[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference vaxaxp::vmsnotes

Title:	VAX and Alpha VMS
Notice:	This is a new VMSnotes, please read note 2.1
Moderator:	VAXAXP::BERNARDO

Created:	Wed Jan 22 1997
Last Modified:	Fri Jun 06 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	703
Total number of notes:	3722

464.0. "Pseudoterminal question -- PTD$WRITE says success but commands never execute" by HYDRA::NEWMAN (Chuck Newman, 508/467-5499 (DTN 297), MRO1-3/F26) Mon Apr 14 1997 10:54

I'm working with an ISV who is using pseudoterminals on OpenVMS Alpha V6.1 and
OpenVMS VAX V5.5-2.

They call PTD$SET_EVENT_NOTIFICATION(mumble,,,PTD$C_START_READ)
They then queue an asynchronous read to the terminal (the READ's completion AST
re-issues the READ)

From the event notification AST, they disable AST delivery, call PTD$WRITE with
a buffer full of commands, then reenable ASTs.  Occasionally on SMP systems
(both Alpha and VAX systems), the WRITE's return status and the status in the
write buffer indicate normal completion, but the commands in the buffer don't
get executed.

They said that this happens most frequently when the pseudoterminal has just
been created.


Any ideas?  Anything else I should ask them?

Thanks!

								-- Chuck Newman

T.R Title User Personal
Name Date Lines

464.1 Try middle read STAR::KENNEY Mon Apr 14 1997 11:49 14

T.R	Title	User	Personal Name	Date	Lines
464.1	Try middle read	STAR::KENNEY		`Mon Apr 14 1997 11:49`	14
	More than likely it only happens on the first read. The problem is that the fork to create the type-ahead buffer may not have run. When that happens the data in the write buffer can end up in the bit bucket. The pseudo terminal does not know this as it gets no indications of problems from the terminal class driver. Have them try changing start read to middle read. I am not really positive that doing that will cure the problem but it may. Gee I hardly remember when I wrote the drivers for that stuff must have been 1989 or so. Forrest

    
    	More than likely it only happens on the first read.  The problem is
    that the fork to create the type-ahead buffer may not have run.  When
    that happens the data in the write buffer can end up in the bit bucket. 
    The pseudo terminal does not know this as it gets no indications of
    problems from the terminal class driver.  Have them try changing start
    read to middle read.  I am not really positive that doing that will
    cure the problem but it may.
    
    	Gee I hardly remember when I wrote the drivers for that stuff must
    have been 1989 or so.
    
    
    Forrest