abf497938b
Revert "VM: cast to fixnum looks suspicious"
...
This reverts commit 5027830b05
2016-11-16 05:23:54 +01:00
5027830b05
VM: cast to fixnum looks suspicious
2016-11-15 05:58:11 +01:00
5f869f8181
VM: the else clause is not necessary
2016-11-15 05:57:40 +01:00
69acbbbefd
VM: next_block_after can be removed
2016-11-15 03:49:46 +01:00
94b274490c
Revert "VM: undo a8aaa42882 ( #1513 )"
...
This reverts commit e50ef532ca
2016-11-12 00:19:52 +01:00
e50ef532ca
VM: undo a8aaa42882 ( #1513 )
2016-11-11 17:36:04 +01:00
0b6ffdc77e
Revert "VM: undoing 0274e889b9 ( #1513 )"
...
This reverts commit a4802aba68
2016-11-09 02:07:56 +01:00
a4802aba68
VM: undoing 0274e889b9 ( #1513 )
2016-11-08 13:37:52 +01:00
2308ca06e0
Revert "VM: undoing 22bb0cfecf ( #1513 )"
...
This reverts commit e9c72baf2e
2016-11-07 23:43:50 +01:00
e9c72baf2e
VM: undoing 22bb0cfecf ( #1513 )
2016-11-07 14:18:10 +01:00
4fee8bd2cb
vm: change where we set safepoints_fep to false to allow old Ctrl-C behavior.
2016-11-01 19:21:47 -07:00
0245a93bee
VM: add enable-ctrl-break and disable-ctrl-break primitives
...
These are for the Listener to wrap around user-supplied quotations. Nesting
is not supported at the moment.
2016-11-01 16:53:00 -07:00
526a995d64
VM: implement a ctrl-break handler thread ( #1573 )
2016-11-01 16:53:00 -07:00
e9cd20604c
VM: factor_vm::compile_inline_cache() can be removed
2016-10-31 16:16:42 +01:00
ffab2bb7e0
VM: code_block::optimized_p() method isn't needed
2016-10-31 14:18:39 +01:00
55fdf79afd
vm: fix missing CAPS enum.
2016-10-26 19:52:18 -07:00
f070a47ec0
VM: add marking as a timed gc phase
...
I think that makes sense because most time when doing a full gc is spent
marking.
2016-10-20 07:46:21 +02:00
752c895d23
VM: refactor gc_event so that the phase times are stored in an array
...
So that you don't need a new method for each gc phase to time.
2016-10-20 00:34:18 +02:00
51408b66cb
VM: the jit class doesn't need to store the code block type
2016-10-19 09:28:59 +02:00
453eca66c4
VM: changing case of code_block_type
2016-10-19 09:14:58 +02:00
f147bd4404
VM: uppercasing gc_op
2016-10-19 09:09:32 +02:00
df171b0485
VM: changing case of enums values bignum_comparison and generation
...
It is consistent if all enum values have all uppercase names.
2016-10-19 08:57:33 +02:00
4f6381587c
VM: removing update_code_roots_for_sweep()
2016-10-19 08:12:40 +02:00
7d9bad465c
VM: refactoring allot_code_block() to work like allot_large_object()
...
To make it possible to eventually grow the code heap
2016-10-17 08:03:48 +02:00
026b626203
VM: move allot_code_block() to the allot.hpp file
2016-10-17 08:03:48 +02:00
3911f7b88c
VM: init_data_heap() isn't needed
2016-10-14 13:52:44 +02:00
67bb9a7c71
VM: checking addresses differ before memmoving
...
It appears to speed up compaction a bit in cases where only a few
objects need to move.
2016-10-14 13:52:44 +02:00
aace892527
VM: rename of collect_growing_heap_op and collect_growing_heap()
...
The idea is to "make room" for collect_growing_code_heap_op and
collect_growing_code_heap() to enable code heap growth.
2016-10-13 02:13:09 +02:00
55e6ccb708
VM: remove the reversing from sampling profiler callstack collection #452
2016-10-13 01:34:55 +02:00
46105b62f6
VM: change callstack>array primitive so it doesn't reverse anymore
...
This fixes one part of the #452 bug.
2016-10-13 01:34:55 +02:00
98db13fe9d
VM: merge of clone_object() and primitive_clone()
2016-10-13 01:34:55 +02:00
bbd71ae975
math.parser: (format-float) doesn't need to zero-terminate returned byte-array
2016-10-09 00:47:37 +02:00
ac44f3069a
VM: indent comment and change error message
2016-10-08 07:13:08 +02:00
ae9818db87
VM: simpler code for find_object_containing_card
2016-10-08 07:11:47 +02:00
8cacf7bf8d
VM: simpler code for reallot_array
2016-10-08 07:10:26 +02:00
c2f4fdb172
VM: merge of the free_list and free_list_allocator classes
...
Seem simpler to have all the free list stuff in one class rather than
split it over two classes.
2016-10-08 07:04:28 +02:00
f0eec26f3c
VM: reorg and more comments
2016-10-08 07:04:28 +02:00
78d9065db0
VM: UAP_SET_TOC_POINTER appears to be a noop on all platforms
2016-09-29 07:14:30 +02:00
9eab4b5652
VM: make visit_instruction_operands() skip over uninitialized blocks
...
It will crash otherwise when compacting the code heap.
2016-09-29 06:35:35 +02:00
603f5f9bbf
VM: better? comment for uninitialized_blocks
2016-09-28 20:38:52 +02:00
91843ef22d
VM: removing end_gc()
2016-09-28 05:22:32 +02:00
8899914697
VM: better start_gc_again() code
2016-09-28 05:20:12 +02:00
c6faf98c86
VM: return the loop, it is needed for best performance
2016-09-26 05:16:33 +02:00
a6df4168af
bootstrap.layouts: sync note in comments
2016-09-25 20:10:16 +02:00
d71bd83f57
VM: removing the 1-arg overload of iterate() (better to be explicit about the fixup)
2016-09-22 17:03:38 +02:00
764a604467
VM: moving initialize_code_blocks() into primitive_modify_code_heap()
2016-09-22 15:16:07 +02:00
f8b94c2679
VM: allot_large_object fits better in the allot.hpp file
2016-09-22 13:13:55 +02:00
c02d913579
VM: big refactoring removing the gc_workhorse
...
Each collector were a slot_visitor containing a gc_workhorse containing
a policy class. This commit removes the gc_workhorse and the policies
and instead "inlines" the common code in the collector bodies. So there
is more code duplication, but the control flow doesn't "jump around" so
much so it is easier to understand. It also makes the nursery gc faster
because its collection code can be optimized better without the
workhorse/policy system.
2016-09-22 13:12:59 +02:00
a0df88f20b
VM: renaming verify_memory_protection_error to set_memory_protection_error
...
and sets the signal_fault_addr and signal_fault_pc in the function.
2016-09-22 13:12:59 +02:00
b2bb1f8ff4
VM: simplifying the load_image() code
2016-09-16 06:27:46 +02:00
Björn Lindqvist
John Benediktsson
Alexander Iljin