;;;; This file contains some parameterizations of various VM
;;;; attributes for the PPC. This file is separate from other stuff so
;;;; that it can be compiled and loaded earlier.
;;;; This software is part of the SBCL system. See the README file for
;;;; more information.
;;;;
;;;; This software is derived from the CMU CL system, which was
;;;; written at Carnegie Mellon University and released into the
;;;; public domain. The software is in the public domain and is
;;;; provided with absolutely no warranty. See the COPYING and CREDITS
;;;; files for more information.
(in-package "SB-VM")
; ok, so we're supposed to use our instruction scheduler, but we don't,
; because of "needs a little more work in the assembler"
(defconstant sb-assem:assem-scheduler-p nil)
(defconstant sb-assem:+inst-alignment-bytes+ 4)
(defconstant +backend-fasl-file-implementation+ :ppc)
;; On Linux, the ABI specifies the page size to be 4k-64k, use the
;; maximum of that range. FIXME: it'd be great if somebody would
;; find out whether using exact multiples of the page size actually
;; matters in the few places where that's done, or whether we could
;; just use 4k everywhere.
(defconstant +backend-page-bytes+ #+linux 65536 #-linux 4096)
;;; The size in bytes of GENCGC cards, i.e. the granularity at which
;;; writes to old generations are logged. With mprotect-based write
;;; barriers, this must be a multiple of the OS page size.
(defconstant gencgc-card-bytes +backend-page-bytes+)
;;; The minimum size of new allocation regions. While it doesn't
;;; currently make a lot of sense to have a card size lower than
;;; the alloc granularity, it will, once we are smarter about finding
;;; the start of objects.
(defconstant gencgc-alloc-granularity 0)
;;; The minimum size at which we release address ranges to the OS.
;;; This must be a multiple of the OS page size.
(defconstant gencgc-release-granularity +backend-page-bytes+)
;;; number of bits per word where a word holds one lisp descriptor
(defconstant n-word-bits 64)
;;; the natural width of a machine word (as seen in e.g. register width,
;;; address space)
(defconstant n-machine-word-bits 64)
;;; flags for the generational garbage collector
(defconstant pseudo-atomic-interrupted-flag 1)
(defconstant pseudo-atomic-flag 4)
(defconstant float-inexact-trap-bit (ash 1 0))
(defconstant float-divide-by-zero-trap-bit (ash 1 1))
(defconstant float-underflow-trap-bit (ash 1 2))
(defconstant float-overflow-trap-bit (ash 1 3))
(defconstant float-invalid-trap-bit (ash 1 4))
(defconstant float-round-to-nearest 0)
(defconstant float-round-to-zero 1)
(defconstant float-round-to-positive 2)
(defconstant float-round-to-negative 3)
(defconstant-eqx float-rounding-mode (byte 2 0) #'equalp) ; RD
;;; FIXME I: Beware, all ye who trespass here. Despite its name,
;;; FLOAT-STICKY-BITS is not the byte specifier for sticky bits in the
;;; floating point control word. It is more like "accrued exceptions"
;;; where FLOAT-EXCEPTIONS-BYTE is "current exceptions". Consequently,
;;; on architectures where there is no "current exceptions"
;;; FLOAT-EXCEPTIONS-BYTE and FLOAT-STICKY-BITS had better be the
;;; same.
;;;
;;; FIXME II: So, I've now documented this in comments in the PowerPC
;;; tree. This may not make it easy to find for when new architectures
;;; get backends written...
;;;
;;; CSR, 2002-06-11
(defconstant-eqx float-sticky-bits (byte 5 25) #'equalp)
(defconstant-eqx float-traps-byte (byte 5 3) #'equalp)
(defconstant-eqx float-exceptions-byte (byte 5 25) #'equalp) ; cexc
(defconstant-eqx float-invalid-byte (byte 6 19) #'equalp)
(defconstant float-fast-bit 2) ; Non-IEEE mode
;;;; Where to put the different spaces.
;;; On non-gencgc we need large dynamic and static spaces for PURIFY
#-gencgc
(progn
(defconstant read-only-space-start #x04000000)
(defconstant read-only-space-end #x07ff8000)
(defconstant static-space-start #x08000000)
(defconstant static-space-end #x097fff00)
(defconstant linkage-table-space-start #x0a000000)
(defconstant linkage-table-space-end #x0b000000))
;;; While on gencgc we don't.
#+gencgc (!gencgc-space-setup #x04000000
:read-only-space-size 0
:dynamic-space-start #x1000000000)
(defconstant linkage-table-growth-direction :up)
(defconstant linkage-table-entry-size #+little-endian 28 #+big-endian 24)
#+linux
(progn
#-gencgc
(progn
(defparameter dynamic-0-space-start #x4f000000)
(defparameter dynamic-0-space-end #x66fff000)))
#+netbsd
(progn
#-gencgc
(progn
(defparameter dynamic-0-space-start #x4f000000)
(defparameter dynamic-0-space-end #x66fff000)))
;;; Text and data segments start at #x01800000. Range for randomized
;;; malloc() starts #x20000000 (MAXDSIZ) after end of data seg and
;;; extends 256 MB. Use 512 - 64 MB for dynamic space so we can run
;;; under default resource limits.
;;; FIXME: MAXDSIZ is a kernel parameter, and can vary as high as 1GB.
;;; These parameters should probably be tested under such a configuration,
;;; as rare as it might or might not be.
#+openbsd
(progn
#-gencgc
(progn
(defparameter dynamic-0-space-start #x4f000000)
(defparameter dynamic-0-space-end #x5cfff000)))
#+darwin
(progn
#-gencgc
(progn
(defparameter dynamic-0-space-start #x10000000)
(defparameter dynamic-0-space-end #x3ffff000)))
(defenum (:start 8)
halt-trap
pending-interrupt-trap
cerror-trap
breakpoint-trap
fun-end-breakpoint-trap
after-breakpoint-trap
single-step-around-trap
single-step-before-trap
error-trap)
;;;; Static symbols.
;;; These symbols are loaded into static space directly after NIL so
;;; that the system can compute their address by adding a constant
;;; amount to NIL.
;;;
;;; The fdefn objects for the static functions are loaded into static
;;; space directly after the static symbols. That way, the raw-addr
;;; can be loaded directly out of them by indirecting relative to NIL.
;;;
(defconstant-eqx +static-symbols+
`#(,@+common-static-symbols+)
#'equalp)
(defconstant-eqx +static-fdefns+
#(length
two-arg-+
two-arg--
two-arg-*
two-arg-/
two-arg-<
two-arg->
two-arg-=
two-arg-<=
two-arg->=
two-arg-/=
eql
%negate
two-arg-and
two-arg-ior
two-arg-xor
two-arg-eqv
two-arg-gcd
two-arg-lcm)
#'equalp)