| Age | Commit message (Collapse) | Author |
|---|
|
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.
|
|
|
|
The 64-bit atomic ops API is implemented by
* native word size atomic ops on 64-bit architectures, and
* native double word size atomic ops on 32-bit architectures
when available. When native double word size atomic is not
available, the fallback using modification counters is
used.
|
|
* sverk/port_get_data-race/OTP-12208:
erts: Fix port data memory allocation bug
|
|
* rickard/maint-17/activate-runq/OTP-12287:
Do not let non-empty run-queue prevent activation of scheduler
|
|
OTP-12323
* sverk/cpool-search-improvement:
erts: Add internal docs for super carrier and new cpool search.
erts: Fix bug causing mbc to be deleted from cpool before it was inserted
erts: Fix bug causing mbc removed from cpool to be used at pool entrance
erts: Add pooled_list and traitor_list
|
|
for non-immediate port data >= sizeof(Eterm)*2 words.
|
|
OTP-12300
* vinoski/dirty-nif-return-gc:
Fix gc-related problem with dirty NIFs
|
|
OTP-12298
* lemenkov/expose_nif_version:
Expose NIF version
|
|
* vinoski/fix-smp-disable-with-ds/OTP-12295:
Fix "-smp disable" for emulator with dirty schedulers
|
|
* lukas/erts/fix_undefined_behaviour/OTP-12290:
erts: Fix ub in list_to_integer and bignum div
|
|
rickard/maint-17/activate-runq/OTP-12287
* rickard/activate-runq/OTP-12287:
Do not let non-empty run-queue prevent activation of scheduler
Conflicts:
erts/emulator/beam/erl_process.c
|
|
Conflicts:
erts/emulator/beam/erl_process.c
|
|
Ensure that the return value from a dirty NIF call is made part of the GC
rootset. Add a new regression test to nif_SUITE.
Thanks to Daniel Goertzen for reporting the error and providing a test
case, and to Sverker Eriksson for making test case reproducible and finding
the fix.
|
|
* sverk/beam-ranges-bug/OTP-12273:
erts: Fix bug in beam_ranges
|
|
This patch allows checking for NIF API version in a way similar to
driver version. E.g. by calling erlang:system_info(nif_version).
Signed-off-by: Peter Lemenkov <[email protected]>
|
|
Symptom: VM on OSX (darwin11.4.2) with +Meamin running sasl tests,
crashing when init:reboot() does erlang:purge_module(installer).
Problem: Off-by-one bug in beam_ranges:find_range, returning the wrong
range if the 'end' of one module is the 'start' of the next. This is only
possible if using sys_alloc (+Meamin) as our own allocators always put
block headers between allocated payload data.
|
|
* sverk/cpool_fetch-dc_list-bug/OTP-12249:
erts: Fix bug when delayed deallocated carrier is reused by cpool_fetch
|
|
* sv/isfinite/OTP-12268:
Use isfinite() instead of finite() when available
|
|
OS X Mavericks builds result in a number of warnings about finite() being
deprecated, like these:
beam/erl_arith.c:451:7: warning: 'finite' is deprecated: first deprecated in OS X 10.9 [-Wdeprecated-declarations]
ERTS_FP_ERROR(p, f1.fd, goto badarith);
^
sys/unix/erl_unix_sys.h:319:33: note: expanded from macro 'ERTS_FP_ERROR'
^
sys/unix/erl_unix_sys.h:244:51: note: expanded from macro '__ERTS_FP_ERROR'
^
/usr/include/math.h:718:12: note: 'finite' has been explicitly marked deprecated here
extern int finite(double) __OSX_AVAILABLE_BUT_DEPRECATED(__MAC_10_0, __MAC_10_9, __IPHONE_NA, __IPHONE_NA);
Add checks to use isfinite() instead of finite() where available. Verified
on OS X Mavericks 10.9.5 and Ubuntu 12.04.
|
|
OTP-12266
* vinoski/erl-nif-init:
Fix missing field initializer in ERL_NIF_INIT macro
|
|
Running "erl -smp disable" on an emulator built with dirty scheduler
support caused problems such as segmentation violations and emulator status
line outputs containing garbage. For example:
$ erl -smp disable
Segmentation fault (core dumped)
and:
$ erl -smp disable
Erlang/OTP 17 [DEVELOPMENT] [erts-6.2] [source] [64-bit] [ds:10:4297895689:4299948152] [async-threads:280]
This problem also caused the emulator smoke_test_SUITE to hit these same
problems if run in an emulator started with the "-smp disable" option.
