Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wuarchive!uunet!mcsun!ukc!icdoc!inmos!brac!davidb
From: davidb@brac.inmos.co.uk (David Boreham)
Newsgroups: comp.sys.transputer
Subject: Re: Too many acknowledges
Message-ID: <15079@ganymede.inmos.co.uk>
Date: 26 Mar 91 16:29:34 GMT
References: <57*.S=iselin.OU=iam.O=unibe.PRMD=SWITCH.ADMD=ARCOM.C=CH.@MHS>
Sender: news@inmos.co.uk
Reply-To: davidb@inmos.co.uk (David Boreham)
Organization: none
Lines: 36

In article <57*.S=iselin.OU=iam.O=unibe.PRMD=SWITCH.ADMD=ARCOM.C=CH.@MHS> iselin@IAM.UNIBE.CH (Nicolas Iselin) writes:
>To all hardware specialists !
>
>We have here quite a tricky problem which seems to be hardware 
>dependent. Can anyone tell me what a transputer does, when a
>link receives an acknowledge packet without having sent a 
>data byte ?
>
>Our guess is that it brings the whole Transputer to halt without
>setting the errorflag.
>
>Who knows more ?
>
The link interface is either ``expecting'' an ACK or
it isn't. If it is and it receives and ACK, it takes
it as a indication that transmission can proceed. If
it isn't expecting an ACK, it ignores it.

ACK's are only ``expected'' when a byte has been
transmitted and before the corresponding ACK has been
received.

So, for example merely injecting ACK packets onto
a quiet link should not cause a problem. Injecting
ACKs onto an active link could cause data overrun at
the receiver, depending upon where the ACK falls in
time. That could certainly lock up a network.



Note that the error flag has nothing whatsoever
do to with links. 

David Boreham, INMOS Limited | mail(uk): davidb@inmos.co.uk or ukc!inmos!davidb
Bristol,  England            |     (us): uunet!inmos.com!davidb
+44 454 616616 ex 547        | Internet: davidb@inmos.com