| Age | Commit message (Collapse) | Author |
|---|
|
Previously, all ETS tables used centralized counter variables to keep
track of the number of items stored and the amount of memory
consumed. These counters can cause scalability problems (especially on
big NUMA systems). This commit adds an implementation of a
decentralized counter and modifies the implementation of ETS so that
ETS tables of type ordered_set with write_concurrency enabled use the
decentralized counter. [Experiments][1] indicate that this change
substantially improves the scalability of ETS ordered_set tables with
write_concurrency enabled in scenarios with frequent `ets:insert/2`
and `ets:delete/2` calls.
The new counter is implemented in the module erts_flxctr
(`erts_flxctr.h` and `erts_flxctr.c`). The module has the suffix
flxctr as it contains the implementation of a flexible counter (i.e.,
counter instances can be configured to be either centralized or
decentralized). Counters that are configured to be centralized are
implemented with a single counter variable which is modified with
atomic operations. Decentralized counters are spread over several
cache lines (how many can be configured with the parameter
`+dcg`). The scheduler threads are mapped to cache lines so that there
is no single point of contention when decentralized counters are
updated. The thread progress functionality of the Erlang VM is
utilized to implement support for linearizable snapshots of
decentralized counters. The snapshot functionality is used by the
`ets:info/1` and `ets:info/2` functions.
[1]: http://winsh.me/ets_catree_benchmark/flxctr_res.html
|
|
This reverts commit f4c121b1d98bf3db7e6eecbb9fb5b292f2bc3bb0.
|
|
|
|
|
|
|
|
* maint:
Remove an unused variable
Spawn prim_file helper as a system process
|
|
|
|
The guard BIF `length/1` would calculate the length of the list in one
go without yielding, even if the list was were long. To make it even
worse, the call to `length/1` would only cost a single reduction.
This commit reimplements `length/1` so that it eats a number of
reductions proportional to the length of the list, and yields if the
available reductions run out.
|
|
Conflicts:
erts/emulator/beam/erl_process.c
|
|
* lukas/OTP-21.1.1/scheduler_pollset/OTP-15475:
erts: Move fds with active true behaviour to own pollset
erts: Fix lists_member_2 reduction count
erts: Allow code_model_small to be set in xcomp setting
erts: Implement delay_send using timer instead of poll
erts: Optimize driver_set_timer(0) to fire at once
erts: Optimize the inet driver multi timers for one timer
erts: Move all inet tcp CONNECTED timers to multi timer
erts: Add erts_io_notify_port_task_executed to check_io msacc state
erts: Add pre-alloc to ALLOC msacc state
erts: Make thr prgr wakeup current or sched 1
erts: Pass thread progress data where possible
|
|
* maint:
Implement a tab for persistent terms in crashdump viewer
Add tests of persistent terms for crashdump_viewer
Add a persistent term storage
Refactor releasing of literals
Extend the sharing-preserving routines to optionally copy literals
Conflicts:
erts/emulator/Makefile.in
erts/emulator/beam/erl_process_dump.c
erts/preloaded/ebin/erts_internal.beam
erts/preloaded/ebin/init.beam
lib/sasl/src/systools_make.erl
|
|
Persistent terms are useful for storing Erlang terms that are never
or infrequently updated. They have the following advantages:
* Constant time access. A persistent term is not copied when it is
looked up. The constant factor is lower than for ETS, and no locks
are taken when looking up a term.
* Persistent terms are not copied in garbage collections.
* There is only ever one copy of a persistent term (until it is
deleted). That makes them useful for storing configuration data
that needs to be easily accessible by all processes.
Persistent terms have the following drawbacks:
* Updates are expensive. The hash table holding the keys for the
persistent terms are updated whenever a persistent term is added,
updated or deleted.
* Updating or deleting a persistent term triggers a "global GC", which
will schedule a heap scan of all processes to search the heap of all
processes for the deleted term. If a process still holds a reference
to the deleted term, the process will be garbage collected and the
term copied to the heap of the process. This global GC can make the
system less responsive for some time.
Three BIFs (implemented in C in the emulator) is the entire
interface to the persistent term functionality:
* put(Key, Value) to store a persistent term.
* get(Key) to look up a persistent term.
* erase(Key) to delete a persistent term.
There are also two additional BIFs to obtain information about
persistent terms:
* info() to return a map with information about persistent terms.
* get() to return a list of a {Key,Value} tuples for all persistent
terms. (The values are not copied.)
|
|
The poll thread does a lot of waking up and then going
back to sleep. A large part of the waking up is managing
thread progress and a large part of that was using thread
specific data to get the thread progress data pointer.
