2018-06-20 19:17:06 +02:00
|
|
|
|
/* Copyright 2001,2009-2015,2018
|
2018-06-20 20:01:49 +02:00
|
|
|
|
Free Software Foundation, Inc.
|
|
|
|
|
|
|
|
|
|
|
|
This file is part of Guile.
|
|
|
|
|
|
|
|
|
|
|
|
Guile is free software: you can redistribute it and/or modify it
|
|
|
|
|
|
under the terms of the GNU Lesser General Public License as published
|
|
|
|
|
|
by the Free Software Foundation, either version 3 of the License, or
|
|
|
|
|
|
(at your option) any later version.
|
|
|
|
|
|
|
|
|
|
|
|
Guile is distributed in the hope that it will be useful, but WITHOUT
|
|
|
|
|
|
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
|
|
|
|
|
|
FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
|
|
|
|
|
|
License for more details.
|
|
|
|
|
|
|
|
|
|
|
|
You should have received a copy of the GNU Lesser General Public
|
|
|
|
|
|
License along with Guile. If not, see
|
|
|
|
|
|
<https://www.gnu.org/licenses/>. */
|
2001-04-22 02:13:48 +00:00
|
|
|
|
|
2008-05-20 11:46:52 +02:00
|
|
|
|
#if HAVE_CONFIG_H
|
2018-06-20 18:31:24 +02:00
|
|
|
|
# include <config.h>
|
2008-05-20 11:46:52 +02:00
|
|
|
|
#endif
|
|
|
|
|
|
|
2009-04-05 20:15:11 +02:00
|
|
|
|
#include <stdlib.h>
|
2001-04-22 02:13:48 +00:00
|
|
|
|
#include <string.h>
|
2018-06-20 18:31:24 +02:00
|
|
|
|
|
2018-06-18 21:38:44 +02:00
|
|
|
|
#include "boolean.h"
|
2018-06-14 22:16:01 +02:00
|
|
|
|
#include "eval.h"
|
|
|
|
|
|
#include "extensions.h"
|
2018-06-20 18:31:24 +02:00
|
|
|
|
#include "gsubr.h"
|
2018-08-06 17:00:45 +02:00
|
|
|
|
#include "instructions.h"
|
2018-06-20 17:19:31 +02:00
|
|
|
|
#include "modules.h"
|
2018-06-19 15:37:23 +02:00
|
|
|
|
#include "numbers.h"
|
2018-08-06 17:00:45 +02:00
|
|
|
|
#include "pairs.h"
|
2018-06-19 12:01:38 +02:00
|
|
|
|
#include "ports.h"
|
2018-06-20 17:19:31 +02:00
|
|
|
|
#include "symbols.h"
|
2018-06-20 11:52:29 +02:00
|
|
|
|
#include "threads.h"
|
2018-06-20 17:19:31 +02:00
|
|
|
|
#include "variable.h"
|
2018-10-07 15:15:02 +02:00
|
|
|
|
#include "version.h"
|
2013-11-21 11:20:19 +01:00
|
|
|
|
#include "vm.h"
|
2001-04-22 02:13:48 +00:00
|
|
|
|
|
2018-06-20 18:31:24 +02:00
|
|
|
|
#include "frames.h"
|
|
|
|
|
|
|
|
|
|
|
|
|
2001-04-22 02:13:48 +00:00
|
|
|
|
SCM
|
2014-04-14 15:14:26 +02:00
|
|
|
|
scm_c_make_frame (enum scm_vm_frame_kind kind, const struct scm_frame *frame)
|
2001-04-22 02:13:48 +00:00
|
|
|
|
{
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
struct scm_frame *p = scm_gc_malloc (sizeof (struct scm_frame),
|
|
|
|
|
|
"vmframe");
|
2014-04-14 15:14:26 +02:00
|
|
|
|
p->stack_holder = frame->stack_holder;
|
|
|
|
|
|
p->fp_offset = frame->fp_offset;
|
|
|
|
|
|
p->sp_offset = frame->sp_offset;
|
|
|
|
|
|
p->ip = frame->ip;
|
|
|
|
|
|
return scm_cell (scm_tc7_frame | (kind << 8), (scm_t_bits)p);
|
2001-04-22 02:13:48 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2010-01-05 19:45:56 +01:00
|
|
|
|
void
|
|
|
|
|
|
scm_i_frame_print (SCM frame, SCM port, scm_print_state *pstate)
|
2009-02-05 18:13:27 +01:00
|
|
|
|
{
|
2016-04-26 23:07:28 +02:00
|
|
|
|
scm_puts ("#<frame ", port);
|
2009-02-05 18:13:27 +01:00
|
|
|
|
scm_uintprint (SCM_UNPACK (frame), 16, port);
|
2015-11-27 12:31:16 +01:00
|
|
|
|
if (scm_module_system_booted_p)
|
|
|
|
|
|
{
|
|
|
|
|
|
SCM name = scm_frame_procedure_name (frame);
|
|
|
|
|
|
|
|
|
|
|
|
if (scm_is_true (name))
|
|
|
|
|
|
{
|
2016-04-26 23:01:14 +02:00
|
|
|
|
scm_putc (' ', port);
|
2015-11-27 12:31:16 +01:00
|
|
|
|
scm_write (name, port);
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
/* Don't write args, they can be ridiculously long. */
|
2016-04-26 23:07:28 +02:00
|
|
|
|
scm_puts (">", port);
|
2009-02-05 18:13:27 +01:00
|
|
|
|
}
|
|
|
|
|
|
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
static union scm_vm_stack_element*
|
|
|
|
|
|
frame_stack_top (enum scm_vm_frame_kind kind, const struct scm_frame *frame)
|
2013-11-21 11:20:19 +01:00
|
|
|
|
{
|
2014-04-14 14:54:14 +02:00
|
|
|
|
switch (kind)
|
2013-11-21 17:21:37 +01:00
|
|
|
|
{
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
case SCM_VM_FRAME_KIND_CONT:
|
|
|
|
|
|
{
|
|
|
|
|
|
struct scm_vm_cont *cont = frame->stack_holder;
|
|
|
|
|
|
return cont->stack_bottom + cont->stack_size;
|
|
|
|
|
|
}
|
2013-11-21 11:20:19 +01:00
|
|
|
|
|
2013-11-21 17:21:37 +01:00
|
|
|
|
case SCM_VM_FRAME_KIND_VM:
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
return ((struct scm_vm *) frame->stack_holder)->stack_top;
|
2013-11-21 17:21:37 +01:00
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
|
abort ();
|
|
|
|
|
|
}
|
2013-11-21 11:20:19 +01:00
|
|
|
|
}
|
2014-04-14 14:54:14 +02:00
|
|
|
|
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
union scm_vm_stack_element*
|
|
|
|
|
|
scm_i_frame_stack_top (SCM frame)
|
|
|
|
|
|
#define FUNC_NAME "frame-stack-top"
|
2014-04-14 14:54:14 +02:00
|
|
|
|
{
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
|
|
|
|
|
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
return frame_stack_top (SCM_VM_FRAME_KIND (frame),
|
|
|
|
|
|
SCM_VM_FRAME_DATA (frame));
|
2014-04-14 14:54:14 +02:00
|
|
|
|
}
|
2013-11-21 11:20:19 +01:00
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2009-03-22 15:46:46 +01:00
|
|
|
|
|
2001-04-22 02:13:48 +00:00
|
|
|
|
/* Scheme interface */
|
|
|
|
|
|
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
SCM_DEFINE (scm_frame_p, "frame?", 1, 0, 0,
|
2001-04-22 02:13:48 +00:00
|
|
|
|
(SCM obj),
|
|
|
|
|
|
"")
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
#define FUNC_NAME s_scm_frame_p
|
2001-04-22 02:13:48 +00:00
|
|
|
|
{
|
2009-11-28 15:11:31 +01:00
|
|
|
|
return scm_from_bool (SCM_VM_FRAME_P (obj));
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2014-04-14 16:31:02 +02:00
|
|
|
|
/* Retrieve the local in slot 0, which may or may not actually be a
|
|
|
|
|
|
procedure, and may or may not actually be the procedure being
|
|
|
|
|
|
applied. If you want the procedure, look it up from the IP. */
|
|
|
|
|
|
SCM
|
|
|
|
|
|
scm_c_frame_closure (enum scm_vm_frame_kind kind, const struct scm_frame *frame)
|
|
|
|
|
|
{
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
union scm_vm_stack_element *fp, *sp;
|
2014-04-16 19:17:38 +02:00
|
|
|
|
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
fp = frame_stack_top (kind, frame) - frame->fp_offset;
|
|
|
|
|
|
sp = frame_stack_top (kind, frame) - frame->sp_offset;
