[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gpsd-users] gpsd fails with Illegal instruction on old Intel CPU
From: |
Eric S. Raymond |
Subject: |
Re: [gpsd-users] gpsd fails with Illegal instruction on old Intel CPU |
Date: |
Tue, 10 Jan 2012 12:24:20 -0500 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
HÃ¥kan Johansson <address@hidden>:
> Probably the illegal instruction is
>
> asm volatile("mfence");
>
> in barrier() in gpsd.h (from gpsd.h-tail). That instruction does
> not exist on senior x86 CPUs. Where senior is something like age >=
> Pentium II/III. There was a discussion on the gpsd-dev mailing list
> end of March about this ("Restrict the barrier() function to x86 and
> x86_64").
Klaus, I'm certain Hakan is right about this. The fact that you were
able to prevent the problem by compiling with the shared-memory export
off is pretty conclusive.
What I'd like to do is change the build script so it forces shm_export
off if the proccessor is as old as yours. What do platform.processor()
and platform.machine() return in your Python?
--
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- Re: [gpsd-users] gpsd fails with Illegal instruction on old Intel CPU,
Eric S. Raymond <=