compiler, runtime: add layout parameter to runtime.alloc

This layout parameter is currently always nil and ignored, but will
eventually contain a pointer to a memory layout.

This commit also adds module verification to the transform tests, as I
found out that it didn't (and therefore didn't initially catch all
bugs).

6 files changed, 35 insertions, 35 deletions

diff --git a/transform/testdata/allocs.ll b/transform/testdata/allocs.ll
index 42932f4fe..58af2ea82 100644
--- a/transform/testdata/allocs.ll
+++ b/transform/testdata/allocs.ll
@@ -3,11 +3,11 @@ target triple = "armv7m-none-eabi"
 
 @runtime.zeroSizedAlloc = internal global i8 0, align 1
 
-declare nonnull i8* @runtime.alloc(i32)
+declare nonnull i8* @runtime.alloc(i32, i8*)
 
 ; Test allocating a single int (i32) that should be allocated on the stack.
 define void @testInt() {
-  %1 = call i8* @runtime.alloc(i32 4)
+  %1 = call i8* @runtime.alloc(i32 4, i8* null)
   %2 = bitcast i8* %1 to i32*
   store i32 5, i32* %2
   ret void
@@ -16,7 +16,7 @@ define void @testInt() {
 ; Test allocating an array of 3 i16 values that should be allocated on the
 ; stack.
 define i16 @testArray() {
-  %1 = call i8* @runtime.alloc(i32 6)
+  %1 = call i8* @runtime.alloc(i32 6, i8* null)
   %2 = bitcast i8* %1 to i16*
   %3 = getelementptr i16, i16* %2, i32 1
   store i16 5, i16* %3
@@ -28,14 +28,14 @@ define i16 @testArray() {
 ; Call a function that will let the pointer escape, so the heap-to-stack
 ; transform shouldn't be applied.
 define void @testEscapingCall() {
-  %1 = call i8* @runtime.alloc(i32 4)
+  %1 = call i8* @runtime.alloc(i32 4, i8* null)
   %2 = bitcast i8* %1 to i32*
   %3 = call i32* @escapeIntPtr(i32* %2)
   ret void
 }
 
 define void @testEscapingCall2() {
-  %1 = call i8* @runtime.alloc(i32 4)
+  %1 = call i8* @runtime.alloc(i32 4, i8* null)
   %2 = bitcast i8* %1 to i32*
   %3 = call i32* @escapeIntPtrSometimes(i32* %2, i32* %2)
   ret void
@@ -43,7 +43,7 @@ define void @testEscapingCall2() {
 
 ; Call a function that doesn't let the pointer escape.
 define void @testNonEscapingCall() {
-  %1 = call i8* @runtime.alloc(i32 4)
+  %1 = call i8* @runtime.alloc(i32 4, i8* null)
   %2 = bitcast i8* %1 to i32*
   %3 = call i32* @noescapeIntPtr(i32* %2)
   ret void
@@ -51,7 +51,7 @@ define void @testNonEscapingCall() {
 
 ; Return the allocated value, which lets it escape.
 define i32* @testEscapingReturn() {
-  %1 = call i8* @runtime.alloc(i32 4)
+  %1 = call i8* @runtime.alloc(i32 4, i8* null)
   %2 = bitcast i8* %1 to i32*
   ret i32* %2
 }
@@ -61,7 +61,7 @@ define void @testNonEscapingLoop() {
 entry:
   br label %loop
 loop:
-  %0 = call i8* @runtime.alloc(i32 4)
+  %0 = call i8* @runtime.alloc(i32 4, i8* null)
   %1 = bitcast i8* %0 to i32*
   %2 = call i32* @noescapeIntPtr(i32* %1)
   %3 = icmp eq i32* null, %2
@@ -72,7 +72,7 @@ end:
 
 ; Test a zero-sized allocation.
 define void @testZeroSizedAlloc() {
-  %1 = call i8* @runtime.alloc(i32 0)
+  %1 = call i8* @runtime.alloc(i32 0, i8* null)
   %2 = bitcast i8* %1 to i32*
   %3 = call i32* @noescapeIntPtr(i32* %2)
   ret void
diff --git a/transform/testdata/allocs.out.ll b/transform/testdata/allocs.out.ll
index 4543ac177..ec7c4c59b 100644
--- a/transform/testdata/allocs.out.ll
+++ b/transform/testdata/allocs.out.ll
@@ -3,7 +3,7 @@ target triple = "armv7m-none-eabi"
 