|
2014-04-14 16:31:02 +02:00
|
|
|
|
|
2014-04-16 19:17:38 +02:00
|
|
|
|
if (SCM_FRAME_NUM_LOCALS (fp, sp) > 0)
|
|
|
|
|
|
return SCM_FRAME_LOCAL (fp, 0);
|
|
|
|
|
|
|
|
|
|
|
|
return SCM_BOOL_F;
|
2014-04-14 16:31:02 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
2015-11-27 12:17:36 +01:00
|
|
|
|
static SCM frame_procedure_name_var;
|
|
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
|
init_frame_procedure_name_var (void)
|
|
|
|
|
|
{
|
|
|
|
|
|
frame_procedure_name_var
|
|
|
|
|
|
= scm_c_private_lookup ("system vm frame", "frame-procedure-name");
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
SCM_DEFINE (scm_frame_procedure_name, "frame-procedure-name", 1, 0, 0,
|
|
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"")
|
|
|
|
|
|
#define FUNC_NAME s_scm_frame_procedure_name
|
|
|
|
|
|
{
|
|
|
|
|
|
static scm_i_pthread_once_t once = SCM_I_PTHREAD_ONCE_INIT;
|
|
|
|
|
|
scm_i_pthread_once (&once, init_frame_procedure_name_var);
|
|
|
|
|
|
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
|
|
|
|
|
|
|
|
|
|
|
return scm_call_1 (scm_variable_ref (frame_procedure_name_var), frame);
|
|
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2014-01-23 11:37:36 -05:00
|
|
|
|
static SCM frame_arguments_var;
|
|
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
|
init_frame_arguments_var (void)
|
|
|
|
|
|
{
|
|
|
|
|
|
frame_arguments_var
|
|
|
|
|
|
= scm_c_private_lookup ("system vm frame", "frame-arguments");
|
|
|
|
|
|
}
|
|
|
|
|
|
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
SCM_DEFINE (scm_frame_arguments, "frame-arguments", 1, 0, 0,
|
|
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"")
|
|
|
|
|
|
#define FUNC_NAME s_scm_frame_arguments
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
{
|
2014-01-23 11:37:36 -05:00
|
|
|
|
static scm_i_pthread_once_t once = SCM_I_PTHREAD_ONCE_INIT;
|
|
|
|
|
|
scm_i_pthread_once (&once, init_frame_arguments_var);
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
2014-01-23 11:37:36 -05:00
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
2014-01-23 11:37:36 -05:00
|
|
|
|
return scm_call_1 (scm_variable_ref (frame_arguments_var), frame);
|
2001-04-22 02:13:48 +00:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2014-04-16 14:33:20 +02:00
|
|
|
|
static SCM frame_call_representation_var;
|
|
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
|
init_frame_call_representation_var (void)
|
|
|
|
|
|
{
|
|
|
|
|
|
frame_call_representation_var
|
|
|
|
|
|
= scm_c_private_lookup ("system vm frame", "frame-call-representation");
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
SCM scm_frame_call_representation (SCM frame)
|
|
|
|
|
|
#define FUNC_NAME "frame-call-representation"
|
|
|
|
|
|
{
|
|
|
|
|
|
static scm_i_pthread_once_t once = SCM_I_PTHREAD_ONCE_INIT;
|
|
|
|
|
|
scm_i_pthread_once (&once, init_frame_call_representation_var);
|
|
|
|
|
|
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
|
|
|
|
|
|
|
|
|
|
|
return scm_call_1 (scm_variable_ref (frame_call_representation_var), frame);
|
|
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2011-02-11 12:43:05 +01:00
|
|
|
|
SCM_DEFINE (scm_frame_source, "frame-source", 1, 0, 0,
|
|
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"")
|
|
|
|
|
|
#define FUNC_NAME s_scm_frame_source
|
2001-04-22 02:13:48 +00:00
|
|
|
|
{
|
2011-02-11 12:43:05 +01:00
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
2013-11-07 23:03:45 +01:00
|
|
|
|
return scm_find_source_for_addr (scm_frame_instruction_pointer (frame));
|
2001-04-22 02:13:48 +00:00
|
|
|
|
}
|
2011-02-11 12:43:05 +01:00
|
|
|
|
#undef FUNC_NAME
|
2001-04-22 02:13:48 +00:00
|
|
|
|
|
2016-01-30 17:47:03 +01:00
|
|
|
|
static const char s_scm_frame_num_locals[] = "frame-num-locals";
|
|
|
|
|
|
static SCM
|
|
|
|
|
|
scm_frame_num_locals (SCM frame)
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
#define FUNC_NAME s_scm_frame_num_locals
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
{
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
union scm_vm_stack_element *fp, *sp;
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
|
|
|
|
|
|
Frame pointer points to local 0 instead of local 1
* libguile/frames.h: Change so that fp points at local 0 instead of
local 1, and clean up a bit.
(struct scm_vm_frame): Remove program, and rename stack to locals.
(SCM_FRAME_DATA_ADDRESS): Remove; it was redundant with
SCM_FRAME_LOWER_ADDRESS.
(SCM_FRAME_STACK_ADDRESS): Remove; replace with the new
SCM_FRAME_LOCALS_ADDRESS.
(SCM_FRAME_UPPER_ADDRESS): Remove; unused.
(SCM_FRAME_NUM_LOCALS, SCM_FRAME_PREVIOUS_SP): New defines.
(SCM_FRAME_BYTE_CAST, SCM_FRAME_STACK_CAST): Remove; unused;
(SCM_FRAME_LOCAL): New define, replaces SCM_FRAME_VARIABLE.
(SCM_FRAME_PROGRAM): Add cautionary commentary.
* libguile/frames.c: Adapt static asserts.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_local_set_x):
Adapt. This means that frame-local-ref 0 now returns the procedure.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME)
(FRAME_LOCALS_COUNT, LOCAL_REF, LOCAL_SET, RETURN_VALUE_LIST): Adapt
to change in fp.
(LOCAL_ADDRESS): New helper.
(POP_CONTINUATION_HOOK): Reimplement, taking the previous FP as an
argument.
(ABORT_CONTINUATION_HOOK): Reimplement, taking no arguments.
(RETURN_ONE_VALUE): Reimplement.
(RETURN_VALUE_LIST): Adapt to FP change.
(halt, return-values, subr-call, foreign-call, prompt)
(continuation-call, compose-continuation, call/cc, abort): Adapt to FP
change, mostly via using LOCAL_ADDRESS, etc abstractions instead of
using the raw frame pointer.
* libguile/control.c (reify_partial_continuation): Update for fp
change.
* libguile/vm.c (vm_reinstate_partial_continuation): Adapt to removal of
SCM_FRAME_UPPER_ADDRESS.
* module/system/vm/frame.scm (frame-call-representation): Adapt to
frame-local-ref change.
* module/system/vm/trace.scm (print-return): Remove unused
frame-num-locals call.
2013-11-17 22:07:44 +01:00
|
|
|
|
fp = SCM_VM_FRAME_FP (frame);
|
2013-05-23 14:52:29 +02:00
|
|
|
|
sp = SCM_VM_FRAME_SP (frame);
|
|
|
|
|
|
|
Frame pointer points to local 0 instead of local 1
* libguile/frames.h: Change so that fp points at local 0 instead of
local 1, and clean up a bit.
(struct scm_vm_frame): Remove program, and rename stack to locals.
(SCM_FRAME_DATA_ADDRESS): Remove; it was redundant with
SCM_FRAME_LOWER_ADDRESS.
