steve
/
factor
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
factor/basis/cpu/x86/32/bootstrap.factor

40 lines
1.1 KiB
Factor
Raw Normal View History

JIT now supports multiple relocations per code template. This simplifies non-optimizing compiler backends
2009-05-06 21:04:49 -04:00
! Copyright (C) 2007, 2009 Slava Pestov.
Initial import
2007-09-20 18:09:08 -04:00
! See http://factorcode.org/license.txt for BSD license.
USING: bootstrap.image.private kernel namespaces system
cpu.x86: update non-optimizing compiler backends for assembler vocab split
2009-07-30 03:22:37 -04:00
cpu.x86.assembler cpu.x86.assembler.operands layouts
vocabs parser compiler.constants ;
More changes
2007-09-21 23:30:47 -04:00
IN: bootstrap.x86
Initial import
2007-09-20 18:09:08 -04:00
4 \ cell set
Starting work on Win64 port
2008-11-07 21:33:32 -05:00
: stack-frame-size ( -- n ) 4 bootstrap-cells ;
fixnum-shift-fast and fixnum-mod subprimitives
2008-11-03 02:18:35 -05:00
: shift-arg ( -- reg ) ECX ;
: div-arg ( -- reg ) EAX ;
: mod-arg ( -- reg ) EDX ;
x86 bootstrap cleanup: renamed arg to arg1
2009-09-23 14:45:18 -04:00
: arg1 ( -- reg ) EAX ;
passing vm ptr to lazy_jit_compile mostly working
2009-08-19 14:02:12 -04:00
: arg2 ( -- reg ) EDX ;
Clean up JIT backend a little
2008-12-09 18:14:17 -05:00
: temp0 ( -- reg ) EAX ;
: temp1 ( -- reg ) EDX ;
: temp2 ( -- reg ) ECX ;
: temp3 ( -- reg ) EBX ;
vm: code heap compaction at runtime using compact-gc primitive
2009-10-16 12:39:22 -04:00
: safe-reg ( -- reg ) EAX ;
Fix compile errors
2008-06-11 05:52:19 -04:00
: stack-reg ( -- reg ) ESP ;
: ds-reg ( -- reg ) ESI ;
Non-optimizing compiler now inlow inlines some primitives, this improves bootstrap time
2008-07-07 11:39:04 -04:00
: rs-reg ( -- reg ) EDI ;
Clean up JIT backend a little
2008-12-09 18:14:17 -05:00
: fixnum>slot@ ( -- ) temp0 1 SAR ;
Fix compile errors
2008-06-11 05:52:19 -04:00
: rex-length ( -- n ) 0 ;
Initial import
2007-09-20 18:09:08 -04:00
New calling convention for VM primitives: instead of the Factor side passing the stack pointer as the first parameter, and having the VM save it to stack_chain->top, we instead have the Factor side save it. Eliminates a lot of crud in the VM
2008-11-13 04:20:34 -05:00
[
bootstrap.image: clean up jit-define and define-sub-primitive so that
2009-04-25 23:35:19 -04:00
! load stack_chain
temp0 0 [] MOV rc-absolute-cell rt-stack-chain jit-rel
! save stack pointer
temp0 [] stack-reg MOV
vm ptr passed to primitives on X86.32 (other cpus still use singleton vm ptr)
2009-08-18 15:49:22 -04:00
! pass vm ptr to primitive
x86 bootstrap cleanup: renamed arg to arg1
2009-09-23 14:45:18 -04:00
arg1 0 MOV rc-absolute-cell rt-vm jit-rel
JIT now supports multiple relocations per code template. This simplifies non-optimizing compiler backends
2009-05-06 21:04:49 -04:00
! call the primitive
Better separation of concerns: cpu.{x86,ppc}.assembler no longer depends on compiler.codegen.fixup and cpu.architecture. Rename rt-xt-direct to rt-xt-pic to better explain its purpose
2009-05-06 17:14:53 -04:00
0 JMP rc-relative rt-primitive jit-rel
bootstrap.image: clean up jit-define and define-sub-primitive so that
2009-04-25 23:35:19 -04:00
] jit-primitive jit-define
New calling convention for VM primitives: instead of the Factor side passing the stack pointer as the first parameter, and having the VM save it to stack_chain->top, we instead have the Factor side save it. Eliminates a lot of crud in the VM
2008-11-13 04:20:34 -05:00
Add vocab: for vocab-relative paths
2009-02-15 20:53:21 -05:00
<< "vocab:cpu/x86/bootstrap.factor" parse-file parsed >>
Fix bootstrap for a different cell size
2008-01-26 22:47:52 -05:00
call