Andy Wingo
1b3bc2f3b5
Use "stack-conservative" and "heap-conservative" terms
...
Thanks to Steve Blackburn for the terms.
2022-12-02 09:07:24 +01:00
Andy Wingo
c614c2e40b
Refactor embedder interface for conservative GC
...
Now users don't have to #ifdef on conservative vs precise tracing; it's
just a generic embedder concern.
2022-10-26 11:59:56 +02:00
Andy Wingo
910b62af8f
Add conservative heap tracing (not just roots)
...
Also accelerate mark_space_live_object_granules.
2022-10-26 10:37:55 +02:00
Andy Wingo
053dbf0b61
Pass heap to tracer functions
...
This will allow conservative intra-heap edges. Hopefully no overhead?
2022-10-25 14:25:55 +02:00
Andy Wingo
5e986e84e9
Update README
2022-10-03 16:12:06 +02:00
Andy Wingo
703bb30e19
add conservative makefile targets
2022-10-03 16:09:21 +02:00
Andy Wingo
1944b54a19
Whippet can trace conservative roots
...
Next up, enabling it via the makefiles.
2022-10-03 16:09:21 +02:00
Andy Wingo
deed415a06
Whippet captures stack when stopping mutators
...
This is part of work to enable conservative GC.
2022-10-03 16:09:21 +02:00
Andy Wingo
d2bde8319f
Add conservative stack capture
...
This isn't really wired up yet anywhere, but add a precursor to
conservative stack scanning.
2022-10-03 16:09:21 +02:00
Andy Wingo
a5b1a66d21
Add platform abstraction
...
This will allow us to iterate conservative roots from stacks and static
data segments.
2022-10-03 16:09:21 +02:00
Andy Wingo
05d2c95950
mt-gcbench: Only crash when tracing holes for precise GC
2022-10-03 16:09:21 +02:00
Andy Wingo
e328346bbd
Refactor alignment utilities in whippet.c
...
Add align_up and align_down helpers.
2022-10-03 15:37:16 +02:00
Andy Wingo
24bd94d9f7
Fix race condition in computation of mark-while-stopping
...
Choose the ragged stop strategy when the GC kind is determined, so that
we do so with respect to a single measurement of pending unavailable
bytes.
Also remove assert in heap_should_mark_while_stopping, as it can be
called after stopping too, when evacuation is enabled.
2022-10-03 15:37:16 +02:00
Andy Wingo
1e3122d054
trace_worker_steal first does a try_pop on its own deque
...
Before asking other threads for values, see if there is any pending data
that overflowed from the local mark stack.
2022-10-03 15:37:16 +02:00
Andy Wingo
8b8ddaf6e2
work-stealing optimization: stay with last-stolen worker
...
Previously we were always going round-robin. Now a thief tries to
plunder its victim again directly. Should result in less churn.
2022-10-03 15:37:16 +02:00
Andy Wingo
56aad402c9
Fix bug in try_pop on chase-lev deque
...
The counters are unsigned, so that they can overflow. (Is that really
necessary though?) In any case try_pop can decrement a counter, leading
to a situation where you can think you have (size_t)-1 elements; not
good. Instead when computing the queue size, use a signed value.
Limits total queue size to half the unsigned space; fine.
2022-10-03 15:37:16 +02:00
Andy Wingo
f77cf923c1
Fix parallel tracer for gc_ref API change
2022-10-03 15:37:01 +02:00
Andy Wingo
1228e346fa
Fix semi-space collector for refactor
2022-10-03 15:37:01 +02:00
Andy Wingo
8a51117763
Rework pinning, prepare for conservative tracing
...
We don't need a pin bit: we just need to mark pinned objects before
evacuation starts. This way we can remove the stopping / marking race
so that we can always mark while stopping.
2022-08-22 21:11:30 +02:00
Andy Wingo
2199d5f48d
Excise struct gcobj
2022-08-16 23:21:16 +02:00
Andy Wingo
6ecf226570
More typesafety, more gc_ref
2022-08-16 22:48:46 +02:00
Andy Wingo
92b8f1e917
Add gc_ prefix to struct heap, struct mutator
2022-08-16 21:35:16 +02:00
Andy Wingo
b082f5f50d
Separate compilation!!!!!
2022-08-16 17:54:15 +02:00
Andy Wingo
fe9bdf6397
Separate out embedder API from mt-gcbench, quads
2022-08-16 16:09:36 +02:00
Andy Wingo
112f27b77b
Simplify GC attributes for the inline allocator
...
Don't require pulling in all of gc-api.h.
