Re: virtual serial drive, so how is serial port settings negotiated

[Kevin <kevin@hackaday.com>]

On Wednesday, October 3, 2012 9:14:52 PM UTC-5, schmidtd wrote:
> On 10/3/2012 9:33 PM, Kevin wrote:
> 
> > On Wednesday, October 3, 2012 7:49:13 AM UTC-5, schmidtd wrote:
> 
> >> One complicating factor - the serial library on the II side is blocking.
> 
> >>    You can hit escape to stop it when communications are going nowhere,
> 
> >> but if there's no connection, it'll just sit until you do.  I changed
> 
> >> that when I implemented the driver for the /// so there's also a
> 
> >> timeout.  I intend to move that code over to the II as well, but there
> 
> >> will be a ripple effect throughout the family that will need quite a bit
> 
> >> of retesting.
> 
> >
> 
> > the current system seems ok to me, but dont let me stop you making it more user friendly!
> 
> 
> 
> Ok!  Now it's non-blocking.  Neat side-effect: speediboot doesn't have 
> 
> to pause or get manually restarted when your SSC takes more time than I 
> 
> expect to recover after the initial reset.
> 
> 
> 
> > next grilling er question
> 
> >
> 
> > delays!
> 
> >
> 
> > I have a good idea what read from client delays should be, but I havent found any write from host delays yet, skimming over your source on SF.net
> 
> > as much as I like a good treasure hunt, I like a good spoiler too =) so whats the hard numbers for
> 
> >
> 
> > host reading bytes from client
> 
> > host sending bytes to client
> 
> 
> 
> I'm not sure what you're asking.  Do you mean how long to wait before 
> 
> deciding the client isn't there any more?  Like timeouts?  For each byte 
> 
> that is being requested from the client, there is a timeout - look for 
> 
> the readByte() calls.  Various places use different values.  But you 
> 
> shouldn't need more than a second or two in any worst case.  And of 
> 
> course sending you just blast as fast as you can at the port; the baud 
> 
> rate will pace you.
> 
> 
> 
> > thank you very much for your help sofar
> 
> 
> 
> You bet!

the timeouts are 5ms on the VSDRIVE parts, I can shave them to like 2 but didnt want to push it, 5 is fine

"blast as fast as you can at the port"

no problem!