Report 1620 Actions
Problem Report Number |
1620 |
Submitter's Classification |
Test Suite problem |
State |
Resolved |
Resolution |
Permanent Interpretation (PIN) |
Problem Resolution ID |
PIN.X.0177 |
Raised |
1996-11-14 08:00 |
Updated |
2003-03-13 08:00 |
Published |
1997-01-08 08:00 |
Product Standard |
Internationalised System Calls and Libraries Extended (UNIX 95) |
Certification Program |
The Open Brand certification program |
Test Suite |
VSU version 4.1.0 |
Test Identification |
CAPI.os/procprim/sigstack 4 |
Specification |
System Interfaces and Libraries Issue 4 Version 2 |
Location in Spec |
See Problem Text |
Problem Summary |
PIN4U.00043 This test may fail on some IUTs because because it makes an assumption as to the proper value for ss_onstack when no value is mandated by the spec. |
Problem Text |
Passing this test requires that one fail the requirements of sigaction() with respect to the sa_flags flag SA_ONSTACK. In XSH4v2, page 545, the following requirements are specified for SA_ONSTACK:
SA_ONSTACK If set and an alternate signal stack has been declared with sigaltstack() or sigstack(), the signal will be delivered to the calling process on that stack. Otherwise, the signal will be delivered on the current stack.
Concerning TSD4U.00023, we disagree with the consultants response which resulted in the change to this test.
Concerning PG4U.00098, we agree with the submitter of the waiver request. That is, the BSD man page which is in direct conflict with the sigaction() SA_ONSTACK description is not normative and must not be allowed to override the normative requirements stated in XSH4v2. We disagree with the consultants initial response stating that the sigstack() semantics as stated in BSD are what was intended in XSH4v2. We believe that the spec1170 team intentionally fixed a BSD bug and as per the sigaction() requirement, that the test should be verifying the alternate stack \fIis\fR in use.
Further appeal -------------- Given that PG4U.00098 was rejected, if our comment pointing out the inconsistencies this creates in XSH4v2 is not obvious and immediately accepted, we request that this be sent to the XoTGbase working group for interpretation without delay.
|
Test Output |
TEST CASE: sigstack
TEST PURPOSE #4 A call to int sigstack(struct sigstack *ss , struct sigstack *oss) when the ss_onstack member of ss is non-zero shall indicate to the system that the process is executing on the specified stack. PREP: Call sigstack(ss, oss) with ss_onstack non-zero PREP: Call sigaction() with SA_ONSTACK set PREP: Send self SIGTERM TEST: Verify alternate stack not used ERROR: Auto declaration is on the alternate stack and shouldn't be Declared Top of Stack = 39c88 Declared End of Stack = 35c88 Received Auto Address = 3799f 4 FAIL
|
Review Information
Review Type |
TSMA Review |
Start Date |
null |
Completed |
null |
Status |
Complete |
Review Recommendation |
No Resolution Given |
Review Response |
We recommend this request be refused because previous rulings regarding the issue exist.
|
Review Type |
SA Review |
Start Date |
null |
Completed |
null |
Status |
Complete |
Review Resolution |
No Resolution Given |
Review Conclusion |
The specification does not state a value for the user setting ss_onstack in the sigstack structure pointed to by the first argument to sigstack() It states: "The value of the ss_onstack member indicates whether the process wants the system to use an alternate signal stack when delivering signals." There is no indication of what value or values signify using the alternate signal stack and what value or values signify not using the alternate signal stack.
The test assumes that by setting ss_onstack to non-zero, an alternate stack registered with sigstack() will be disabled from use, and that a corresponding call to sigaction() when SA_ONSTACK is set should not use the alternate stack registered by sigstack().
It is recommended that a PIN should be granted.
This interface is marked to be withdrawn, and portable applications should use sigaltstack(). Since the concept of signal stacks is not in POSIX, and existing code using BSD signals will be required to be ported to the Single UNIX Specification this is not thought to be a problem. It is also debatable whether applications would want to use sigstack() in this way to disable an alternate stack.
It is recommended that this be forwarded to the Base Working group for consideration.
|
Review Type |
Expert Group Review |
Start Date |
null |
Completed |
null |
Status |
Complete |
Review Resolution |
No Resolution Given |
Review Conclusion |
A Permanent Interpretation (PIN) should be granted.
|
Review Type |
SA Review |
Start Date |
null |
Completed |
null |
Status |
Complete |
Review Resolution |
Permanent Interpretation (PIN) |
Review Conclusion |
A Permanent Interpretation is granted.
|
Problem Reporting System Options:
|