This is the mail archive of the systemtap@sourceware.org mailing list for the systemtap project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: loc2-test and stap disagree with each other


Roland McGrath wrote:
Looks like loc2c-test indicates that the *prev* and *next* variables are available.

It says they are in scope, i.e. it knows what variables you're talking about. That doesn't mean their values are available at that PC location. If that is the distinction that matters, then certainly stap should give a more clear error message about it.

Did you try "loc2c-test -k 0x... prev pid"?  You should always use that
(loc2c-test with identifier args) to debug $identifier resolution problems.
That step should thoroughly disambiguate loc2c bugs vs translator proper
bugs.


Thanks,
Roland
Indeed, eventhough the variable is in scope, its value is not available.
I tried
./loc2c-test -k 0xc00000000036d048 prev pid
loc2c-test: not accessible at this address

It looks like a compiler issue, does it? I built my own kernel-debuginfo on a SLES10 box with gcc 4.1.0.

Thanks, Hien.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]