| Age | Commit message (Collapse) | Author |
|---|
|
* vinoski/ds-enif-send:
enable enif_send to work from a dirty scheduler
|
|
* bjorn/erts/fix-lingering-tracer:
Teach the call_time trace to notice when the trace dies (non-SMP system)
|
|
The call_time trace is a special kind of tracing that requires
a tracer process just like ordinary call trace, but it never
actually sends anything to the tracer. It merely use the existence
of a trace process (and call trace flags) as an indication that
call_time tracing is active for the process.
If the tracer dies in a non-SMP run-time system, processes with
call_time tracing would not notice that the tracer had
died. Furthermore, if the set_on_spawn flag was active, the dead
tracer could be propagaged to newly spawned processes.
Before accumulating trace information in a non-SMP system, always
validate the tracer process. (In an SMP system, a reference to a
dead tracer will be cleared away each time a process is scheduled.)
While we could put all of the new code beam_bp.c, we have chosen to
make a function call from beam_bp.c to a function in erl_trace.c for
clarity's sake and to ease further maintenance. In the future, we
might want to handle tracing in more similar ways in the SMP and
non-SMP system.
|
|
* cobusc/correct_comment:
Fix comment that differs from code
OTP-11793
|
|
|
|
This is just a preparation to allow detection of older nodes
that do not understand maps (R16 and older).
|
|
to be: 116,Arity, K1,V1,K2,V2,...,Kn,Vn
instead of: 116,Arity, K1,K2,...,Kn, V1,V2,....,Vn
We think this will be better for future internal map structures
like HAMT. Would be bad if we need to iterate twice over HAMT
in term_to_binary, one for keys and one for values.
|
|
|
|
The comment in the code state that the tree is traversed to the left,
when in fact it is traversed to the right.
|
|
The expression,
#{} =:= M
where M was any Map, would always result in 'true'.
This commit fixes this issue by first comparing sizes for
both terms and then checking for size zero.
|
|
* sverk/hipe-disable-gc-bug:
erts: Fix heap overwrite by hipe "trap frames" when GC is disabled
ASSERT that GC is not tried with "need" when GC is disabled
|
|
by trapping BIFs like term_to_binary and binary_to_term.
|
|
* lukas/erts/float_encoding/OTP-11738:
erts: Set default external enc to use new float scheme
|
|
* lukas/ose/master/OTP-11334: (71 commits)
erts: Fix unix efile assert
ose: Use -O2 when building
ose: Expand OSE docs
ose: Add dummy ttsl driver
ose: Cleanup cleanup of mutex selection defines
ose: Polish mmap configure checks
ose: Add ose specific x-compile flags
ose: Updating fd_driver and spawn_driver for OSE
ose: Updating event and signal API for OSE
ose: Cleanup of mutex selection defines
win32: Compile erl_log.exe
ose: Remove uneccesary define
ose: Fix ssl configure test for osx
erts: Fix sys_msg_dispatcher assert
ose: Fix broken doc links
ose: Thread priorities configurable from lmconf
ose: Yielding the cpu is done "the OSE" way
ose: Start using ppdata for tse key
ose: Do not use spinlocks on OSE
ose: Fix support for crypto
...
Conflicts:
lib/crypto/c_src/crypto.c
|
|
|
|
Schedulers is too restrictive. Managed threads should be
able to clean this up.
|
|
The pattern used for getting the priority from the lmconf
is based on the name of the process created. The pattern is:
ERTS_%%PROCESS_NAME%%_PRIO
with the %%PROCESS_NAME%% replaced by the prefix of the process
the priority applies to. eg:
ERTS_SCHEDULER_PRIO=24
applies to processes with name SCHEDULER_1, SCHEDULER_2 etc.
|
|
Now schedulers 2..N make sure to wake sched 1 if they
find that all io has been consumed and sched 1 is
sleeping. Before sched 1 was spinning in sys_schedule
waiting for sched 2..N to finish consuming io jobs
|
|
This is needed on OSs that do not do round robin scheduling
of threads.
|
|
This new API has less impact on the check_io code and
also removes the callback from ErlDrvEntry. The downside
is that you have to give the resolve function when creating
each event.
Also the mode if the resolve was removed as this mimics the
win32 code and decreases complexity.
|
|
This simplified debugging on OSE and also limits the number of ppdata
keys that are created when beam is restarted.
|
|
Make it possible to change the number of write_concurrency locks to use. This
is usefull to change when you for some reason want to use more/less locks
per write_concurrency ets table. eg. OSs with a limit on how many mutexes
can exist at once.
|
|
This is needed because a file has to be opened and operated on
in the same process at all times. Using async threads guarantee this.
|
|
|
|
This port has support for both non-smp and smp.
