x86-64: Allow immediate SC to cell-set for both operands.

[sbcl.git] / base-target-features.lisp-expr

;;;; -*- Lisp -*-

;;;; tags which are set during the build process and which end up in
;;;; CL:*FEATURES* in the target SBCL, plus some comments about other
;;;; CL:*FEATURES* tags which have special meaning to SBCL or which
;;;; have a special conventional meaning
;;;;
;;;; Note that the recommended way to customize the features of a
;;;; local build of SBCL is not to edit this file, but instead to
;;;; tweak customize-target-features.lisp. (You must create this file
;;;; first; it is not in the SBCL distribution, and is in fact
;;;; explicitly excluded from the distribution in places like
;;;; .cvsignore.) If you define a function in
;;;; customize-target-features.lisp, it will be used to transform the
;;;; target features list after it's read and before it's used. E.g.,
;;;; you can use code like this:
;;;;    (lambda (list)
;;;;      (flet ((enable (x) (pushnew x list))
;;;;             (disable (x) (setf list (remove x list))))
;;;;        #+nil (enable :sb-show)
;;;;        (enable :sb-after-xc-core)
;;;;        #+nil (disable :sb-doc)
;;;;        list))
;;;; By thus editing a local file (one which is not in the source
;;;; distribution, and which is in .cvsignore) your customizations
;;;; will remain local even if you do things like "cvs update",
;;;; will not show up if you try to submit a patch with "cvs diff",
;;;; and might even stay out of the way if you use other non-CVS-based
;;;; methods to upgrade the files or store your configuration.

;;;; 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.

(
 ;;
 ;; features present in all builds
 ;;

 ;; our standard
 :ansi-cl :common-lisp
 ;; FIXME: Isn't there a :x3jsomething feature which we should set too?
 ;; No. CLHS says ":x3j13 [...] A conforming implementation might or
 ;; might not contain such a feature." -- CSR, 2002-02-21

 ;; our dialect
 :sbcl

 ;; Douglas Thomas Crosher's conservative generational GC (the only one
 ;; we currently support for X86).
 ;; :gencgc used to be here; CSR moved it into
 ;; local-target-features.lisp-expr via make-config.sh, as alpha,
 ;; sparc and ppc ports don't currently support it. -- CSR, 2002-02-21

 ;;
 ;; features present in this particular build
 ;;

 ;; Setting this enables the compilation of documentation strings
 ;; from the system sources into the target Lisp executable.
 ;; Traditional Common Lisp folk will want this option set.
 ;; I (WHN) made it optional because I came to Common Lisp from
 ;; C++ through Scheme, so I'm accustomed to asking
 ;; Emacs about things that I'm curious about instead of asking
 ;; the executable I'm running.
 :sb-doc

 ;; Make more debugging information available (for debugging SBCL
 ;; itself). If you aren't hacking or troubleshooting SBCL itself,
 ;; you probably don't want this set.
 ;;
 ;; At least two varieties of debugging information are enabled by this
 ;; option:
 ;;   * SBCL is compiled with a higher level of OPTIMIZE DEBUG, so that
 ;;     the debugger can tell more about the state of the system.
 ;;   * Various code to print debugging messages, and similar debugging code,
 ;;     is compiled only when this feature is present.
 ;;
 ;; Note that the extra information recorded by the compiler at
 ;; this higher level of OPTIMIZE DEBUG includes the source location
 ;; forms. In order for the debugger to use this information, it has to
 ;; re-READ the source file. In an ordinary installation of SBCL, this
 ;; re-READing may not work very well, for either of two reasons:
 ;;   * The sources aren't present on the system in the same location that
 ;;     they were on the system where SBCL was compiled.
 ;;   * SBCL is using the standard readtable, without the added hackage
 ;;     which allows it to handle things like target features.
 ;; If you want to be able to use the extra debugging information,
 ;; therefore, be sure to keep the sources around, and run with the
 ;; readtable configured so that the system sources can be read.
 ; :sb-show

