* lib/jit_arm.c: Correct use of wrong argument offset
variable in armv7l or float/double argument for varargs
function in armv7hl.
Correct jit_getarg* logic in software float mode to
match expected behavior in other backends, that is, if
a function is not called, it is safe to use a few lightning
calls before a next jit_getarg* call, as done in the test
case check/stack.tst. The proper solution should be to
extend the parser in lib/lightning.c to check if there is
some float operation that will call some (libgcc?) function,
but software float arm should be a very uncommon backend for
lightning, so, just load the already in place arguments
saved to stack, assuming the register argument was clobbered
(what should not be the case most times...).
* check/clobber.ok, check/clobber.tst: New test case doing
extensive validation tests to ensure registers not used in
a operation are not clobbered.
* check/Makefile.am: Update for new test case.
* lib/lightning.c: Partially rewrite/revert code to compute
initial register live state at the start of a basic block.
The original logic was corrupted when adding optimizations
to do as few computations as possible in jit_update. The
reglive field must be always a known set of live registers
at the start of a basic block. The value that was incorrect
was the regmask field, that must be the set of registers
that are in unknown state, because they are not known live,
neither set (or possibly not set) in the basic block, and
*must* store the state at the start of the basic block.
* include/lightning/jit_ppc.h: Correct mismatch of JIT_F{1,5}
with enum codes, that were correct, and returned by jit_f().
* lib/jit_ppc-cpu.c, lib/jit_ppc-fpu.c, lib/jit_ppc.c: Properly
implement and better describe values when generating stack
frames.
* check/stack.ok, check/stack.tst: New files to test data
integrity on a deep chain of stack frames.
* lib/jit_arm.c, lib/jit_arm-cpu.c, lib/jit_mips.c,
lib/jit_mips-cpu.c, lib/jit_ppc.c, lib/jit_ppc-cpu.c,
lib/jit_x86.c, lib/jit_x86-cpu.c: Calculate _jit->function->stack
in the emit stage, otherwise it will calculate it wrong if
need to jit_allocai space to spill registers.
* lib/lightning.c: Correct wrong offset when updating the
"current" jit function pointer in the code that may need to
allocate stack space to spill registers.
* check/lightning.c: Correct off by one data space check.
* check/Makefile.am: Update for new test case.
* check/fop_abs.ok, check/fop_abs.tst, check/fop_sqrt.ok,
check/fop_sqrt.tst: New files implementing simple test cases
for the extra float operations.
* check/Makefile.am: Update for new test cases.
* check/alu.inc: Add an extra macro to check for unordered
equality on tests where it is expected to use NaN as an
argument.
* check/lightning.c: Minor change for proper/common argument
syntax handling ommiting arguments to options.
* check/Makefile.am: Automatically generate pattern list
of tests with alternate jit generation options. This should
prevent typos and needing to change multiple places after
a change.
* check/lightning.c: Remove the ".cpu name value" syntax,
as it was not able to do proper changes before the jit
internal data structure was initialized. Now it supports
several getopt options to force using different jit
generation options, effectively replacing the previous
syntax.
* check/run-test: Add simple extra logic to handle differently
named test scripts, used to test things like x87 coprocessor
in ix86, and arm instruction set or software float in armv7l.
* configure.ac: Add some AC_RUN_IFELSE calls to figure at
compile time if can test different code generation options,
and update Makefile generation accordingly.
* check/Makefile.am, lib/jit_arm.c, lib/jit_x86.c: Update to
properly work with the test tool updating the jit_cpu global
information.
* check/check.arm.sh, check/check.swf.sh, check/check.x87.sh:
New wrapper files passing -mthumb=0, mvfp=0 and -mx87=1 to
the test tool, if applicable, so that it can validate alternate
code generation options on test hosts that support them.
