| Age | Commit message (Collapse) | Author |
|---|
|
* sverk/deprecate-nif-reload:
erts: Deprecate the NIF reload mechanism
OTP-9771
|
|
The second function erts_unblock_fpe is not needed in here.
|
|
|
|
Almost all uses of the 'long' datatype is removed from VM and tests
Emulator test now runs w/o drivers crashing
Nasty abs bug fixed in VM as well as type errors in allocator debug functions
Still one allocator test that fails, domain knowledge is needed to fix that.
Fix type inconsistency in beam_load causing crashes
|
|
Removed symbolic links from repository.
|
|
Can still not setup -a, but cerl works.
|
|
|
|
Still does not run, just compiles.
|
|
Commit 8781932b3b8769b6f208ac7c00471122ec7dd055 broke
erlang:system_info(system_version) in the non-SMP emulator so
that it would typically dump core.
"rq:%d" was removed from the format string for erts_printf(),
but the corresponding argument in the argument list was not removed,
which ultimately caused "kernel-poll:%s" to be passed an integer
(typically 0).
|
|
* rickard/deprecate-runtime-binding/OTP-9749:
Deprecate erlang:system_flag/2 arguments scheduler_bind_type and cpu_topology
|
|
* rickard/rm-common-runq/OTP-9727:
Remove common run-queue in SMP case
Fix scheduler suspend bug
Conflicts:
erts/emulator/beam/erl_init.c
|
|
The use of erlang:system_flag(scheduler_bind_type, _) and
erlang:system_flag(cpu_topology, _) have been deprecated and scheduled
for removal in erts-5.10/OTP-R16. For more information see the
documentation of erlang:system_flag/2.
|
|
The common run-queue implementation is removed since it is unused,
untested, undocumented, unsupported, and only complicates the code.
A spinlock used by the run-queue management sometimes got heavily
contended. This code has now been rewritten, and the spinlock
has been removed.
|
|
Calls to erlang:system_flag(schedulers_online, N) and/or
erlang:system_flag(multi_scheduling, block|unblock) could cause
internal data used by this functionality to get into an inconsistent
state. When this happened various problems occurred. This bug was
quite hard to trigger, so hopefully no-one has been effected by it.
|
|
* rickard/default-unbound/OTP-9726:
Use unbound schedulers as default
|
|
* rickard/pix-mutex/OTP-9723:
Fix warning when lock-checker is enabled
Replace spinlock with mutex as pix_lock implementation
|
|
* bjorn/deprecate-tuple-funs/OTP-9649:
erts: Warn the first time a tuple fun is called
otp_mibs: Eliminate use of tuple fun
os_mon: Eliminate use of tuple fun
asn1: Eliminate use of tuple fun
parsetools: Eliminate use of tuple fun
mnesia tests: Eliminate use of tuple fun
snmp: Eliminate use of tuple fun
wrap_log_reader_SUITE: Eliminate use of tuple fun
big_SUITE: Eliminate use of tuple fun
file_SUITE: Eliminate use of tuple fun
fprof: Eliminate use of tuple fun
xref_compiler: Eliminate use of tuple fun
shell: Eliminate use of tuple funs
erl_eval: Eliminate use of tuple funs
user_sup: Eliminate use of tuple fun
|
|
* bjorn/llvm-issues/OTP-9712:
INSTALL.md: Update build instructions for Lion
Fix clang linking problem
configure: Define NO_JUMP_TABLE if all we have is llvm-gcc
beam_emu.c: Eliminate warnings when NO_JUMP_TABLE is defined
beam_emu.c: Use the correct void* type for computed gotos
MacOS X: Completely remove obsolete -no-cpp-precomp option
|
|
|
|
The reload mechanism is an odd feature that does not harmonize
with how module upgrade is otherwise done. We do not want its presence
to hinder future development of code loading and upgrade.
|
|
* bjorn/erts/remove-useless-variable:
erl_process.c: Remove the redundant variable 'processes_busy'
Update primary bootstrap
otp_internal: Deprecate ssl:pid/1
|
|
There is a static variable called 'processes_busy' in erl_process.c,
which will be incremented but never used.
To confuse things, there is also a global variable with the same
name, but it it is only defined and used if BM_COUNTERS is defined.
In erl_process.c, the global version of 'processes_busy' will be
seen and incremented if BM_COUNTERS is defined.
Remove the static version of 'processes_busy' in erl_process.c,
but keep the global version if BM_COUNTERS is defined.
Noticed-by: Jovi Zhang
|
|
* lukas/erts/large_float_cmp/OTP-9497:
Use CMP_TMP_HEAD_SIZE for C-stack
|
|
Thanks to Tuncer Ayaz
|
|
Add init_iff_file() for verifying the IFF header. Also let it handle
compressed BEAM files so that it will be done in one place. That means
that code:get_chunk/2 and code:module_md5/1 will now support compressed
BEAM files.
|
|
|
|
The undocumented code:get_chunk/2 BIF is supposed to be a fast way
to extract a chunk from a BEAM file when loading native code. In
practice, it might not have been faster because it happened to
calculate an MD5 checksum for the chunk it extracted because it
shared the scan_iff_file() function with the erlang:load_module/2
BIF.
