1# The svr4 reference port for the i860 contains an alloca.o routine
2# in /usr/ucblib/libucb.a, but we can't just try to get that by
3# setting CLIB to /usr/ucblib/libucb.a because (unfortunately)
4# there are a lot of other routines in libucb.a which are supposed
5# to be the Berkeley versions of library routines normally found in
6# libc.a and many of these Berkeley versions are badly broken.  Thus,
7# if we try to link programs with libucb.a before libc.a, those
8# programs tend to crash.
9
10# Also, the alloca() routine supplied in early version of svr4 for
11# the i860 is non-ABI compliant.  It doesn't keep the stack aligned
12# to a 16-byte boundary as the ABI requires.
13
14# More importantly however, even a fully ABI compliant alloca() routine
15# would fail to work correctly with some versions of the native svr4 C
16# compiler currently being distributed for the i860 (as of 1/29/92).
17# The problem is that the native C compiler generates non-ABI-compliant
18# function epilogues which cut back the stack (upon function exit) in
19# an incorrect manner.  Specifically, they cut back the stack by adding
20# the nominal *static* frame size (determined statically at compile-time)
21# to the stack pointer rather than setting the stack pointer based upon
22# the current value of the frame pointer (as called for in the i860 ABI).
23# This can cause serious trouble in cases where you repeatedly call a
24# routine which itself calls alloca().  In such cases, the stack will
25# grow continuously until you finally run out of swap space or exceed
26# the system's process size limit.  To avoid this problem (which can
27# arise when a stage1 gcc is being used to build a stage2 gcc) you
28# *must* link in the C language version of alloca() which is supplied
29# with gcc to your stage1 version of gcc.  The following definition
30# forces that to happen.
31
32ALLOCA=alloca.o
33
34# We build all stages *without* shared libraries because that may make
35# debugging the compiler easier (until there is a GDB which supports
36# both Dwarf *and* svr4 shared libraries).
37
38# Note that the native C compiler for the svr4 reference port on the
39# i860 recognizes a special -gg option.  Using that option causes *full*
40# Dwarf debugging information to be generated, whereas using only -g
41# causes only limited Dwarf debugging information to be generated.
42# (This is an undocumented feature of the native svr4 C compiler.)
43
44CCLIBFLAGS=-Bstatic -dn -gg
45