_VERSION_GE(14, 0)
   thinLTOFinalizeInModule(Mod, DefinedGlobals, /*PropagateAttrs=*/true);
-#else
-  thinLTOResolvePrevailingInModule(Mod, DefinedGlobals);
-#endif
   return true;
 }
 

compiler/rustc_llvm/llvm-wrapper/RustWrapper.cpp

@@ -63,11 +63,7 @@ static LLVM_THREAD_LOCAL char *LastError;
 //
 // Notably it exits the process with code 101, unlike LLVM's default of 1.
 static void FatalErrorHandler(void *UserData,
-#if LLVM_VERSION_LT(14, 0)
-                              const std::string& Reason,
-#else
                               const char* Reason,
-#endif
                               bool GenCrashDiag) {
   // Do the same thing that the default error handler does.
   std::cerr << "LLVM ERROR: " << Reason << std::endl;
@@ -249,18 +245,10 @@ static Attribute::AttrKind fromRust(LLVMRustAttribute Kind) {
 template<typename T> static inline void AddAttributes(T *t, unsigned Index,
                                                       LLVMAttributeRef *Attrs, size_t AttrsLen) {
   AttributeList PAL = t->getAttributes();
-  AttributeList PALNew;
-#if LLVM_VERSION_LT(14, 0)
-  AttrBuilder B;
-  for (LLVMAttributeRef Attr : makeArrayRef(Attrs, AttrsLen))
-    B.addAttribute(unwrap(Attr));
-  PALNew = PAL.addAttributes(t->getContext(), Index, B);
-#else
   AttrBuilder B(t->getContext());
   for (LLVMAttributeRef Attr : ArrayRef<LLVMAttributeRef>(Attrs, AttrsLen))
     B.addAttribute(unwrap(Attr));
-  PALNew = PAL.addAttributesAtIndex(t->getContext(), Index, B);
-#endif
+  AttributeList PALNew = PAL.addAttributesAtIndex(t->getContext(), Index, B);
   t->setAttributes(PALNew);
 }
 

compiler/rustc_target/src/spec/aarch64_be_unknown_linux_gnu.rs

@@ -8,7 +8,7 @@ pub fn target() -> Target {
         data_layout: "E-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128".into(),
         arch: "aarch64".into(),
         options: TargetOptions {
-            features: "+outline-atomics".into(),
+            features: "+v8a,+outline-atomics".into(),
             max_atomic_width: Some(128),
             mcount: "\u{1}_mcount".into(),
             endian: Endian::Big,

compiler/rustc_target/src/spec/aarch64_be_unknown_linux_gnu_ilp32.rs

@@ -12,7 +12,7 @@ pub fn target() -> Target {
         arch: "aarch64".into(),
         options: TargetOptions {
             abi: "ilp32".into(),
-            features: "+outline-atomics".into(),
+            features: "+v8a,+outline-atomics".into(),
             mcount: "\u{1}_mcount".into(),
             endian: Endian::Big,
             ..base

compiler/rustc_target/src/spec/aarch64_kmc_solid_asp3.rs

@@ -9,7 +9,7 @@ pub fn target() -> Target {
         arch: "aarch64".into(),
         options: TargetOptions {
             linker: Some("aarch64-kmc-elf-gcc".into()),
-            features: "+neon,+fp-armv8".into(),
+            features: "+v8a,+neon,+fp-armv8".into(),
             relocation_model: RelocModel::Static,
             disable_redzone: true,
             max_atomic_width: Some(128),

compiler/rustc_target/src/spec/aarch64_linux_android.rs

@@ -13,7 +13,7 @@ pub fn target() -> Target {
             max_atomic_width: Some(128),
             // As documented in https://developer.android.com/ndk/guides/cpu-features.html
             // the neon (ASIMD) and FP must exist on all android aarch64 targets.
-            features: "+neon,+fp-armv8".into(),
+            features: "+v8a,+neon,+fp-armv8".into(),
             supported_sanitizers: SanitizerSet::CFI
                 | SanitizerSet::HWADDRESS
                 | SanitizerSet::MEMTAG

compiler/rustc_target/src/spec/aarch64_nintendo_switch_freestanding.rs

@@ -10,6 +10,7 @@ pub fn target() -> Target {
         data_layout: "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128".into(),
         arch: "aarch64".into(),
         options: TargetOptions {
+            features: "+v8a".into(),
             linker_flavor: LinkerFlavor::Gnu(Cc::No, Lld::Yes),
             linker: Some("rust-lld".into()),
             link_script: Some(LINKER_SCRIPT.into()),