It contains a new way to do io checking in which erts_poll_wait
receives the payload of the polled entity. This has implications
for all linked-in drivers.
|
|
* egil/maps/fix-cerl-inlining:
compiler: Update map_SUITE to handle inlining
erts: Maps src instructions can't be literals
compiler: Fix map inlining
compiler: Add variable coverage of map in cerl
|
|
* rickard/assumed-cacheline-size/OTP-11742:
Introduce configure option --with-assumed-cache-line-size=SIZE
|
|
* vinoski/ds2:
further enhancements for dirty schedulers
allow optional whitespace in dirty scheduler erl options
|
|
Add support for setting the number of dirty CPU schedulers online via
erlang:system_flag/2. Assuming the emulator is built with dirty schedulers
enabled, the number of dirty CPU schedulers online may not be less than 1,
nor greater than the number of dirty CPU schedulers available, nor greater
than the number of normal schedulers online. Dirty CPU scheduler threads
that are taken offline via system_flag/2 are suspended. The number of dirty
CPU schedulers online may be adjusted independently of the number of normal
schedulers online, but if system_flag/2 is used to set the number of normal
schedulers online to a value less than the current number of normal
schedulers online, the number of dirty CPU schedulers online is decreased
proportionally. Likewise, if the number of normal schedulers online is
increased, the number of dirty CPU schedulers online is increased
proportionally. For example, if 8 normal schedulers and 4 dirty CPU
schedulers are online, and system_flag/2 is called to set the number of
normal schedulers online to 4, the number of dirty CPU schedulers online is
also decreased by half, to 2. Subsequently setting the number of normal
schedulers online back to 8 also sets the number of dirty CPU schedulers
online back to 4. Augment the system_flag/2 documentation in the erlang man
page to explain this relationship between schedulers_online and
dirty_cpu_schedulers_online.
Also ensure that all dirty CPU and I/O schedulers are suspended when
multi-scheduling is blocked via system_flag/2, and brought back online when
multi-scheduling is unblocked.
Add Rickard Green's rewritten check_enqueue_in_prio_queue() function that
inspects process state more thoroughly to determine if to enqueue it and if
so on what queue, including dirty queues when appropriate.
Make sure dirty NIF jobs do not trigger erlang:system_monitor long_schedule
messages.
Add more dirty scheduler testing to the scheduler test suite.
Remove the erts_no_dirty_cpu_schedulers_online global variable, since it's
no longer needed.
Execute dirty NIFs on a normal scheduler thread while multi-scheduling
blocking is in effect. Evacuate any dirty jobs residing in the dirty run
queues over to a normal run queue when multi-scheduling is blocked.
Allow dirty schedulers to execute aux work.
Set the dirty run queues halt_in_progress flag when halting the normal
schedulers.
Change dirty scheduler numbers to a structure including both scheduler
number and type, either dirty CPU or dirty I/O. Add some assertions to
ensure that dirty CPU schedulers operate only on dirty CPU scheduler
process flags, and the same for dirty I/O schedulers.
|
|
This change was triggered by the OSE float printing function
not working exactly the same way as linux/win32. But it is
also a good one in general as it cuts size in more than half
for floats.
|
|
|
|
Move src to a register if it is a literal.
|
|
* bjorn/erts/restore-cmp-arguments:
beam/utils.c: Change back cmp() to only take two arguments
|
|
* egil/compiler/maps-get_map_elements:
compiler: Strengthen Maps compile tests
compiler: Remove dead warning
erts: Fix erts_debug:disassemble/1
compiler: Transform list of Args to exact literal type
compiler: Test Maps aliasing
compiler: Use aliasing in map pair patterns
compiler: Check literal order in beam_validator
erts: Introduce new instructions for combined key fetches
compiler: Change map instructions for fetching values
|
|
* rickard/otp_version/OTP-11615:
Misc adjustments of OTP version
|
|
|
|
Now handles map instructions correctly.
|
|
OTP-11726
* sverk/ets-all-race:
erts: Fix faulty asserts in erts_sys_aligned_alloc/free
erts: Fix harmless (?) typo in beam_load.c
erts: Fix race bug in ets:all/0
|
|
|
|
causing recently created/deleted tables to not be included/excluded.
|
|
OTP-11722
OTP-11724
* sverk/crypto/hmac-context-reuse-bug:
crypto: Fix bug when using old hmac context
erts: Fix NIF bug when load/upgrade fails after enif_open_resource_type
Conflicts:
erts/emulator/test/nif_SUITE.erl
|
|
|
|
|
|
Commit d5c238473b9cec819d93faaef4ccc00ddb60465f added a third
argument to the cmp() function, but did not change the code generation
of native code that called cmp().
Name the new functions with an "erts_" prefix, and reinstate the
old cmp() function with two arguments.
We don't have to keep a cmp() function in the halfword emulator,
since HiPE does not support the halfword emulator.
|
|
The +SDcpu, +SDPcpu, and +SDio options did not properly handle having their
arguments immediately following them without intervening whitespace,
e.g. +SDio20 was treated as an error. Fix all the dirty scheduler command
line options so they handle optional whitespace between them and their
associated arguments.
|
|
* rickard/default_acul_mod/OTP-11604:
Fix +Mea config
|
|
Exact updates on empty Maps must fail directly.
|
|
..has been successfully called.
Opened resource types (created or taken-over) were left "hanging"
leading both to memory leakage and other more strange and serious behavior.
Now a proper rollback is done.
|
|
Map source may be anything, not only registers.
|
|
'maps_head' was not restored when yielding. Risk for crash increases with
size and number of maps in term.
|
