| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Add functions size_shared, copy_shared_calculate and copy_shared_perform.
Add the infrastructure for making these communicate with each other.
Add debug information to other places in the VM, to watch interaction
with the sharing-preserving copy.
CAUTION: If you define the SHCOPY_DEBUG macro (after SHCOPY is actually
used in the VM) and make the whole OTP, there will be a lot of debugging
messages during make (it will also be enabled in erlc). You have been
warned...
|
|
* rickard/gc-bump-reds/OTP-13097:
Bump reductions on GC
|
|
* rickard/ohmq/OTP-13047:
Fragmented young heap generation and off_heap_message_queue option
Refactor GC
Introduce literal tag
Conflicts:
erts/doc/src/erlang.xml
erts/emulator/beam/erl_gc.c
|
|
* sverk/literal-memory-range:
erts: Refactor line table in loaded beam code
erts: Refactor header of loaded beam code
fix check_process_code for separate literal area
erts: Add support for fast erts_is_literal()
erts: Refactor erl_mmap to allow several mapper instances
erts: Add new allocator LITERAL
erts: Fix strangeness in treatment of MSEG_ALIGN_BITS
erts: Cleanup main carrier creation
erts: Remove unused erts_have_erts_mmap
erts: Refactor config test for posix_memalign
|
|
|
|
* The youngest generation of the heap can now consist of multiple
blocks. Heap fragments and message fragments are added to the
youngest generation when needed without triggering a GC. After
a GC the youngest generation is contained in one single block.
* The off_heap_message_queue process flag has been added. When
enabled all message data in the queue is kept off heap. When
a message is selected from the queue, the message fragment (or
heap fragment) containing the actual message is attached to the
youngest generation. Messages stored off heap is not part of GC.
|
|
|
|
|
|
|
|
Binary vheap mature is not necessary for binary gc.
|
|
* Removed COMPRESS_POINTER and EXPAND_POINTER
|
|
|
|
|
|
* hamt_bin2term:
erts: Add erts_factory_trim_and_close
erts: Optimize driver_deliver_term
erts: Remove hashmap probabilistic heap overestimation
Conflicts:
erts/emulator/beam/beam_load.c
|
|
by adding a dynamic heap factory.
"binary_to_term" is now a hybrid solution with both
a call to decoded_size() to calculate needed heap space
AND possible dynamic allocation of more heap space
if needed for big maps.
The heap size returned from decoded_size() is guaranteed
to be sufficient for all term heap data except for hashmap
nodes. All hashmap nodes are created at the end of dec_term()
by invoking the heap factory interface that may allocate more
heap space on process heap or in fragments.
With this commit it is no longer guaranteed that a message
is confined to only one heap fragment.
|
|
* sverk/map-merge-trap:
erts: Optimize maps:merge
erts: Yield in maps:merge
erts: Refactor arg swapping for maps:merge
erts: Add save/restore for PSTACK
erts: Fix magic binary alignment on 32-bit
erts: Add maps to send_term_SUITE
erts: Fix calculation of reclaimed data during full gc
erts: Fix warning about const pointer to make_boxed and make_list
erts: Fix typo in etp-carrier-blocks
|
|
* egil/fix-purge-literals/OTP-12821:
erts: Fix garbage collect literals in code purge
|
|
During code purging and check_process_code, the checking of the binary reference
embedded in the match binary state was omitted for the tracing tests. This would cause
the binary match state to reference deallocated memory.
|
|
The old code did not take take the old-heap into acount.
|
|
* rickard/timer-optimization/OTP-12650:
Optimized timer implementation
Reusable red-black tree implementation
Conflicts:
erts/emulator/beam/erl_bif_timer.c
|
|
* rickard/time-improvement/OTP-11997:
Allow execution of estone suite on pre OTP-18 systems
Add parallel time monotonicity test-case
Replace usage of erlang:now() in line-tracing
Replace erlang:now() usage in emulator suite
Replace erlang:now() usage in system suite
Misc time improvements
|
|
|
|
- Possibility to chose different clock sources
- Improved mach clock usage
- Improved linux clock_gettime() usage
- ...