Fix this segmentation violation by ensuring that dirty scheduler
information is printed in the status line only when the emulator is
compiled with ERTS_SMP enabled.
With this fix in place, the smoke_test_SUITE now passes when the "-smp
disable" option is used, and the emulator status line prints correctly for
both "-smp enable" and "-smp disable":
$ erl -smp enable
Erlang/OTP 17 [DEVELOPMENT] [erts-6.2] [source] [64-bit] [smp:8:8] [ds:8:8:10] [async-threads:10] [kernel-poll:false]
and:
$ erl -smp disable
Erlang/OTP 17 [DEVELOPMENT] [erts-6.2] [source] [64-bit] [async-threads:10] [kernel-poll:false]
|
|
* rickard/maint-17/eager-check-io/OTP-12117:
No eager check I/O on OSE
Introduce support for eager check I/O scheduling
|
|
Set IN_POOL flag _after_ mbc has been actually inserted.
Earlier it did not matter if IN_POOL was set early as it was not possible
to find a carrier before is was fully inserted. Now when searching
pooled_list and traitor_list we must make sure a found carrier has been
fully inserted into cpool before removing it.
|
|
Clear both IN_POOL and BUSY flags when empty carrier is removed
is removed from pool to be destroyed.
Earlier it was enough to leave BUSY flag set but now with pooled_list
we must clear IN_POOL to avoid using it as cpool_entrance in cpool_fetch().
|
|
|
|
The delayed dealloc queue destroyes one word but cpool_fetch() is expected
to return healthy carriers.
So we restore that overwritten word with a little bit of hackish code.
|
|
|
|
In the ERL_NIF_INIT macro, add a missing static initializer for the new
options field which was added to the ErlNifEntry struct for 17.3. This fix
prevents compile-time errors in NIFs when the compiler is instructed to
treat missing field initializers as errors.
|
|
Always update prt->data with atomic xchg-op.
Check for NULL data to detect racing port terminator.
Use NULL, as THE_NON_VALUE can be a valid pointer on debug VM.
|
|
hsize field was not set leading to VM crash
|
|
|
|
rickard/maint-17/eager-check-io/OTP-12117
* rickard/eager-check-io/OTP-12117:
Introduce support for eager check I/O scheduling
Conflicts:
erts/emulator/beam/erl_bif_info.c
erts/emulator/beam/erl_init.c
erts/emulator/beam/erl_process.c
erts/preloaded/ebin/erlang.beam
|
|
|
|
|
|
and add 'dirty_scheduler_support' to ErlNifSysInfo
|
|
enif_schedule_nif() put LAST of the unconditional functions to keep the
order which is vital for ABI compatibility on Windows.
The conditional dirty scheduler stuff moved down at the end of the list
to keep them out of the way. We don't want them mess things up then they
become unconditional some day.
|
|
OTP-12128
* vinoski/enif-schedule-nif:
Fix leak of NIF exports
Use separate allocation type for NIF export
|
|
|
|
|
|
* sverk/term_to_binary-realloc-size/OTP-12141:
erts: Fix bug in term_to_binary that reallocates binary with wrong size
|
|
* sverk/MIN_SMALL-to-integer:
erts: Correct conversion of MIN_SMALL numeral to fixnum
OTP-12140
|
|
list_to_integer and binary_to_integer returned un-normalized bignum for
-134217728 on 32-bit and -576460752303423488 on 64-bit.
Thanks to Jesper Louis Andersen, Mikael Pettersson
and Anthony Ramine for report, initial patch and optimization suggestion.
|
|
* sverk/nif-inspect-copy-bug/OTP-9828:
erts: Fix bug with enif_make_copy reallocating writable binary
Conflicts:
erts/emulator/test/nif_SUITE.erl
|
|
that could invalidate a pointer received from an earlier
call to enif_inspect_binary.
Solution: Emasculate writable binary at enif_inspect_binary. There are
room for optimizations here as we now do an unconditional emasculation
even though enif_make_copy is not called later in the NIF.
|
|
|
|
* rickard/runnable-trace-ooo-bug/OTP-12105:
Fix busy_port_SUITE:io_to_busy test-case
Ensure "runnable port" trace messages are not sent out of order
Ensure "runnable proc" trace messages are not sent out of order
|
|
* rickard/port-emigrate-bug/maint/OTP-12084:
Verify run-queue asserts
Fix emigrate bug in erts_port_task_schedule()
|
|
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().
|
|
|
