Re: [Sablevm-developer] Runaway memory usage

[Archie C. <ar...@de...>]

Archie Cobbs wrote:
> Note the absurd 'frame_size=178286532' value being passed to
> _svmf_ensure_stack_capacity() (because of the first bug, it actually
> tries to allocate that much memory, but nevermind that for a minute)...
> This value results from a corrupted 'method' structure. It appears
> that the 'method->frame_info' pointer is bogus:

Further progress report on this...

It appears that SableVM is trying to run the START_PREPARE_INVOKEVIRTUAL
"bytecode" using a '_svmt_CONSTANT_Methodref_info' which describes
an interface method instead of an instance method.

The result is that the computed method_id is 616 which of course
is supposed to be an interface method index but SableVM tries to
use it as a virtual method index to the vtable, hence the 'corrupted
method' and crash.

The method being executed is soot.AbstractValueBox.setValue().
The interface method that this method is trying to invoke is
soot.ValueBox.canContainValue(). So in summary, a concrete method
(setValue()) in an abstract class (AbstractValueBox) that implements
an interface (ValueBox) is trying to invoke an abstract interface
method (canContainValue()) using the 'this' object.

Hopefully this rings a bell.. ? I haven't been able to reproduce this
with a simple example.

-Archie

__________________________________________________________________________
Archie Cobbs     *     Packet Design     *     http://www.packetdesign.com