With this refactor the tpd is passed to each of the functions
which greatly decreases the number of ethr_get_tsd calls
which in turn halves the CPU usage of the poller thread in
certain scenarios.
|
|
* richcarl/erts/erl_init-cleanup/OTP-15336:
sasl: Order systools_make:preloaded modules alphabetically
Update preloaded modules
Move calling on_load for preloaded modules to erl_init
Make erl_init.c pass the boot module to erl_init.beam
Remove obsolete comment text
Remove undocumented and unused -# display_items emulator option
Remove broken and undocumented boot function emulator option
Replace remaining references to otp_ring0 with erl_init
Drop otp_ring0, using erl_init instead
Update preloaded modules
Add erl_init module
Conflicts:
erts/emulator/beam/erl_init.c
erts/preloaded/ebin/erl_prim_loader.beam
erts/preloaded/ebin/erl_tracer.beam
erts/preloaded/ebin/erlang.beam
erts/preloaded/ebin/erts_code_purger.beam
erts/preloaded/ebin/erts_dirty_process_signal_handler.beam
erts/preloaded/ebin/erts_internal.beam
erts/preloaded/ebin/erts_literal_area_collector.beam
erts/preloaded/ebin/init.beam
erts/preloaded/ebin/otp_ring0.beam
erts/preloaded/ebin/prim_buffer.beam
erts/preloaded/ebin/prim_eval.beam
erts/preloaded/ebin/prim_file.beam
erts/preloaded/ebin/prim_inet.beam
erts/preloaded/ebin/prim_zip.beam
erts/preloaded/ebin/zlib.beam
|
|
This introduces a similar optimisation for normal funs
to what was introduced for external funs in #1725.
It is possible to allocate the fun as a literal, if it does not capture
the environment (i.e. it does not close over any variables).
Unfortunately it's not possible to do this in the compiler due to
problems with representation of such functions in the `.beam` files.
Fortunately, we can do this in the loader.
Simple evaluation shows that functions that don't capture the
enviornment consistute over 60% of all funs in the source code of
Erlang/OTP itself.
The only downside is that we lose a meningful value in the `pid` field
of the fun. The goal of this field, beyond debugging, was to be
able to identify the original node of a function. To be able to still do
this, the functions that are created in the loader are assigned the init
pid as the creator.
To solve issues with staryp, initially set the `erts_init_process_id`
to `ERTS_INVALID_PID` and skip the described optimisation if the value
is still uninitialised.
|
|
|
|
|
|
|
|
|
|
|
|
This commit replaces the old memory instrumentation with a new
implementation that scans carriers instead of wrapping
erts_alloc/erts_free. The old implementation could not extract
information without halting the emulator, had considerable runtime
overhead, and the memory maps it produced were noisy and lacked
critical information.
Since the new implementation walks through existing data structures
there's no longer a need to start the emulator with special flags to
get information about carrier utilization/fragmentation. Memory
fragmentation is also easier to diagnose as it's presented on a
per-carrier basis which eliminates the need to account for "holes"
between mmap segments.
To help track allocations, each allocation can now be tagged with
what it is and who allocated it at the cost of one extra word per
allocation. This is controlled on a per-allocator basis with the
+M<S>atags option, and is enabled by default for binary_alloc and
driver_alloc (which is also used by NIFs).
|
|
Sharing these settings for all schedulers can degrade performance,
so it makes sense to be able to configure them separately.
This also changes the default busy-wait time to "short" for both
kinds of dirty schedulers.
|
|
* john/erts/async-thread-defaults/OTP-14928:
Change default async thread count to 1
|
|
All uses of async threads in the built-in drivers have been
replaced with dirty IO, so it no longer makes sense to leave the
default at 10.
|
|
Communication between Erlang processes has conceptually always been
performed through asynchronous signaling. The runtime system
implementation has however previously preformed most operation
synchronously. In a system with only one true thread of execution, this
is not problematic (often the opposite). In a system with multiple threads
of execution (as current runtime system implementation with SMP support)
it becomes problematic. This since it often involves locking of structures
when updating them which in turn cause resource contention. Utilizing
true asynchronous communication often avoids these resource contention
issues.
The case that triggered this change was contention on the link lock due
to frequent updates of the monitor trees during communication with a
frequently used server. The signal order delivery guarantees of the
language makes it hard to change the implementation of only some signals
to use true asynchronous signaling. Therefore the implementations
of (almost) all signals have been changed.
Currently the following signals have been implemented as true
asynchronous signals:
- Message signals
- Exit signals
- Monitor signals
- Demonitor signals
- Monitor triggered signals (DOWN, CHANGE, etc)
- Link signals
- Unlink signals
- Group leader signals
All of the above already defined as asynchronous signals in the
language. The implementation of messages signals was quite
asynchronous to begin with, but had quite strict delivery constraints
due to the ordering guarantees of signals between a pair of processes.