|
|
|
|
* rickard/time_api/OTP-11997: (22 commits)
Update primary bootstrap
inets: Suppress deprecated warning on erlang:now/0
inets: Cleanup of multiple copies of functions Add inets_lib with common functions used by multiple modules
inets: Update comments
Suppress deprecated warning on erlang:now/0
Use new time API and be back-compatible in inets Remove unused functions and removed redundant test
asn1 test SUITE: Eliminate use of now/0
Disable deprecated warning on erlang:now/0 in diameter_lib
Use new time API and be back-compatible in ssh
Replace all calls to now/0 in CT with new time API functions
test_server: Replace usage of erlang:now() with usage of new API
Replace usage of erlang:now() with usage of new API
Replace usage of erlang:now() with usage of new API
Replace usage of erlang:now() with usage of new API
Replace usage of erlang:now() with usage of new API
otp_SUITE: Warn for calls to erlang:now/0
Replace usage of erlang:now() with usage of new API
Multiple timer wheels
Erlang based BIF timer implementation for scalability
Implement ethread events with timeout
...
Conflicts:
bootstrap/bin/start.boot
bootstrap/bin/start_clean.boot
bootstrap/lib/compiler/ebin/beam_asm.beam
bootstrap/lib/compiler/ebin/compile.beam
bootstrap/lib/kernel/ebin/auth.beam
bootstrap/lib/kernel/ebin/dist_util.beam
bootstrap/lib/kernel/ebin/global.beam
bootstrap/lib/kernel/ebin/hipe_unified_loader.beam
bootstrap/lib/kernel/ebin/inet_db.beam
bootstrap/lib/kernel/ebin/inet_dns.beam
bootstrap/lib/kernel/ebin/inet_res.beam
bootstrap/lib/kernel/ebin/os.beam
bootstrap/lib/kernel/ebin/pg2.beam
bootstrap/lib/stdlib/ebin/dets.beam
bootstrap/lib/stdlib/ebin/dets_utils.beam
bootstrap/lib/stdlib/ebin/erl_tar.beam
bootstrap/lib/stdlib/ebin/escript.beam
bootstrap/lib/stdlib/ebin/file_sorter.beam
bootstrap/lib/stdlib/ebin/otp_internal.beam
bootstrap/lib/stdlib/ebin/qlc.beam
bootstrap/lib/stdlib/ebin/random.beam
bootstrap/lib/stdlib/ebin/supervisor.beam
bootstrap/lib/stdlib/ebin/timer.beam
erts/aclocal.m4
erts/emulator/beam/bif.c
erts/emulator/beam/erl_bif_info.c
erts/emulator/beam/erl_db_hash.c
erts/emulator/beam/erl_init.c
erts/emulator/beam/erl_process.h
erts/emulator/beam/erl_thr_progress.c
erts/emulator/beam/utils.c
erts/emulator/sys/unix/sys.c
erts/preloaded/ebin/erlang.beam
erts/preloaded/ebin/erts_internal.beam
erts/preloaded/ebin/init.beam
erts/preloaded/src/erts_internal.erl
lib/common_test/test/ct_hooks_SUITE_data/cth/tests/empty_cth.erl
lib/diameter/src/base/diameter_lib.erl
lib/kernel/src/os.erl
lib/ssh/test/ssh_basic_SUITE.erl
system/doc/efficiency_guide/advanced.xml
|
|
|
|
The old time API is based on erlang:now/0. The major issue with
erlang:now/0 is that it was intended to be used for so many
unrelated things. This tied these unrelated operations together
and unnecessarily caused performance, scalability as well as
accuracy, and precision issues for operations that do not need
to have such issues. The new API spreads different functionality
over multiple functions in order to improve on this.
The new API consists of a number of new BIFs:
- erlang:convert_time_unit/3
- erlang:monotonic_time/0
- erlang:monotonic_time/1
- erlang:system_time/0
- erlang:system_time/1
- erlang:time_offset/0
- erlang:time_offset/1
- erlang:timestamp/0
- erlang:unique_integer/0
- erlang:unique_integer/1
- os:system_time/0
- os:system_time/1
and a number of extensions of existing BIFs:
- erlang:monitor(time_offset, clock_service)
- erlang:system_flag(time_offset, finalize)
- erlang:system_info(os_monotonic_time_source)
- erlang:system_info(time_offset)
- erlang:system_info(time_warp_mode)
- erlang:system_info(time_correction)
- erlang:system_info(start_time)
See the "Time and Time Correction in Erlang" chapter of the
ERTS User's Guide for more information.