* include/lightning.h, include/lightning/jit_private.h,
lib/jit_arm.c, lib/jit_mips.c, lib/jit_ppc.c, lib/jit_x86.c,
lib/lightning.c: Make jit_ellipsis implementation not
backend specific. It is not intended to handle va_list
like objects at runtime, as jit_arg* and jit_getarg*
return constant values resolved at parse time, so, effectively
it is not possible to create printf like jit functions, as
there is no va_start, va_arg, va_end, etc, abstraction. This
limitation should be kept for the sake of making new ports
easier.
* include/lightning.h, lib/jit_arm.c, lib/jit_mips.c,
lib/jit_ppc.c, lib/jit_x86.c, lib/lightning.c: Change jit_prepare
to no longer receive an argument. If receiving an argument, it
should be an ABI specifier, not a boolean if varargs or not,
and add the new jit_ellipsis call, to specify where the
ellipsis is in the C prototype of the function being called.
Note that currently it is not supported to define varargs
functions and it will be ignored if calling jit_ellipsis not
in a prepare/finish* block, but this should be addressed.
* check/allocai.tst, check/alu_add.tst, check/alu_and.tst,
check/alu_com.tst, check/alu_div.tst, check/alu_lsh.tst,
check/alu_mul.tst, check/alu_neg.tst, check/alu_or.tst,
check/alu_rem.tst, check/alu_rsh.tst, check/alu_sub.tst,
check/alu_xor.tst, check/alux_add.tst, check/alux_sub.tst,
check/bp.tst, check/branch.tst, check/cvt.tst, check/divi.tst,
check/fib.tst, check/ldsti.tst, check/ldstr-c.tst,
check/ldstr.tst, check/ldstxi-c.tst, check/ldstxi.tst,
check/ldstxr-c.tst, check/ldstxr.tst, check/rpn.tst,
check/lightning.c: Update for the change to jit_prepare and
addition of jit_ellipsis.
* lib/jit_ppc-cpu.c: Make movr a function that checks arguments
so that other code can safely assume it is a noop if src and dst
are the same register.
Implement rem{r,i}{,_u} as a div{,u}/mul/sub.
Correct ANDIS, ORIS and XORIS calls to cast the argument to
unsigned before the shift to avoid an assertion if the argument
had the topmost bit set.
Implement lshi, rshi and rshi_u as functions to test for a
zero argument, that would otherwise trigger an assertion when
computing the shift value.
Do a simple implementation of bm{s,c}{r,i} with a temporary,
"andr" of arguments and jump based on comparison with zero.
Correct typo in ldxi_c.
* lib/jit_ppc-fpu.c: Correct wrong arguments to FDIV* and STF*.
* lib/jit_ppc.c: Correct wrong check for 6 instead of 8 integer
arguments in registers. If calling a varargs function and
passing a float or double argument, also either store the
value in the stack or in integer registers, as varargs functions
do not fetch it from float registers.
Add "case" for new functions and incorrectly missing ones.
Call libgcc's __clear_cache, that should know what to do
if the hardware needs flushing cache before execution.
* lib/lightning.c: Do a simple/trivial logic in jit_regset_scan1,
that should make it easier for the compiler to optimize it, and
that also corrects the previously wrong code for big endian, and
that was causing problems in ppc due to not saving all callee save
registers as it was not "finding" them in the regset due to the
little endian assumption bug.
* configure.ac: Only default to using the builtin disassembler
if on GNU/Linux. This should be temporary, due to requiring
/proc/self/exe.
Correctly check $target_cpu for powerpc.
* include/lightning/jit_ppc.h: Correctly implement jit_v_num.
* include/lightning/jit_private.h: Declare proper prototype
for jit_init_debug and jit_finish_debug.
* lib/jit_ppc-cpu.c: Remove code to save/restore callee save
float registers, as it is not required since those float
registers are not usable currently.
Change prolog and epilog generation to, at least comparing
code, match what gcc generates in "gcc -O0", but it is still
failing in Darwin PPC, apparently due to the __clear_cache
call not being enough, as frequently it will also fail to
execute, and the code buffer is all zeroes.
* lib/lightning.c: Do not fail in jit_regset_scan1 calls due
to passing 64 as argument on computers with 64 registers.
