If argument registers are in incremental order in the global _rvs
vector, the check if a register argument is reversed.
Adapted from a regression found in sparc 64.
* include/lightning/jit_private.h: Add new register classes to
flag float registers and double only registers, required for sparc64
where only low 32 bit fpr registers can be used for single precision
operations.
Add new 128 bit jit_regset_t type for sparc64 register set.
* include/lightning/jit_sparc.h, lib/jit_sparc-cpu.c, lib/jit_sparc-fpu.c,
lib/jit_sparc-sz.c, lib/jit_sparc.c: Update for 64 bits sparc.
* lib/lightning.c: Update for new jit_regset_t required for sparc64.
* libguile/intrinsics.c (scm_to_uint64_truncate): New intrinsic.
(scm_bootstrap_intrinsics): Init new intrinsics.
* libguile/intrinsics.h: Add scm->u64, scm->u64/truncate, and scm->s64
as intrinsics, with their corresponding types.
* libguile/vm-engine.c (call-u64<-scm, call-s64<-scm): New intrinsic
calling ops.
(scm->u64, scm->s64, scm->u64/truncate): Disable opcodes.
* module/language/cps/reify-primitives.scm (compute-known-primitives):
Add intrinsics as new macroinstructions.
* module/system/vm/assembler.scm: Declare new intrinsic assemblers.
* libguile/numbers.h (struct scm_t_double, struct scm_t_complex): Avoid
adding an extra padding word on systems with 8-byte pointers.
* module/system/base/types.scm (cell->object): Update to compute correct
offset of embedded double.
* module/language/cps/effects-analysis.scm (string-ref): Remove effects
declaration, given that the primitive is exploded now.
* module/language/cps/reify-primitives.scm (compute-known-primitives):
Add string-set!.
* libguile/vm-engine.c (string-set!): Disable opcode.
* module/language/cps/types.scm (string-ref, string-set!): Remove type
checker and inferrers for string-ref and string-set!, as both are
exploded. In the case of string-set! there are still type-check
effects in the intrinsic call but they can't be elided by the checker,
as we don't track when strings are read-only.
* module/language/tree-il/compile-cps.scm (ensure-char): New helper.
(string-set!): New primcall exploded converter.
* module/language/cps/types.scm (annotation->type):
* module/language/cps/effects-analysis.scm (annotation->memory-kind):
Add case for string memory kinds. Remove special type and effect
inferrers for string-length.
* module/language/cps/slot-allocation.scm (compute-var-representations):
Remove string-length.
* module/language/tree-il/compile-cps.scm (ensure-string): New helper.
(string-length): Add custom converter.
* module/language/cps/reify-primitives.scm (compute-known-primitives):
Treat add, sub, and so on as known primitives.
* libguile/vm-engine.c: Remove code for now-unused instructions.
* libguile/intrinsics.h: Rename intrinsic types added in previous
commit.
* libguile/vm-engine.c (call-scm<-scm-scm, call-scm<-scm-uimm): New
instructions.
* libguile/vm.c: Include intrinsics.h.
* module/language/bytecode.scm
* module/language/bytecode.scm (*intrinsic-codes*, *intrinsic-names*):
New internal definitions.
(intrinsic-name->index, intrinsic-index->name): New exported
definitions.
* module/system/vm/assembler.scm (encode-X8_S8_S8_S8-C32<-/shuffle):
(encode-X8_S8_S8_C8-C32<-/shuffle): New shuffling encoders.
(shuffling-encoder-name): Add case for new shuffling encoders.
(define-scm<-scm-scm-intrinsic, define-scm<-scm-uimm-intrinsic): New
helpers. Define encoders for "add", etc.
Fixes <https://bugs.gnu.org/29520>.
Reported by Stefan Israelsson Tampe <stefan.itampe@gmail.com>.
* module/language/tree-il/peval.scm (peval)[lift-applied-lambda]: Before
performing eta-conversion, check that the variable(s) passed to the
inner 'apply' are not referenced from the inner lambda, and that the
number of required arguments would not be reduced by the conversion.
* module/system/base/target.scm (cpu-endianness): Add case for "riscv" variants.
Signed-off-by: Shea Levy <shea@shealevy.com>
Signed-off-by: Mark H Weaver <mhw@netris.org>
* check/lightning.c, include/lightning.h: Add the new jit_va_push
interface. That should be called when passing a va_list to a C
function. This is required because on Alpha a va_list is passed
by value, and lightning does not know about data types, so, cannot
understand it is pushing a va_list as argument.
* lib/jit_names.c, lib/lightning.c: Minor changes for the new
jit_code_va_push.
* check/cva_list.c: Update only test case using jit_va_push, to
pass a va_list to a C function.
doc/body.texi: Better documentation of the varargs interface.
* jit_alpha.c, jit_alpha-cpu.c: Update to properly push a
C va_list and correctly calculate varargs offset.
* lib/jit_aarch64-sz.c, lib/jit_aarch64.c, lib/jit_alpha-sz.c,
lib/jit_arm-sz.c, lib/jit_arm.c, lib/jit_hppa-sz.c, lib/jit_hppa.c,
lib/jit_ia64-sz.c, lib/jit_ia64.c, lib/jit_mips-sz.c, lib/jit_mips.c,
lib/jit_ppc-sz.c, lib/jit_ppc.c, lib/jit_s390-sz.c, lib/jit_s390.c,
lib/jit_sparc-sz.c, lib/jit_sparc.c, lib/jit_x86-sz.c, lib/jit_x86.c:
Update for the new jit_va_push interface.
* lib/jit_alpha-cpu.c: Always set t12 to the address of the
current function, to properly work on all systems. Previously
the shortcut did only work on Tru64. For Linux and glibc the
change is required.
* lib/jit_aarch64.c, lib/jit_alpha.c, lib/jit_arm.c,
lib/jit_mips.c, lib/jit_ppc.c, lib/jit_sparc.c, lib/jit_x86.c:
Correct wrong logic in usage of jit_live in jit_retr. The
problem is that if a temporary is required during epilog,
the return register might be allocated, so, jit_live must always
be used.
This is a followup to <https://bugs.gnu.org/29704>.
* module/srfi/srfi-18.scm (thread-sleep!): When TIMEOUT is a number,
keep it as-is.
* test-suite/tests/srfi-18.test ("thread sleep with number"): Pass 0 as
the timeout.
("thread sleeps fractions of a second"): Pass 0.5 as the timeout.
Fixes <https://bugs.gnu.org/29704>.
Reported by David Beswick <dlbeswick@gmail.com>.
* module/srfi/srfi-18.scm (timeout->absolute-time): New procedure.
(mutex-lock!): Use it in 'thread:lock-mutex' call.
(mutex-unlock!): Use it.
* test-suite/tests/srfi-18.test ("mutex-lock! returns false on timeout")
("mutex-lock! returns true when lock obtained within timeout")
("recursive lock waits")
("mutex unlock is false when condition times out"): Adjust cases where
the 'timeout' parameter is a number so that it's a relative number.
