Re: ptrace(2) PT_STEP changes and gdb

[Paul Koning <paul_koning%dell.com@localhost>]

On Aug 17, 2010, at 5:33 PM, Eduardo Horvath wrote:

> On Tue, 17 Aug 2010, Paul Koning wrote:
> 
>>> (3) if the process only has one thread ignore he argument.
>> 
>> That means the semantics of ptrace can change out from under the program 
>> using it.  If it passes an argument to PT_STEP because there are two 
>> threads, and one thread exits before the ptrace() call actually runs, then 
>> the argument would be ignored producing an unexpected result.
> 
> Is that really a problem?  If you are using PT_STEP then presumably the 
> thread in question is stopped and can only die if something else sends a 
> signal.  If the other threads die then you are still single-stepping the 
> thread you thought you would be single-stepping.  If there is only one 
> thread then he question about what to do about other threads is moot.

Yes, I think you're right.

> OTOH versioning the interface is probably cleaner over all.

I agree.

        paul