* lib/jit_mips-cpu.c: Correct all current test cases.
Call the "xori" not the "XORI" macro for jit_xori implementation,
as the XORI macro handles only 16 bit unsigned values.
Call the "movr" macro, not the "movi" macro in the special
case of adding or subtracting zero.
Use the proper temporary register in the jit_andr implementation.
* check/alu.inc, check/alu_add.ok, check/alu_add.tst,
check/alu_and.ok, check/alu_and.tst, check/alu_com.ok,
check/alu_com.tst, check/alu_div.ok, check/alu_div.tst,
check/alu_lsh.ok, check/alu_lsh.tst, check/alu_mul.ok,
check/alu_mul.tst, check/alu_neg.ok, check/alu_neg.tst,
check/alu_or.ok, check/alu_or.tst, check/alu_rem.ok,
check/alu_rem.tst, check/alu_rsh.ok, check/alu_rsh.tst,
check/alu_sub.ok, check/alu_sub.tst, check/alu_xor.ok,
check/alu_xor.tst, check/alux_add.ok, check/alux_add.tst,
check/alux_sub.ok, check/alux_sub.tst, check/branch.ok,
check/branch.tst: New test cases for arithmetic and branch
tests.
* check/Makefile.am: Update for new test cases.
* include/lightning/jit_private.h: Make the jit_reg_free_p
macro shared by all backends. Previously was added for the
arm backend, but is useful in the x86_64 backend when checking
state of "special purpose register".
Also add the new jit_class_named register class, that must be
or'ed with the register value if calling jit_get_reg expecting
an specific value, because the specific register value may be
zero, that previously was treated as no register requested.
* lib/jit_arm-cpu.c: Correct argument order for T2_MVN.
* lib/jit_arm-swf.c: Call the proper function for double
divide. The "software float" implementation just calls
libgcc functions.
* lib/jit_arm.c: Return float/double values in the float
register if using the hard float ABI.
* lib/jit_x86-cpu.c: Change the can_sign_extend_int_p macro
to not include -0x80000000L, because there is code that
"abuses" it and thinks it can negate the immediate value
after calling that macro.
Correct implementation of jit_subi that had a wrong code
patch logic doing subtraction with reversed arguments.
Correct REX prefix calculation in the jit_muli implementation.
Correct logic to get/unget %*ax and %*dx registers in divremr
and divremi.
Correct divremi that was using the symbolic, unique %*ax
value in on place (not using the _REGNO name suffix).
Correct cut&paste error causing it to use "xor" instead of
"or" in one code path of the jit_ori implementation.
Correct several flaws when clobbering registers and/or when
one of the arguments was %*cx in the rotshr wrapper function
implementing most shift operations.
* lib/lightning.c: No longer expect that the backend be smart
enough to know what to do when asking for a named register
if that register is already an argument or is live. It fails
if it is an argument, or if register is live, fails if cannot
spill.
No longer incorrectly assume that eqr_{f,d} and ltgr_{f,d} are
safe to inverse value tests in jump thread optimization.
* check/Makefile.am, check/cvt.ok, check/cvt.tst: Add new
"cvt" test case to test conversion from/to int/float types.
* check/lightning.c: Only define truncr_{f,d}_l in 64 bit mode.
* include/lightning.h: Correct typo that caused it to define
jit_truncr_{f,d}_l in 32 bit mode.
* lib/jit_arm-cpu.c: Avoid assertion failure in the signed/unsigned
extend opcodes generation as it shares an interface for 3 argument
opcode generation.
* lib/jit_x86-cpu.c: Correct wrong argument passed to
jit_unget_reg in the andi implementation and wrong byte
unsigned extend code generation.
* lib/jit_x86-sse.c: Correct conversion from "word" to float or
double as is dependent on wordsize.
