[LTP] [lkp] [signal] 98ea270c46: ltp.float_exp_log.fail

[Thomas Gleixner]

On Tue, 28 Jun 2016, Cyril Hrubis wrote:
> Looking at the test it creates a separate thread to handle signals which
> sleeps in sigwait() which is implemented in glibc via the sigtimedwait
> syscall with NULL timeout.
> 
> Looking at the patch the ktime_t timeout is initialized to 0 instead of
> MAX_SCHEDULE_TIMEOUT now so I guess that the call with NULL timeout does
> poll instead of sleeping in kernel waiting for signals to arrive. That
> should be fixed.

You're right. I misread the code...