SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


    [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

    Re: ISCSI: Urgent Flag requirement violates TCP.



    Douglas Otis wrote:
    
    > There is little assurance of pointer accuracy as
    > most machine's pointers do not resolve to the byte
    
    The "urgent pointer" is not a pointer to memory.  It is a 16 bit integer that is
    added to the 32 bit tcp sequence number (also an integer) to indicate the
    position in the byte stream where the urgent data is.  Last time I checked, most
    all machines are able to do integer arithmetic.
    
    
    > and TCP never intended a urgent pointer to be used as a record mark.
    
    Sure it did. From the glossary of rfc793:
    
    urgent pointer - A control field meaningful only when the URG bit is on. This
    field communicates the value of the urgent pointer which indicates the data
    octet associated with the sending user's urgent call.
    
    Did you get that? "indicates the data octet associated with the sending user's
    urgent call" sure sounds like a marker to me.
    
    Also note rfc1122, section 4.2.2.4 Urgent Pointer:
    
    ... the Urgent mechanism may be used for any application...
    
    > This urgent pointer simply
    > defines a change in state where the actual pointer is not directly presented
    > to the application but used to expedite processing.
    
    You're right.  The "pointer" is not presented to the application.  But the data
    byte the urgent pointer points to is presented to the application.
    
    Matt Wakeley
    Agilent Technologies
    
    


Home

Last updated: Tue Sep 04 01:06:26 2001
6315 messages in chronological order