 ;; Build SBCL with the old CMU CL low level debugger, "ldb". In the
 ;; ideal world you would not need this unless you are messing with
 ;; SBCL at a very low level (e.g., trying to diagnose GC problems, or
 ;; trying to debug assembly code for a port to a new CPU). However,
 ;; experience shows that sooner or later everyone lose()'s, in which
 ;; case SB-LDB can at least provide an informative backtrace.
 :sb-ldb

 ;; This isn't really a target Lisp feature at all, but controls
 ;; whether the build process produces an after-xc.core file. This
 ;; can be useful for shortening the edit/compile/debug cycle when
 ;; you modify SBCL's own source code, as in slam.sh. Otherwise
 ;; you don't need it.
 ; :sb-after-xc-core

 ;; Enable extra debugging output in the assem.lisp assembler/scheduler
 ;; code. (This is the feature which was called :DEBUG in the
 ;; original CMU CL code.)
 ; :sb-show-assem

 ;; Compile the C runtime with support for low-level debugging output
 ;; through FSHOW and FSHOW_SIGNAL.  If enabled, this feature allows
 ;; users to turn on such debugging output using environment variables at
 ;; run-time.
 ; :sb-qshow

 ;; Setting this makes SBCL more "fluid", i.e. more amenable to
 ;; modification at runtime, by suppressing various INLINE declarations,
 ;; compiler macro definitions, FREEZE-TYPE declarations; and by
 ;; suppressing various burning-our-ships-behind-us actions after
 ;; initialization is complete; and so forth. This tends to clobber the
 ;; performance of the system, so unless you have some special need for
 ;; this when hacking SBCL itself, you don't want this set.
 ; :sb-fluid

 ;; Enable code for collecting statistics on usage of various operations,
 ;; useful for performance tuning of the SBCL system itself. This code
 ;; is probably pretty stale (having not been tested since the fork from
 ;; base CMU CL) but might nonetheless be a useful starting point for
 ;; anyone who wants to collect such statistics in the future.
 ; :sb-dyncount

 ;; Enable code for detecting concurrent accesses to the same hash-table
 ;; in multiple threads. Note that this implementation is currently
 ;; (2007-09-11) somewhat too eager: even though in the current implementation
 ;; multiple readers are thread safe as long as there are no writers, this
 ;; code will also trap multiple readers.
 ; :sb-hash-table-debug

 ;; Enabled automatically by make-config.sh for platforms which implement
 ;; short vector SIMD intrinsics.
 ;;
 ; :sb-simd-pack

 ;; Enabled automatically by make-config.sh for platforms which implement
 ;; the %READ-CYCLE-COUNTER VOP. Can be disabled manually: affects TIME.
 ;;
 ;; FIXME: Should this be :SB-CYCLE-COUNTER instead? If so, then the same goes
 ;; for :COMPARE-AND-SWAP-VOPS as well, and a bunch of others. Perhaps
 ;; built-time convenience features like this should all live in eg. SB!INT
 ;; instead?
 ;;
 ; :cycle-counter

 ;; Build with support for an additional dynamic heap
 ;; differing from the main dynamic heap in two ways:
 ;;  1. it is guaranteed to reside below 4GB so that all pointers
 ;;      into it fit in 32 bits. (Only an issue for >32 bit address space)
 ;;  2. all objects therein are immovable, and space is reclaimed
 ;;     by a mark-and-sweep collector.
 ;; That combination of aspects potentially allows various efficiencies
 ;; in code generation, especially for the x86-64 backend.
 ;; The extra space has a fixed size which can only be changed by a rebuild,
 ;; and out-of-space conditions are not easily preventable, so the space
 ;; is sized rather generously to sidestep the issue.
 ;; Additionally, it is assumed that for all objects in the immobile heap,
 ;; speed of allocation of those objects is relatively unimportant.
 ;; If unexpected performance regressions are observed,
 ;; consider disabling this feature and reporting a bug.
 ; :immobile-space

 ;; Allocate most functions in the immobile space.
 ;; Enabled by default if supported.
 ;; The down-side of this feature is that the allocator is significantly
 ;; slower than the allocator for movable code. If a particular application
 ;; is performance-constrained by speed of creation of compiled functions
 ;; (not including closures), the feature can be disabled.
 ; :immobile-code

 ;; Combine the layout pointer, instance-length, and widetag of INSTANCE
 ;; into a single machine word. This represents a space savings of anywhere
 ;; from 4% to 8% in typical applications. (Your mileage may vary).
 ; :compact-instance-header

 ;; Enabled automatically for platforms which implement complex arithmetic
 ;; VOPs. Such platforms should implement real-complex, complex-real and
 ;; complex-complex addition and subtractions (for complex-single-float
 ;; and complex-double-float). They should also also implement complex-real
 ;; and real-complex multiplication, complex-real division, and
 ;; sb!vm::swap-complex, which swaps the real and imaginary parts.
 ;; Finally, they should implement conjugate and complex-real, real-complex
 ;; and complex-complex CL:= (complex-complex EQL would usually be a good
 ;; idea).
 ;;
 ; :complex-float-vops

 ;; Enabled automatically for platforms which implement VOPs for EQL
 ;; of single and double floats.
 ;;
 ; :float-eql-vops

 ;; Enabled automatically for platform that can implement inline constants.
 ;;
 ;; Such platform must implement 5 functions, in SB!VM:
 ;; * canonicalize-inline-constant: converts a constant descriptor (list) into
 ;;    a canonical description, to be used as a key in an EQUAL hash table
 ;;    and to guide the generation of the constant itself.
 ;; * inline-constant-value: given a canonical constant descriptor, computes
 ;;    two values:
 ;;     1. A label that will be used to emit the constant (usually a
 ;;         sb!assem:label)
 ;;     2. A value that will be returned to code generators referring to
 ;;         the constant (on x86oids, an EA object)
 ;; * sort-inline-constants: Receives a vector of unique constants;
 ;;    the car of each entry is the constant descriptor, and the cdr the
 ;;    corresponding label. Destructively returns a vector of constants
 ;;    sorted in emission order. It could actually perform arbitrary
 ;;    modifications to the vector, e.g. to fuse constants of different
 ;;    size.
 ;; * emit-constant-segment-header: receives the vector of sorted constants
 ;;    and a flag (true iff speed > space). Expected to emit padding
 ;;    of some sort between the ELSEWHERE segment and the constants, or some
 ;;    metadata.
 ;; * emit-inline-constant: receives a constant descriptor and its associated
 ;;    label. Emits the constant.
 ;;
 ;; Implementing this features lets VOP generators use sb!c:register-inline-constant
 ;; to get handles (as returned by sb!vm:inline-constant-value) from constant
 ;; descriptors.
 ;;
 ; :inline-constants

 ;; Peter Van Eynde's increase-bulletproofness code for CMU CL
 ;;
 ;; Some of the code which was #+high-security before the fork has now
 ;; been either made unconditional, deleted, or rewritten into
 ;; unrecognizability, but some remains. What remains is not maintained
 ;; or tested in current SBCL, but I haven't gone out of my way to
 ;; break it, either.
 ;;
 ; :high-security
 ; :high-security-support

 ;; low-level thread primitives support
 ;;
 ;; As of SBCL 1.0.33.26, threads are part of the default build on
 ;; x86oid Linux. Other platforms that support them include
 ;; x86oid Darwin, FreeBSD, and Solaris.
 ; :sb-thread

 ;; futex support
 ;;
 ;; While on linux we are able to use futexes for our locking
 ;; primitive, on other platforms we don't have this luxury.
 ;;
 ; :sb-futex