 @runtime.zeroSizedAlloc = internal global i8 0, align 1
 
-declare nonnull i8* @runtime.alloc(i32)
+declare nonnull i8* @runtime.alloc(i32, i8*)
 
 define void @testInt() {
   %stackalloc.alloca = alloca [1 x i32], align 4
@@ -25,14 +25,14 @@ define i16 @testArray() {
 }
 
 define void @testEscapingCall() {
-  %1 = call i8* @runtime.alloc(i32 4)
+  %1 = call i8* @runtime.alloc(i32 4, i8* null)
   %2 = bitcast i8* %1 to i32*
   %3 = call i32* @escapeIntPtr(i32* %2)
   ret void
 }
 
 define void @testEscapingCall2() {
-  %1 = call i8* @runtime.alloc(i32 4)
+  %1 = call i8* @runtime.alloc(i32 4, i8* null)
   %2 = bitcast i8* %1 to i32*
   %3 = call i32* @escapeIntPtrSometimes(i32* %2, i32* %2)
   ret void
@@ -47,7 +47,7 @@ define void @testNonEscapingCall() {
 }
 
 define i32* @testEscapingReturn() {
-  %1 = call i8* @runtime.alloc(i32 4)
+  %1 = call i8* @runtime.alloc(i32 4, i8* null)
   %2 = bitcast i8* %1 to i32*
   ret i32* %2
 }
diff --git a/transform/testdata/coroutines.ll b/transform/testdata/coroutines.ll
index 94462bbcb..c5f711d25 100644
--- a/transform/testdata/coroutines.ll
+++ b/transform/testdata/coroutines.ll
@@ -9,7 +9,7 @@ declare void @"internal/task.Pause"(i8*, i8*)
 
 declare void @runtime.scheduler(i8*, i8*)
 
-declare i8* @runtime.alloc(i32, i8*, i8*)
+declare i8* @runtime.alloc(i32, i8*, i8*, i8*)
 declare void @runtime.free(i8*, i8*, i8*)
 
 declare %"internal/task.Task"* @"internal/task.Current"(i8*, i8*)
diff --git a/transform/testdata/coroutines.out.ll b/transform/testdata/coroutines.out.ll
index fa6eb287d..d4a49a5e4 100644
--- a/transform/testdata/coroutines.out.ll
+++ b/transform/testdata/coroutines.out.ll
@@ -10,7 +10,7 @@ declare void @"internal/task.Pause"(i8*, i8*)
 
 declare void @runtime.scheduler(i8*, i8*)
 
-declare i8* @runtime.alloc(i32, i8*, i8*)
+declare i8* @runtime.alloc(i32, i8*, i8*, i8*)
 
 declare void @runtime.free(i8*, i8*, i8*)
 
