|
|
@14111
|
11 years |
gb |
Yet another scheme for subprim calls. Go ahead and laugh.
Details:
- …
|
|
|
@14109
|
11 years |
gb |
props
|
|
|
@14108
|
11 years |
gb |
arm-lapmacros: define LRI, which loads a 32-bit constant into a GPR …
|
|
|
@14107
|
11 years |
gb |
More subprim changes.
|
|
|
@14106
|
11 years |
gb |
Current image.
|
|
|
@14105
|
11 years |
gb |
Deal with exception frames better, especially for C.
|
|
|
@14104
|
11 years |
gb |
Don't use "ba" pseudo-instruction.
|
|
|
@14103
|
11 years |
gb |
Don't use "ba" pseudo-instruction.
|
|
|
@14102
|
11 years |
gb |
Don't use "ba" pseudo-instruction.
|
|
|
@14101
|
11 years |
gb |
un-botch interrupt check.
|
|
|
@14100
|
11 years |
gb |
Finish bootstrapping STRUCTURE-TYPEP. (Started 2 years ago.)
|
|
|
@14099
|
11 years |
gb |
Register usage in ARM2-ASET2; subprim changes.
|
|
|
@14098
|
11 years |
gb |
Fix braino in UNBOX-S8; new subprim call/jump scheme.
|
|
|
@14097
|
11 years |
gb |
Fix COMPOSE-DIGIT; add LOAD-SUBPRIM.
|
|
|
@14096
|
11 years |
gb |
No more arm::*called-subprim-jmp-labels*.
|
|
|
@14095
|
11 years |
gb |
Stop supporting the "BA" pseudoinstruction (which was really a
"mov …
|
|
|
@14094
|
11 years |
gb |
Make a built-in-class for FAKE-STACK-FRAME.
|
|
|
@14093
|
11 years |
gb |
SIMPLE-1D-ARRAY-SUBSEQ: fix conditionalization.
|
|
|
@14092
|
11 years |
gb |
_SPbuiltin_ash: right shift via asr, not lsr. Sheesh.
_SPsdiv32: call …
|
|
|
@14091
|
11 years |
gb |
Skip over 0 word pairs when walking cstack. (Leftovers from bignum …
|
|
|
@14090
|
11 years |
gb |
Handle DIVISION-BY-ZERO via #'%ERR-DISP.
|
|
|
@14089
|
11 years |
gb |
%MULTIPLY-AND-ADD-HARDER-LOOP: need to preserve X, which is kind of …
|
|
|
@14088
|
11 years |
gb |
In %FIXNUM-TRUNCATE, set fn register so we can ref constant.
|
|
|
@14087
|
11 years |
gb |
Typo in REQUIRE-NUMBER.
|
|
|
@14084
|
11 years |
gb |
Typo in bitvector case of .SPmisc-set.
|
|
|
@14083
|
11 years |
gb |
In errors involving 3-bit tag, order of CONS and NULL changed
a few …
|
|
|
@14082
|
11 years |
gb |
Constant too large in %MAKE-SHORT-FLOAT-FROM-FIXNUMS.
|
|
|
@14081
|
11 years |
gb |
Fix the SINGLE-FLOAT-VECTOR case in ... %INIT-MISC.
|
|
|
@14080
|
11 years |
gb |
Register targeting in ARM2-AREF2. (May need to do more of this …
|
|
|
@14079
|
11 years |
gb |
REQUIRE-REAL, REQUIRE-NUMBER: use arm::real-tags-mask, …
|
|
|
@14078
|
11 years |
gb |
arm::real-tags-mask, arm::numeric-tags-mask.
|
|
|
@14077
|
11 years |
gb |
svn property changes only, AFAIK.
|
|
|
@14076
|
11 years |
gb |
Don't disable EGC or enable integrity checks in the Makefile.
(Have …
|
|
|
@14075
|
11 years |
gb |
Straighten out keyword_flags definitions.
Don't consider …
|
|
|
@14074
|
11 years |
gb |
handle_uuo(): pass more kinds of UUOs to handle_error().
Make sure …
|
|
|
@14073
|
11 years |
gb |
APPLY+: just restore lr from saved lisp frame.
|
|
|
@14072
|
11 years |
gb |
Bug du Jour in %INIT-MISC.
|
|
|
@14071
|
11 years |
gb |
%SLOT-REF: keep dest, slot-vector, index args disjoint.
|
|
|
@14070
|
11 years |
gb |
SET-OR-CLEAR-BIT had it backwards (should clear bit if E flag set). …
|
|
|
@14069
|
11 years |
gb |
Encoding of fnegs/fnegd instructions was wrong.
|
|
|
@14067
|
11 years |
gb |
Fixes to _SPdebind (lots of 'em) and _SPstack_misc_alloc_init.
