RFR(S): 8016131: nsk/sysdict/vm/stress/chain tests crash the VM in 'entry_frame_is_first()'

[Rickard Bäckman]

Hi,

can I please have a couple of reviews for this change?

The problem in this crash was that we were given an incorrect fp (in this case 0x0) and had a pc that matched the C -> Java entry frame. The code then dereferenced fp +- offset.

This change verifies that the fp +- offset is actually on the stack of the thread before doing the derefencing.

Webrev: http://cr.openjdk.java.net/~rbackman/8016131/

Thanks
/R