(SCM_FRAME_STACK_ADDRESS): Remove; replace with the new
SCM_FRAME_LOCALS_ADDRESS.
(SCM_FRAME_UPPER_ADDRESS): Remove; unused.
(SCM_FRAME_NUM_LOCALS, SCM_FRAME_PREVIOUS_SP): New defines.
(SCM_FRAME_BYTE_CAST, SCM_FRAME_STACK_CAST): Remove; unused;
(SCM_FRAME_LOCAL): New define, replaces SCM_FRAME_VARIABLE.
(SCM_FRAME_PROGRAM): Add cautionary commentary.
* libguile/frames.c: Adapt static asserts.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_local_set_x):
Adapt. This means that frame-local-ref 0 now returns the procedure.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME)
(FRAME_LOCALS_COUNT, LOCAL_REF, LOCAL_SET, RETURN_VALUE_LIST): Adapt
to change in fp.
(LOCAL_ADDRESS): New helper.
(POP_CONTINUATION_HOOK): Reimplement, taking the previous FP as an
argument.
(ABORT_CONTINUATION_HOOK): Reimplement, taking no arguments.
(RETURN_ONE_VALUE): Reimplement.
(RETURN_VALUE_LIST): Adapt to FP change.
(halt, return-values, subr-call, foreign-call, prompt)
(continuation-call, compose-continuation, call/cc, abort): Adapt to FP
change, mostly via using LOCAL_ADDRESS, etc abstractions instead of
using the raw frame pointer.
* libguile/control.c (reify_partial_continuation): Update for fp
change.
* libguile/vm.c (vm_reinstate_partial_continuation): Adapt to removal of
SCM_FRAME_UPPER_ADDRESS.
* module/system/vm/frame.scm (frame-call-representation): Adapt to
frame-local-ref change.
* module/system/vm/trace.scm (print-return): Remove unused
frame-num-locals call.
2013-11-17 22:07:44 +01:00
|
|
|
|
return scm_from_ptrdiff_t (SCM_FRAME_NUM_LOCALS (fp, sp));
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2015-10-28 17:03:42 +00:00
|
|
|
|
enum stack_item_representation
|
|
|
|
|
|
{
|
|
|
|
|
|
STACK_ITEM_SCM = 0,
|
2015-11-12 21:44:24 +01:00
|
|
|
|
STACK_ITEM_F64 = 1,
|
Add support for unboxed s64 values
* libguile/frames.c (enum stack_item_representation):
(scm_to_stack_item_representation):
(scm_frame_local_ref, scm_frame_local_set_x): Support for S64
representations.
* libguile/frames.h (union scm_vm_stack_element): Add signed 64-bit
integer field.
* libguile/vm-engine.c (scm->s64, s64->scm, load-s64): New
instructions.
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/cse.scm (compute-equivalent-subexpressions):
* module/language/cps/effects-analysis.scm:
* module/language/cps/slot-allocation.scm (compute-var-representations)
(compute-needs-slot, allocate-slots):
* module/language/cps/utils.scm (compute-constant-values):
* module/language/cps/specialize-primcalls.scm (specialize-primcalls):
Add support for new primcalls.
* module/language/cps/types.scm (&s64): New type.
(&s64-min, &s64-max, &u64-max): New convenience definitions.
(&range-min, &range-max): Use &s64-min and &u64-max names.
(scm->s64, load-s64, s64->scm): Add support for new primcalls.
* module/system/vm/assembler.scm (emit-scm->s64, emit-s64->scm)
(emit-load-s64): New exports.
* module/system/vm/assembler.scm (write-arities): Support for s64
slots.
* module/system/vm/debug.scm (arity-definitions): Support for s64
slots.
2015-11-21 10:32:33 +01:00
|
|
|
|
STACK_ITEM_U64 = 2,
|
|
|
|
|
|
STACK_ITEM_S64 = 3
|
2015-10-28 17:03:42 +00:00
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
static enum stack_item_representation
|
|
|
|
|
|
scm_to_stack_item_representation (SCM x, const char *subr, int pos)
|
|
|
|
|
|
{
|
|
|
|
|
|
if (scm_is_eq (x, scm_from_latin1_symbol ("scm")))
|
|
|
|
|
|
return STACK_ITEM_SCM;
|
|
|
|
|
|
if (scm_is_eq (x, scm_from_latin1_symbol ("f64")))
|
|
|
|
|
|
return STACK_ITEM_F64;
|
2015-11-12 21:44:24 +01:00
|
|
|
|
if (scm_is_eq (x, scm_from_latin1_symbol ("u64")))
|
|
|
|
|
|
return STACK_ITEM_U64;
|
Add support for unboxed s64 values
* libguile/frames.c (enum stack_item_representation):
(scm_to_stack_item_representation):
(scm_frame_local_ref, scm_frame_local_set_x): Support for S64
representations.
* libguile/frames.h (union scm_vm_stack_element): Add signed 64-bit
integer field.
* libguile/vm-engine.c (scm->s64, s64->scm, load-s64): New
instructions.
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/cse.scm (compute-equivalent-subexpressions):
* module/language/cps/effects-analysis.scm:
* module/language/cps/slot-allocation.scm (compute-var-representations)
(compute-needs-slot, allocate-slots):
* module/language/cps/utils.scm (compute-constant-values):
* module/language/cps/specialize-primcalls.scm (specialize-primcalls):
Add support for new primcalls.
* module/language/cps/types.scm (&s64): New type.
(&s64-min, &s64-max, &u64-max): New convenience definitions.
(&range-min, &range-max): Use &s64-min and &u64-max names.
(scm->s64, load-s64, s64->scm): Add support for new primcalls.
* module/system/vm/assembler.scm (emit-scm->s64, emit-s64->scm)
(emit-load-s64): New exports.
* module/system/vm/assembler.scm (write-arities): Support for s64
slots.
* module/system/vm/debug.scm (arity-definitions): Support for s64
slots.
2015-11-21 10:32:33 +01:00
|
|
|
|
if (scm_is_eq (x, scm_from_latin1_symbol ("s64")))
|
|
|
|
|
|
return STACK_ITEM_S64;
|
2015-10-28 17:03:42 +00:00
|
|
|
|
|
|
|
|
|
|
scm_wrong_type_arg (subr, pos, x);
|
|
|
|
|
|
return 0; /* Not reached. */
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2016-01-30 17:47:03 +01:00
|
|
|
|
static const char s_scm_frame_local_ref[] = "frame-local-ref";
|
|
|
|
|
|
static SCM
|
|
|
|
|
|
scm_frame_local_ref (SCM frame, SCM index, SCM representation)
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
#define FUNC_NAME s_scm_frame_local_ref
|
2001-04-22 02:13:48 +00:00
|
|
|
|
{
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
union scm_vm_stack_element *fp, *sp;
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
unsigned int i;
|
2015-10-28 17:03:42 +00:00
|
|
|
|
enum stack_item_representation repr;
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
SCM_VALIDATE_UINT_COPY (2, index, i);
|
2015-10-28 17:03:42 +00:00
|
|
|
|
repr = scm_to_stack_item_representation (representation, FUNC_NAME, SCM_ARG3);
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
Frame pointer points to local 0 instead of local 1
* libguile/frames.h: Change so that fp points at local 0 instead of
local 1, and clean up a bit.
(struct scm_vm_frame): Remove program, and rename stack to locals.
(SCM_FRAME_DATA_ADDRESS): Remove; it was redundant with
SCM_FRAME_LOWER_ADDRESS.
(SCM_FRAME_STACK_ADDRESS): Remove; replace with the new
SCM_FRAME_LOCALS_ADDRESS.
(SCM_FRAME_UPPER_ADDRESS): Remove; unused.
(SCM_FRAME_NUM_LOCALS, SCM_FRAME_PREVIOUS_SP): New defines.
(SCM_FRAME_BYTE_CAST, SCM_FRAME_STACK_CAST): Remove; unused;
(SCM_FRAME_LOCAL): New define, replaces SCM_FRAME_VARIABLE.
(SCM_FRAME_PROGRAM): Add cautionary commentary.
