Reviewers: Karl Klose, Description: Change kPointerSize to kIntSize in ia32 specific stack check patching.
The call uses a 32 bit displacement relative to the next instruction. Althoug this has no actual impact on the 32 bit platform I think this is more correct
(and will make the x64 and ia32 code more alike). Please review this at http://codereview.chromium.org/6515011/ SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M src/ia32/deoptimizer-ia32.cc Index: src/ia32/deoptimizer-ia32.cc =================================================================== --- src/ia32/deoptimizer-ia32.cc (revision 6765) +++ src/ia32/deoptimizer-ia32.cc (working copy) @@ -137,7 +137,7 @@ void Deoptimizer::PatchStackCheckCodeAt(Address pc_after, Code* check_code, Code* replacement_code) { - Address call_target_address = pc_after - kPointerSize; + Address call_target_address = pc_after - kIntSize; ASSERT(check_code->entry() == Assembler::target_address_at(call_target_address)); // The stack check code matches the pattern: @@ -169,7 +169,7 @@ void Deoptimizer::RevertStackCheckCodeAt(Address pc_after, Code* check_code, Code* replacement_code) { - Address call_target_address = pc_after - kPointerSize; + Address call_target_address = pc_after - kIntSize; ASSERT(replacement_code->entry() == Assembler::target_address_at(call_target_address)); // Replace the nops from patching (Deoptimizer::PatchStackCheckCode) to -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