* check/ldstr-c.ok, check/ldstr-c.tst, check/ldstxi-c.ok,
check/ldstxi-c.tst, check/ldstxr-c.ok, check/ldstxr-c.tst:
New test case files testing load clobbering the base and/or
index register;
* check/ldst.inc: New file with common definition for all the
ldst* test cases.
check/Makefile.am, check/ldsti.tst, check/ldstr.tst,
check/ldstxi.tst, check/ldstxr.tst: Update for new common
definitions file and new register clobber ldst tests.
* check/ldstr.ok, check/ldstr.tst, check/ldsti.ok,
check/ldsti.tst, check/ldstxr.ok, check/ldstxr.tst,
check/ldstxi.ok, check/ldstxi.tst:
New test case files exercising a very large amount of
register combinations to verify load/store implementation.
* check/Makefile.am: Update for new test cases.
* lib/jit_x86-cpu.c: Correct wrong argument order when
computing REX prefix for {ld,st}r_T codes;
* lib/jit_mips-fpu.c, lib/jit_mips.c: Implement missing mips
jit_sqrtr_{f,d} codes.
* check/all.tst, include/lightning.h, lib/jit_print.c: Change
declaration order and call order in all.tst of {add,sub}c and
{add,sub}x. *c must be called before to set the carry and *x
second to use the carry and keep it set. The wrong call order
was causing all.tst to fail in mips, where a register is
allocated to keep a global carry state.
* include/lightning/jit_mips.h, lib/jit_mips-cpu.c,
lib/jit_mips-fpu.c, lib/jit_mips.c: Correct float/double
argument handling and make the mips backend pass the initial
test cases.
* include/lightning.h, ib/jit_print.c, lib/lightning.c:
Add extra enum values for argument handling functions that
could not be abstracted to the current codes, that is, when
float values need to move from/to gpr from/to fpr. It would
be more tempting to add such primitives, but they would have
wordsize limitations, and it is not expected to add codes
with one gpr argument for 64 bit and two for 32 bit.
* lib/jit_ppc.c: Check _jit->function before calling jit_epilog()
to avoid a runtime exception.
* include/lightning/jit_mips.h, lib/jit_mips.c: Update to
make the mips backend compile in a qemu image.
* lib/jit_ppc.c: Minor adaptations to help in having the
ppc backend compilable.
* configure.ac, include/lightning/jit_private.h, lib/jit_arm-cpu.c,
lib/jit_arm-swf.c, lib/jit_arm.c, check/Makefile.am: Correct
implementation of the arm backend port to build and pass the
current test cases. Tested on armv7 with softfp abi.
* lib/jit_disasm.c: Rename and change prototype of static
disassemble function as in the arm backend it is required
to access state information stored in the jit_state_t object.
* check/3to2.tst, check/add.tst: Correct test case code assuming
JIT_RO and JIT_RET are the same, and even if they are the same,
the logic was incorrect because it must always call jit_retval*
to fetch a function call return before any other instruction.
The arm backend hash a special condition if jit_retval is not
called, because "r0" is not JIT_R0, but is JIT_RET and *also*
the first argument for a called function, so JIT_RET must be
only used as an argument to jit_retval.
* TODO: New file listing important tasks to be resolved.
* check/all.tst, check/lightning.c: Only declare or use 64 bit
interfaces on 64 bit builds.
* check/fib.tst: Use simpler logic to not need preprocessor
conditionals for 32 or 64 bit.
* include/lightning.h: Only declare 64 bit macros on a 64 bit
build. Code using lightning must know about wordsize and the
jit generation limitations, also, this way it generates a
compile time failure, not a runtime assertion.
* include/lightning/jit_x86.h: Correct typo in macro name.
* lib/jit_arm.c, lib/jit_arm-cpu.c, lib/jit_mips.c,
lib/jit_mips-cpu.c, lib/jit_ppc.c, lib/jit_ppc-cpu.c,
lib/jit_x86.c, lib/jit_x86-cpu.c: Correct wrong code to get
current jit function pointer.
* lib/lightning.c: Move call to the simplify() optimization
to after register liveness is known. Previous code did work
by accident but now with proper test cases the problem was
noticed.