* libguile/frames.c: Adapt static asserts.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_local_set_x):
Adapt. This means that frame-local-ref 0 now returns the procedure.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME)
(FRAME_LOCALS_COUNT, LOCAL_REF, LOCAL_SET, RETURN_VALUE_LIST): Adapt
to change in fp.
(LOCAL_ADDRESS): New helper.
(POP_CONTINUATION_HOOK): Reimplement, taking the previous FP as an
argument.
(ABORT_CONTINUATION_HOOK): Reimplement, taking no arguments.
(RETURN_ONE_VALUE): Reimplement.
(RETURN_VALUE_LIST): Adapt to FP change.
(halt, return-values, subr-call, foreign-call, prompt)
(continuation-call, compose-continuation, call/cc, abort): Adapt to FP
change, mostly via using LOCAL_ADDRESS, etc abstractions instead of
using the raw frame pointer.
* libguile/control.c (reify_partial_continuation): Update for fp
change.
* libguile/vm.c (vm_reinstate_partial_continuation): Adapt to removal of
SCM_FRAME_UPPER_ADDRESS.
* module/system/vm/frame.scm (frame-call-representation): Adapt to
frame-local-ref change.
* module/system/vm/trace.scm (print-return): Remove unused
frame-num-locals call.
2013-11-17 22:07:44 +01:00
|
|
|
|
fp = SCM_VM_FRAME_FP (frame);
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
sp = SCM_VM_FRAME_SP (frame);
|
2013-11-15 17:13:10 +01:00
|
|
|
|
|
Frame pointer points to local 0 instead of local 1
* libguile/frames.h: Change so that fp points at local 0 instead of
local 1, and clean up a bit.
(struct scm_vm_frame): Remove program, and rename stack to locals.
(SCM_FRAME_DATA_ADDRESS): Remove; it was redundant with
SCM_FRAME_LOWER_ADDRESS.
(SCM_FRAME_STACK_ADDRESS): Remove; replace with the new
SCM_FRAME_LOCALS_ADDRESS.
(SCM_FRAME_UPPER_ADDRESS): Remove; unused.
(SCM_FRAME_NUM_LOCALS, SCM_FRAME_PREVIOUS_SP): New defines.
(SCM_FRAME_BYTE_CAST, SCM_FRAME_STACK_CAST): Remove; unused;
(SCM_FRAME_LOCAL): New define, replaces SCM_FRAME_VARIABLE.
(SCM_FRAME_PROGRAM): Add cautionary commentary.
* libguile/frames.c: Adapt static asserts.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_local_set_x):
Adapt. This means that frame-local-ref 0 now returns the procedure.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME)
(FRAME_LOCALS_COUNT, LOCAL_REF, LOCAL_SET, RETURN_VALUE_LIST): Adapt
to change in fp.
(LOCAL_ADDRESS): New helper.
(POP_CONTINUATION_HOOK): Reimplement, taking the previous FP as an
argument.
(ABORT_CONTINUATION_HOOK): Reimplement, taking no arguments.
(RETURN_ONE_VALUE): Reimplement.
(RETURN_VALUE_LIST): Adapt to FP change.
(halt, return-values, subr-call, foreign-call, prompt)
(continuation-call, compose-continuation, call/cc, abort): Adapt to FP
change, mostly via using LOCAL_ADDRESS, etc abstractions instead of
using the raw frame pointer.
* libguile/control.c (reify_partial_continuation): Update for fp
change.
* libguile/vm.c (vm_reinstate_partial_continuation): Adapt to removal of
SCM_FRAME_UPPER_ADDRESS.
* module/system/vm/frame.scm (frame-call-representation): Adapt to
frame-local-ref change.
* module/system/vm/trace.scm (print-return): Remove unused
frame-num-locals call.
2013-11-17 22:07:44 +01:00
|
|
|
|
if (i < SCM_FRAME_NUM_LOCALS (fp, sp))
|
2015-10-28 17:03:42 +00:00
|
|
|
|
{
|
|
|
|
|
|
union scm_vm_stack_element *item = SCM_FRAME_SLOT (fp, i);
|
|
|
|
|
|
switch (repr)
|
|
|
|
|
|
{
|
|
|
|
|
|
case STACK_ITEM_SCM:
|
|
|
|
|
|
return item->as_scm;
|
|
|
|
|
|
case STACK_ITEM_F64:
|
2015-10-28 17:47:48 +00:00
|
|
|
|
return scm_from_double (item->as_f64);
|
2015-11-12 21:44:24 +01:00
|
|
|
|
case STACK_ITEM_U64:
|
|
|
|
|
|
return scm_from_uint64 (item->as_u64);
|
Add support for unboxed s64 values
* libguile/frames.c (enum stack_item_representation):
(scm_to_stack_item_representation):
(scm_frame_local_ref, scm_frame_local_set_x): Support for S64
representations.
* libguile/frames.h (union scm_vm_stack_element): Add signed 64-bit
integer field.
* libguile/vm-engine.c (scm->s64, s64->scm, load-s64): New
instructions.
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/cse.scm (compute-equivalent-subexpressions):
* module/language/cps/effects-analysis.scm:
* module/language/cps/slot-allocation.scm (compute-var-representations)
(compute-needs-slot, allocate-slots):
* module/language/cps/utils.scm (compute-constant-values):
* module/language/cps/specialize-primcalls.scm (specialize-primcalls):
Add support for new primcalls.
* module/language/cps/types.scm (&s64): New type.
(&s64-min, &s64-max, &u64-max): New convenience definitions.
(&range-min, &range-max): Use &s64-min and &u64-max names.
(scm->s64, load-s64, s64->scm): Add support for new primcalls.
* module/system/vm/assembler.scm (emit-scm->s64, emit-s64->scm)
(emit-load-s64): New exports.
* module/system/vm/assembler.scm (write-arities): Support for s64
slots.
* module/system/vm/debug.scm (arity-definitions): Support for s64
slots.
2015-11-21 10:32:33 +01:00
|
|
|
|
case STACK_ITEM_S64:
|
|
|
|
|
|
return scm_from_int64 (item->as_s64);
|
2015-10-28 17:03:42 +00:00
|
|
|
|
default:
|
|
|
|
|
|
abort();
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2013-11-15 17:13:10 +01:00
|
|
|
|
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
SCM_OUT_OF_RANGE (SCM_ARG2, index);
|
2001-04-23 04:28:13 +00:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
2001-04-22 02:13:48 +00:00
|
|
|
|
|
2016-01-30 17:47:03 +01:00
|
|
|
|
static const char s_scm_frame_local_set_x[] = "frame-local-set!";
|
|
|
|
|
|
static SCM
|
|
|
|
|
|
scm_frame_local_set_x (SCM frame, SCM index, SCM val, SCM representation)
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
#define FUNC_NAME s_scm_frame_local_set_x
|
2001-04-23 04:28:13 +00:00
|
|
|
|
{
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
union scm_vm_stack_element *fp, *sp;
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
unsigned int i;
|
2015-10-28 17:03:42 +00:00
|
|
|
|
enum stack_item_representation repr;
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
SCM_VALIDATE_UINT_COPY (2, index, i);
|
2015-10-28 17:03:42 +00:00
|
|
|
|
repr = scm_to_stack_item_representation (representation, FUNC_NAME, SCM_ARG3);
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
Frame pointer points to local 0 instead of local 1
* libguile/frames.h: Change so that fp points at local 0 instead of
local 1, and clean up a bit.
(struct scm_vm_frame): Remove program, and rename stack to locals.
(SCM_FRAME_DATA_ADDRESS): Remove; it was redundant with
SCM_FRAME_LOWER_ADDRESS.
(SCM_FRAME_STACK_ADDRESS): Remove; replace with the new
SCM_FRAME_LOCALS_ADDRESS.
(SCM_FRAME_UPPER_ADDRESS): Remove; unused.
(SCM_FRAME_NUM_LOCALS, SCM_FRAME_PREVIOUS_SP): New defines.
(SCM_FRAME_BYTE_CAST, SCM_FRAME_STACK_CAST): Remove; unused;
(SCM_FRAME_LOCAL): New define, replaces SCM_FRAME_VARIABLE.
(SCM_FRAME_PROGRAM): Add cautionary commentary.