@@ -66,7 +66,7 @@ entry:
 define void @ditchTail(i32 %0, i64 %1, i8* %2, i8* %parentHandle) {
 entry:
   %task.current = bitcast i8* %parentHandle to %"internal/task.Task"*
-  %ret.ditch = call i8* @runtime.alloc(i32 4, i8* undef, i8* undef)
+  %ret.ditch = call i8* @runtime.alloc(i32 4, i8* null, i8* undef, i8* undef)
   call void @"(*internal/task.Task).setReturnPtr"(%"internal/task.Task"* %task.current, i8* %ret.ditch, i8* undef, i8* undef)
   %3 = call i32 @delayedValue(i32 %0, i64 %1, i8* undef, i8* %parentHandle)
   ret void
@@ -85,7 +85,7 @@ entry:
   %ret.ptr = call i8* @"(*internal/task.Task).getReturnPtr"(%"internal/task.Task"* %task.current, i8* undef, i8* undef)
   %ret.ptr.bitcast = bitcast i8* %ret.ptr to i32*
   store i32 %0, i32* %ret.ptr.bitcast, align 4
-  %ret.alternate = call i8* @runtime.alloc(i32 4, i8* undef, i8* undef)
+  %ret.alternate = call i8* @runtime.alloc(i32 4, i8* null, i8* undef, i8* undef)
   call void @"(*internal/task.Task).setReturnPtr"(%"internal/task.Task"* %task.current, i8* %ret.alternate, i8* undef, i8* undef)
   %4 = call i32 @delayedValue(i32 %1, i64 %2, i8* undef, i8* %parentHandle)
   ret i32 undef
@@ -96,7 +96,7 @@ entry:
   %call.return = alloca i32, align 4
   %coro.id = call token @llvm.coro.id(i32 0, i8* null, i8* null, i8* null)
   %coro.size = call i32 @llvm.coro.size.i32()
-  %coro.alloc = call i8* @runtime.alloc(i32 %coro.size, i8* undef, i8* undef)
+  %coro.alloc = call i8* @runtime.alloc(i32 %coro.size, i8* null, i8* undef, i8* undef)
   %coro.state = call i8* @llvm.coro.begin(token %coro.id, i8* %coro.alloc)
   %task.current2 = bitcast i8* %parentHandle to %"internal/task.Task"*
   %task.state.parent = call i8* @"(*internal/task.Task).setState"(%"internal/task.Task"* %task.current2, i8* %coro.state, i8* undef, i8* undef)
@@ -143,7 +143,7 @@ entry:
   %a = alloca i8, align 1
   %coro.id = call token @llvm.coro.id(i32 0, i8* null, i8* null, i8* null)
   %coro.size = call i32 @llvm.coro.size.i32()
-  %coro.alloc = call i8* @runtime.alloc(i32 %coro.size, i8* undef, i8* undef)
+  %coro.alloc = call i8* @runtime.alloc(i32 %coro.size, i8* null, i8* undef, i8* undef)
   %coro.state = call i8* @llvm.coro.begin(token %coro.id, i8* %coro.alloc)
   %task.current = bitcast i8* %parentHandle to %"internal/task.Task"*
   %task.state.parent = call i8* @"(*internal/task.Task).setState"(%"internal/task.Task"* %task.current, i8* %coro.state, i8* undef, i8* undef)
@@ -180,7 +180,7 @@ entry:
   %a = alloca i8, align 1
   %coro.id = call token @llvm.coro.id(i32 0, i8* null, i8* null, i8* null)
   %coro.size = call i32 @llvm.coro.size.i32()
-  %coro.alloc = call i8* @runtime.alloc(i32 %coro.size, i8* undef, i8* undef)
+  %coro.alloc = call i8* @runtime.alloc(i32 %coro.size, i8* null, i8* undef, i8* undef)
   %coro.state = call i8* @llvm.coro.begin(token %coro.id, i8* %coro.alloc)
   %task.current = bitcast i8* %parentHandle to %"internal/task.Task"*
   %task.state.parent = call i8* @"(*internal/task.Task).setState"(%"internal/task.Task"* %task.current, i8* %coro.state, i8* undef, i8* undef)
@@ -225,7 +225,7 @@ define i8 @usePtr(i8* %0, i8* %1, i8* %parentHandle) {
 entry:
   %coro.id = call token @llvm.coro.id(i32 0, i8* null, i8* null, i8* null)
   %coro.size = call i32 @llvm.coro.size.i32()
-  %coro.alloc = call i8* @runtime.alloc(i32 %coro.size, i8* undef, i8* undef)
+  %coro.alloc = call i8* @runtime.alloc(i32 %coro.size, i8* null, i8* undef, i8* undef)
   %coro.state = call i8* @llvm.coro.begin(token %coro.id, i8* %coro.alloc)
   %task.current = bitcast i8* %parentHandle to %"internal/task.Task"*
   %task.state.parent = call i8* @"(*internal/task.Task).setState"(%"internal/task.Task"* %task.current, i8* %coro.state, i8* undef, i8* undef)
diff --git a/transform/testdata/gc-stackslots.ll b/transform/testdata/gc-stackslots.ll
index 8b43a83e8..130b0c990 100644
--- a/transform/testdata/gc-stackslots.ll
+++ b/transform/testdata/gc-stackslots.ll
@@ -8,7 +8,7 @@ target triple = "wasm32-unknown-unknown-wasm"
 
 declare void @runtime.trackPointer(i8* nocapture readonly)
 