2022-08-16 16:00:06 +02:00
Andy Wingo
8a111256c6
Compile with -fvisibility=hidden; will be good for separate compilation
2022-08-16 12:04:56 +02:00
Andy Wingo
9e8940e59f
Get handles out of collectors
2022-08-16 11:53:32 +02:00
Andy Wingo
607585e7f0
Add whippet-inline.h
2022-08-16 10:25:23 +02:00
Andy Wingo
33aa5230da
Add bdw-inline.h
2022-08-15 18:30:42 +02:00
Andy Wingo
8f2f4f7c69
API-ify gc_print_stats; add semi-inline.h
2022-08-15 18:17:18 +02:00
Andy Wingo
a00c83878e
Inline post-allocation actions
2022-08-15 16:06:52 +02:00
Andy Wingo
a75842be90
Mostly implementation-independent inline allocation
...
This is a step towards separate compilation of the GC without losing
performance. Only remaining task is the write barrier.
2022-08-15 11:17:15 +02:00
Andy Wingo
4d8a7169d0
Add inline to gc-api.h
2022-08-14 09:18:21 +02:00
Andy Wingo
fb71c4c363
Separate tagging from collector
...
The collector now has an abstract interface onto the embedder. The
embedder has to supply some functionality, such as tracing and
forwarding. This is a pretty big change in terms of lines but it's
supposed to have no functional or performance change.
2022-08-12 16:44:38 +02:00
Andy Wingo
cacc28b577
Always add a header onto objects
...
We're targetting systems that need to be able to inspect the kind of an
object, so this information has to be somewhere. If it's out-of-line,
we might save memory, but we would lose locality. Concretely in Guile
the tag bits are in the object itself.
2022-08-09 16:14:47 +02:00
Andy Wingo
d8bcbf2d74
More API-ification
2022-08-09 11:35:31 +02:00
Andy Wingo
4ccb489869
Set fixed heap size, parallelism via explicit options
2022-08-09 11:21:02 +02:00
Andy Wingo
2e6dde66b3
Attempt to start creating a proper API
2022-08-09 09:49:51 +02:00
Andy Wingo
c824f17bd9
Rename gc-types.h to gc-api.h
2022-08-08 11:08:36 +02:00
Andy Wingo
67f9c89f2a
Use fragmentation_low_threshold for venerable_threshold
...
This way fragmentation from venerable blocks doesn't cause the collector
to keep evacuating.
2022-08-04 11:32:06 +02:00
Andy Wingo
0450a282dd
Skip mostly-tenured blocks during sweep/allocate after minor GC
2022-08-04 09:04:27 +02:00
Andy Wingo
0fe13e1cab
Accelerate scanning of remembered set
2022-08-03 21:25:18 +02:00
Andy Wingo
47c07dd0eb
Fix embarassing ctz issue
2022-08-03 16:40:34 +02:00
Andy Wingo
8f6a2692ab
Update README
2022-08-03 12:13:25 +02:00
Andy Wingo
96b68095b7
Fix mark pattern updating for generational whippet
...
After a minor collection, we were erroneously failing to sweep dead
objects with the survivor tag.
2022-08-03 12:06:19 +02:00
Andy Wingo
0210a8caf0
Refactor out-of-memory detection
...
Firstly, we add a priority evacuation reserve to prioritize having a few
evacuation blocks on hand. Otherwise if we give them all to big
allocations first and we have a fragmented heap, we won't be able to
evacuate that fragmented heap to give more blocks to the large
allocations.
Secondly, we remove `enum gc_reason`. The issue is that with multiple
mutator threads, the precise thread triggering GC does not provide much
information. Instead we should make choices on how to collect based on
the state of the heap.
Finally, we move detection of out-of-memory inside the collector,
instead of the allocator.
Together, these changes let mt-gcbench (with fragmentation) operate in
smaller heaps.
2022-08-03 10:10:33 +02:00
Andy Wingo
1358d99abc
Fix yield calculation after evacuating collections
2022-08-02 22:15:13 +02:00
Andy Wingo
a4e1f55f37
Implement generational collection
...
Not really battle-tested but it seems to work. Need to implement
heuristics for when to do generational vs full-heap GC.
2022-08-02 15:37:02 +02:00
Andy Wingo
13b3bb5b24
Update barrier functions to also have the object being written
...
Also remove read barriers, as they were unused, and we have no plans to
use them.
2022-08-02 15:37:02 +02:00
Andy Wingo
7f405c929e
Initial live mask does not include young allocations
...
After rotation, the young bit wasn't being included anyway. This just
improves the first collection.
2022-08-02 15:37:02 +02:00
