At 2002/3/27 13:08-0500 Joey Hess writes:
> Of course the whole point of that test it to make sure the implementation
> fails with ENOMEM if there is not enough room to mmap the requested space.
> The testcase's pre-mapping of the space to try to fill up a bit of it so that
> the second mmap will fail is here failing.
> I suspect that something else earlier in the mmap_P test is consuming some
> space (would have to be near to half of the address space though) and not
> freeing it or so. Has anyone else seen this testcase fail?
Are any of the other mmap tests failing? (especially 34?) Its
possible that VSRT_ADDR_SPACE_PAGES is set too high for your system,
so it might be worthwhile trying to change it in
/home/tet/test_sets/TESTROOT/tetexec.cfg and rerun the tests.
You can also run an individual testcase which should clarify if another
test is causing the problem:
(as vsx0, you may need to a '. profile' if you're working off the
source based test suites)
and the results will be left in tet_xres.
IBM OzLabs Linux Development Group
To UNSUBSCRIBE, email to email@example.com
with subject of "unsubscribe". Trouble? Email firstname.lastname@example.org