Reviewers: ulan, Benedikt Meurer,

Message:
Use proper call sequence for calls.

Description:
ARM: clean-up some stub call sequences.

BUG=

Please review this at https://codereview.chromium.org/203263016/

SVN Base: https://v8.googlecode.com/svn/branches/bleeding_edge

Affected files (+5, -11 lines):
  M src/arm/code-stubs-arm.cc


Index: src/arm/code-stubs-arm.cc
diff --git a/src/arm/code-stubs-arm.cc b/src/arm/code-stubs-arm.cc
index cad641738c7a682310eb4b2a752ea35cf98dcdf5..5609cd0ccbf5009ac5106d224649e6cf0f89d6c5 100644
--- a/src/arm/code-stubs-arm.cc
+++ b/src/arm/code-stubs-arm.cc
@@ -1575,9 +1575,9 @@ void CEntryStub::GenerateCore(MacroAssembler* masm,
   {
     // Prevent literal pool emission before return address.
     Assembler::BlockConstPoolScope block_const_pool(masm);
-    masm->add(lr, pc, Operand(4));
+    __ add(lr, pc, Operand(4));
     __ str(lr, MemOperand(sp, 0));
-    masm->Jump(r5);
+    __ Call(r5);
   }

   __ VFPEnsureFPSCRState(r2);
@@ -1865,16 +1865,10 @@ void JSEntryStub::GenerateBody(MacroAssembler* masm, bool is_construct) {
     __ mov(ip, Operand(entry));
   }
   __ ldr(ip, MemOperand(ip));  // deref address
+  __ add(ip, ip, Operand(Code::kHeaderSize - kHeapObjectTag));

- // Branch and link to JSEntryTrampoline. We don't use the double underscore
-  // macro for the add instruction because we don't want the coverage tool
- // inserting instructions here after we read the pc. We block literal pool
-  // emission for the same reason.
-  {
-    Assembler::BlockConstPoolScope block_const_pool(masm);
-    __ mov(lr, Operand(pc));
-    masm->add(pc, ip, Operand(Code::kHeaderSize - kHeapObjectTag));
-  }
+  // Branch and link to JSEntryTrampoline.
+  __ Call(ip);

   // Unlink this frame from the handler chain.
   __ PopTryHandler();


--
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
--- You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to