Re: libssl update problem
On Monday 09 May 2016 03:39:49 Jörg-Volker Peetz wrote:
> If no amd64 packages needed on this machine, maybe the amd64
> architecture could be removed:
>
> dpkg --remove-architecture amd64
>
> Then it won't install another amd64 package in the future.
>
Which is how I got into this mess in the first place. Because the RTAI
patched 64 bit kernels still don't have anywhere near the low latency
needed to run machinery, as in cnc machinery, that a 32 bit kernel can
achieve, I had added that architecture so I could run a 64 bit amd64
kernel so I could do my own latency measurements. This also had th
added advantage of using all 8Gb in this machine, whereas the RTAI
patches to a 32 bit kernel also wipe out the PAE, so I am 500 megs into
swap in a days runtime. Performance of course goes in the toilet.
That worked although the latency was up some, but I only run the
simulated LCNC on this machine. Then I got the brilliant (not) idea of
converting this 32 bit install into a full 64 bit amd64, but quickly ran
into dependency hell since you can't easily replace libc6 on a running
system. Everything was cool and I had removed most of the :amd64 stuff
until the latest libssl fix hit the repo's. It wouldn't configure as
the amd64 kit was still installed. Anyway, the procedure I posted fixed
it all back up. The above command in your post had already been done.
So overall, this 64 bit kernel runs much better, but the latency of this
kernel,
gene@coyote:~/bin$ uname -a
Linux coyote 3.16.0-0.bpo.4-amd64 #1 SMP Debian 3.16.7-ckt25-2~bpo70+1
(2016-04-12) x86_64 GNU/Linux
is such that I could never run a stepper motor with this machine, the 25
micro-second base-thread has up to an 83 millisecond jitter, whereas the
RTAI patched 32 bit kernel is, on the right cpu, capable of a 3
microsecond jitter, and the 1 millisecond servo-thread will show above
100 milliseconds without too long a wait. But that doesn't prevent me
from writing gcode and testing it for correctness by running a simulated
virtual machine on this machine from a comfy chair. My eyes cannot see
the timing jitter as the code is being simulated, but that would totally
destroy both the workpiece, and maybe even the real machinery itself as
it tries to follow orders that could be up to 100+ milliseconds late.
> Regards,
> jvp.
Cheers, Gene Heskett
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
Reply to: