steve
/
factor
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
factor/basis/compiler/constants/constants.factor

85 lines
3.3 KiB
Factor
Raw Normal View History

Change how non-volatile register preservation is done in alien callbacks, with the aim of fixing callbacks on PowerPC, and to eventually eliminate assembly code from VM - Simplify calculation of offset in relocation table - Open-code %alien-callback - Remove magic_frame hack from context objects - Move magical return instruction from optimizing compiler backend into callback entry stub
2010-01-02 07:03:30 -05:00
! Copyright (C) 2008, 2010 Slava Pestov.
Redo the profiler
2008-01-02 19:36:36 -05:00
! See http://factorcode.org/license.txt for BSD license.
Non-optimizing compiler now open-codes megamorphic dispatch fast path
2009-04-30 04:37:07 -04:00
USING: math kernel layouts system strings words quotations byte-arrays
Remove ENUM: f and replace uses with CONSTANTs. Fix bootstrap and load-all errors from enum classes.
2010-04-18 16:34:18 -04:00
alien arrays literals sequences ;
Redo the profiler
2008-01-02 19:36:36 -05:00
IN: compiler.constants
! These constants must match vm/memory.h
use CONSTANT: in basis
2009-02-22 20:13:08 -05:00
CONSTANT: card-bits 8
CONSTANT: deck-bits 18
PowerPC backend almost functional; some new compiler unit tests added, better compilation of 'f eq?'; f becomes an immediate operand move aux-offset to compiler.constants
2008-11-06 07:27:27 -05:00
: card-mark ( -- n ) HEX: 40 HEX: 80 bitor ; inline
Redo the profiler
2008-01-02 19:36:36 -05:00
! These constants must match vm/layouts.h
compiler: tweak ##write-barrier-imm
2009-10-15 03:40:23 -04:00
: slot-offset ( slot tag -- n ) [ bootstrap-cells ] dip - ; inline
vm: 4 bit tags, new representation of alien objects makes unbox-any-c-ptr more efficient (work in progress)
2009-11-02 04:25:39 -05:00
: float-offset ( -- n ) 8 float type-number - ; inline
: string-offset ( -- n ) 4 string type-number slot-offset ; inline
: string-aux-offset ( -- n ) 2 string type-number slot-offset ; inline
: profile-count-offset ( -- n ) 8 \ word type-number slot-offset ; inline
: byte-array-offset ( -- n ) 16 byte-array type-number - ; inline
: alien-offset ( -- n ) 4 alien type-number slot-offset ; inline
: underlying-alien-offset ( -- n ) 1 alien type-number slot-offset ; inline
: tuple-class-offset ( -- n ) 1 tuple type-number slot-offset ; inline
vm: strip out call-counting profiler This makes the separate "code" and "entry_point" fields in word and quotation redundant, so also remove them to reclaim an additional cell per word and quotation object, which should help with #318.
2011-11-10 19:00:47 -05:00
: word-entry-point-offset ( -- n ) 9 \ word type-number slot-offset ; inline
Stop wearing monocle and use the term "entry point" instead of "XT" throughout VM and compiler; also remove two unused relocation types
2010-01-18 02:51:27 -05:00
: quot-entry-point-offset ( -- n ) 4 quotation type-number slot-offset ; inline
vm: strip out call-counting profiler This makes the separate "code" and "entry_point" fields in word and quotation redundant, so also remove them to reclaim an additional cell per word and quotation object, which should help with #318.
2011-11-10 19:00:47 -05:00
: code-block-entry-point-offset ( -- n ) 4 bootstrap-cells ; inline
vm: 4 bit tags, new representation of alien objects makes unbox-any-c-ptr more efficient (work in progress)
2009-11-02 04:25:39 -05:00
: array-start-offset ( -- n ) 2 array type-number slot-offset ; inline
compiler.constants: update compiled-header-size
2009-05-05 16:06:05 -04:00
: compiled-header-size ( -- n ) 4 bootstrap-cells ; inline
vm: move c_to_factor, lazy_jit_compile_impl, throw_impl, set_callstack assembly routines into non-optimizing compiler for x86-64
2010-01-05 21:47:36 -05:00
: callstack-length-offset ( -- n ) 1 \ callstack type-number slot-offset ; inline
: callstack-top-offset ( -- n ) 2 \ callstack type-number slot-offset ; inline
Update x86-32 for assembly entry point changes
2010-01-05 23:55:20 -05:00
: vm-context-offset ( -- n ) 0 bootstrap-cells ; inline
vm: actually use context callstacks when running code
2010-03-26 22:44:43 -04:00
: vm-spare-context-offset ( -- n ) 1 bootstrap-cells ; inline
vm: resumably handle signals from leaf procedures
2011-10-25 02:46:34 -04:00
: vm-signal-handler-addr-offset ( -- n ) 8 bootstrap-cells ; inline
cpu.*.bootstrap: clear faulting flag directly No reason to go through a safepoint in unwind-native-frames really.
