[WebAssembly] Create a __stack_limit variable

[sunfish]

Summary

Create a stack_limit variable, similar to stack_pointer, to mark
the end of the stack. This will support stack overflow checks.

[sbc100]

Does this need to be wasm global? Or can it just be regular data symbol like globalBase? i.e. do we want fast access to it without a load?

lld/wasm/Writer.cpp

221

Maybe avoid the extra local and just do this 4 lines above. You can also remove the now redundant log("mem: stack base = " + Twine(memoryPtr)); line?

[sunfish]

In D70515#1754288, @sbc100 wrote:

Does this need to be wasm global? Or can it just be regular data symbol like globalBase? i.e. do we want fast access to it without a load?

I made it a global because I was thinking about multiple threads, where each thread has its own stack. But that's true, we could do a data symbol in the single-threaded case. I'll look into that.

[sbc100]

I think if you want it to be visible as a C symbol, with an address etc, then it would need to be data symbol. It can be TLS for multi-threading I guess?

[sbc100]

Sorry, I have somehow forgotten about this change when I created https://reviews.llvm.org/D135910.

I'm not sure if something like this also needed? Unlike the two variables I created this global is thread local which means it could be more useful in mutli-thread environments. In emscripten we currently inject a (TLS) wasm global to aid with stack checking: https://github.com/emscripten-core/emscripten/blob/af961ad5c4c278ec510f0b7f7d522a95ee5a90f8/system/lib/compiler-rt/stack_limits.S#L21-L24.. but I suppose having the linker so it could also make sense.

Do you think this change is still useful and/or better than https://reviews.llvm.org/D135910?

[sunfish]

In D70515#3857014, @sbc100 wrote:

Do you think this change is still useful and/or better than https://reviews.llvm.org/D135910?

I agree that D135910 is the better approach here.

lld/test/wasm/data-layout.ll

 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           66624
 ; CHECK-NEXT:       - Index:           1
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
-; CHECK-NEXT:           Value:           1080
+; CHECK-NEXT:           Value:           1088
 ; CHECK-NEXT:       - Index:           2
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
+; CHECK-NEXT:           Value:           1080
+; CHECK-NEXT:       - Index:           3
+; CHECK-NEXT:         Type:            I32
+; CHECK-NEXT:         Mutable:         false
+; CHECK-NEXT:         InitExpr:
+; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           66624
 
 ; CHECK:        - Type:            DATA
 ; CHECK-NEXT:     Segments:
 ; CHECK-NEXT:       - SectionOffset:   7
 ; CHECK-NEXT:         InitFlags:       0
 ; CHECK-NEXT:         Offset:
 ; CHECK-NEXT:           Opcode:          I32_CONST

lld/test/wasm/gc-sections.ll

 ; NO-GC-NEXT:     Globals:
 ; NO-GC-NEXT:       - Index:           0
 ; NO-GC-NEXT:         Type:            I32
 ; NO-GC-NEXT:         Mutable:         true
 ; NO-GC-NEXT:         InitExpr:
 ; NO-GC-NEXT:           Opcode:          I32_CONST
 ; NO-GC-NEXT:           Value:           66576
 ; NO-GC-NEXT:       - Index:       1
+; NO-GC-NEXT:         Type:        I32
+; NO-GC-NEXT:         Mutable:         false
+; NO-GC-NEXT:         InitExpr:
+; NO-GC-NEXT:           Opcode:          I32_CONST
+; NO-GC-NEXT:           Value:           1040
+; NO-GC-NEXT:       - Index:       2
 ; NO-GC-NEXT:         Type:        I64
 ; NO-GC-NEXT:         Mutable:     true
 ; NO-GC-NEXT:         InitExpr:
 ; NO-GC-NEXT:           Opcode:          I64_CONST
 ; NO-GC-NEXT:           Value:           123
-; NO-GC-NEXT:       - Index:       2
+; NO-GC-NEXT:       - Index:       3
 ; NO-GC-NEXT:         Type:        I64
 ; NO-GC-NEXT:         Mutable:     true
 ; NO-GC-NEXT:         InitExpr:
 ; NO-GC-NEXT:           Opcode:          I64_CONST
 ; NO-GC-NEXT:           Value:           456
 
 ; NO-GC:        - Type:            DATA
 ; NO-GC-NEXT:     Segments:

lld/test/wasm/no-tls.test

 ; __stack_pointer
 CHECK-NEXT:      - Index:           0
 CHECK-NEXT:        Type:            I32
 CHECK-NEXT:        Mutable:         true
 CHECK-NEXT:        InitExpr:
 CHECK-NEXT:          Opcode:          I32_CONST
 CHECK-NEXT:          Value:           66576
 