|
|
and usage
|
|
* sverk/hipe-inline-reserve-trap-frame:
erts: Extend usage of ASM macro to avoid including asm macros in C code
erts: Make hipe_{un}reserve_beam_trap_frame INLINE
|
|
|
|
except the reference counter 'refc', as different callers
have different strategies regarding the lifetime of the binary.
|
|
In the #erlang IRC channel Anthony Ramine once mentioned the idea of
allowing a NIF to use an emulator trap, similar to a BIF trap, to schedule
another NIF for execution. This is exactly how dirty NIFs were implemented
for Erlang/OTP 17.0, so this commit refactors and generalizes that dirty
NIF code to support a new enif_schedule_nif() API function.
The enif_schedule_nif() function allows a long-running NIF to be broken
into separate NIF invocations. The NIF first executes part of the
long-running task, then calls enif_schedule_nif() to schedule a NIF for
later execution to continue the task. Any number of NIFs can be scheduled
in this manner, one after another. Since the emulator regains control
between invocations, this helps avoid problems caused by native code tying
up scheduler threads for too long.
The enif_schedule_nif() function also replaces the original experimental
dirty NIF API. The function takes a flags parameter that a caller can use
to indicate the NIF should be scheduled onto either a dirty CPU scheduler
thread, a dirty I/O scheduler thread, or scheduled as a regular NIF on a
regular scheduler thread. With this change, the original experimental
enif_schedule_dirty_nif(), enif_schedule_dirty_nif_finalizer() and
enif_dirty_nif_finalizer() API functions are no longer needed and have been
removed. Explicit scheduling of a dirty NIF finalization function is no
longer necessary; if an application wants similar functionality, it can
have a dirty NIF just invoke enif_schedule_nif() to schedule a non-dirty
NIF to complete its task.
Lift the restriction that dirty NIFs can't call enif_make_badarg() to raise
an exception. This was a problem with the original dirty NIF API because it
forced developers to get and check all incoming arguments in a regular NIF,
and then schedule the dirty NIF which then had to get all the arguments
again. Now, the argument checking can be done in the dirty NIF and it can
call enif_make_badarg() itself to flag incorrect arguments.
Extend the ErlNifFunc struct with a new flags field that allows NIFs to be
declared as dirty. The default value for this field is 0, indicating a
regular NIF, so it's backwards compatible with all existing statically
initialized ErlNifFunc struct instances, and so such instances require no
code changes. Defining the flags field with a value of
ERL_NIF_DIRTY_JOB_CPU_BOUND indicates that the NIF should execute on a
dirty CPU scheduler thread, or defining it with a value of
ERL_NIF_DIRTY_JOB_IO_BOUND indicates that the NIF should execute on a dirty
I/O scheduler thread. Any other flags field value causes a NIF library
loading error.
Extend the ErlNifEntry struct with a new options field that indicates
whether a NIF library was built with support for optional features such as
dirty NIFs. When a NIF library is loaded, the runtime checks the options
field to ensure compatibility. If a NIF library built with dirty NIF
support is loaded into a runtime that does not support dirty NIFs, and the
library defines one or more ErlNifFunc entries with non-zero flags fields
indicating dirty NIFs, a NIF library loading error results. There is no
error if a NIF library built with dirty NIF support is loaded into a
runtime that does not support dirty NIFs but the library does not have any
dirty NIFs. It is also not an error if a library without dirty NIF support
is loaded into a runtime built with dirty NIF support.
Add documentation and tests for enif_schedule_nif().
|
|
|
|
|
|
* egil/gc-increase-tenure-rate/OTP-11617:
erts: Increase gc tenure rate
|
|
as we don't use it and instead have the feature to disable GC
during trapping BIFs.
|
|
The garbage collector tries to maintain the previous heap block size
during a minor gc, i.e. 'need' is not utilized in determining the size of
the new heap, instead it relies on tenure and garbage to be sufficiently large.
In instances during intense growing with exlusively live data on the heap
coupled with delayed tenure, fullsweeps would be triggered directly after
a minor gc to make room for 'need' since the new heap would be full.