(Some …
|
|
|
@14066
|
11 years |
gb |
Lots of changes to support stack-overflow detection on ARM Linux. …
|
|
|
@14065
|
11 years |
gb |
debind subprim bitmask constants.
|
|
|
@14064
|
11 years |
gb |
add call_handler_on_main_stack, for sigaltstack nonsense.
|
|
|
@14063
|
11 years |
gb |
%ERR-DISP callback: report stack overflow.
|
|
|
@14062
|
11 years |
gb |
Fix runtime versions of REALP, NUMBERP, too.
|
|
|
@14061
|
11 years |
gb |
BIGNUM-REM: missed refs to *TRUNCATE-X*.
|
|
|
@14060
|
11 years |
gb |
Typechecking, other fixes in %INIT-MISC.
|
|
|
@14059
|
11 years |
gb |
arm2.lisp: in ARM2-SPREAD-LAMBDA-LIST, get the bits in the doadlword …
|
|
|
@14043
|
11 years |
gb |
Today's image.
|
|
|
@14042
|
11 years |
gb |
Split ARM instructions into two 16-bit halves, to try to reduce …
|
|
|
@14041
|
11 years |
gb |
resurrect .SPjmpsym.
|
|
|
@14040
|
11 years |
gb |
movc16 macro expands to movw.
|
|
|
@14039
|
11 years |
gb |
Need to clear FPSCR status before calling into math lib on ARM.
|
|
|
@14038
|
11 years |
gb |
Compare both halves of DOUBLE-FLOATs in EQL.
|
|
|
@14037
|
11 years |
gb |
FFI FP exception stuff.
|
|
|
@14036
|
11 years |
gb |
Better ISTRUCT-TYPEP; other small changes.
|
|
|
@14035
|
11 years |
gb |
use movw, movt.
|
|
|
@14034
|
11 years |
gb |
movw, movt, & support for them.
|
|
|
@14033
|
11 years |
gb |
movw, movt, & support for them.
|
|
|
@14032
|
11 years |
gb |
Add (resurrect) SPjmpsym.
|
|
|
@14031
|
11 years |
gb |
In %APPLY-WITH-METHOD-CONTEXT, just load LR from the saved frame
and …
|
|
|
@14027
|
11 years |
gb |
Wrong comment. Clean up handling of code-vector locatives (PC/LR, …
|
|
|
@14026
|
11 years |
gb |
About 7 years late: use the current (GC) thread's binding of *PACKAGE* …
|
|
|
@14025
|
11 years |
gb |
About 7 years late: use the current (GC) thread's binding of *PACKAGE* …
|
|
|
@14024
|
11 years |
gb |
Define a new UUO, reserved for use as a forward marker (not a valid …
|
|
|
@14023
|
11 years |
gb |
Check the result of the strex correctly in .SPset_hash_key_conditional.
|
|
|
@14022
|
11 years |
gb |
D'oh!
|
|
|
@14019
|
11 years |
gb |
Latest, greatest.
|
|
|
@14018
|
11 years |
gb |
t_offset's been wrong for a few weeks.
|
|
|
@14017
|
11 years |
gb |
Handle new array UUOs.
Conditionalize for ARM, provide some missing …
|
|
|
@14016
|
11 years |
gb |
In STRING, don't do vector initialization to set a single element in …
|
|
|
@14015
|
11 years |
gb |
More UUOs (for N-dimensional AREF.)
vinsns for N-d aref; backend …
|
|
|
@14014
|
11 years |
gb |
More UUOs (for N-dimensional AREF.)
vinsns for N-d aref; backend …
|
|
|
@14013
|
11 years |
gb |
BIGNUM-SHIFT-LEFT-LOOP: keep fixing bugs …
|
|
|
@14011
|
11 years |
gb |
Current.
|
|
|
@14009
|
11 years |
gb |
BIGNUM-SHIFT-LEFT-LOOP: last digit set by ashift right.
|
|
|
@14008
|
11 years |
gb |
.SPadd_values and .SPrecover_values weren't finding forward/backward …
|
|
|
@14007
|
11 years |
gb |
ARM stack range changes.
|
|
|
@14006
|
11 years |
gb |
%FIXNUM-REF-NATURAL doesn't need a 0-valued offset arg.
|
|
|
@14005
|
11 years |
gb |
Funcall's argument goes straight to arm::nfn in ARM2-MVCALL.
|
|
|
@14004
|
11 years |
gb |
Fix braino in POP-ARGUMENT-REGISTERS.
|
|
|
@14003
|
11 years |
gb |
Show source info (if any) in disassembly.
|
|
|
@14002
|
11 years |
gb |
Forgot to check this in …
|
|
|
@14001
|
11 years |
gb |
Current image.
|
|
|
@14000
|
11 years |
gb |
Better .SPcheck-fpu-exception.
New UUO for FP exceptions.
|
|
|
@13999
|
11 years |
gb |
Handle FPU exceptions in non FFI case.
|
|
|
@13998
|
11 years |
gb |
Fix in %MAKE-FLOAT-FROM-FIXNUMS.
Get control bits right in …
|
|
|
@13997
|
11 years |
gb |
"safe" FP operations need to clear fpscr status bits first.
|
|
|
@13996
|
11 years |
gb |
UUO-ERROR-FPU-EXCEPTION.
Parse FPUX operands (fpscr, etc.) in vinsns.
|
|
|
@13995
|
11 years |
gb |
EXTRACT-ARM-@RN-OPERAND.
|
|
|
@13993
|
11 years |
gb |
ARM Linux support.
|
|
|
@13992
|
11 years |
gb |
Keep trying to guess how to identify a write fault on ARM Linux.
|
|
|
@13991
|
11 years |
gb |
Some fixes in check_refmap_consistency().
|
|
|
@13989
|
11 years |
gb |
Let callback functions decide whether/how to advance PC on return.
|
|
|
@13988
|
11 years |
gb |
Handle return from callbacks.
|
|
|
@13987
|
11 years |
gb |
Hopefully more useful.
|
|
|
@13986
|
11 years |
gb |
Conditionalize for ARM (and maybe PPC, finally.)
|
|
|
@13985
|
11 years |
gb |
Small changes/additions.
|
|
|
@13984
|
11 years |
gb |
Make %ERR-DISP callback return number of bytes to advance PC on return …
|
|
|