shallow thoughts on SHM
jamesb.fe80 at gmail.com
Sat Oct 26 20:00:00 UTC 2019
On Sat, Oct 26, 2019 at 7:49 AM Eric S. Raymond <esr at thyrsus.com> wrote:
> James Browning via devel <devel at ntpsec.org>:
> > I would like to propose a new SHM implementation
> The trouble with any new SHM proposal is that the underlayer is not
> POSIX and we theefore can't count on it continuing to exist.
> We need that kind of funcrionality, but any new design should be dome
> over an IPC layer that's POSIX/SuSv2.
I do not have access to a copy of POSIX and the SuSv2 seems to have
SHM support. Changing it to use say a UNIX socket would allow for
simpler packet design. Up to six bits for the version another two for
leap second status and then an octet for precision. Then the body
which could be a pair of timestamps. Then possibly refclock and
hostname overriding text fields.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the devel