2011-11-12 15:57:24 -05:00
: vm-fault-flag-offset ( -- n ) 9 bootstrap-cells ; inline
Update x86-32 for assembly entry point changes
2010-01-05 23:55:20 -05:00
: context-callstack-top-offset ( -- n ) 0 bootstrap-cells ; inline
: context-callstack-bottom-offset ( -- n ) 1 bootstrap-cells ; inline
: context-datastack-offset ( -- n ) 2 bootstrap-cells ; inline
: context-retainstack-offset ( -- n ) 3 bootstrap-cells ; inline
vm: actually use context callstacks when running code
2010-03-26 22:44:43 -04:00
: context-callstack-save-offset ( -- n ) 4 bootstrap-cells ; inline
Get green threads working on Windows - store stack base and limit in TIB - set up a frame-based structured exception handler in each context's callstack - boot.x86.32.image has now been replaced by boot.winnt-x86.32.image and boot.unix-x86.32.image
2010-04-03 20:24:33 -04:00
: context-callstack-seg-offset ( -- n ) 7 bootstrap-cells ; inline
: segment-start-offset ( -- n ) 0 bootstrap-cells ; inline
: segment-size-offset ( -- n ) 1 bootstrap-cells ; inline
: segment-end-offset ( -- n ) 2 bootstrap-cells ; inline
Updating x86 backend for new codegen
2008-09-17 01:46:38 -04:00
! Relocation classes
Remove ENUM: f and replace uses with CONSTANTs. Fix bootstrap and load-all errors from enum classes.
2010-04-18 16:34:18 -04:00
CONSTANT: rc-absolute-cell 0
CONSTANT: rc-absolute 1
CONSTANT: rc-relative 2
CONSTANT: rc-absolute-ppc-2/2 3
CONSTANT: rc-absolute-ppc-2 4
32 and 64 bit Linux PPC support
2011-05-20 18:11:50 -04:00
CONSTANT: rc-relative-ppc-2-pc 5
CONSTANT: rc-relative-ppc-3-pc 6
Remove ENUM: f and replace uses with CONSTANTs. Fix bootstrap and load-all errors from enum classes.
2010-04-18 16:34:18 -04:00
CONSTANT: rc-relative-arm-3 7
CONSTANT: rc-indirect-arm 8
CONSTANT: rc-indirect-arm-pc 9
CONSTANT: rc-absolute-2 10
CONSTANT: rc-absolute-1 11
32 and 64 bit Linux PPC support
2011-05-20 18:11:50 -04:00
CONSTANT: rc-absolute-ppc-2/2/2/2 12
Updating x86 backend for new codegen
2008-09-17 01:46:38 -04:00
! Relocation types
Remove ENUM: f and replace uses with CONSTANTs. Fix bootstrap and load-all errors from enum classes.
2010-04-18 16:34:18 -04:00
CONSTANT: rt-dlsym 0
CONSTANT: rt-entry-point 1
CONSTANT: rt-entry-point-pic 2
CONSTANT: rt-entry-point-pic-tail 3
CONSTANT: rt-here 4
CONSTANT: rt-this 5
CONSTANT: rt-literal 6
CONSTANT: rt-untagged 7
CONSTANT: rt-megamorphic-cache-hits 8
CONSTANT: rt-vm 9
CONSTANT: rt-cards-offset 10
CONSTANT: rt-decks-offset 11
CONSTANT: rt-exception-handler 12
32 and 64 bit Linux PPC support
2011-05-20 18:11:50 -04:00
CONSTANT: rt-dlsym-toc 13
Add RT_INLINE_CACHE_MISS relocation type to avoid frequent dlsym lookups when compiling PICs. Fixes #123
2011-09-14 04:08:32 -04:00
CONSTANT: rt-inline-cache-miss 14
compiler: rt-safepoint and rel-safepoint words
2011-10-17 18:23:05 -04:00
CONSTANT: rt-safepoint 15
Updating x86 backend for new codegen
2008-09-17 01:46:38 -04:00
: rc-absolute? ( n -- ? )
vm: add a new rc-absolute-1 relocation class to allow storing values in 8-bit operands, and optimized code sequences for inline caches using this
2010-04-12 17:22:41 -04:00
${
Remove ENUM: f and replace uses with CONSTANTs. Fix bootstrap and load-all errors from enum classes.
2010-04-18 16:34:18 -04:00
$ rc-absolute-ppc-2/2
$ rc-absolute-cell
$ rc-absolute
$ rc-absolute-2
$ rc-absolute-1
vm: add a new rc-absolute-1 relocation class to allow storing values in 8-bit operands, and optimized code sequences for inline caches using this
2010-04-12 17:22:41 -04:00
} member? ;