[Bug build/20967] Several Tests failed on MIPS64 N64

["joseph at codesourcery dot com" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=20967

--- Comment #2 from joseph at codesourcery dot com <joseph at codesourcery dot com> ---
Omnibus bugs like this aren't useful and should be closed as INVALID, 
because there is no good way of telling whether or in what version an 
issue is fixed, and Bugzilla can only track bugs as fixed or not fixed, 
not as partly fixed.  Bugs for test failures are only useful once triaged 
to be a single issue, not a problem with your test environment, likely to 
have a single fix and apparently distinct from all other open bugs.  At 
that point the bug subject should describe the issue in logical terms 
rather than in terms of the failing tests.

> FAIL: elf/check-execstack
> FAIL: elf/check-localplt
> FAIL: elf/tst-execstack
> FAIL: elf/tst-execstack-needed
> 
> These are serious security issues and toolchain configuration issues. Either a
> misconfiguration or a bootstrap with an executable stack (bad). Might be one
> object which is being added from the compiler which has the wrong attributes.

See the comments on the XFAIL I added for check-execstack.  I haven't 
investigated whether the others should be XFAILed for the same reasons.

> FAIL: nptl/tst-cancel24-static
> 
> Could be anything. Maybe related to unwind information or compiler issues

It's bug 19826.  Illustrating again why testing master is more useful than 
testing releases.

> FAIL: nptl/tst-stack4
> 
> Potentially related to the above execstack failures.

nptl/tst-stack4 fails on many platforms.  See 
<https://sourceware.org/ml/libc-alpha/2016-11/msg00565.html>.

-- 
You are receiving this mail because:
You are on the CC list for the bug.