-declare noalias nonnull i8* @runtime.alloc(i32)
+declare noalias nonnull i8* @runtime.alloc(i32, i8*)
 
 ; Generic function that returns a pointer (that must be tracked).
 define i8* @getPointer() {
@@ -18,7 +18,7 @@ define i8* @getPointer() {
 define i8* @needsStackSlots() {
   ; Tracked pointer. Although, in this case the value is immediately returned
   ; so tracking it is not really necessary.
-  %ptr = call i8* @runtime.alloc(i32 4)
+  %ptr = call i8* @runtime.alloc(i32 4, i8* null)
   call void @runtime.trackPointer(i8* %ptr)
   ; Restoring the stack pointer can happen at this position, before the return.
   ; This avoids issues with tail calls.
@@ -41,7 +41,7 @@ define i8* @needsStackSlots2() {
   call void @runtime.trackPointer(i8* %ptr2)
 
   ; Here is finally the point where an allocation happens.
-  %unused = call i8* @runtime.alloc(i32 4)
+  %unused = call i8* @runtime.alloc(i32 4, i8* null)
   call void @runtime.trackPointer(i8* %unused)
 
   ret i8* %ptr1
@@ -59,7 +59,7 @@ define i8* @fibNext(i8* %x, i8* %y) {
   %x.val = load i8, i8* %x
   %y.val = load i8, i8* %y
   %out.val = add i8 %x.val, %y.val
-  %out.alloc = call i8* @runtime.alloc(i32 1)
+  %out.alloc = call i8* @runtime.alloc(i32 1, i8* null)
   call void @runtime.trackPointer(i8* %out.alloc)
   store i8 %out.val, i8* %out.alloc
   ret i8* %out.alloc
@@ -67,9 +67,9 @@ define i8* @fibNext(i8* %x, i8* %y) {
 
 define i8* @allocLoop() {
 entry:
-  %entry.x = call i8* @runtime.alloc(i32 1)
+  %entry.x = call i8* @runtime.alloc(i32 1, i8* null)
   call void @runtime.trackPointer(i8* %entry.x)
-  %entry.y = call i8* @runtime.alloc(i32 1)
+  %entry.y = call i8* @runtime.alloc(i32 1, i8* null)
   call void @runtime.trackPointer(i8* %entry.y)
   store i8 1, i8* %entry.y
   br label %loop
@@ -95,7 +95,7 @@ define void @testGEPBitcast() {
   %arr = call [32 x i8]* @arrayAlloc()
   %arr.bitcast = getelementptr [32 x i8], [32 x i8]* %arr, i32 0, i32 0
   call void @runtime.trackPointer(i8* %arr.bitcast)
-  %other = call i8* @runtime.alloc(i32 1)
+  %other = call i8* @runtime.alloc(i32 1, i8* null)
   call void @runtime.trackPointer(i8* %other)
   ret void
 }
diff --git a/transform/testdata/gc-stackslots.out.ll b/transform/testdata/gc-stackslots.out.ll
index af474534c..15fab17ed 100644
--- a/transform/testdata/gc-stackslots.out.ll
+++ b/transform/testdata/gc-stackslots.out.ll
@@ -8,7 +8,7 @@ target triple = "wasm32-unknown-unknown-wasm"
 
 declare void @runtime.trackPointer(i8* nocapture readonly)
 