* libguile/frames.c: Adapt static asserts.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_local_set_x):
Adapt. This means that frame-local-ref 0 now returns the procedure.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME)
(FRAME_LOCALS_COUNT, LOCAL_REF, LOCAL_SET, RETURN_VALUE_LIST): Adapt
to change in fp.
(LOCAL_ADDRESS): New helper.
(POP_CONTINUATION_HOOK): Reimplement, taking the previous FP as an
argument.
(ABORT_CONTINUATION_HOOK): Reimplement, taking no arguments.
(RETURN_ONE_VALUE): Reimplement.
(RETURN_VALUE_LIST): Adapt to FP change.
(halt, return-values, subr-call, foreign-call, prompt)
(continuation-call, compose-continuation, call/cc, abort): Adapt to FP
change, mostly via using LOCAL_ADDRESS, etc abstractions instead of
using the raw frame pointer.
* libguile/control.c (reify_partial_continuation): Update for fp
change.
* libguile/vm.c (vm_reinstate_partial_continuation): Adapt to removal of
SCM_FRAME_UPPER_ADDRESS.
* module/system/vm/frame.scm (frame-call-representation): Adapt to
frame-local-ref change.
* module/system/vm/trace.scm (print-return): Remove unused
frame-num-locals call.
2013-11-17 22:07:44 +01:00
|
|
|
|
fp = SCM_VM_FRAME_FP (frame);
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
sp = SCM_VM_FRAME_SP (frame);
|
2013-11-15 17:13:10 +01:00
|
|
|
|
|
Frame pointer points to local 0 instead of local 1
* libguile/frames.h: Change so that fp points at local 0 instead of
local 1, and clean up a bit.
(struct scm_vm_frame): Remove program, and rename stack to locals.
(SCM_FRAME_DATA_ADDRESS): Remove; it was redundant with
SCM_FRAME_LOWER_ADDRESS.
(SCM_FRAME_STACK_ADDRESS): Remove; replace with the new
SCM_FRAME_LOCALS_ADDRESS.
(SCM_FRAME_UPPER_ADDRESS): Remove; unused.
(SCM_FRAME_NUM_LOCALS, SCM_FRAME_PREVIOUS_SP): New defines.
(SCM_FRAME_BYTE_CAST, SCM_FRAME_STACK_CAST): Remove; unused;
(SCM_FRAME_LOCAL): New define, replaces SCM_FRAME_VARIABLE.
(SCM_FRAME_PROGRAM): Add cautionary commentary.
* libguile/frames.c: Adapt static asserts.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_local_set_x):
Adapt. This means that frame-local-ref 0 now returns the procedure.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME)
(FRAME_LOCALS_COUNT, LOCAL_REF, LOCAL_SET, RETURN_VALUE_LIST): Adapt
to change in fp.
(LOCAL_ADDRESS): New helper.
(POP_CONTINUATION_HOOK): Reimplement, taking the previous FP as an
argument.
(ABORT_CONTINUATION_HOOK): Reimplement, taking no arguments.
(RETURN_ONE_VALUE): Reimplement.
(RETURN_VALUE_LIST): Adapt to FP change.
(halt, return-values, subr-call, foreign-call, prompt)
(continuation-call, compose-continuation, call/cc, abort): Adapt to FP
change, mostly via using LOCAL_ADDRESS, etc abstractions instead of
using the raw frame pointer.
* libguile/control.c (reify_partial_continuation): Update for fp
change.
* libguile/vm.c (vm_reinstate_partial_continuation): Adapt to removal of
SCM_FRAME_UPPER_ADDRESS.
* module/system/vm/frame.scm (frame-call-representation): Adapt to
frame-local-ref change.
* module/system/vm/trace.scm (print-return): Remove unused
frame-num-locals call.
2013-11-17 22:07:44 +01:00
|
|
|
|
if (i < SCM_FRAME_NUM_LOCALS (fp, sp))
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
{
|
2015-10-28 17:03:42 +00:00
|
|
|
|
union scm_vm_stack_element *item = SCM_FRAME_SLOT (fp, i);
|
|
|
|
|
|
switch (repr)
|
|
|
|
|
|
{
|
|
|
|
|
|
case STACK_ITEM_SCM:
|
|
|
|
|
|
item->as_scm = val;
|
|
|
|
|
|
break;
|
|
|
|
|
|
case STACK_ITEM_F64:
|
2015-10-28 17:47:48 +00:00
|
|
|
|
item->as_f64 = scm_to_double (val);
|
|
|
|
|
|
break;
|
2015-11-12 21:44:24 +01:00
|
|
|
|
case STACK_ITEM_U64:
|
|
|
|
|
|
item->as_u64 = scm_to_uint64 (val);
|
|
|
|
|
|
break;
|
Add support for unboxed s64 values
* libguile/frames.c (enum stack_item_representation):
(scm_to_stack_item_representation):
(scm_frame_local_ref, scm_frame_local_set_x): Support for S64
representations.
* libguile/frames.h (union scm_vm_stack_element): Add signed 64-bit
integer field.
* libguile/vm-engine.c (scm->s64, s64->scm, load-s64): New
instructions.
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/cse.scm (compute-equivalent-subexpressions):
* module/language/cps/effects-analysis.scm:
* module/language/cps/slot-allocation.scm (compute-var-representations)
(compute-needs-slot, allocate-slots):
* module/language/cps/utils.scm (compute-constant-values):
* module/language/cps/specialize-primcalls.scm (specialize-primcalls):
Add support for new primcalls.
* module/language/cps/types.scm (&s64): New type.
(&s64-min, &s64-max, &u64-max): New convenience definitions.
(&range-min, &range-max): Use &s64-min and &u64-max names.
(scm->s64, load-s64, s64->scm): Add support for new primcalls.
* module/system/vm/assembler.scm (emit-scm->s64, emit-s64->scm)
(emit-load-s64): New exports.
* module/system/vm/assembler.scm (write-arities): Support for s64
slots.
* module/system/vm/debug.scm (arity-definitions): Support for s64
slots.
2015-11-21 10:32:33 +01:00
|
|
|
|
case STACK_ITEM_S64:
|
|
|
|
|
|
item->as_s64 = scm_to_int64 (val);
|
|
|
|
|
|
break;
|
2015-10-28 17:03:42 +00:00
|
|
|
|
default:
|
|
|
|
|
|
abort();
|
|
|
|
|
|
}
|
2013-11-15 17:13:10 +01:00
|
|
|
|
return SCM_UNSPECIFIED;
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
}
|
2013-11-15 17:13:10 +01:00
|
|
|
|
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
SCM_OUT_OF_RANGE (SCM_ARG2, index);
|
|
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
2018-08-06 17:00:45 +02:00
|
|
|
|
static const char s_scm_frame_return_values[] = "frame-return-values";
|
|
|
|
|
|
static SCM
|
|
|
|
|
|
scm_frame_return_values (SCM frame)
|
|
|
|
|
|
#define FUNC_NAME s_scm_frame_return_values
|
|
|
|
|
|
{
|
|
|
|
|
|
const uint32_t *ip;
|
|
|
|
|
|
union scm_vm_stack_element *fp, *sp;
|
|
|
|
|
|
SCM vals = SCM_EOL;
|
|
|
|
|
|
size_t n;
|
|
|
|
|
|
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
|
|
|
|
|
|
|
|
|
|
|
ip = SCM_VM_FRAME_IP (frame);
|
|
|
|
|
|
fp = SCM_VM_FRAME_FP (frame);
|
|
|
|
|
|
sp = SCM_VM_FRAME_SP (frame);
|
|
|
|
|
|
|
|
|
|
|
|
if ((*ip & 0xff) != scm_op_return_values)
|
|
|
|
|
|
scm_wrong_type_arg_msg (FUNC_NAME, 1, frame, "not a return frame");
|
|
|
|
|
|
|
|
|
|
|
|
n = SCM_FRAME_NUM_LOCALS (fp, sp);
|
|
|
|
|
|
while (n--)
|
|
|
|
|
|
vals = scm_cons (SCM_FRAME_LOCAL (fp, n), vals);
|
|
|
|
|
|
|
|
|
|
|
|
return vals;
|
|
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2010-09-23 13:19:49 +02:00
|
|
|
|
SCM_DEFINE (scm_frame_address, "frame-address", 1, 0, 0,
|
|
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"Return the frame pointer for @var{frame}.")
