For sse2, we don't need to do the fldz/fstp dance when doing a
foreign function call.

A cross-compile is needed!

 * src/bootfiles/20e/boot-2014-04-cross-x86.lisp:
   * Cross-compile script to enable these changes.
 * src/compiler/x86/c-call.lisp:
   * Make the float result tn be xmm registers, not x87 fpu registers.
 * src/compiler/x86/sse2-c-call.lisp:
   * Simplify call-out not to use fldz/fstp except for the minimum
     required for the 32-bit ABI.
 * src/lisp/x86-assem.S
   * Comment out the fstp/fldz insts that aren't needed for SSE2.

stack space to hold the SSE2 state.  (Forgot to add extra bytes for
16-byte alignment!)

src/lisp/x86-assem.S:
o Remove the (now) unused alloc_to_<reg> functions.

src/compiler/x86/macros.lisp:
o Simplify the not-inline-allocation to use the same routines as
  inline allocation.

src/bootfiles/20c/boot-2012-07-1-x86-cross.lisp:
o X86 cross-compile script

src/compiler/x86/macros.lisp:
o New inline-allocation to call our new overflow routine.

src/lisp/x86-assem.S:
o New alloc_overflow routine.

src/tools/cross-scripts/cross-x86-x86.lisp:
o Export all symbols in VM that are external symbols in OLD-VM with
  the corresponding symbol name.

fine with gcc (on darwin and linux).

lisp/x86-assem.S:
o Make sure the stack is 16-byte aligned in the alloc_overflow_foo and
  alloc_to_foo routines.  These eventually call into C code, and the
  stack is required to be 16-byte aligned on Darwin.  We apply this to
  all x86 implementations since it's harmless.
o Did not update the alloc_8/16_to_foo routines because they are going
  to be deleted.

x86/macros.lisp:
o Don't call the alloc_8/16_to_foo routines when we're not doing
  inline allocation.  I don't think there's much to be gained with
  these special functions and maintainence is a pain with assembly
  code.

Make the size of all spaces configurable from the command line.

code/commandline.lisp:
o Define the switches so lisp doesn't complain.

general-info/lisp.1:
o Document new switches.

general-info/release-20c.txt:
o Update

lisp/Darwin-os.c
lisp/FreeBSD-os.c
lisp/Linux-os.c
lisp/NetBSD-os.c
lisp/backtrace.c
lisp/cgc.c
lisp/gencgc.c
lisp/os-common.c
lisp/ppc-validate.h
lisp/purify.c
lisp/solaris-os.c
lisp/sparc-arch.c
lisp/sparc-validate.h
lisp/x86-validate.h:
lisp/validate.c
o Basically rename CONTROL_STACK_SIZE, BINDING_STACK_SIZE,
  READ_ONLY_SPACE_SIZE, and STATIC_SPACE_SIZE to control_stack_size,
  binding_stack_size, read_only_space_size, and static_space_size,
  respectively.

lisp/coreparse.c
o Check the space size in the core file against the allocated size.
  If the allocated space is too small, print an error message and
  exit.

lisp/globals.c
o Define new variables for the allocated size of spaces.

lisp/globals.h
o Declare the new variables for the size of the spaces.

lisp/lisp.c
o Set the default values of the spaces.
o Parse the new command-line flags and set the space sizes.

lisp/sunos-os.c:
o Added os_init0 (that we forgot last time).

lisp/x86-assem.S:
o Change $CONTROL_STACK_END to GNAME(control_stack_end) to get the
  correct end of the stack.

Solaris/x86.  There should be no functional changes for either other
x86 ports or for the sparc port.

closure_tramp and undefined_tramp from the C side to the Lisp side via
Lisp assembly routines.  (Simplifies things quite a bit for sparc.)

Also includes some GC fixes for sparc and ppc to handle moving the pc,
npc, lr, and ctr registers.  They need to be handled like the lip
register.

See bootfiles/20a/boot-2010-07-1-cross.lisp for instructions on how to
bootstrap this change.  A basic cross-compile with a specific cross
bootstrap file is needed.

