Hello Tony

Sorry for my late response, been quite busy.

>From: Tony Reix [mailto:[EMAIL PROTECTED]
>
>Hi Inaky, Dinakar,
>
>I'm studying the effort that would require the port of PTT for tracing
>the RobustMutexes (fusyn).

You would have to start with cleaning up and making the 
glibc patch work with the latest kernel tip. So far seems 
nobody had time to step up to the task and thus, it remains 
badly broken.

>Could you provide me the following information:
> - how many new routines have been added to NPTL for RobustMutexes ?

Just a few--top of my head no more than six small routines.

> - and how many old routines have been modified for RobustMutexes ?

Most of the ones dealing with synchronization; pthread_mutex_*lock(),
pthread_rwmutex_*lock(), semaphore routines, etc

> - are there test code and large real applications that use them ?

There is test code, POSIX microbenchmaks in the fusyn-support
package; any big application that uses mutexes heavily (like
MySQL with its benchmarks) can be used to test and as a big
application.

>Do you see other problems/complexity/... that have an impact when
>porting PTT to the Robust Mutexes ?

Not really, other than the clean up.

-- Inaky

_______________________________________________
robustmutexes mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/robustmutexes

Reply via email to