solaris_bm.link

#==============================================================================
# FILE:       solaris_bm.link
# OVERVIEW:   This is the prototype solaris linker map file (or a copy of it)
#               used to build Solaris binaries with the data sections at the
#               correct (original) addresses. The start addresses of the
#               sections in this file are strings starting with "$", e.g.
#               $ dstartro for the start of the read only data section; these
#               strings are changed to the actual values using linker.sed and
#               the "sed" stream editing tool.
# NOTE:       If a block move is not required (only needed for about half of
#               Pentium to Sparc translations) then solaris.link will be used
#               instead. In this configuration, both the read-only and read-
#               write sections are combined into one read-write section, and the
#               read/write portion will be block moved in some startup code
#
# Copyright (C) 1999, The University of Queensland, BT group
#============================================================================*/

# $Revision: 1.5 $

# These 2 lines are for the first segment. The first segment will always have
# the elf header at the start of it, and this would otherwise muck up the load
# address of the first section. The .interp section is forced to this first
# segment (its address doesn't matter), and the other segments and sections
# can float after that.
# Address 0 is used, because otherwise the segment with the lowest address will
# get the elf header. Thus, segments at low addresses should have .align
# statements (especially pa-risc code segments).
dummy = V0;
dummy : .interp;

seg.code = LOAD V0x$cstart;
seg.code: .uqbt.code;
seg.data = LOAD V0x$dstartro;
seg.data: .uqbt.data;