force-inline a few much-used functions

these functions are called a whole lot, and they're all comparatively small.
always inlining them gives ~0.7% performance boost on eval.

before:

  Benchmark 1: nix flakes search --no-eval-cache --offline ../nixpkgs hello
    Time (mean ± σ):      6.935 s ±  0.052 s    [User: 5.852 s, System: 0.853 s]
    Range (min … max):    6.808 s …  7.026 s    20 runs

  Benchmark 2: nix flakes eval -f ../nixpkgs/pkgs/development/haskell-modules/hackage-packages.nix
    Time (mean ± σ):     329.8 ms ±   2.7 ms    [User: 299.0 ms, System: 30.8 ms]
    Range (min … max):   326.6 ms … 336.5 ms    20 runs

  Benchmark 3: nix flakes eval --raw --impure --file expr.nix
    Time (mean ± σ):      2.655 s ±  0.038 s    [User: 2.364 s, System: 0.220 s]
    Range (min … max):    2.574 s …  2.737 s    20 runs

after:

  Benchmark 1: nix flakes search --no-eval-cache --offline ../nixpkgs hello
    Time (mean ± σ):      6.912 s ±  0.036 s    [User: 5.823 s, System: 0.856 s]
    Range (min … max):    6.849 s …  6.980 s    20 runs

  Benchmark 2: nix flakes eval -f ../nixpkgs/pkgs/development/haskell-modules/hackage-packages.nix
    Time (mean ± σ):     325.1 ms ±   2.5 ms    [User: 293.2 ms, System: 31.8 ms]
    Range (min … max):   322.2 ms … 332.8 ms    20 runs

  Benchmark 3: nix flakes eval --raw --impure --file expr.nix
    Time (mean ± σ):      2.636 s ±  0.024 s    [User: 2.352 s, System: 0.226 s]
    Range (min … max):    2.574 s …  2.681 s    20 runs

3 files changed, 77 insertions, 68 deletions

diff --git a/src/libexpr/eval-inline.hh b/src/libexpr/eval-inline.hh
index aef1f6351..3331a7643 100644
--- a/src/libexpr/eval-inline.hh
+++ b/src/libexpr/eval-inline.hh
@@ -24,6 +24,76 @@ LocalNoInlineNoReturn(void throwTypeError(const Pos & pos, const char * s, const
 }
 
 
+/* Note: Various places expect the allocated memory to be zeroed. */
+[[gnu::always_inline]]
+inline void * allocBytes(size_t n)
+{
+    void * p;
+#if HAVE_BOEHMGC
+    p = GC_MALLOC(n);
+#else
+    p = calloc(n, 1);
+#endif
+    if (!p) throw std::bad_alloc();
+    return p;
+}
+
+
+[[gnu::always_inline]]
+Value * EvalState::allocValue()
+{
+    /* We use the boehm batch allocator to speed up allocations of Values (of which there are many).
+       GC_malloc_many returns a linked list of objects of the given size, where the first word
+       of each object is also the pointer to the next object in the list. This also means that we
+       have to explicitly clear the first word of every object we take. */
+    if (!*valueAllocCache) {
+        *valueAllocCache = GC_malloc_many(sizeof(Value));
+        if (!*valueAllocCache) throw std::bad_alloc();
+    }
+
+    /* GC_NEXT is a convenience macro for accessing the first word of an object.
+       Take the first list item, advance the list to the next item, and clear the next pointer. */
+    void * p = *valueAllocCache;
+    *valueAllocCache = GC_NEXT(p);
+    GC_NEXT(p) = nullptr;
+
+    nrValues++;
+    return (Value *) p;
+}
+
+
+[[gnu::always_inline]]
+Env & EvalState::allocEnv(size_t size)
+{
+    nrEnvs++;
+    nrValuesInEnvs += size;
+
+    Env * env;
+
+    if (size != 1)
+        env = (Env *) allocBytes(sizeof(Env) + size * sizeof(Value *));
+    else {
+        /* see allocValue for explanations. */
+        if (!*env1AllocCache) {
+            *env1AllocCache = GC_malloc_many(sizeof(Env) + sizeof(Value *));
+            if (!*env1AllocCache) throw std::bad_alloc();
+        }
+
+        void * p = *env1AllocCache;
+        *env1AllocCache = GC_NEXT(p);
+        GC_NEXT(p) = nullptr;
+        env = (Env *) p;
+    }
+
+    env->type = Env::Plain;
+
+    /* We assume that env->values has been cleared by the allocator; maybeThunk() and lookupVar fromWith expect this. */
+
+    return *env;
+}
+
+
+[[gnu::always_inline]]
 void EvalState::forceValue(Value & v, const Pos & pos)
 {
     forceValue(v, [&]() { return pos; });
@@ -52,6 +122,7 @@ void EvalState::forceValue(Value & v, Callable getPos)
 }
 
