Reviewers: Erik Corry,
Message:
I am quite sure that the push of 0 in the CEntryStub::GenerateCore might as
well
be a decrement of sp, as pushing 1 instead does not break any tests, so most
likely the GC does not look at this part of the stack. However I will like
to
check this more throughly before making that change.
I am in the process of running the tests on ARM hardware.
Description:
Ensure stack alignment in ARM port
In the CEntryStub keep track of the actual skew from the correct alignment
to
ensure proper alignment before calling the runtime. Add checks to test the
expected skew when running on hardware.
Use the PrepareCallCFunction/CallCFunction in a few other places.
Add check to the ARM simulator for correct alignment.
Add option --sim-stack-alignment to set the stack alignment in the
simulator.
Currntly only values 4 and 8 makes sense. 8 is the default to make snapshots
generated with the simulator work on hardware requiring 8 byte stack
alignment.
Please review this at http://codereview.chromium.org/1530043/show
SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/
Affected files:
M src/arm/codegen-arm.cc
M src/arm/macro-assembler-arm.h
M src/arm/macro-assembler-arm.cc
M src/arm/simulator-arm.cc
M src/codegen.h
M src/flag-definitions.h
M src/ia32/codegen-ia32.cc
M src/x64/codegen-x64.cc
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
To unsubscribe, reply using "remove me" as the subject.