(defun testfn ()
  (let* ((i-gc-ed-u nil)
         (ext:*gc-notify-before* (lambda (a) (/ a 0.34d0))))
    (dotimes (i 100000)
      (setf i-gc-ed-u nil)
      (let* ((v1 (list (* 1d0 (random 10)) (* 1d0 (random 10))))
             (v2 (list (* 1d0 (random 10)) (* 1d0 (random 10))))
             (dot1 (reduce #'+ (mapcar #'* v1 v2)))
             (dot2 (reduce #'+ (mapcar #'* v1 v2))))
        (when (/= dot1 dot2)
          (print `(gc ,i-gc-ed-u v1 ,v1 v2 ,v2 dot1 ,dot1 dot2 ,dot2)))))))

Running this with sse2 would cuase dot1 and dot2 to sometimes be
different.  We forgot to save the SSE2 state in call_into_lisp.

(This bug was mentioned on comp.lang.lisp, http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/828371aa4800272c?hl=en#

x86-assem.S:
o Save and restore SSE2 state when running with SSE2 core.  (This
  might need more tweaking.  Should we use cpuid instead of looking at
  fpu_mode?)

gencgc.c:
o Save the SSE2 state, along with X87 state for SSE2 cores.

works on Darwin and Linux.  No more hard-coded jmp instruction on
Darwin!

function_end_breakpoint_guts needs to be two bytes long but the
assembler was using a five byte version.

sse2-packed-2008-11-12).

register as an argument to FNSTSW.  Make it so.

pre-processor macros.  Also, use the more compatible .globl rather than
.global and enliminate the unused eight byte alignment macro.

FreeBSD.

demonstrably slower than extant compiler intrinsics or C runtime
memset() implementations for blocks of memory greater than or equal to
an x86 VM page.

	* x86-assem.S (multiple_value_return): Add new functions to handle
	heap overflow warnings and errors.

	* x86-arch.c (sigtrap_handler): Add cases for the two new heap
	overflow traps.

	* interrupt.c (interrupt_handle_space_overflow): Add support for
	x86.

	* parms.lisp (static-symbols): Add new static symbols for heap
	overflow checking.  Add new trap types for heap overflow
	checking.

o Conditionalized some code that is only used in linkage-table builds
  on :linkage-table and LINKAGE_TABLE respectively.
o Modernized undefined_ff_tramp based on recent changes, and renamed
  it to undefined_foreign_symbol_trap, because it really isn't a
  trampoline.
o Conditionalized a stray puts(...,dlerror()) in Linux-os.c on DEBUG
o Installed an error reporter for UNDEFINED_FOREIGN_SYMBOL_ERROR,
  which for now signals a simple-program-error.  We might want to
  introduced a new condition UNDEFINED-FOREIGN-SYMBOL or something,
  which could also be used by the old foreign linkage code.

On x86 FreeBSD and Linux, change the way foreign symbol addresses are resolved.
They now go through a table -- effectively a new space in the core file.
Function references are resolved lazily, data references are resolved on startup
and when a .so is loaded.  The end result is that cores can be dumped that
contain references to symbols in shared libraries.  Also, the dependence of the core on addresses in the Lisp runtime is broken.

The linkage table feature is controlled by :linkage-table and LINKAGE_TABLE in C
runtime.  Several foreign symbols are now Lisp static symbols, so a cross
compile is required whether or not the new stuff is used.  I've checked in
boot4-cross-foreign-linkage.lisp that builds the compiler for linkage table; do whatever you usually do for the non-linkage table case:)  Seriously, lets start
a discussion on standardizing "cross compilation," not to mention the general
build procedure.

platform assembler code can use those, instead of magic numbers and
bogus conditionalization on type_Long_Float.  Inspired/Stolen from
SBCL.  Also added some more comments in places, and a generic
SC_OFFSET macro that matches lisp/debug-info.lisp.

was already cleaned up with the OpenBSD port, this doesn't require
massive changes.

Since current NetBSD is ELF-based by default, we don't make use of the
ELF feature, or in other words:  The presence of the NetBSD feature
implies ELF, since there is no support for non-ELF NetBSD.

bits as used by the long-float build. This corrects the handling of
double-float-epsilon for the double-float build, and should also
improve performance on some processors such as the AMD Athlon.

'jmp *offset(%eax)' rather than 'jmp offset(%eax)'. Peter Van Eynde
reports that the older syntax is incorrectly assembled on recent linux
systems. The new syntax does work on recent FreeBSD systems.

(like Linux) rather than bits.

that this may need to be undone if someone wants to build
with the aout compatibility package. From Martin Cracauer.

monitor can use the NPX (for printing float objects), and restore the
state when returning. This may be unnecessary on Linux?

been returning two values.

since purify was patched some time ago.