* lib/jit_disasm.c: Always cast bfd_vma to long long when
passing it as printf argument.
* lib/jit_ppc.c: Assign copyright ownership to FSF.
* lib/jit_x86-cpu.c: Correct integer multiplication that was
generating code with reversed register arguments.
* check/rpn.ok, check/rpn.tst: New test case file.
2012-12-02 Paulo Andrade <pcpa@gnu.org>
* lib/jit_x86-cpu.c, lib/jit_x86-sse.c, lib/jit_x86-x87.c:
Actually change copyright owner to FSF as avertised.
* lib/jit_arm-cpu.c, lib/jit_arm-swf.c,
lib/jit_arm-vfp.c, lib/jit_arm.c,
lib/jit_mips-cpu.c, lib/jit_mips-fpu.c, lib/jit_mips.c,
lib/jit_ppc-cpu.c, lib/jit_ppc-fpu.c, lib/jit_ppc.c: New
files implementing initial code different jit backends.
* include/lightning/jit_private.h: Add extra field to the
private jit_patch_t type, required by the arm port.
* lib/Makefile.am: Update for the new backend implementation
files.
* .gitignore: Update pattern of ignored files.
* check/Makefile.am: Add rule to build liblightning.la dependency
in case of running "make check" before building the library.
* lightning/Makefile.am, lightning/asm-common.h,
lightning/core-common.h, lightning/fp-common.h,
lightning/funcs-common.h, lightning/i386/Makefile.frag,
lightning/i386/asm-32.h, lightning/i386/asm-64.h,
lightning/i386/asm.h, lightning/i386/core-32.h,
lightning/i386/core-64.h, lightning/i386/core.h,
lightning/i386/fp-32.h, lightning/i386/fp-64.h,
lightning/i386/fp.h, lightning/i386/funcs.h,
lightning/ppc/asm.h, lightning/ppc/core.h,
lightning/ppc/fp.h, lightning/ppc/funcs.h,
lightning/sparc/asm.h, lightning/sparc/core.h,
lightning/sparc/fp.h, lightning/sparc/funcs.h:
Removed. The core logic is used in the new code, and new mips
and arm ports will be added. At first, sparc will not be
supported as it has not yet been ported to the new engine.
2012-12-02 Paulo Andrade <pcpa@gnu.org>
* tests/Makefile.am, tests/3to2.c, tests/3to2.ok, tests/add.c,
tests/add.ok, tests/allocai.c, tests/allocai.ok, tests/bp.c,
tests/bp.ok, tests/divi.c, tests/divi.ok, tests/fib.c, tests/fib.ok,
tests/fibdelay.c, tests/fibdelay.ok, tests/fibit.c, tests/fibit.ok,
tests/funcfp.c, tests/funcfp.ok, tests/incr.c, tests/incr.ok,
tests/ldst.c, tests/ldst.ok, tests/ldxi.c, tests/ldxi.ok,
tests/modi.c, tests/modi.ok, tests/movi.c, tests/movi.ok,
tests/printf.c, tests/printf.ok, tests/printf2.c, tests/printf2.ok,
tests/ret.c, tests/ret.ok, tests/rpn.c, tests/rpn.ok, tests/rpnfp.c,
tests/rpnfp.ok, tests/sete.c, tests/sete.ok, tests/testfp.c,
tests/testfp.ok, tests-run-test: Removed previous test suite, in
favor of a newer one in the check subdirectory.
* check/3to2.ok, check/3to2.tst, check/add.ok, check/add.tst,
check/allocai.ok, check/allocai.tst, check/bp.ok, check/bp.tst,
check/divi.ok, check/divi.tst, check/fib.ok, check/fib.tst:
New sample input for the new test program, loosely matching
several of the previous test cases.
* check/Makefile.am: New test suite makefile.
* check/check.sh, check/run-test: New wrapper files for the
new test suite.
* check/lightning.c: New file. The main driver of the new test
suite, that compiles to a parser of a very simple assembly like
language, generates jit and executes it.
