| Age | Commit message (Collapse) | Author |
|---|
|
Symptom: open_port() succeeds, port program never starts but instead
seem to terminate with exit_status 150 (EINVAL+128).
When: If the command payload buffer (including the entire environment)
for the port program is between 1 to 4 bytes above the pipe capacity,
which is 65536 bytes on newer Linux, but can be as low as 4096 on older
32-bit Linux.
Since: OTP-19.0-rc1.
|
|
* rickard/mach-clock-fix/OTP-13686:
Fix mach clock usage on Sierra
|
|
|
|
|
|
|
|
A fix for running 19.0-rc2 on FreeBSD with HiPE enabled.
* erl_mmap.h: disregard MAP_NORESERVE for FreeBSD
MAP_NORESERVE is undefined in FreeBSD 10.x and later.
This is to enable 64bit HiPE experimentally on FreeBSD.
Note that on FreeBSD MAP_NORESERVE was "never implemented"
even before 11.x (and the flag does not exist in /usr/src/sys/vm/mmap.c
of 10.3-STABLE r301478 either), and HiPE was working on OTP 18.3.3,
so mandating MAP_NORESERVE on FreeBSD might not be needed.
See the following message on how MAP_NORESERVE was treated on FreeBSD:
<http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20150202/122958.html>
* erl_mmap.c: disable MAP_NORESERVE for FreeBSD
* See also <https://github.com/erlang/otp/pull/925/commits/b02ce940c8738785ad018c9f758ca0d05d256bbb>
|
|
Similar bug that was fixed for unix in 7bbb207b30360c60fb99653.
|
|
|
|
and make erts_mmap unavailable at compile time
if not supported.
|
|
by (only) moving around stuff in the file.
in order to make it easier to ifdef away
all erts_mmap_* if not supported.
|
|
will now return
[{instance,0,[{segments_size,9961472,9961472,11010048}]},
{instance,1,[{segments_size,6291456,6291456,6815744}]},
{instance,2,[{segments_size,524288,524288,786432}]},
{instance,3,[{segments_size,1048576,1048576,1835008}]},
{instance,4,[{segments_size,0,0,262144}]}]
and not just empty lists.
|
|
from {allocator,mseg_alloc} to {allocator,erts_mmap}
|
|
* mikpe/fp-exceptions-cleanups/PR-1019/OTP-13531:
clean up FP exception code in sys_float.c
|
|
When FP exceptions are used, matherr() forces a fake FP exception,
which is interpreted as an error by the checking code after a math
routine call. This is wrong for several reasons:
- it's not necessary for error checking: when FP exceptions aren't
used, matherr() is a stub and no information is derived from it
being called
- it's not necessary for FPU maintenance: the FPU only needs to be
reprogrammed after an actual FP exception
- it causes false negatives: matherr() may be called even though
Erlang doesn't consider the case to be an error (exp() and pow()
underflows on Solaris for instance); with FP exceptions enabled
they are incorrectly considered errors
The fix is to remove all FP exception related code from matherr().
|
|
|
|
This performs a number of cleanups in the FP exception code:
- inline the body of unmask_fpe_conditional() in its only caller,
then delete the duplicated and identical definitions of it
- start the big processor-specific block with a comment describing
the two functions that must be defined, then delete redundant
comments at all the mask_*() functions
- add a comment before fpe_sig_action() explaining exactly what
processor-specific action is required of it
- flatten #ifdef nesting in fpe_sig_action()
- move common code in the x86 unmask_fpe() and erts_restore_fpu()
function into a subroutine
- minor tweaks: drop a redundant L suffix on a 0, add a comment
after an #else, bump the size of a debug buffer,
There should be no change in behaviour from these changes.
|
|
that uses its own super carrier (erts_exec_mmapper)
to guarantee low addressed and executable memory (PROT_EXEC).
Currently only used on x86_64 that needs low memory
for HiPE/AMD64's small code model.
By initializing erts_exec_mapper early we secure
its low memory area before erts_literal_mmapper might
steal it.
|
|
to prepare for hipe native code allocation.
|
|
Make the callbacks more general to be usable for any allocator
that that uses its own ErtsMemMapper.
|
|
Reduce main carrier size
and number of free descriptors.
|
|
* henrik/update-copyrightyear:
update copyright-year
|
|
* egil/erts/tracing-beam-lttng/OTP-10282:
erts: Don't use ratio in carrier lttng tracepoints
Add lttng testcases
erts: Extend erlang:system_info/1 with lttng
Refactor and fix dtrace define in erl_message
erts: Add lttng tracepoints for async pool queue
erts: Add lttng tracepoints for drivers
erts: Add lttng tracepoints for scheduler events
erts: Add lttng tracepoints for memory carriers
erts: Update lttng-wrapper with mfa conversion
erts: Teach lttng to configure and build system
|
|
* driver_event
* driver_flush
* driver_finish
* driver_init
* driver_output
* driver_outputv
* driver_process_exit
* driver_ready_async
* driver_ready_input
* driver_ready_output
* driver_start
* driver_stop
* driver_stop_select
* driver_timeout
|
|
* lukas/erts/inline_tagvaldef/OTP-13440:
erts: inline tag_val_def
|
|
The tag_val_def function was called and multiple switch statements
had to be traversed in term.c, and then a big switch in the calling
code to branch on the term types. By inlining the switches are
merged by the compiler and a lot fewer branches have to be taken.
