Re: Robust mutexes (-rt)

[David Singleton <daviado@gmail.com>]

Thomas,
     here is a patch that fixes the -EINTR problem and a fix for Dinakar's
simple futex deadlock test program.  When Esben's new deadlock detection
is ready I plan to use it for robust futex deadlock detection, both simple and
circular deadlock detection.


     http://source.mvista.com:~dsingleton/patch-2.6.15-rt16-rf1

 include/linux/futex.h |    3 +++
 kernel/futex.c          |    3 +++
 kernel/rt.c               |   19 ++++++++++++++++++-
 3 files changed, 24 insertions(+), 1 deletion(-)

David

> Just another problem:
>
> pthread_mutex_lock() of a robust mutex returns -EINTR, which is wrong by
> the specification. We encountered this in an application. Not sure what
> to do about that. It needs either a fix in glibc or the kernel could
> handle it via a restart function. That happens when a signal is
> delivered to the thread while the thread is blocked on the mutex.
>
>         tglx
>
>
>
>
>