OpenSolaris

During snv_68 testing in ON PIT Ireland we have seen testcase tst.dlclose1.ksh failing twice on a Pentium 4 machine with the following log :

 dtrace: buffer size lowered to 2m
 dtrace: failed to match test_prov*:::: No probe matches description
    ID   PROVIDER            MODULE                          FUNCTION NAME
                      FUNCTION NAME
 
Note that the same testcase passed fine on other P4 machines.
[ahl 7.2.2007]

The test is correct as written as the expectation is that the attempt to list probes will fail. If you look at tst.dlclose1.ksh.out you can see this in the expected output. It sounds like the issue is that there's some spurious difference on this small memory configuration machine. Can you please send or point me to the failure.N directory so I can verify this?
As said above, it turns out that the testcase actually ran fine. The failure is generated because when dtrace buffer is re-sized it generates some unexpected output which does not match the content of tst.dlclose1.ksh.out.

We expect :
dtrace: failed to match test_prov*:::: No probe matches description
   ID   PROVIDER            MODULE                          FUNCTION NAME 

But we get : 
dtrace: buffer size lowered to 2m
dtrace: failed to match test_prov*:::: No probe matches description
    ID   PROVIDER            MODULE                          FUNCTION NAME

So it looks like dtrt harness should remove buffer resizing message before comparing outputs.

N/A

N/A