[#RVM-599] New poisoned test failures on x86

RVM

New poisoned test failures on x86

Created: 24/Jul/08 04:19 AM Updated: 24/Jul/08 05:49 AM

Component/s:

None

Affects Version/s:

None

Fix Version/s:

3.0

Time Tracking:

Not Specified

Description

« Hide

~~RVM-593~~ added extra checks to see if references were valid coming from the boot image. These checks are failing in poisoned configurations, for example:

http://jikesrvm.anu.edu.au/cattrack/results/rvmx86lnx32b.anu.edu.au/sanity/4494/BaseBasePoisoned/default/dacapo/fop/default/Output.txt

The general theme is that the TIB looks invalid, in particular the top byte is clear.

All

Comments

Work Log

Change History

Sort Order:

[ Permlink | « Hide ]

Ian Rogers added a comment - 24/Jul/08 05:06 AM

The problem is that object references are becoming poisoned (so the bottom bit of the reference is set to 1), this causes the TIB to be off by a byte.

[ Permlink | « Hide ]

Ian Rogers added a comment - 24/Jul/08 05:49 AM

Fix in r14785.