| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Since commit 385b18de6fd72672ed7d6736b30f56d6691d4433, the emulator path was not
copied anymore before pushing it to the args vector (before it was done within
the `push_words` function. Since on Windows `free_env_val` is not a NOP as it is
on Unix systems, the string is freed and afterwards used, leading to strange
errors like this:
escript: Error executing 'àyI': 2
This is fixed by removing the `free_env_val` call.
|
|
* maint:
Updated OTP version
Update release notes
Update version numbers
ssl: Prepare for release
ssl: Proper handling of clients that choose to send an empty answer to a certificate request
heart: Use ntohs instead of manual conversion
|
|
into maint-20
* john/erts/fix-heart-command-overflow/OTP-15034/ERIERL-166:
heart: Use ntohs instead of manual conversion
# Conflicts:
# lib/kernel/test/heart_SUITE.erl
|
|
Multiplying a signed char by 256 is undefined behavior and caused
problems on some platforms when the length was long enough. We
could cast it to an unsigned int to make it work, but it's best not
to reinvent the wheel.
Fixes OTP-15034
|
|
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.
|
|
* raimo/stop-encouraging-v4-mapped/ERL-503/OTP-13716:
Stop translating V4MAPPED addresses
Stop returning V4MAPPED addresses
Implement function for IPv4-mapped IPv6 addresses
|
|
|
|
into 'sverker/master/alloc-n-migration/ERIERL-88'
|
|
into 'sverker/maint-20/alloc-n-migration/ERIERL-88'
OTP-14915
OTP-14916
OTP-14917
OTP-14918
|
|
acnl: Abandon Carrier Nr Limit
acfml: Abandon Carrier Free block Min Limit
|
|
|
|
erlexec adds $ROOT/erts-<vsn>/bin and $ROOT/bin first in $PATH on
startup, but didn't do so if they were in the $PATH already, which
meant that the bindir didn't point to the current release after a
downgrade.
|
|
|
|
* lukas/erts/win_user_home_dir/OTP-14691:
erts: Use PROFILE dir as home on windows
|
|
* dgud/dot_erlang/OTP-14439:
fixup! Do not load .erlang from current dir
erlc: Do not load .erlang
escript: Do not load .erlang
dialyzer: Do not load .erlang
reltool: Add no_dot_erlang bootfiles
Enable usage of no_dot_erlang in bootstrap
Do not load .erlang from current dir
|
|
Instead of using C:\Windows we use the profile path as the
home path on windows. The profile path normally resolves
to C:\Users\%USERNAME%. This fixes an issue where the default
path to the .erlang.cookie was not the same for jinterface as
for erl.
|
|
|
|
Use 'no_dot_erlang' start script for tools. Options and
paths should be stated on the command line, via a
Makefile, rebar.config etc.
|
|
Previously you could not opt out on loading .erlang, change the
default to not load the resource file. The escript author
can invoke c:erlangrc(PathList) to find and load .erlang if needed.
|
|
Use 'no_dot_erlang' start script for tools.
Options and paths should be stated on the command line.
|
|
|
|
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
|
|
|
|
CFLAGS weren't updated to include the sys/common directory for this build
target.
|
|
* 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
|
|
|
|
* rickard/escript-space-path-fix/OTP-14433:
escript: Do not split path to Erlang system into multiple words
|
|
Fix erlang system lookup from escript bug introduced in PR #1293
OTP-14201
|
|
escript failed to start Erlang systems with spaces in the
absolute path (when absolute path was used).
|
|
Erlang system found in PATH was used even when explicitly pointing out
the escript binary in another Erlang system.
|
|
Changing argv[0] from emu name to progname if ESCRIPT_NAME was not set
caused to many failing scripts, revert that part of the change.
See 86f6a9856
|
|
Add the option -emu_type to start an emulator of a certain
type. For exampe, "-emu_type lcnt" will start beam.lcnt.smp.
Any string will be accepted after -emu_type. If there is no
corresponding emulator, there will be an error similar to:
erlexec: The emulator '.../bin/x86_64-unknown-linux-gnu/beam.foo.smp' does not exist.
On Windows, there is an undocumented option "-debug". Remove
that option, because -emu_type can be used instead.
|
|
|
|
* dgud/erts/erlscript_name:
Rename argv[0] from beam to invoking program name
OTP-14381
|
|
OTP-14380
* rickard/ds-stack-size:
Suggested stack size options for dirty schedulers
|
|
Allows ps and htop to display the invoking program/script name
instead of beam[.smp].
|
|
|
|
* bjorn/dialyzer/add-typer/OTP-14336:
Add smoke test for TypER
Add back TypEr to the main OTP repository
|
|
It turned out that the dependencies between Dialyzer
and TypEr makes it impractical to have TypEr in a
separate repository.
Add it back to the OTP repository, but put the Erlang
module 'typer' in the dialyzer application.
|
|
|
|
|
|
* lukas/erts/hide-debug-consoles:
erts: Only show debug consoles if ERL_CONSOLE_MODE is defined
|
|
The application now has an own repo, https://github.com/erlang/typer
|
|
|
|
|
|
The code has been rearranged to make use of the actual path
"get_default_emulator(scriptname)" to the escript instead of
the given one "get_default_emulator(argv[0])".
TL;DR
Assume a source system with some Erlang applications (app1, app2 etc.)
and an escript called "mytool". When generating a standalone target
system (with reltool for example), the escript(s) are located in the
same top bin directory as "erl". See mytool* below.
In such a system the original "mytool" escript is given the extension
".escript" and the file with the same name as the original escript is
a copy of the "escript" executable. One purpose of the escript
executable is to determine which "erl" to use to start the system.
In a standalone system we want it to find the runtime system bundled
with the escript(s). This is done by analyzing the path in order to
find the "erl" located in the same directory as the escript.
A dilemma here is that we do not want to put the top bin directory
in the execution path (PATH env var) as we then would cause other
Erlang applications to make use of our bundled run-time system.
One way of solving this is to choose some suitable bin directory in
the execution path (such as /user/local/bin) and put a symbolic link
there to our mytool executable.
Unfortunately this did not work as the escript executable (in this
case called mytool) would try to find "erl" in /usr/local/bin and when
it did not find such a file it resorted to use the command "erl" which
would find some (unwanted) "erl" in the execution path.
My fix solves that problem.
├── bin/
│ ├── erl* (dyn_erl.c)
│ ├── mytool* (escript.c)
│ ├── mytool.escript* (original mytool escript)
│ └── start_clean.boot
├── erts-vsn/
│ └── bin/
│ ├── beam*
│ ├── beam.smp*
│ ├── erl*
│ ├── erl_child_setup*
│ ├── erlexec*
│ └── inet_gethost*
└── lib/
├── app1-vsn
├── app2-vsn
└── ...
|
|
Hardware watchdog support was removed from heart in R13A, but there were
still some vestiges in the code and the documentation.
- Remove mentions of the HW_WD_DISABLE variable, as it's no longer used.
- Remove the HEART_BEAT_BOOT_DELAY variable, as it was only used for the
hardware watchdog.
|