 ;; On some operating systems the FS segment register (used for SBCL's
 ;; thread local storage) is not reliably preserved in signal
 ;; handlers, so we need to restore its value from the pthread thread
 ;; local storage.
 ; :restore-fs-segment-register-from-tls

 ;; On some x86oid operating systems (darwin) SIGTRAP is not reliably
 ;; delivered for the INT3 instruction, so we use the UD2 instruction
 ;; which generates SIGILL instead.
 ; :ud2-breakpoints

 ;; Support for detection of unportable code (when applied to the
 ;; COMMON-LISP package, or SBCL-internal pacakges) or bad-neighbourly
 ;; code (when applied to user-level packages), relating to material
 ;; alteration to packages or to bindings in symbols in packages.
 :sb-package-locks

 ;; Support for the entirety of the 21-bit character space defined by
 ;; the Unicode consortium, rather than the classical 8-bit ISO-8859-1
 ;; character set.
 :sb-unicode

 ;; Support for a full evaluator that can execute all the CL special
 ;; forms, as opposed to the traditional SBCL evaluator which called
 ;; COMPILE for everything complicated.
 :sb-eval
 ;; Support for a different evaluator (interpreter) with improved performance.
 ;; You can't have both.
 ; :sb-fasteval

 ;; Record source location information for variables, classes, conditions,
 ;; packages, etc. Gives much better information on M-. in Slime, but
 ;; increases core size by about 100kB.
 :sb-source-locations

 ;; Record xref data for SBCL internals. This can be rather useful for
 ;; people who want to develop on SBCL itself because it'll make M-?
 ;; (slime-edit-uses) work which lists call/expansion/etc. sites.
 ;; It'll increase the core size by major 5-6mB, though.
 ; :sb-xref-for-internals

 ;; We support package local nicknames. No :sb-prefix here as we vainly
 ;; believe our API is worth copying to other implementations as well.
 ;; This doesn't affect the build at all, merely declares how things are.
 :package-local-nicknames

 ;; This affects the definition of a lot of things in bignum.lisp. It
 ;; doesn't seem to be documented anywhere what systems it might apply
 ;; to. It doesn't seem to be needed for X86 systems anyway.
 ; :32x16-divide

 ;; This is set in classic CMU CL, and presumably there it means
 ;; that the floating point arithmetic implementation
 ;; conforms to IEEE's standard. Here it definitely means that the
 ;; floating point arithmetic implementation conforms to IEEE's standard.
 ;; I (WHN 19990702) haven't tried to verify
 ;; that it does conform, but it should at least mostly conform (because
 ;; the underlying x86 hardware tries).
 :ieee-floating-point

 ;; CMU CL had, and we inherited, code to support 80-bit LONG-FLOAT on the x86
 ;; architecture. Nothing has been done to actively destroy the long float
 ;; support, but it hasn't been thoroughly maintained, and needs at least
 ;; some maintenance before it will work. (E.g. the LONG-FLOAT-only parts of
 ;; genesis are still implemented in terms of unportable CMU CL functions
 ;; which are not longer available at genesis time in SBCL.) A deeper
 ;; problem is SBCL's bootstrap process implicitly assumes that the
 ;; cross-compilation host will be able to make the same distinctions
 ;; between floating point types that it does. This assumption is
 ;; fundamentally sleazy, even though in practice it's unlikely to break down
 ;; w.r.t. distinguishing SINGLE-FLOAT from DOUBLE-FLOAT; it's much more
 ;; likely to break down w.r.t. distinguishing DOUBLE-FLOAT from LONG-FLOAT.
 ;; Still it's likely to be quite doable to get LONG-FLOAT support working
 ;; again, if anyone's sufficiently motivated.
 ; :long-float