To remedy this, the tenure of terms on the minor heap will always happen
(if it is below the high watermark) instead of every other minor gc.
|
|
* rickard/garbage_collect/OTP-11388:
Parallel check_process_code when code_server purge a module
Functionality for disabling garbage collection
Use asynchronous check_process_code in code_parallel_SUITE
Execution of system tasks in context of another process
Conflicts:
bootstrap/lib/kernel/ebin/hipe_unified_loader.beam
erts/preloaded/ebin/erlang.beam
erts/preloaded/ebin/erts_internal.beam
|
|
The test needs to be false to fail, not true. This was noticed with the
following warning, where marking erl_assert_error as non-returning didn't
help:
beam/erl_gc.c:2772:2: warning: variable 'refc' is used uninitialized whenever switch default is taken [-Wsometimes-uninitialized]
default:
^~~~~~~
beam/erl_gc.c:2775:26: note: uninitialized use occurs here
ERTS_CHK_OFFHEAP_ASSERT(refc >= 1);
^~~~
beam/erl_gc.c:2738:11: note: expanded from macro 'ERTS_CHK_OFFHEAP_ASSERT'
if (!(EXP)) \
^
beam/erl_gc.c:2759:18: note: initialize the variable 'refc' to silence this warning
erts_aint_t refc;
^
= 0
|
|
Being able to disable garbage collection over context
switches vastly simplifies implementation of yielding
native code that builds large or complex data structures
on the heap. This since the heap can be left in an
inconsistent state over the context switch.
|
|
|
|
|
|
Rewrite and extend of Happi's initial work
Extra_root to process structure to enable GC of state - Changed the
process structure to point to a separate struct, the struct also
contains a destructor function to allow for proper cleanup.
Rewrote encode_size_struct and enc_term to have internal versions with
reduction counters which will result in interrupt for later restart
when the counter reaches zero - removed the EWA_STACK from Happis
version and directly save the ESTACK's and WSTACK's in the above
mentioned struct (or array thereof) that are pointed out from the
process structure. The destructor will take care of the deallocation
in case of process death.
Added ESTACK and WSTACK macros to save and restore stack and to change
allocator, which makes the previously mentioned stack-save work.
Rewrote enc_term to not store pointers on the stack, and use one WSTACK
for commands etc and another ESTACK for Eterms - Slightly different than
Happis version to make halfword code simpler.
Rewrote encode_size_struct2 so that it does not store pointers on the
stack, also switched to ESTACK instead of WSTACK, this also handles
halfword correctly.
Added interfaces for chunkwise compression, that are
used from term_to_binary/2 when the compressed option is given.
|
|
|
|
|
|
|
|
Using %p, %bpx and %bex for pointers, UWords and Uint.
|
|
OTP-10415
* sverk/egil/r16/new-alloc-header-scheme/OTP-10273: (42 commits)
erts: Make ll main mbc fit into 2pow size
erts: Clear entire mseg cache upon request
erts: Reduce max heap sizes
tests: Refactor away ?line macro in beam_SUITE
tests: Fix heap_sizes check
tests: Refactor away ?line macro in process_SUITE
tests: Use new correct min_bin_vheap_size in test
erts: Set super alignment (256kb) and limits for sbct (8Mb) and lmbcs (128Mb)
erts: Do not cache segments that are misaligned
erts: Add mseg cache for large sbc segments
erts: Reintroduce mseg options amcbf and rmcbf
erts: Optimize erl_alloc_util.c by substitute MBC_BLK_SZ
erts: Fix bug when allocating size near sbc_threshold
erts: Make gc sizes fit into MB Carrier blocks
erts: Force sbmbc to be disabled in a crude way
erts: Fix new header scheme for win64
erts: Fix mseg cache. Malplaced NULL pointer
erts: Remove unused mseg options amcbf and rmcbf
erts: Use aligned bits as constant in mseg_alloc
erts: Don't let zero be considered a power of two
...
Conflicts:
erts/emulator/test/process_SUITE.erl
|
|
Reduces max heap sizes to max addressable memory space. In reality this could be
even less since we need at least twice as much adressable memory to do a garbage
collection.
The rest of the system also uses memory thus further constraining heap memory
space and in the 64 bit case we really only have 48 bits mappable memory.
|
