Issue 1982:  GIOP fragment alignment issue for CORBA 2.3 (interop)
Source:  (, )
Nature:  Uncategorized Issue
Severity: 
Summary: Summary: 
 The following issue should be fixed prior to publishing CORBA 2.3.
 
 Section 15.4.8 Fragment Message, page 15-40, of ptc/98-08-03, says:
 
 > A primitive data type of 8 bytes or smaller should never be broken across two
 > fragments.
 
 This text has been in the specification since GIOP 1.1.
 
 Section 15.4.1 GIOP Message Header, page 15-29, of ptc/98-08-03,
 contains new text under the "message size" bullet:
 
 > For GIOP version 1.2, if the second least significant bit of flags is 1, the
 > message_size value must be evenly divisible by 8.
 

Resolution: resolved
Revised Text: In section 15.4.1 (pg 15-29), change the last paragraph to be 
" 
For Giop version 1.2, if the second least significant bit of Flags is 1, the sum of the message size value and 12 must be evenly divisible by 8. 
" 
    In section 15.4.8 on the fragment message, in page 15-40,  Change 
" 
For GIOP version 1.2, fragments other than the final fragment ... 
" 
    to 
" 
For GIOP version 1.2, the total length (including the message header) of a fragment other than the final fragment ... 
" 

Actions taken:
September 20, 1998: received issue
September 30, 1998: moved from core to interop
March 8, 1999: closed issue
Discussion: 

End of Annotations:=====
Return-Path: <kukura@iona.com>
Date: Sun, 20 Sep 1998 17:25:34 -0400
From: Bob Kukura <kukura@iona.com>
Organization: IONA Technologies
To: issues@omg.org, interop@omg.org, orb_revision@omg.org
CC: cryan@iona.com
Subject: GIOP fragment alignment issue for CORBA 2.3

The following issue should be fixed prior to publishing CORBA 2.3.

Section 15.4.8 Fragment Message, page 15-40, of ptc/98-08-03, says:

> A primitive data type of 8 bytes or smaller should never be broken
across two
> fragments.

This text has been in the specification since GIOP 1.1.

Section 15.4.1 GIOP Message Header, page 15-29, of ptc/98-08-03,
contains new text under the "message size" bullet:

> For GIOP version 1.2, if the second least significant bit of flags
is 1, the
> message_size value must be evenly divisible by 8.

The problem is that these two quoted statements are contradictory, and
I
believe the second must be modified.  The message_size field does not
include the 12 byte GIOP MessageHeader, but this header does effect
alignment.  The byte following the MessageHeader has an alignment of
4. 
If message_size is divisible by 8, the message will also end with a 4
byte alignment, forcing an 8 byte primitive to be broken across two
fragments.  This can be fixed by changing the new 15.4.1 text quoted
above to read:

> For GIOP version 1.2, if the second least significant bit of flags
is 1, the	> remainder after dividing the message_size value by 8
must be 4.

Note that the first 4 bytes following the MessageHeader of a fragment
message are occupied by the 4 byte FragmentHeader_1_2, so the actual
fragment payload data thus begins and ends on 8 byte boundaries, as
was
the intent of the RTF.

-Bob