|
|
|
|
|
|
#define FUNC_NAME s_scm_frame_address
|
|
|
|
|
|
{
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
2014-05-04 11:46:18 +02:00
|
|
|
|
return scm_from_ptrdiff_t (SCM_VM_FRAME_FP_OFFSET (frame));
|
2010-09-23 13:19:49 +02:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2010-09-30 21:29:20 +02:00
|
|
|
|
SCM_DEFINE (scm_frame_stack_pointer, "frame-stack-pointer", 1, 0, 0,
|
|
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"")
|
|
|
|
|
|
#define FUNC_NAME s_scm_frame_stack_pointer
|
|
|
|
|
|
{
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
|
|
|
|
|
|
2014-05-04 11:46:18 +02:00
|
|
|
|
return scm_from_ptrdiff_t (SCM_VM_FRAME_SP_OFFSET (frame));
|
2010-09-30 21:29:20 +02:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
SCM_DEFINE (scm_frame_instruction_pointer, "frame-instruction-pointer", 1, 0, 0,
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"")
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
#define FUNC_NAME s_scm_frame_instruction_pointer
|
runtime and debugging support for callee-parsed procedure args
* libguile/objcodes.h: Bump for metadata format change.
* libguile/frames.h: Rework so we don't frob the program's nargs, nlocs,
etc at runtime. Instead we don't really know what's a local var, an
argument, or an intermediate value. It's a little unfortunate, but
this will allow for case-lambda, and eventually for good polymorphic
generic dispatch; and the nlocs etc can be heuristically
reconstructed. Such a reconstruction would be better done at the
Scheme level, though.
(SCM_FRAME_STACK_ADDRESS): New macro, the pointer to the base of the
stack elements (not counting the program).
(SCM_FRAME_UPPER_ADDRESS): Repurpose to be the address of the last
element in the bookkeeping part of the stack -- i.e. to point to the
return address.
* libguile/vm-engine.h:
* libguile/vm-i-system.c: Adapt to removal of stack_base. Though we
still detect stack-smashing underflow, we don't do so as precisely as
we did before, because now we only detect overwriting of the frame
metadata.
* libguile/vm-engine.c (vm_engine): Remove the stack_base variable. It
is unnecessary, and difficult to keep track of in the face of
case-lambda. Also fix miscommented "ra" and "mvra" pushes. Push the
vp->ip as the first ra...
* libguile/vm-i-system.c (halt): ...because here we can restore the
vp->ip instead of setting ip to 0. Allows us to introspect ips all
down the stack, including in recursive VM invocations.
* libguile/frames.h:
* libguile/frames.c (scm_vm_frame_stack): Removed, because it's getting
more difficult to tell what's an argument and what's a temporary stack
element.
(scm_vm_frame_num_locals): New accessor.
(scm_vm_frame_instruction_pointer): New accessor.
(scm_vm_frame_arguments): Defer to an implementation in Scheme.
(scm_vm_frame_num_locals scm_vm_frame_local_ref)
(scm_vm_frame_local_set_x): Since we can get not-yet-active frames on
the stack now, with our current calling convention, we have to add a
heuristic here to jump over those frames -- because frames have
pointers in them, not Scheme values.
* libguile/programs.h:
* libguile/programs.c (scm_program_arity): Remove, in favor of..
(scm_program_arities): ...this, which a list of arities, in a new
format, occupying a slot in the metadata.
* module/language/assembly/decompile-bytecode.scm (decode-load-program):
Fix mv-call decompilation.
* module/system/vm/frame.scm (vm-frame-bindings, vm-frame-binding-ref)
(vm-frame-binding-set!): New functions, to access bindings by name in
a frame.
(vm-frame-arguments): Function now implemented in Scheme. Commented
fairly extensively.
* module/system/vm/program.scm (program-bindings-by-index)
(program-bindings-for-ip): New accessors, parsing the program bindings
metadata into something more useful.
(program-arities, program-arguments): In a case-lambda world, we have
to assume that programs can have multiple arities. But it's tough to
detect this algorithmically; instead we're going to require that the
program metadata include information about the arities, and the parts
of the program that that metadata applies to.
(program-lambda-list): New accessor.
(write-program): Show multiple arities.
* module/language/glil/compile-assembly.scm (glil->assembly): Add
"arities" to the state of the compiler, and add arities entries as
appropriate.
2009-09-27 19:25:58 -04:00
|
|
|
|
{
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
2012-05-10 12:43:33 +02:00
|
|
|
|
|
2018-06-21 08:39:03 +02:00
|
|
|
|
return scm_from_uintptr_t ((uintptr_t) SCM_VM_FRAME_IP (frame));
|
2001-04-22 02:13:48 +00:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
SCM_DEFINE (scm_frame_return_address, "frame-return-address", 1, 0, 0,
|
2001-04-22 02:13:48 +00:00
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"")
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
#define FUNC_NAME s_scm_frame_return_address
|
2001-04-22 02:13:48 +00:00
|
|
|
|
{
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
2018-07-15 09:50:52 +02:00
|
|
|
|
return scm_from_uintptr_t ((uintptr_t) (SCM_FRAME_VIRTUAL_RETURN_ADDRESS
|
|
|
|
|
|
(SCM_VM_FRAME_FP (frame))));
|
2001-04-22 02:13:48 +00:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
SCM_DEFINE (scm_frame_dynamic_link, "frame-dynamic-link", 1, 0, 0,
|
2001-04-22 02:13:48 +00:00
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"")
|
replace frame implementation with VM frames
* libguile/stacks.h: Rework so that a stack doesn't copy information out
of VM frames, it just holds onto a VM frame, along with the stack id
and length. VM frames are now the only representation of frames in
Guile.
(scm_t_info_frame, SCM_FRAME_N_SLOTS, SCM_FRAME_REF, SCM_FRAME_NUMBER)
(SCM_FRAME_FLAGS, SCM_FRAME_SOURCE, SCM_FRAME_PROC, SCM_FRAME_ARGS)
(SCM_FRAME_PREV, SCM_FRAME_NEXT)
(SCM_FRAMEF_VOID, SCM_FRAMEF_REAL, SCM_FRAMEF_PROC)
(SCM_FRAMEF_EVAL_ARGS, SCM_FRAMEF_OVERFLOW)
(SCM_FRAME_VOID_P, SCM_FRAME_REAL_P, SCM_FRAME_PROC_P)
(SCM_FRAME_EVAL_ARGS_P, SCM_FRAME_OVERFLOW_P): Remove these macros
corresponding to the old frame implementation.
(scm_frame_p scm_frame_source, scm_frame_procedure)
(scm_frame_arguments): These definitions are now in frames.h.
(scm_last_stack_frame): Remove declaration of previously-removed
constructor. Probably should re-instate it though.
(scm_frame_number, scm_frame_previous, scm_frame_next)
(scm_frame_real_p, scm_frame_procedure_p, scm_frame_evaluating_args_p)
(scm_frame_overflow_p) : Remove these procedures corresponding to the
old stack implementation.
* libguile/stacks.c: Update for new frames implementation.
* libguile/frames.h:
* libguile/frames.c: Rename functions operating on VM frames to have a
scm_frame prefix, not scm_vm_frame -- because they really are the only
frames we have. Rename corresponding Scheme functions too, from
vm-frame-foo to frame-foo.
* libguile/deprecated.h: Remove scm_stack and scm_info_frame data types.
* libguile/vm.c (vm_dispatch_hook): Adapt to scm_c_make_frame name
change.
* module/system/vm/frame.scm: No need to export functions provided
frames.c now, as we load those procedures into the default environment
now. Rename functions, and remove a couple of outdated, unused
functions. The bottom half of this file is still bitrotten, though.
* libguile/backtrace.c: Rework to operate on the new frame
representation. Also fix a bug displaying file names for compiled
procedures.