 ;; Some platforms don't use a 32-bit off_t by default, and thus can't
 ;; handle files larger than 2GB. This feature will control whether
 ;; we'll try to use platform-specific compilation options to enable a
 ;; 64-bit off_t. The intent is for this feature to be automatically
 ;; enabled by make-config.sh on platforms where it's needed and known
 ;; to work, you shouldn't be enabling it manually. You might however
 ;; want to disable it, if you need to pass file descriptors to
 ;; foreign code that uses a 32-bit off_t.
 ; :largefile

 ;; Enabled automatically on platforms that have VOPs to compute the
 ;; high half of a full word-by-word multiplication.  When disabled,
 ;; SB-KERNEL:%MULTIPLY-HIGH is implemented in terms of
 ;; SB-BIGNUM:%MULTIPLY.
 ; :multiply-high-vops

 ;; SBCL has optional support for zlib-based compressed core files.  Enable
 ;; this feature to compile it in.  Obviously, doing so adds a dependency
 ;; on zlib.
 ; :sb-core-compression

 ;; On certain thread-enabled platforms, synchronization between threads
 ;; for the purpose of stopping and starting the world around GC can be
 ;; performed using safepoints instead of signals.  Enable this feature
 ;; to compile with safepoints and to use them for GC.
 ;; (Replaces use of SIG_STOP_FOR_GC.)
 ; :sb-safepoint

 ;; When compiling with safepoints, the INTERRUPT-THREAD mechanism can
 ;; also use safepoints to roll the target thread to a point at which it
 ;; can be interrupted safely, instead of using a signal for this
 ;; purpose.  Enable this feature in addition to :SB-SAFEPOINT to enable
 ;; such behaviour.
 ;; (Replaces use of SIGPIPE, except to wake up syscalls.)
 ; :sb-thruption

 ;; When compiling with safepoints and thruptions, the TIMER facility
 ;; can replace its use of setitimer with a background thread.
 ;; (Replaces use of SIGALRM.)
 ; :sb-wtimer

 ;; This platform implements VOPs for %ash/right, variable-width shift right
 ; :ash-right-vops

 ;;
 ;; miscellaneous notes on other things which could have special significance
 ;; in the *FEATURES* list
 ;;

 ;; Any target feature which affects binary compatibility of fasl files
 ;; needs to be recorded in *FEATURES-POTENTIALLY-AFFECTING-FASL-FORMAT*
 ;; (elsewhere).

 ;; notes on the :NIL and :IGNORE features:
 ;;
 ;; #+NIL is used to comment out forms. Occasionally #+IGNORE is used
 ;; for this too. So don't use :NIL or :IGNORE as the names of features..

 ;; notes on :SB-XC and :SB-XC-HOST features (which aren't controlled by this
 ;; file, but are instead temporarily pushed onto *FEATURES* or
 ;; *TARGET-FEATURES* during some phases of cross-compilation):
 ;;
 ;; :SB-XC-HOST stands for "cross-compilation host" and is in *FEATURES*
 ;; during the first phase of cross-compilation bootstrapping, when the
 ;; host Lisp is being used to compile the cross-compiler.
 ;;
 ;; :SB-XC stands for "cross compiler", and is in *FEATURES* during the second
 ;; phase of cross-compilation bootstrapping, when the cross-compiler is
 ;; being used to create the first target Lisp.

 ;; notes on the :SB-ASSEMBLING feature (which isn't controlled by
 ;; this file):
 ;;
 ;; This is a flag for whether we're in the assembler. It's
 ;; temporarily pushed onto the *FEATURES* list in the setup for
 ;; the ASSEMBLE-FILE function. It would be a bad idea
 ;; to use it as a name for a permanent feature.