compiler/rustc_target/src/spec/aarch64_pc_windows_gnullvm.rs

@@ -3,7 +3,7 @@ use crate::spec::Target;
 pub fn target() -> Target {
     let mut base = super::windows_gnullvm_base::opts();
     base.max_atomic_width = Some(128);
-    base.features = "+neon,+fp-armv8".into();
+    base.features = "+v8a,+neon,+fp-armv8".into();
     base.linker = Some("aarch64-w64-mingw32-clang".into());
 
     Target {

compiler/rustc_target/src/spec/aarch64_pc_windows_msvc.rs

@@ -3,7 +3,7 @@ use crate::spec::Target;
 pub fn target() -> Target {
     let mut base = super::windows_msvc_base::opts();
     base.max_atomic_width = Some(128);
-    base.features = "+neon,+fp-armv8".into();
+    base.features = "+v8a,+neon,+fp-armv8".into();
 
     Target {
         llvm_target: "aarch64-pc-windows-msvc".into(),

compiler/rustc_target/src/spec/aarch64_unknown_freebsd.rs

@@ -7,6 +7,7 @@ pub fn target() -> Target {
         data_layout: "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128".into(),
         arch: "aarch64".into(),
         options: TargetOptions {
+            features: "+v8a".into(),
             max_atomic_width: Some(128),
             supported_sanitizers: SanitizerSet::ADDRESS
                 | SanitizerSet::CFI

compiler/rustc_target/src/spec/aarch64_unknown_fuchsia.rs

@@ -7,6 +7,7 @@ pub fn target() -> Target {
         data_layout: "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128".into(),
         arch: "aarch64".into(),
         options: TargetOptions {
+            features: "+v8a".into(),
             max_atomic_width: Some(128),
             supported_sanitizers: SanitizerSet::ADDRESS
                 | SanitizerSet::CFI

compiler/rustc_target/src/spec/aarch64_unknown_hermit.rs

@@ -3,7 +3,7 @@ use crate::spec::Target;
 pub fn target() -> Target {
     let mut base = super::hermit_base::opts();
     base.max_atomic_width = Some(128);
-    base.features = "+strict-align,+neon,+fp-armv8".into();
+    base.features = "+v8a,+strict-align,+neon,+fp-armv8".into();
 
     Target {
         llvm_target: "aarch64-unknown-hermit".into(),

compiler/rustc_target/src/spec/aarch64_unknown_linux_gnu.rs

@@ -7,7 +7,7 @@ pub fn target() -> Target {
         data_layout: "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128".into(),
         arch: "aarch64".into(),
         options: TargetOptions {
-            features: "+outline-atomics".into(),
+            features: "+v8a,+outline-atomics".into(),
             mcount: "\u{1}_mcount".into(),
             max_atomic_width: Some(128),
             supported_sanitizers: SanitizerSet::ADDRESS

compiler/rustc_target/src/spec/aarch64_unknown_linux_gnu_ilp32.rs

@@ -8,7 +8,7 @@ pub fn target() -> Target {
         arch: "aarch64".into(),
         options: TargetOptions {
             abi: "ilp32".into(),
-            features: "+outline-atomics".into(),
+            features: "+v8a,+outline-atomics".into(),
             max_atomic_width: Some(128),
             mcount: "\u{1}_mcount".into(),
             ..super::linux_gnu_base::opts()

compiler/rustc_target/src/spec/aarch64_unknown_linux_musl.rs

@@ -4,6 +4,7 @@ pub fn target() -> Target {
     let mut base = super::linux_musl_base::opts();
     base.max_atomic_width = Some(128);
     base.supports_xray = true;
+    base.features = "+v8a".into();
 
     Target {
         llvm_target: "aarch64-unknown-linux-musl".into(),

compiler/rustc_target/src/spec/aarch64_unknown_netbsd.rs

@@ -7,6 +7,7 @@ pub fn target() -> Target {
         data_layout: "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128".into(),
         arch: "aarch64".into(),
         options: TargetOptions {
+            features: "+v8a".into(),
             mcount: "__mcount".into(),
             max_atomic_width: Some(128),
             ..super::netbsd_base::opts()