Re: Please publish attached draft-ietf-secsh-break-01

[Chris Lonvick <clonvick%cisco.com@localhost>]

Hi,

I've been told that some chipsets may be assigned a timelength for sending
a BREAK.  And, as the draft says, if they are not assigned a value then
they use their default value.  Over the years, that value has moved around
with a low end of 100ms and a high end of 1 second.  Pairing up a server
that sends a 100ms BREAK to a console (or whatever) that expects a break
to last 1000ms will probably not achieve the expectations.  Setting a
default to 500ms at least gives everyone a point to shoot at.  Allowing
the value to be sent in the protocol gives SSH terminal server
manufacturors a chance to meet the requirements of their customers with
specific needs.

{Speaking for myself and not my company.  ..as in; I don't even _know_ the
people in my company who do serial port chips or programming of them but
I think it's a good feature to have.  :-}

Thanks,
Chris

On Wed, 13 Aug 2003, Jeffrey Altman wrote:

> I would assume in most cases the break-length would be ignored.
> However, if the device is a serial port the application will need to
> determine the break length
>
> Jeffrey Altman
>
>
> Markus Friedl wrote:
>
> > On Sat, Mar 29, 2003 at 09:48:21AM -0700, Joseph Galbraith wrote:
> >
> >>           uint32             break-length in milliseconds
> >
> >
> > what's the reason for including this in the protocol?
> >
> > shouldn't the server decide what break-length is reasonable?  e.g.
> > depending on what kind of devices is attached?
>