Read trigger meaning

ITEM: RTA000088327

In the smit panel for hardware characteristics of a tty, I have a               
question about the read trigger parameter. The default  is 'waiting for         
14 chracters in the FIFO' before signalling data have been received.            
First, is this signal for the system, and allows the transfer of data           
from the hardware adapter to kernel buffer, or from the kernel buffer           
to the user read?                                                               
If I receive only ten characters when will this FIFO be flushed?  Is            
there a timeout that allows this signal to take place? What is its              
My problem is  I have to aknowledge (by software) packets received in           
a certain amount of time and it happens some miss in this.  I was               
wondering if this parameter in smit could explain this problem. For             
example the number of characters used is 14 (the default); what                 
happens if there are only 10 characters in the hardware buffer?                
Will this queue be flushed after a certain time? What is this time?             
We already check the points:  only one                                          
programs reading the tty vmin already set to 1 raw mode with no                 
processing at all.  So for the UNIX/tty point of view, I think every            
points are corrects the only place I do not know ( that's why I have            
this QA)is what happens in raw mode if only 10 characters arrive on             
the hardware buffer when will this characters been transfered in the            
memory and available for reading.  Am I correct if I say the read               
trigger qualify the transfer from hardware adapter to tty driver and            
not from tty driver to user space?  Is there a timeout after which if           
there are characters in the hardware bu buffer they are automatically           
transfered to memory.  This problem only appears time to                        
time and didn't appear since a while so may be it's the read trigger            
who works.                                                                     
The Read Trigger (rtrig) in AIX is used to make the Standard,                   
8-Port and 16-Port Async Adapters more efficient.  The value                    
of the Read Trigger determines how much data the Adapter will                   
read before interrupting AIX.                                                   
If you believe the Read Trigger is causing the problem, then                    
change to Read Trigger to 0 (/usr/lbin/tty/stty-rs rtrig 0                      
< /dev/tty#).  This will interrupt AIX for every character.                     
There are several other things to check also.                                   
Make sure there is only ONE process (program) running on the                    
tty port.  If there are multiple process running on the same                   
tty port (like a getty and your program), then the input data                   
would be split for the different process.                                       
If you are in line mode (-icanon with the 'stty -a /dev/tty#'                  
command) then a New Line Character will be needed to read any                   
of the data.                                                                    
If you are in raw mode (icanon), then the Vmin Value determines                 
how many characters will need to be available to AIX before                     
your process will read them (tty POSIX structure).  The default                 
for Vmin is FOUR and can be changed to ONE with the 'stty eof .A                
/dev/tty#' command (Control A = 1, Control B = 2, Control C = 3,               
Control D = 4, etc).                                                            
Have you set the Read Trigger to 0?  Please set the Read                       
Trigger to 0 (/usr/lbin/tty/stty-rs rtrig 0 
WWQA: ITEM: RTA000088327 ITEM: RTA000088327
Dated: 01/1996 Category: ITSAI6000HW
This HTML file was generated 99/06/24~12:43:29
Comments or suggestions? Contact us