-; __tls_base
+; __stack_limit
 CHECK-NEXT:      - Index:           1
 CHECK-NEXT:        Type:            I32
+CHECK-NEXT:        Mutable:         false
+CHECK-NEXT:        InitExpr:
+CHECK-NEXT:          Opcode:          I32_CONST
+CHECK-NEXT:          Value:           1040
+
+; __tls_base
+CHECK-NEXT:      - Index:           2
+CHECK-NEXT:        Type:            I32
 CHECK-NEXT:        Mutable:         true
 CHECK-NEXT:        InitExpr:
 CHECK-NEXT:          Opcode:          I32_CONST
 CHECK-NEXT:          Value:           0
 
 ; __tls_size
-CHECK-NEXT:      - Index:           2
+CHECK-NEXT:      - Index:           3
 CHECK-NEXT:        Type:            I32
 CHECK-NEXT:        Mutable:         false
 CHECK-NEXT:        InitExpr:
 CHECK-NEXT:          Opcode:          I32_CONST
 CHECK-NEXT:          Value:           0
 
 ; __tls_align
-CHECK-NEXT:      - Index:           3
+CHECK-NEXT:      - Index:           4
 CHECK-NEXT:        Type:            I32
 CHECK-NEXT:        Mutable:         false
 CHECK-NEXT:        InitExpr:
 CHECK-NEXT:          Opcode:          I32_CONST
 CHECK-NEXT:          Value:           1

lld/test/wasm/pic-static.ll

 ; __stack_pointer
 ; CHECK-NEXT:       - Index:           0
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         true
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           66576
 
-; GOT.func.ret32
+; __stack_limit
 ; CHECK-NEXT:       - Index:           1
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
+; CHECK-NEXT:           Value:           1040
+
+; GOT.func.ret32
+; CHECK-NEXT:       - Index:           2
+; CHECK-NEXT:         Type:            I32
+; CHECK-NEXT:         Mutable:         false
+; CHECK-NEXT:         InitExpr:
+; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           1
 
 ; GOT.func.missing_function
-; CHECK-NEXT:       - Index:           2
+; CHECK-NEXT:       - Index:           3
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           2
 
 ; __table_base
-; CHECK-NEXT:       - Index:           3
+; CHECK-NEXT:       - Index:           4
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           1
 
 ; GOT.mem.missing_float
-; CHECK-NEXT:       - Index:           4
+; CHECK-NEXT:       - Index:           5
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           0
 
 ; GOT.mem.global_float
-; CHECK-NEXT:       - Index:           5
+; CHECK-NEXT:       - Index:           6
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           1024
 
 ; GOT.mem.ret32_ptr
-; CHECK-NEXT:       - Index:           6
+; CHECK-NEXT:       - Index:           7
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           1032
 
 ; __memory_base
-; CHECK-NEXT:       - Index:           7
+; CHECK-NEXT:       - Index:           8
 ; CHECK-NEXT:         Type:            I32
 ; CHECK-NEXT:         Mutable:         false
 ; CHECK-NEXT:         InitExpr:
 ; CHECK-NEXT:           Opcode:          I32_CONST
 ; CHECK-NEXT:           Value:           0

lld/test/wasm/tls-align.ll

 ; CHECK-NEXT:   Globals:
 ; CHECK-NEXT:     - Index:           0
 ; CHECK-NEXT:       Type:            I32
 ; CHECK-NEXT:       Mutable:         true
 ; CHECK-NEXT:       InitExpr:
 ; CHECK-NEXT:         Opcode:          I32_CONST
 ; CHECK-NEXT:         Value:           66592
 
-; __tls_base
+; __stack_limit
 ; CHECK-NEXT:     - Index:           1
 ; CHECK-NEXT:       Type:            I32
+; CHECK-NEXT:       Mutable:         false
+; CHECK-NEXT:       InitExpr:
+; CHECK-NEXT:         Opcode:          I32_CONST
+; CHECK-NEXT:         Value:           1056
+
+; __tls_base
+; CHECK-NEXT:     - Index:           2
+; CHECK-NEXT:       Type:            I32
 ; CHECK-NEXT:       Mutable:         true
 ; CHECK-NEXT:       InitExpr:
 ; CHECK-NEXT:         Opcode:          I32_CONST
 ; CHECK-NEXT:         Value:           0
 
 ; __tls_size
-; CHECK-NEXT:     - Index:           2
+; CHECK-NEXT:     - Index:           3
 ; CHECK-NEXT:       Type:            I32
 ; CHECK-NEXT:       Mutable:         false
 ; CHECK-NEXT:       InitExpr:
 ; CHECK-NEXT:         Opcode:          I32_CONST
 ; CHECK-NEXT:         Value:           20
 
