[Bug testsuite/14073] New: setjmp.exp testcase broken

[Bug testsuite/14073] New: setjmp.exp testcase broken

[mjw at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=14073

             Bug #: 14073
           Summary: setjmp.exp testcase broken
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: testsuite
        AssignedTo: systemtap@sourceware.org
        ReportedBy: mjw@redhat.com
    Classification: Unclassified


The setjmp.exp testcase is broken in several ways:

- It uses -D_FORTIFY_SOURCE=2 without -O1 or -O2, which basically does nothing.
- When adding -O1 or -O2 (with or without -D_FORTIFY_SOURCE=2) the testcase
fails because of backtrace mismatch.
- At least on Fedora 17 glibc doesn't have setjmp/longjmp markers, so the whole
test doesn't actually test anything

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug testsuite/14073] setjmp.exp testcase broken

[mjw at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=14073

--- Comment #1 from Mark Wielaard <mjw at redhat dot com> 2012-05-07 19:31:23 UTC ---
(In reply to comment #0)
> - At least on Fedora 17 glibc doesn't have setjmp/longjmp markers, so the whole
> test doesn't actually test anything

I take that back. The issue is:

$ stap -l 'process("/lib/libc.so.*").mark("*")'
process("/usr/lib/libc-2.15.so").mark("lll_futex_wake")?
process("/usr/lib/libc-2.15.so").mark("lll_lock_wait_private")?
process("/usr/lib/libc-2.15.so").mark("longjmp")?
process("/usr/lib/libc-2.15.so").mark("longjmp_target")?
process("/usr/lib/libc-2.15.so").mark("setjmp")?

and

$ stap -l 'process("/lib64/libc.so.*").mark("*")'
process("/usr/lib64/libc-2.15.so").mark("lll_futex_wake")?
process("/usr/lib64/libc-2.15.so").mark("lll_lock_wait_private")?
process("/usr/lib64/libc-2.15.so").mark("longjmp")?
process("/usr/lib64/libc-2.15.so").mark("longjmp_target")?
process("/usr/lib64/libc-2.15.so").mark("setjmp")?

but...

$ stap -l 'process("/lib*/libc.so.*").mark("*")'

Nothing...

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug testsuite/14073] setjmp.exp testcase broken

[mjw at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=14073

--- Comment #2 from Mark Wielaard <mjw at redhat dot com> 2012-05-08 09:31:36 UTC ---
Issue 3, as explained in comment #1, is a completely unrelated bug. See
https://bugzilla.redhat.com/show_bug.cgi?id=819680 ("librpmio.so exports
bundled broken glob(3) clone"). Building with configure --without-rpm "fixes"
it.

So that only leaves items 1) and 2) from the description, setjmp.exp:

- It uses -D_FORTIFY_SOURCE=2 without -O1 or -O2, which basically does nothing.
- When adding -O1 or -O2 (with or without -D_FORTIFY_SOURCE=2) the testcase
fails because of backtrace mismatch.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug testsuite/14073] setjmp.exp testcase broken with -D_FORTIFY_SOURCE=2 and/or -O1 or -O2

[mjw at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=14073

Mark Wielaard <mjw at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|setjmp.exp testcase broken  |setjmp.exp testcase broken
                   |                            |with -D_FORTIFY_SOURCE=2
                   |                            |and/or -O1 or -O2

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.