* libguile/init.c: Load the VM much earlier, just because we can. Also
it allows us to have frames.[ch] loaded in time for stacks to be
initialized, so that scm_frame_arguments can do the right thing.
2009-12-03 13:09:58 +01:00
|
|
|
|
#define FUNC_NAME s_scm_frame_dynamic_link
|
2001-04-22 02:13:48 +00:00
|
|
|
|
{
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
|
|
|
|
|
/* fixme: munge fp if holder is a continuation */
|
2013-11-07 18:11:00 +01:00
|
|
|
|
return scm_from_uintptr_t
|
2018-06-21 08:39:03 +02:00
|
|
|
|
((uintptr_t)
|
2015-10-18 22:59:23 +02:00
|
|
|
|
SCM_FRAME_DYNAMIC_LINK (SCM_VM_FRAME_FP (frame)));
|
2001-04-22 02:13:48 +00:00
|
|
|
|
}
|
|
|
|
|
|
#undef FUNC_NAME
|
|
|
|
|
|
|
2014-04-14 14:54:14 +02:00
|
|
|
|
int
|
|
|
|
|
|
scm_c_frame_previous (enum scm_vm_frame_kind kind, struct scm_frame *frame)
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
{
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
union scm_vm_stack_element *this_fp, *new_fp, *new_sp;
|
|
|
|
|
|
union scm_vm_stack_element *stack_top = frame_stack_top (kind, frame);
|
2009-12-15 00:20:47 +01:00
|
|
|
|
|
|
|
|
|
|
again:
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
this_fp = stack_top - frame->fp_offset;
|
2014-04-16 19:17:38 +02:00
|
|
|
|
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
if (this_fp == stack_top)
|
2014-04-16 19:17:38 +02:00
|
|
|
|
return 0;
|
|
|
|
|
|
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
new_fp = SCM_FRAME_DYNAMIC_LINK (this_fp);
|
2014-04-16 19:17:38 +02:00
|
|
|
|
|
2015-10-18 22:59:23 +02:00
|
|
|
|
if (new_fp >= stack_top)
|
2014-04-14 14:54:14 +02:00
|
|
|
|
return 0;
|
2014-04-16 19:17:38 +02:00
|
|
|
|
|
|
|
|
|
|
new_sp = SCM_FRAME_PREVIOUS_SP (this_fp);
|
VM stack grows downward
Adapt VM stack to grow downward. This will make native compilation look
more like the VM code, as we will be able to use native CALL
instructions, taking proper advantage of the return address buffer.
* libguile/continuations.c (scm_i_continuation_to_frame): Record offsets
from stack top.
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Adapt for reversed
order of arguments, and instead of relying on the abort to push on the
number of arguments, make the caller save the stack depth, which
allows us to compute the number of arguments ourselves.
(reify_partial_continuation, scm_c_abort): Adapt to reversed stack
order.
* libguile/dynstack.c (scm_dynstack_wind_prompt): Since we wind the
stack in a downward direction, subtract the reloc instead of adding
it.
* libguile/dynstack.h (SCM_F_DYNSTACK_PROMPT_ESCAPE_ONLY): Remove flag;
instead rely on prompt-establishing code to save the stack depth.
* libguile/eval.c (eval): Remove extraneous "volatile" declarations for
variables that are not re-set between the setjmp and any longjmp.
Adapt to save stack depth before instating the prompt.
* libguile/foreign.c (scm_i_foreign_call): Adapt to receive arguments in
reverse order.
* libguile/frames.c (frame_stack_top, scm_i_frame_stack_top): Adapt to
compute stack top instead of stack bottom.
(scm_c_frame_closure): Adapt to stack growth change.
(scm_frame_num_locals, scm_frame_local_ref, scm_frame_set_x): Use
union data type to access stack.
(RELOC): Reformat.
(scm_c_frame_previous): Adapt to stack growth change.
* libguile/frames.h: Adapt stack diagram to indicate that the stack
grows up.
(union scm_vm_stack_element): New data type used to access items on
the stack.
(SCM_FRAME_PREVIOUS_SP)
(SCM_FRAME_RETURN_ADDRESS, SCM_FRAME_SET_RETURN_ADDRESS)
(SCM_FRAME_DYNAMIC_LINK, SCM_FRAME_SET_DYNAMIC_LINK)
(SCM_FRAME_LOCAL, SCM_FRAME_NUM_LOCALS): Adapt to stack representation
change.
(SCM_FRAME_SLOT): New helper.
(SCM_VM_FRAME_FP, SCM_VM_FRAME_SP): Adapt to stack growth change.
* libguile/stacks.c (scm_make_stack): Record offsets from top of stack.
* libguile/throw.c (catch): Adapt to scm_i_prompt_pop_abort_args_x
change.
* libguile/vm-engine.c (ALLOC_FRAME, RESET_FRAME):
(FRAME_LOCALS_COUNT_FROM): Adapt to stack growth change.
(LOCAL_ADDRESS): Use SCM_FRAME_SLOT to get the address as the proper
data type.
(RETURN_ONE_VALUE, RETURN_VALUE_LIST): Adapt to stack growth change.
(apply): Shuffling up the SMOB apply args can cause the stack to
expand, so use ALLOC_FRAME instead of RESET_FRAME.
(vm_engine): Adapt for stack growth change.
* libguile/vm.c (vm_increase_sp, vm_push_sp, vm_restore_sp): Adapt to
stack representation change.
(scm_i_vm_cont_to_frame): Adapt to take offsets from the top.
(scm_i_vm_capture_stack): Adapt to capture from the top.
(vm_return_to_continuation_inner): Adapt for data type changes.
(vm_return_to_continuation): Likewise, and instead of looping, just
splat the saved arguments on with memcpy.
(vm_dispatch_hook): Adapt to receive arguments in the reverse order.
Adapt callers.
(vm_abort): There is never a tail argument. Adapt to stack
representation change.
(vm_reinstate_partial_continuation)
(vm_reinstate_partial_continuation_inner): Adapt to stack growth
change.
(allocate_stack, free_stack): Adapt to data type change.
(expand_stack): Don't try to mremap(), as you can't grow a mapping
from the bottom. Without knowing that there's a free mapping space
right below the old stack, which there usually isn't on Linux, we have
to copy. We can't use MAP_GROWSDOWN because Linux is buggy.
(make_vm): Adapt to stack representation changes.
(return_unused_stack_to_os): Round down instead of up, as the stack
grows down.
(scm_i_vm_mark_stack): Adapt to walk up the stack.
(scm_i_vm_free_stack): Adapt to scm_vm changes.
(vm_expand_stack_inner, reset_stack_limit, vm_expand_stack): Adapt to
the stack growing down.
(scm_call_n): Adapt to the stack growing down. Don't allow argv to
point into the stack.
* libguile/vm.h (struct scm_vm, struct scm_vm_cont): Adapt to hold the
stack top and bottom.
2015-09-22 10:24:30 +00:00
|
|
|
|
frame->fp_offset = stack_top - new_fp;
|
|
|
|
|
|
frame->sp_offset = stack_top - new_sp;
|
2018-07-15 09:50:52 +02:00
|
|
|
|
frame->ip = SCM_FRAME_VIRTUAL_RETURN_ADDRESS (this_fp);
|
2014-04-16 19:17:38 +02:00
|
|
|
|
|
2015-11-26 16:36:22 +01:00
|
|
|
|
if (scm_i_vm_is_boot_continuation_code (frame->ip))
|
|
|
|
|
|
goto again;
|
2014-04-16 19:17:38 +02:00
|
|
|
|
|
|
|
|
|
|
return 1;
|
2014-04-14 14:54:14 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
SCM_DEFINE (scm_frame_previous, "frame-previous", 1, 0, 0,
|
|
|
|
|
|
(SCM frame),
|
|
|
|
|
|
"")
|
|
|
|
|
|
#define FUNC_NAME s_scm_frame_previous
|
|
|
|
|
|
{
|
|
|
|
|
|
enum scm_vm_frame_kind kind;
|
|
|
|
|
|
struct scm_frame tmp;
|
|
|
|
|
|
|
|
|
|
|
|
SCM_VALIDATE_VM_FRAME (1, frame);
|
|
|
|
|
|
|
|
|
|
|
|
kind = SCM_VM_FRAME_KIND (frame);
|
|
|
|
|
|
memcpy (&tmp, SCM_VM_FRAME_DATA (frame), sizeof tmp);
|
|
|
|
|
|
|
|
|
|
|
|
if (!scm_c_frame_previous (SCM_VM_FRAME_KIND (frame), &tmp))
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
return SCM_BOOL_F;
|
2014-04-14 14:54:14 +02:00
|
|
|
|
|
2014-04-14 15:14:26 +02:00
|
|
|
|
return scm_c_make_frame (kind, &tmp);
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
}
|
2009-12-15 00:20:47 +01:00
|
|
|
|
#undef FUNC_NAME
|
remove heap links in VM frames, incorporate vm frames into normal backtraces
* doc/ref/vm.texi (Stack Layout): Update to remove references to the
"heap link".