 ;; notes on local features (which are set automatically by the
 ;; configuration script, and should not be set here unless you
 ;; really, really know what you're doing):
 ;;
 ;; machine architecture features:
 ;;   :x86
 ;;      any Intel 386 or better, or compatibles like the AMD K6 or K7
 ;;   :x86-64
 ;;      any x86-64 CPU running in 64-bit mode
 ;;   :alpha
 ;;      DEC/Compaq Alpha CPU
 ;;   :sparc
 ;;      any Sun UltraSPARC (possibly also non-Ultras -- currently untested)
 ;;   :ppc
 ;;      any PowerPC CPU
 ;;   :hppa
 ;;      any PA-RISC CPU
 ;;   :mips
 ;;      any MIPS CPU (in little-endian mode with :little-endian)
 ;;   :arm
 ;;      an ARM CPU (details yet to be determined)
 ;;   :arm64
 ;;     an ARMv8 AArch64 CPU
 ;; (CMU CL also had a :pentium feature, which affected the definition
 ;; of some floating point vops. It was present but not enabled or
 ;; documented in the CMU CL code that SBCL is derived from, and has
 ;; now been moved to the backend-subfeatures mechanism.)
 ;;
 ;; properties derived from the machine architecture
 ;;
 ;;   :64-bit
 ;;     means (= sb-vm:n-word-bits 64) currently true for x86-64 and arm64
 ;;
 ;;   :64-bit-registers
 ;;     means (= sb-vm:n-machine-word-bits 64) currently true for alpha,
 ;;     arm64, and x86-64.
 ;;
 ;;   :control-stack-grows-downward-not-upward
 ;;     On the X86, the Lisp control stack grows downward. On the
 ;;     other supported CPU architectures as of sbcl-0.7.1.40, the
 ;;     system stack grows upward.
 ;;   Note that there are other stack-related differences between the
 ;;   X86 port and the other ports. E.g. on the X86, the Lisp control
 ;;   stack coincides with the C stack, meaning that on the X86 there's
 ;;   stuff on the control stack that the Lisp-level debugger doesn't
 ;;   understand very well. As of sbcl-0.7.1.40 things like that are
 ;;   just parameterized by #!+X86, but it'd probably be better to
 ;;   use new flags like :CONTROL-STACK-CONTAINS-C-STACK.
 ;;
 ;;   :stack-allocatable-closures
 ;;     The compiler can allocate dynamic-extent closures on stack.
 ;;
 ;;   :alien-callbacks
 ;;     Alien callbacks have been implemented for this platform.
 ;;
 ;;   :compare-and-swap-vops
 ;;     The backend implements compare-and-swap VOPs.
 ;;
 ;;   :memory-barrier-vops
 ;;     Memory barriers (for multi-threaded synchronization) have been
 ;;     implemented for this platform.
 ;;
 ;; operating system features:
 ;;   :unix       = We're intended to run under some Unix-like OS. (This is not
 ;;                 exclusive with the features which indicate which particular
 ;;                 Unix-like OS we're intended to run under.)
 ;;   :linux      = We're intended to run under some version of Linux.
 ;;   :bsd        = We're intended to run under some version of BSD Unix. (This
 ;;                 is not exclusive with the features which indicate which
 ;;                 particular version of BSD we're intended to run under.)
 ;;   :freebsd    = We're intended to run under FreeBSD.
 ;;   :openbsd    = We're intended to run under OpenBSD.
 ;;   :netbsd     = We're intended to run under NetBSD.
 ;;   :dragonfly  = We're intended to run under DragonFly BSD.
 ;;   :darwin     = We're intended to run under Darwin (including MacOS X).
 ;;   :sunos      = We're intended to run under Solaris user environment
 ;;                 with the SunOS kernel.
 ;;   :hpux       = We're intended to run under HP-UX 11.11 or later
 ;;   :osf1       = We're intended to run under Tru64 (aka Digital Unix
 ;;                 aka OSF/1).
 ;;   :win32      = We're intended to under some version of Microsoft Windows.
 ;; (No others are supported by SBCL as of 1.0.8, but :hpux or :irix
 ;; support could be ported from CMU CL if anyone is sufficiently
 ;; motivated to do so.)
 )