-declare noalias nonnull i8* @runtime.alloc(i32)
+declare noalias nonnull i8* @runtime.alloc(i32, i8*)
 
 define i8* @getPointer() {
   ret i8* @someGlobal
@@ -22,7 +22,7 @@ define i8* @needsStackSlots() {
   store %runtime.stackChainObject* %1, %runtime.stackChainObject** %2, align 4
   %3 = bitcast { %runtime.stackChainObject*, i32, i8* }* %gc.stackobject to %runtime.stackChainObject*
   store %runtime.stackChainObject* %3, %runtime.stackChainObject** @runtime.stackChainStart, align 4
-  %ptr = call i8* @runtime.alloc(i32 4)
+  %ptr = call i8* @runtime.alloc(i32 4, i8* null)
   %4 = getelementptr { %runtime.stackChainObject*, i32, i8* }, { %runtime.stackChainObject*, i32, i8* }* %gc.stackobject, i32 0, i32 2
   store i8* %ptr, i8** %4, align 4
   store %runtime.stackChainObject* %1, %runtime.stackChainObject** @runtime.stackChainStart, align 4
@@ -49,7 +49,7 @@ define i8* @needsStackSlots2() {
   %ptr2 = getelementptr i8, i8* @someGlobal, i32 0
   %7 = getelementptr { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }, { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }* %gc.stackobject, i32 0, i32 5
   store i8* %ptr2, i8** %7, align 4
-  %unused = call i8* @runtime.alloc(i32 4)
+  %unused = call i8* @runtime.alloc(i32 4, i8* null)
   %8 = getelementptr { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }, { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }* %gc.stackobject, i32 0, i32 6
   store i8* %unused, i8** %8, align 4
   store %runtime.stackChainObject* %1, %runtime.stackChainObject** @runtime.stackChainStart, align 4
@@ -72,7 +72,7 @@ define i8* @fibNext(i8* %x, i8* %y) {
   %x.val = load i8, i8* %x, align 1
   %y.val = load i8, i8* %y, align 1
   %out.val = add i8 %x.val, %y.val
-  %out.alloc = call i8* @runtime.alloc(i32 1)
+  %out.alloc = call i8* @runtime.alloc(i32 1, i8* null)
   %4 = getelementptr { %runtime.stackChainObject*, i32, i8* }, { %runtime.stackChainObject*, i32, i8* }* %gc.stackobject, i32 0, i32 2
   store i8* %out.alloc, i8** %4, align 4
   store %runtime.stackChainObject* %1, %runtime.stackChainObject** @runtime.stackChainStart, align 4
@@ -89,10 +89,10 @@ entry:
   store %runtime.stackChainObject* %0, %runtime.stackChainObject** %1, align 4
   %2 = bitcast { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }* %gc.stackobject to %runtime.stackChainObject*
   store %runtime.stackChainObject* %2, %runtime.stackChainObject** @runtime.stackChainStart, align 4
-  %entry.x = call i8* @runtime.alloc(i32 1)
+  %entry.x = call i8* @runtime.alloc(i32 1, i8* null)
   %3 = getelementptr { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }, { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }* %gc.stackobject, i32 0, i32 2
   store i8* %entry.x, i8** %3, align 4
-  %entry.y = call i8* @runtime.alloc(i32 1)
+  %entry.y = call i8* @runtime.alloc(i32 1, i8* null)
   %4 = getelementptr { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }, { %runtime.stackChainObject*, i32, i8*, i8*, i8*, i8*, i8* }* %gc.stackobject, i32 0, i32 3
   store i8* %entry.y, i8** %4, align 4
   store i8 1, i8* %entry.y, align 1
@@ -131,7 +131,7 @@ define void @testGEPBitcast() {
   %arr.bitcast = getelementptr [32 x i8], [32 x i8]* %arr, i32 0, i32 0
   %4 = getelementptr { %runtime.stackChainObject*, i32, i8*, i8* }, { %runtime.stackChainObject*, i32, i8*, i8* }* %gc.stackobject, i32 0, i32 2
   store i8* %arr.bitcast, i8** %4, align 4
-  %other = call i8* @runtime.alloc(i32 1)
+  %other = call i8* @runtime.alloc(i32 1, i8* null)
   %5 = getelementptr { %runtime.stackChainObject*, i32, i8*, i8* }, { %runtime.stackChainObject*, i32, i8*, i8* }* %gc.stackobject, i32 0, i32 3
   store i8* %other, i8** %5, align 4
   store %runtime.stackChainObject* %1, %runtime.stackChainObject** @runtime.stackChainStart, align 4