 
+[[gnu::always_inline]]
 inline void EvalState::forceAttrs(Value & v, const Pos & pos)
 {
     forceAttrs(v, [&]() { return pos; });
@@ -59,6 +130,7 @@ inline void EvalState::forceAttrs(Value & v, const Pos & pos)
 
 
 template <typename Callable>
+[[gnu::always_inline]]
 inline void EvalState::forceAttrs(Value & v, Callable getPos)
 {
     forceValue(v, getPos);
@@ -67,6 +139,7 @@ inline void EvalState::forceAttrs(Value & v, Callable getPos)
 }
 
 
+[[gnu::always_inline]]
 inline void EvalState::forceList(Value & v, const Pos & pos)
 {
     forceValue(v, pos);
@@ -74,18 +147,5 @@ inline void EvalState::forceList(Value & v, const Pos & pos)
         throwTypeError(pos, "value is %1% while a list was expected", v);
 }
 
-/* Note: Various places expect the allocated memory to be zeroed. */
-inline void * allocBytes(size_t n)
-{
-    void * p;
-#if HAVE_BOEHMGC
-    p = GC_MALLOC(n);
-#else
-    p = calloc(n, 1);
-#endif
-    if (!p) throw std::bad_alloc();
-    return p;
-}
-
 
 }
diff --git a/src/libexpr/eval.cc b/src/libexpr/eval.cc
index ddbbdeaa6..038b6bb7c 100644
--- a/src/libexpr/eval.cc
+++ b/src/libexpr/eval.cc
@@ -867,59 +867,6 @@ inline Value * EvalState::lookupVar(Env * env, const ExprVar & var, bool noEval)
 }
 
 
-Value * EvalState::allocValue()
-{
-    /* We use the boehm batch allocator to speed up allocations of Values (of which there are many).
-       GC_malloc_many returns a linked list of objects of the given size, where the first word
-       of each object is also the pointer to the next object in the list. This also means that we
-       have to explicitly clear the first word of every object we take. */
-    if (!*valueAllocCache) {
-        *valueAllocCache = GC_malloc_many(sizeof(Value));
-        if (!*valueAllocCache) throw std::bad_alloc();
-    }
-
-    /* GC_NEXT is a convenience macro for accessing the first word of an object.
-       Take the first list item, advance the list to the next item, and clear the next pointer. */
-    void * p = *valueAllocCache;
-    *valueAllocCache = GC_NEXT(p);
-    GC_NEXT(p) = nullptr;
-
-    nrValues++;
-    auto v = (Value *) p;
-    return v;
-}
-
-
-Env & EvalState::allocEnv(size_t size)
-{
-    nrEnvs++;
-    nrValuesInEnvs += size;
-
-    Env * env;
-
-    if (size != 1)
-        env = (Env *) allocBytes(sizeof(Env) + size * sizeof(Value *));
-    else {
-        /* see allocValue for explanations. */
-        if (!*env1AllocCache) {
-            *env1AllocCache = GC_malloc_many(sizeof(Env) + sizeof(Value *));
-            if (!*env1AllocCache) throw std::bad_alloc();
-        }
-
-        void * p = *env1AllocCache;
-        *env1AllocCache = GC_NEXT(p);
-        GC_NEXT(p) = nullptr;
-        env = (Env *) p;
-    }
-
-    env->type = Env::Plain;
-
-    /* We assume that env->values has been cleared by the allocator; maybeThunk() and lookupVar fromWith expect this. */
-
-    return *env;
-}
-
-
 void EvalState::mkList(Value & v, size_t size)
 {
     v.mkList(size);
diff --git a/src/libexpr/eval.hh b/src/libexpr/eval.hh
index 41384f044..d2efe8a47 100644
--- a/src/libexpr/eval.hh
+++ b/src/libexpr/eval.hh
@@ -350,8 +350,8 @@ public:
     void autoCallFunction(Bindings & args, Value & fun, Value & res);
 
     /* Allocation primitives. */
-    Value * allocValue();
-    Env & allocEnv(size_t size);
+    inline Value * allocValue();
+    inline Env & allocEnv(size_t size);
 
     Value * allocAttr(Value & vAttrs, const Symbol & name);
     Value * allocAttr(Value & vAttrs, std::string_view name);
@@ -512,3 +512,5 @@ extern EvalSettings evalSettings;
 static const std::string corepkgsPrefix{"/__corepkgs__/"};
 
 }
+
+#include "eval-inline.hh"