Benchmarks show that this increases performance of enc_term by as
much as 10%.
|
|
* lukas/erts/fix_scheduler_suspend/ERL-94/PR-978/OTP-13425:
erts: Fix install of suspend handler
Conflicts:
erts/emulator/sys/unix/erl_unix_sys.h
erts/emulator/sys/unix/sys.c
|
|
Floating-point exception support on MacOS X has never been especially
reliable, and has therefore been disabled by default for a long time.
The fpe support is now broken.
Therefore, take out the unnecessary test for modern mcontext in
configure (whatever that means) and the associated code in sys_float.c.
Add #error directives to sys_float.c to make it clear that
fpe is not supported.
It seems to risky to mess with the mess of #ifdef's, so we will
not attempt to remove all fpe support code for MacOS X.
|
|
|
|
* sverk/literal-alloc-polish:
erts: Add emulator flag +MIscs for literal super carrier size
erts: Refactor init of erts_literal_mmapper
erts: Make literal_alloc documented and configurable
|
|
* sverk/master/halt-INT_MIN:
erts: Make erlang:halt() accept bignums as Status
erts: Change erl_exit into erts_exit
kernel: Remove calls to erl_exit
|
|
OTP-13251
* sverk/halt-INT_MIN:
erts: Make erlang:halt() accept bignums as Status
erts: Change erl_exit into erts_exit
kernel: Remove calls to erl_exit
|
|
|
|
|
|
This is mostly a pure refactoring.
Except for the buggy cases when calling erlang:halt() with a positive
integer in the range -(INT_MIN+2) to -INT_MIN that got confused with
ERTS_ABORT_EXIT, ERTS_DUMP_EXIT and ERTS_INTR_EXIT.
Outcome OLD erl_exit(n, ) NEW erts_exit(n, )
------- ------------------- -------------------------------------------
exit(Status) n = -Status <= 0 n = Status >= 0
crashdump+abort n > 0, ignore n n = ERTS_ERROR_EXIT < 0
The outcome of the old ERTS_ABORT_EXIT, ERTS_INTR_EXIT and
ERTS_DUMP_EXIT are the same as before (even though their values have
changed).
|
|
|
|
Except it cannot be disabled and cannot be multi-threaded.
The bit-vector 'erts_literal_vspace_map' on 32-bit is currently only
protected by the literal allocator mutex. We could allow multiple
instances on 64-bit (I think), but what would be the point?
|
|
This commit makes sure to setup the suspend handler
to matter what +B option is given at the command line.
|
|
|
|
perf counter is now part of the function pointer interface
and also the function returns the value instead of writing
to a memory buffer.
|
|
|
|
|
|
Microstate accounting is a way to track which state the
different threads within ERTS are in. The main usage area
is to pin point performance bottlenecks by checking which
states the threads are in and then from there figuring out
why and where to optimize.
Since checking whether microstate accounting is on or off is
relatively expensive if done in a short loop only a few of the
states are enabled by default and more states can be enabled
through configure.
I've done some benchmarking and the overhead with it turned off
is not noticible and with it on it is a fraction of a percent.
If you enable the extra states, depending on the benchmark,
the ovehead when turned off is about 1% and when turned on
somewhere inbetween 5-15%.
OTP-12345
|
|
The perf_counter is a very very cheap and high resolution timer
that can be used to timestamp system events. It does not have
monoticity guarantees, but should on most OS's expose a monotonous
time.
A special instruction has been created for this counter to further
speed up fetching it.
OTP-12908
|
|
* maint:
Introduce time management in native APIs
Introduce time warp safe replacement for safe_fixed option
Introduce time warp safe trace timestamp formats
Conflicts:
erts/emulator/beam/erl_bif_trace.c
erts/emulator/beam/erl_driver.h
erts/emulator/beam/erl_nif.h
erts/emulator/beam/erl_trace.c
erts/preloaded/ebin/erlang.beam
|
|
|
|
* lukas/erts/forker: (28 commits)
erts: Never abort in the forked child
erts: Mend ASSERT makro for erl_child_setup
erts: Allow enomem failures in port_SUITE
erts: iter_port sleep longer on freebsd
erts: Allow one dangling fd if there is a gethost port
erts: Only use forker StackAck on freebsd
erts: It is not possible to exit the forker driver
erts: Add forker StartAck for port start flowcontrol
erts: Fix large open_port arg segfault for win32
erts: Fix memory leak at async open port
kernel: Remove cmd server for unix os:cmd
erts: Add testcase for huge port environment
erts: Move os_pid to port hash to child setup
erts: Handle all EINTR and EAGAIN cases in child setup
erts: Make child_setup work with large environments
erts: Fix forker driver ifdefs for win32
erts: Fix uds socket handling for os x
erts: Fix dereferencing of unaligned integer for sparc
erts: Flatten too long io vectors in uds write
erts: Add fd count test for spawn_driver
...
Conflicts:
erts/emulator/beam/erl_node_tables.c
erts/preloaded/src/erts_internal.erl
|
|
We always want the error to propagate up to the application
when a child cannot be created.
|
|
when called by hash.c for example.
We use ASSERT from sys.h but erl_child_setup implements
its own erl_assert_error() as it doesn't link with sys.o.
|
|
|
|
An acknowledgement of the Start command has to be managed
as we have to make sure that packages are not dropped and
also that the close calls do not happen too early.
|
