Age | Commit message (Collapse) | Author |
|
OTP-10106
OTP-10107
|
|
* sa/dialyzer-parallel: (54 commits)
Logfile-like statistics (enabled with --resources)
Anonymous SCCtoPID ETS table
Anonymous time server
Regulate all kinds of running workers up to the number of schedulers
Relocate start and stop of timing server
Better names for callgaph ETS tables
Remove needless conversion
Fix types and specs
Inline a function in dialyzer_worker
Remove unused function
Change --time to --statistics and include more info
Better reflect side-effect based code in dialyzer_callgraph
Code simplifications (tidier)
More efficient calculation of module deps and postorder
Solve big SCC constraints in parallel
Coordinator is no longer a separate process
All spawns are now spawn_links
Fix race in coordinator
Typesig and dataflow analyses no longer use ticket regulation
Plain concatenation for typesig not-fixpoint list
...
OTP-10103
|
|
* lukas/fix_windows_release_tests/OTP-10099:
Fix compilation of examples and tc names on windows
Create erl.ini for host emulator
|
|
* sverk/valgrind-async-thr-suppression:
erts: Add valgrind suppression for async thread bound data
|
|
|
|
|
|
|
|
* bjorn/test-cases:
ts: Recognize Windows 7
otp_SUITE: Ignore calls to dialyzer if the entire application is missing
otp_SUITE: Remove broken filtering of megaco calls
|
|
This significantly reduces latency for tcp servers
with high load, as we need not go into poll to get
the next message. Maximum throughput may increase
between 4 to 6 times compared to R15B.
|
|
* sverk/cross_tests/OTP-10074:
Fix include issues in when building tests
Fix cross compile of testcases
Update cross build system for tests to work
|
|
* egil/fix-run_erl/OTP-10076:
run_erl: Set controlling terminal
run_erl: Use openpty slave descriptor
run_erl: Retry interrupted system calls
|
|
|
|
The call to the megaco_filter/1 function is broken. It will cause
a badmatch if the megaco_filter/1 would return a different list than
its argument. Since the test case has not failed, megaco_filter/1
has no effect and can be removed.
|
|
|
|
|
|
* egil/fix-port-io-statistics/OTP-10073:
erts: Add port-I/O statistics for active once/true
|
|
|
|
|
|
|
|
Conflicts:
erts/vsn.mk
|
|
|
|
|
|
* rickard/driver_system_info/OTP-10059:
Clear number of async threads if no thread support
|
|
* rickard/no_ets_write_concurrency-r15b01/OTP-10048:
Remove unnecessary memory barriers in ETS when no write_concurrency is used
|
|
* rickard/sched-busy-wait/OTP-10044:
Add switch controlling scheduler busy wait
Conflicts:
erts/emulator/beam/erl_process.c
erts/emulator/beam/erl_process.h
|
|
This leak is harmless as the beam process is about to exit anyway.
|
|
rickard/sched-wakeup-other-r15b01/OTP-10033
Conflicts:
erts/emulator/beam/erl_process.c
erts/vsn.mk
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
When spawning OS (unix) processes with erlang:open_port, store the
resulting unix pid so that it can be queried later on using
erlang:port_info/1,2.
|
|
* rr/wa-lcnt/OTP-10051:
lcnt: Let runq locks reflect actual call location
lcnt: Enhancement fixups
erts: Add lcnt port and suspend options
Increase ethr event and lock counting constants to allow at least +A 1024.
|
|
|
|
|
|
* Add new suspend, location, and port_locks options to erts_debug:lock_counters.
* Init and destroy lock counting for processes and ports as appropriate.
Conflicts:
erts/emulator/beam/erl_bif_info.c
|
|
|
|
rickard/no_ets_write_concurrency-r15b01/OTP-10048
Conflicts:
erts/emulator/beam/erl_db_hash.c
|
|
|
|
|
|
|
|
* sverk/proc-race-leak/OTP-10041:
erts: Fix memory leak caused by race on exiting process
|
|
* rickard/erts-poll/OTP-10019:
Fix wakeup functionality in no thread support case
OTP-10036 is completely unrelated to this commit. Only here to trigger
info about the OTP-10036 ticket in the release notes script...
OTP-10036
|
|
Seen with valgrind running ets_SUITE:delete_large_tab
or delete_large_named_table.
|
|
|
|
* ta/docs-maint:
Fix typos in ct getting_started_chapter.xml
Fix typos in driver.xml and erl_driver.xml
Fix typos in INSTALL-WIN32.md
OTP-10037
|