 ; __tls_align
-; CHECK-NEXT:     - Index:           3
+; CHECK-NEXT:     - Index:           4
 ; CHECK-NEXT:       Type:            I32
 ; CHECK-NEXT:       Mutable:         false
 ; CHECK-NEXT:       InitExpr:
 ; CHECK-NEXT:         Opcode:          I32_CONST
 ; CHECK-NEXT:         Value:           16

lld/test/wasm/tls.ll

 ; CHECK-NEXT:   Globals:
 ; CHECK-NEXT:     - Index:           0
 ; CHECK-NEXT:       Type:            I32
 ; CHECK-NEXT:       Mutable:         true
 ; CHECK-NEXT:       InitExpr:
 ; CHECK-NEXT:         Opcode:          I32_CONST
 ; CHECK-NEXT:         Value:           66576
 
-; __tls_base
+; __stack_limit
 ; CHECK-NEXT:     - Index:           1
 ; CHECK-NEXT:       Type:            I32
+; CHECK-NEXT:       Mutable:         false
+; CHECK-NEXT:       InitExpr:
+; CHECK-NEXT:         Opcode:          I32_CONST
+; CHECK-NEXT:         Value:           1040
+
+; __tls_base
+; CHECK-NEXT:     - Index:           2
+; CHECK-NEXT:       Type:            I32
 ; CHECK-NEXT:       Mutable:         true
 ; CHECK-NEXT:       InitExpr:
 ; CHECK-NEXT:         Opcode:          I32_CONST
 ; CHECK-NEXT:         Value:           0
 
 ; __tls_size
-; CHECK-NEXT:     - Index:           2
+; CHECK-NEXT:     - Index:           3
 ; CHECK-NEXT:       Type:            I32
 ; CHECK-NEXT:       Mutable:         false
 ; CHECK-NEXT:       InitExpr:
 ; CHECK-NEXT:         Opcode:          I32_CONST
 ; CHECK-NEXT:         Value:           8
 
 ; __tls_align
-; CHECK-NEXT:     - Index:           3
+; CHECK-NEXT:     - Index:           4
 ; CHECK-NEXT:       Type:            I32
 ; CHECK-NEXT:       Mutable:         false
 ; CHECK-NEXT:       InitExpr:
 ; CHECK-NEXT:         Opcode:          I32_CONST
 ; CHECK-NEXT:         Value:           4
 
 
 ; CHECK:      - Type:            CODE
 ; CHECK-NEXT:   Functions:
 ; Skip __wasm_call_ctors and __wasm_init_memory
 ; CHECK:          - Index:           2
 ; CHECK-NEXT:       Locals:          []
-; CHECK-NEXT:       Body:            20002401200041004108FC0800000B
+; CHECK-NEXT:       Body:            20002402200041004108FC0800000B
 
 ; Expected body of __wasm_init_tls:
 ;   local.get 0
 ;   global.set  1
 ;   local.get 0
 ;   i32.const 0
 ;   i32.const 8
 ;   memory.init 1, 0
 ;   end
 
 ; CHECK-NEXT:     - Index:           3
 ; CHECK-NEXT:       Locals:          []
-; CHECK-NEXT:       Body:            2381808080004180808080006A0B
+; CHECK-NEXT:       Body:            2382808080004180808080006A0B
 
 ; Expected body of tls1_addr:
 ;   global.get 1
 ;   i32.const 0
 ;   i32.add
 ;   end
 
 ; CHECK-NEXT:     - Index:           4
 ; CHECK-NEXT:       Locals:          []
-; CHECK-NEXT:       Body:            2381808080004184808080006A0B
+; CHECK-NEXT:       Body:            2382808080004184808080006A0B
 
 ; Expected body of tls1_addr:
 ;   global.get 1
 ;   i32.const 4
 ;   i32.add
 ;   end
 
 ; CHECK-NEXT:     - Index:           5
 ; CHECK-NEXT:       Locals:          []
-; CHECK-NEXT:       Body:            2383808080000B
+; CHECK-NEXT:       Body:            2384808080000B
 
 ; Expected body of tls1_addr:
 ;   global.get 3
 ;   end

lld/wasm/Driver.cpp

WasmSym::memoryBase =
         createUndefinedGlobal("__memory_base", &globalTypeI32);
     WasmSym::tableBase = createUndefinedGlobal("__table_base", &globalTypeI32);
     WasmSym::memoryBase->markLive();
     WasmSym::tableBase->markLive();
   } else {
     // For non-PIC code
     WasmSym::stackPointer = createGlobalVariable("__stack_pointer", true, 0);
     WasmSym::stackPointer->markLive();
+
+    WasmSym::stackLimit = createGlobalVariable("__stack_limit", false, 0);
   }
 
   if (config->sharedMemory && !config->shared) {
     // Passive segments are used to avoid memory being reinitialized on each
     // thread's instantiation. These passive segments are initialized and
     // dropped in __wasm_init_memory, which is registered as the start function
     WasmSym::initMemory = symtab->addSyntheticFunction(
         "__wasm_init_memory", WASM_SYMBOL_VISIBILITY_HIDDEN,

lld/wasm/Symbols.h

struct WasmSym {
   // Symbol marking the start of the global section.
   static DefinedData *globalBase;
 