* check/all.tst: New file. A generic debug and sample test file
with a directive to prevent it from being executed, and useful to
read disassembly of all possible instructions, using a fixed set
of registers.
* include/Makefile.am, include/lightning.h,
include/lightning/Makefile.am, include/lightning/jit_arm.h,
include/lightning/jit_mips.h, include/lightning/jit_ppc.h,
include/lightning/jit_private.h, include/lightning/jit_x86.h,
lib/Makefile.am, lib/jit_disasm.c, lib/jit_print.c,
lib/jit_x86-cpu.c, lib/jit_x86-sse.c, lib/jit_x86-x87.c,
lib/jit_x86.c, lib/lightning.c: New files. These files are
written from scratch, only by <pcpa@gnu.org>, and have now
copyright assignment to the FSF. This is the core of the new
lightning rework. Previously it was integrated in code with
a garbage collector and several custom types like vectors and
hash tables, so this first code merge with lightning converts
that code into a library extracting only the jit bits, and at
first only for x86_64 GNU/Linux.
* lightning.h, m4/lightning.m4: Removed. These are no longer
required in the new lightning code.
.gitignore, Makefile.am, configure.ac: Update for the new
lightning code.
2012-12-02 Paulo Andrade <pcpa@gnu.org>
* .cvsignore: Removed for extra cleanup.
* build-aux: Rename directory to m4.
* m4: Renamed to "default" name and for consistency with merge
with code rework to be imported in lightning.
* .gitignore, configure.ac, Makefile.am, doc/Makefile.am:
Update for build-aux to m4 rename.
2012-12-01 Paulo Andrade <pcpa@gnu.org>
* opcode/Makefile.am, opcode/Makefile.in, opcode/ansidecl.h,
opcode/bfd.h, opcode/dis-asm.h, opcode/dis-buf.c, opcode/disass.c,
opcode/i386-dis.c, opcode/i386.h, opcode/ppc-dis.c, opcode/ppc-opc.c,
opcode/ppc.h, opcode/sparc-dis.c, opcode/sparc-opc.c, opcode/sparc.h,
opcode/sysdep.h: Removed. Do not bundle GNU binutils files.
* aclocal.m4, configure, Makefile.in, config.h.in, doc/Makefile.in,
lightning/Makefile.in, tests/Makefile.in: Removed. Do not maintain
autogenerated files that also generate too much diff noise when
regenerated in git.
* build-aux/help2man, build-aux/texinfo.tex, build-aux/texi2dvi:
Removed. Buildenvironment must have an up to date version from
upstream installed.
* build-aux/config.guess, build-aux/config.sub, build-aux/depcomp,
build-aux/install-sh build-aux/mdate-sh build-aux/missing: Removed.
Do not maintain a copy of automake files in git. Release tarballs
must use an up to date version.
* lightningize.in, doc/lightningize.1: Removed. Do not encourage
bundling lightning in other packages. It should use a system package
or a proper thirdy part subdirectory.
* INSTALL: Removed. Autoreconf removes it and creates a symlink
when regenerating files, so, avoid conflicts in git and let
automake create the symlink.
* .gitignore: Add INSTALL and autogenerated files.
* configure.ac, Makefile.am: Update for removal of opcode subdir,
auto generated files and lightningize.
* tests/Makefile.am, tests/3to2.c, tests/add.c, tests/bp.c,
tests/fib.c, tests/fibdelay.c, tests/fibit.c, tests/funcfp.c,
tests/incr.c, tests/printf.c, tests/rpn.c, tests/rpnfp.c,
tests/sete.c, tests/testfp.c: Update for removal of opcode subdir.
* doc/Makefile.am: Update for removal of lightningize.
* configure.ac, lightning/ppc/funcs.h, lightning/sparc/funcs.h,
lightning/i386/fp.h, lightning/i386/core.h, lightning/i386/asm.h,
tests/3to2.c, tests/add.c, tests/bp.c, tests/fib.c, tests/fibdelay.c,
tests/fibit.c, tests/funcfp.c, tests/incr.c, tests/printf.c,
tests/rpn.c, tests/rpnfp.c, tests/sete.c, tests/testfp.c:
Remove LIGHTNING_CROSS, it is half supported and incomplete.