The previously used message queue partitioned into two halves has been
replaced by a more general signal queue partitioned into three parts
that service all kinds of signals. More details regarding the signal
queue can be found in comments in the erl_proc_sig_queue.h file.
The monitor and link implementations have also been completely replaced
in order to fit the new asynchronous signaling implementation as good
as possible. More details regarding the new monitor and link
implementations can be found in the erl_monitor_link.h file.
|
|
* john/erts/assert-on-memcpy-memset-etc:
Always use sys_memcpy/cmp/etc instead of plain memcpy/cmp/etc
Check the arguments to sys_memcpy and friends
|
|
|
|
This is a fix for a bug introduced in 22cde2bda
|
|
The reduction cost of sending messages is now constant and will no
longer scale according to the length of the receiving process'
message queue.
|
|
putenv(3) and friends aren't thread-safe regardless of how you slice
it; a global lock around all environment operations (like before)
keeps things safe as far as our own operations go, but we have
absolutely no control over what libc or a library dragged in by a
driver/NIF does -- they're free to call getenv(3) or putenv(3)
without honoring our lock.
This commit solves this by setting up an "emulated" environment which
can't be touched without going through our interfaces. Third-party
libraries can still shoot themselves in the foot but benign uses of
os:putenv/2 will no longer risk crashing the emulator.
|
|
It is not longer relevant when using the poll thread
|
|
|
|
|
|
|
|
This refactor was done using the unifdef tool like this:
for file in $(find erts/ -name *.[ch]); do unifdef -t -f defile -o $file $file; done
where defile contained:
#define ERTS_SMP 1
#define USE_THREADS 1
#define DDLL_SMP 1
#define ERTS_HAVE_SMP_EMU 1
#define SMP 1
#define ERL_BITS_REENTRANT 1
#define ERTS_USE_ASYNC_READY_Q 1
#define FDBLOCK 1
#undef ERTS_POLL_NEED_ASYNC_INTERRUPT_SUPPORT
#define ERTS_POLL_ASYNC_INTERRUPT_SUPPORT 0
#define ERTS_POLL_USE_WAKEUP_PIPE 1
#define ERTS_POLL_USE_UPDATE_REQUESTS_QUEUE 1
#undef ERTS_HAVE_PLAIN_EMU
#undef ERTS_SIGNAL_STATE
|
|
* john/erts/runtime-lcnt:
Document rt_mask and add warnings about copy_save
Add an emulator test suite for lock counting
Break erts_debug:lock_counters/1 into separate BIFs
Allow toggling lock counting at runtime
Move lock flags to a common header
Enable register_SUITE for lcnt builds
Enable lcnt smoke test on all builds that have lcnt enabled
Make lock counter info independent of the locks being counted
OTP-14412
OTP-13170
OTP-14413
|
|
The implementation is still hidden behind ERTS_ENABLE_LOCK_COUNT, and
all categories are still enabled by default, but the actual counting can be
toggled at will.
OTP-13170
|
|
|
|
erts: Remove old unused functions
|
|
The functions have been found using: https://github.com/caolanm/callcatcher
|
|
|
|
OTP-14380
* rickard/ds-stack-size:
Suggested stack size options for dirty schedulers
|
|
|
|
|
|
|
|
OTP-14331
* rickard/pcre-8.40:
Update documentation
Update README.pcre_update.md
Stack guard for PCRE
Adjust for incompatibility between PCRE 8.40 and perl 5.22.1
Generate re replacement and split tests with perl vsn 5.22.1
Fix re_SUITE:pcre_compile_workspace_overflow/1
Skip line with lockout of modifiers in PCRE tests
Update tests for PCRE version 8.40
Update PCRE to version 8.40
Conflicts:
erts/emulator/beam/beam_debug.c
|
|
|
|
|
|
|
|
Magic references are *intentionally* indistinguishable from ordinary
references for the Erlang software. Magic references do not change
the language, and are intended as a pure runtime internal optimization.
An ordinary reference is typically used as a key in some table. A
magic reference has a direct pointer to a reference counted magic
binary. This makes it possible to implement various things without
having to do lookups in a table, but instead access the data directly.
Besides very fast lookups this can also improve scalability by
removing a potentially contended table. A couple of examples of
planned future usage of magic references are ETS table identifiers,
and BIF timer identifiers.
Besides future optimizations using magic references it should also
be possible to replace the exposed magic binary cludge with magic
references. That is, magic binaries that are exposed as empty
binaries to the Erlang software.
|