Split scan_iff_file() into scan_iff_file() and verify_chunks(),
so that the unnecessary MD5 calculation can be avoided.
|
|
code:make_stub_module/3 leaked memory if given either a corrupt
BEAM file, or a compressed BEAM file and an error occurred, or
a binary not aligned on byte boundaries.
|
|
Inlining was not done in a portable way. clang follows the C99
semantics for inlining ('inline' essentially implies 'static' in
C99, but not in GCC), so bp_sched2ix() was not visible outside
beam_bp.c. Since the function need to be used from more than one
source file, put the function definition in the beam_bp.h header
file. Also, give it an 'erts_' prefix since it is globally visible.
|
|
|
|
gcc don't care about the type, but clang does and will issue
a warning.
|
|
* pan/binary_match_scope/OTP-9701:
Remove remaining gcc 4.6 assigned-but-not-used warnings from erts
Remove GCC 4.6 set-but-not-used warning from erl_bif_binary
Make binary:match with scope return correct values
|
|
* sverk/hipe-without-fpe/OTP-9724:
otp_build: Disable FPE by default on Linux
stdlib: Make sure qlc_SUITE:otp_6964 restores backtrace_depth
erts: Add test for inf/NaN intermediate float results
hipe,erts: Allow hipe without floating point exceptions
hipe: Fix bug in hipe_rtl_lcm:calc_killed_expr_bb
erts: Rename macros used by float instructions without FPE
|
|
As of ERTS version 5.9 (OTP-R15B) the runtime system will by default
not bind schedulers to logical processors.
If the Erlang runtime system is the only operating system process that
binds threads to logical processors, this improves the performance of
the runtime system. However, if other operating system processes (as
for example another Erlang runtime system) also bind threads to logical
processors, there might be a performance penalty instead. In some cases
this performance penalty might be severe. Due to this, we change the
default so that the user must make an active decision in order to bind
schedulers.
|
|
|
|
The spinlocks used implementing pix-locks have been replaced with
mutexes since they perform better during heavy contention.
|
|
* rickard/sched-compact-load/OTP-9695:
Add switch that can disable scheduler compaction of load
|
|
* bjorn/erts/refc-binary-literals/OTP-9486:
external.c: Remove the option to create over-sized heap binaries
Allow refc binaries in literal pools
code_SUITE clean-up: Remove experimental 'constant_pool' option
|
|
* sverker/valgrind-fixing:
erts: valgrind suppressions for prebuilt terms in os_info_init
Fix dlopen-leak of drivers with incorrect version
erts: Add valgrind suppression files
erts: Remove valgrind limit for erts_alloc_permanent_cache_aligned
erts: Fix write-after-free bug in inet driver
ETS: Fix faulty size calculation SIZEOF_EXTSEG
ETS: Fix valgrind PossiblyLost in ETS hash tables
|
|
|
|
No semantic change, just code clarification.
Use unique macro names instead of redefining the original ones.
|
|
Since refc binaries are now supported in literal pools, there is no
longer any need to allow the creation of over-sized heap binaries.
|
|
To simplify the implementation of literal pools (constant pools)
for the R12 release, a shortcut was taken regarding binaries --
all binaries would be stored as heap binaries regardless of size.
To allow a module containing literals to be unloaded, literal
terms are copied when sent to another process. That means that
huge literal binaries will also be copied if they are sent to
another process, which could be surprising.
Another problem is that the arity field in the header for the heap
object may not be wide enough to handle big binaries.
Therefore, bite the bullet and allow refc binaries to be stored
in literal pools. In short, the following need to be changed:
* Each loaded module needs a MSO list, linking all refc binaries
in the literal pool.
* When check_process_code/2 copies literals to a process heap,
it must link each referenced binary into the MSO list for the
process and increment the reference counter for the binary.
* purge_module/1 must decrement the reference counter for each
refc binary in the literal pool.
|
|
* bjorn/erts/code-loading/OTP-9720:
BEAM loader: Don't stop other schedulers until truly needed
BEAM loader: Refactor code loading into separate functions
beam_load.c: Factor out cleaning up of the loader state into free_state()
BEAM loader: Clean up handling of error reasons
Remove the useless -l option
beam_load.c: Simplify handling of the bs_put_string_II instruction
beam_load.c: Remove the unused struct LoadedCode
|
|
The erlang:module_load/2 BIF will stop all other schedulers
before it starts to process the BEAM file. That is not a problem
when starting the Erlang/OTP system, since the start-up is mostly
sequential. It is a problem, however, when big BEAM files are
loaded in an already running system.
After the refactoring in the previous commits, we can now do most
of the processing of the BEAM file before we stop the other schedulers,
so that we can minimize the time when only one scheduler will be
running.
Suggested-by: Bob Ippolito
|
|
|
|
|
|
|
|
|
|
Break apart code loading into the three functions:
erts_alloc_loader_state()
erts_prepare_loading()
erts_finish_loading()
The erts_alloc_loader_state() and erts_prepare_loading() can be
executed with all schedulers running. Only erts_finish_loading()
needs to be run in a single-scheduling system.
|
