Request for review: 6310967: SA: jstack -m produce failures in output

[David Holmes]

Hi Poonam,

It seems to me that rather than passing the ThreadProxy through to 
f.sender the frame, which has to be a frame of some thread, should 
already know what that thread is and so be able to access it directly.

In other words: shouldn't each CFrame maintain a reference to the thread 
it corresponds to?

David

On 12/06/2012 11:05 PM, Poonam Bajaj wrote:
> Hi,
>
> Please review this fix for bug 6310967
> <http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6310967>.
>
> 6310967 <http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6310967>:
> SA: jstack -m produce failures in output
>
> http://cr.openjdk.java.net/~poonam/6310967/webrev.00/
>
> Problem: jstack -m fails with UnalignedAddressException.The problem is
> that while finding the caller frame of a frame in sender() method, we
> don't check the validity of the frame pointer (rbp / ebp).
>
> These changes add a simple check that the frame pointer(rbp) should be a
> valid pointer on the stack by making sure that it is not less than the
> stack
> pointer(rsp).
>
>
> Thanks,
> Poonam
>