* tests/3to2.c, tests/funcfp.c, tests/rpnfp.c: Remove preprocessor
check on JIT_FPR. If no hardware registers are available, the backend
must provide an alternative for software float.
* lightning/ppc/core.h, lightning/sparc/core.h, tests/Makefile.am:
Remove JIT_NEED_PUSH_POP. It is absolutely not trivial to implement
properly on some backends due to stack alignment constraints, and
whenever it is required, using jit_allocai and using a properly
aligned stack vector, or a heap buffer, is better.
* tests/push-pop.c, tests/push-pop.ok: Removed due to
JIT_NEED_PUSH_POP no longer available.
There is no guarantee of order of jit_function calls, so, it must
reset it for every function. This should correct possible issues in
OS/X, and will be required when i386 code is updated to use MMX/SSE*
registers, possibly as some kind of extended interface, and then,
would require 16 bytes aligned stack. This is already an issue if
calling C code that uses SSE, and currently requires "manually"
padding the stack by using jit_allocai.
The problems would also happen with 32 bit unsigned operands that would
happen to have the topmost bit set, so, now it only uses 32 bit immediates
when bit 31 is not set.
This patch correct all the issues found by the test cases alu_add.tst and
alu_sub.tst in http://code.google.com/p/exl/source/browse/trunk/check/lightning
The checks were moved from i386/asm.h to i386/asm-{32,64}.h, as well
as some macros from core-{32,64}.h. Now it checks if the value is in the
range of a valid register, and in the proper register class, what should
prevent the common mistake of calling a jit*r_x macro passing an immediate
as argument.
Now it pass lightning's make check in i386/x86_64, as well as all test
cases in http://code.google.com/p/exl/source/browse/trunk/check/lightning
when compiled with -D_ASM_SAFETY.
jit_bra_l had the logic reversed, and correcting that also corrected
jit_b{lt,le,eq,ge,gt,ne}i_l.
TESTQir and _ALUQir were not properly working with 64 bit immediates,
that require using a temporary register (JIT_REXTMP) as there are no
related opcodes for 64 bit immediates. This corrected jit_bm{s,c}i_l and
jit_bo{add,sub}i_l.
Now, the tests in
http://code.google.com/p/exl/source/browse/trunk/check/lightning/branch.tst
pass.
The x86_64 correction to keep stack aligned in jit_finish* was only
correct for functions being called, but did not correctly handle alignment
of stack arguments.
This changes the usage of some of the jit_local_state fields, and add
two extra ones. Since jit_arg_x now returns either a register offset or
an stack offset, it does not add _XMM0 to nextarg_putfp to distinguish
from stack offsets.
The nextarg_puti field is used to known when run out of integer register
arguments, and should push arguments on the stack.
The new framesize field is used to keep track of jit_arg_x offset on
the stack, when jit_arg_x runs out of registers.
The argssize field is used in a slightly different way now, and is
used to keep track of number of stack slots used to pass arguments, and
nextarg_puti is used in its previous semantics, to also match nextarg_putfp
logic.
It is required to set %al to 0-8 inclusive, telling the number of
xmm registers used to pass arguments.
This also corrects the calculation of _jitl.nextarg_putfp, that
would be incorrect if calling a function with both, float and double
arguments.
Lightning provides at least 6 float registers, so, in this commit,
reduce from 8 to 7 so that code respecting JIT_FPR_NUM should be
safe again't use of JIT_FPTMP.
To not clobber JIT_R2, JIT_REXTMP is defined to the callee save register
%r12, instead of the expected, scratch register %r11.
To have JIT_V_NUM at least at 3, it now also uses %r14, but does not add
%r15, that could be JIT_V(4) to simplify logic of keeping stack 16 bytes
aligned.