   // __stack_pointer
   // Global that holds the address of the top of the explicit value stack in
   // linear memory.
   static GlobalSymbol *stackPointer;
 
+  // __stack_limit
+  // Global that holds the address of the end of the memory allocated for the
+  // explicit value stack in linear memory.
+  static GlobalSymbol *stackLimit;
+
   // __tls_base
   // Global that holds the address of the base of the current thread's
   // TLS block.
   static GlobalSymbol *tlsBase;
 
   // __tls_size
   // Symbol whose value is the size of the TLS block.
   static GlobalSymbol *tlsSize;

lld/wasm/Symbols.cpp

 DefinedFunction *WasmSym::applyRelocs;
 DefinedFunction *WasmSym::initTLS;
 DefinedData *WasmSym::dsoHandle;
 DefinedData *WasmSym::dataEnd;
 DefinedData *WasmSym::globalBase;
 DefinedData *WasmSym::heapBase;
 DefinedData *WasmSym::initMemoryFlag;
 GlobalSymbol *WasmSym::stackPointer;
+GlobalSymbol *WasmSym::stackLimit;
 GlobalSymbol *WasmSym::tlsBase;
 GlobalSymbol *WasmSym::tlsSize;
 GlobalSymbol *WasmSym::tlsAlign;
 UndefinedGlobal *WasmSym::tableBase;
 DefinedData *WasmSym::definedTableBase;
 UndefinedGlobal *WasmSym::memoryBase;
 DefinedData *WasmSym::definedMemoryBase;
 

lld/wasm/Writer.cpp

void Writer::layoutMemory() {
   auto placeStack = [&]() {
     if (config->relocatable || config->isPic)
       return;
     memoryPtr = alignTo(memoryPtr, stackAlignment);
     if (config->zStackSize != alignTo(config->zStackSize, stackAlignment))
       error("stack size must be " + Twine(stackAlignment) + "-byte aligned");
     log("mem: stack size  = " + Twine(config->zStackSize));
     log("mem: stack base  = " + Twine(memoryPtr));
+    uint32_t limitPtr = memoryPtr;
     memoryPtr += config->zStackSize;
     auto *sp = cast<DefinedGlobal>(WasmSym::stackPointer);
     sp->global->global.InitExpr.Value.Int32 = memoryPtr;
     log("mem: stack top   = " + Twine(memoryPtr));
+    auto *sl = cast<DefinedGlobal>(WasmSym::stackLimit);
+    sl->global->global.InitExpr.Value.Int32 = limitPtr;
+    log("mem: stack limit = " + Twine(limitPtr));
   };
 
   if (config->stackFirst) {
     placeStack();
   } else {
     memoryPtr = config->globalBase;
     log("mem: global base = " + Twine(config->globalBase));
   }

llvm/lib/Target/WebAssembly/WebAssemblyMCInstLower.cpp

MCSymbol *WebAssemblyMCInstLower::GetExternalSymbolSymbol(
   const char *Name = MO.getSymbolName();
   auto *WasmSym = cast<MCSymbolWasm>(Printer.GetExternalSymbolSymbol(Name));
   const WebAssemblySubtarget &Subtarget = Printer.getSubtarget();
 
   // Except for certain known symbols, all symbols used by CodeGen are
   // functions. It's OK to hardcode knowledge of specific symbols here; this
   // method is precisely there for fetching the signatures of known
   // Clang-provided symbols.
-  if (strcmp(Name, "__stack_pointer") == 0 || strcmp(Name, "__tls_base") == 0 ||
+  if (strcmp(Name, "__stack_pointer") == 0 ||
+      strcmp(Name, "__stack_limit") == 0 ||
+      strcmp(Name, "__tls_base") == 0 ||
       strcmp(Name, "__memory_base") == 0 || strcmp(Name, "__table_base") == 0 ||
       strcmp(Name, "__tls_size") == 0 || strcmp(Name, "__tls_align") == 0) {
     bool Mutable =
         strcmp(Name, "__stack_pointer") == 0 || strcmp(Name, "__tls_base") == 0;
     WasmSym->setType(wasm::WASM_SYMBOL_TYPE_GLOBAL);
     WasmSym->setGlobalType(wasm::WasmGlobalType{
         uint8_t(Subtarget.hasAddr64() ? wasm::WASM_TYPE_I64
                                       : wasm::WASM_TYPE_I32),