* gdbinit: Update for "heap link" removal.
* libguile/frames.c:
* libguile/frames.h: Update macros and diagram for removal of "heap
link". As part of this, we also remove "heap frames", replacing them
with "vm frames", which are much like the interpreter's debug objects,
but for VM stacks. That is to say, they don't actually hold the stack
themselves, just the pointers into stack that's held by a continuation
(either captured or current).
* libguile/stacks.c (stack_depth, read_frames): Since a "stack" object is
really a copy of information that comes from somewhere else, it makes
sense to copy over info from the VM, just as `make-stack' does from the
evaluator. The tricky bit is to figure out how to interleave VM and
interpreter frames. We do that by starting in the interpreter, and
whenever the current frame's procedure is actually a program, we switch
to the VM stack, switching back when we reach a "bootstrap frame". The
last bit is hacky, but it does work...
(is_vm_bootstrap_frame): Hacky predicate to see if a VM frame is a
bootstrap frame.
(scm_make_stack): Accept a VM frame in addition to debug frames.
Probably has some bugs in this case. But in the case that the arg is
#t (a common case), do the right thing, capturing the top VM frame as
well, and interleaving those frames appropriately on the stack.
As an accident, we lost the ability to limit the number of frames in
the backtrace. We could add that back, but personally I always want
*all* frames in the trace... Narrowing still works fine, though there
are some hiccups sometimes -- e.g. an outer cut to a procedure that
does a tail-call in VM code will never find the cut, as it no longer
exists in the continuation.
* libguile/vm.h (struct scm_vm): So! Now that we have switched to save
stacks in the normal make-stack, there's no more need for `this_frame'
or `last_frame'. On the other hand, we can take this opportunity to fix
tracing: when we're in a trace hook, we set `trace_frame' on the VM,
so we know not to fire hooks when we're already in a hook.
(struct scm_vm_cont): Expose this, as make-stack needs it to make VM
frames from VM continuations.
* libguile/vm.c (scm_vm_trace_frame): New function, gets the current
trace frame.
(vm_mark, make_vm): Hook up the trace frame.
(vm_dispatch_hook): New hook dispatcher, with a dynwind so it does the
right thing if the hook exits nonlocally.
* libguile/vm-engine.c (vm_run): No more this_frame in the wind data.
* libguile/vm-engine.h (RUN_HOOK): Run hooks through the dispatcher.
(ALIGN_AS_NON_IMMEDIATE, POP_LIST_ON_STACK): Remove unused code.
(NEW_FRAME): Adapt for no HL in the frame.
* libguile/vm-i-system.c (goto/args, mv-call, return, return/values):
Adapt for no HL in the frame.
* module/system/vm/frame.scm:
* module/system/vm/vm.scm: Beginnings of some reworkings, needs more
thought.
2008-12-26 17:59:46 +01:00
|
|
|
|
|
2001-04-22 02:13:48 +00:00
|
|
|
|
|
2016-01-30 17:47:03 +01:00
|
|
|
|
static void
|
|
|
|
|
|
scm_init_frames_builtins (void *unused)
|
|
|
|
|
|
{
|
|
|
|
|
|
scm_c_define_gsubr (s_scm_frame_num_locals, 1, 0, 0,
|
|
|
|
|
|
(scm_t_subr) scm_frame_num_locals);
|
|
|
|
|
|
scm_c_define_gsubr (s_scm_frame_local_ref, 3, 0, 0,
|
|
|
|
|
|
(scm_t_subr) scm_frame_local_ref);
|
|
|
|
|
|
scm_c_define_gsubr (s_scm_frame_local_set_x, 4, 0, 0,
|
|
|
|
|
|
(scm_t_subr) scm_frame_local_set_x);
|
2018-08-06 17:00:45 +02:00
|
|
|
|
scm_c_define_gsubr (s_scm_frame_return_values, 1, 0, 0,
|
|
|
|
|
|
(scm_t_subr) scm_frame_return_values);
|
2016-01-30 17:47:03 +01:00
|
|
|
|
}
|
|
|
|
|
|
|
big reorg of scheme modules -- e.g. programs.c -> (system vm program)
This reorganization kills the ugly module-export-all hacks in
bootstrap.scm and core.scm. In fact, it gets rid of core.scm entirely,
breaking out its functionality into separate files.
* module/system/vm/trace.scm:
* module/system/vm/profile.scm:
* module/system/vm/disasm.scm:
* module/system/vm/debug.scm:
* module/system/vm/conv.scm:
* module/system/vm/assemble.scm:
* module/system/repl/repl.scm:
* module/system/repl/common.scm:
* module/system/base/compile.scm:
* module/system/repl/command.scm: Update for changes, and fix a bug in
procedure-documentation.
* module/system/vm/bootstrap.scm: Just call scm_bootstrap_vm, which
handles setting load-compiled for us.
* module/system/vm/core.scm: Removed, functionality folded into other
modules.
* module/system/vm/frame.scm: Export the C frame procedures here; also
move scheme functions from core.scm here.
* module/system/vm/instruction.scm: New file, exports procedures from
instructions.c.
* module/system/vm/objcode.scm: New file, exports procedures from
objcodes.c.
* module/system/vm/program.scm: New file, exports procedures from
programs.c, and some scheme functions originally from core.scm.
* module/system/vm/vm.scm: New file, from vm.c and core.scm.
* src/Makefile.am (libguile_vm_la_SOURCES): Add bootstrap.h.
* src/bootstrap.h: New file, prototypes scm_bootstrap_vm (), which the
scm_init_* functions call.
* src/frames.h:
* src/frames.c (scm_init_frames):
* src/frames.c (scm_bootstrap_frames):
* src/vm.h:
* src/instructions.h:
* src/instructions.c (scm_init_instructions):
* src/instructions.c (scm_bootstrap_instructions):
* src/objcodes.h:
* src/objcodes.c (scm_bootstrap_objcodes):
* src/objcodes.c (scm_init_objcodes):
* src/programs.h:
* src/programs.c (scm_bootstrap_programs):
* src/programs.c (scm_init_programs):
* src/vm.c (scm_bootstrap_vm):
* src/vm.c (scm_init_vm): Call scm_bootstrap_vm() before doing anything
in an init function. Bootstrap_vm will call bootstrap_instructions(),
etc to initialize types, then set load-compiled to point to
load-compiled/vm.
* src/vm.c (scm_load_compiled_with_vm): Code to load .go files, if
they're present.
2008-08-07 13:11:27 +02:00
|
|
|
|
void
|
|
|
|
|
|
scm_init_frames (void)
|
|
|
|
|
|
{
|
2001-04-22 02:13:48 +00:00
|
|
|
|
#ifndef SCM_MAGIC_SNARFER
|
2018-06-20 17:19:31 +02:00
|
|
|
|
#include "frames.x"
|
2001-04-22 02:13:48 +00:00
|
|
|
|
#endif
|
2016-01-30 17:47:03 +01:00
|
|
|
|
|
|
|
|
|
|
scm_c_register_extension ("libguile-" SCM_EFFECTIVE_VERSION,
|
|
|
|
|
|
"scm_init_frames_builtins",
|
|
|
|
|
|
scm_init_frames_builtins,
|
|
|
|
|
|
NULL);
|
2001-04-22 02:13:48 +00:00
|
|
|
|
}
|