| Age | Commit message (Collapse) | Author |
|---|
|
Conflicts:
erts/aclocal.m4
erts/include/internal/ethread_header_config.h.in
|
|
|
|
The ethread atomics API now also provide double word size atomics.
Double word size atomics are implemented using native atomic
instructions on x86 (when the cmpxchg8b instruction is available)
and on x86_64 (when the cmpxchg16b instruction is available). On
other hardware where 32-bit atomics or word size atomics are
available, an optimized fallback is used; otherwise, a spinlock,
or a mutex based fallback is used.
The ethread library now performs runtime tests for presence of
hardware features, such as for example SSE2 instructions, instead
of requiring this to be determined at compile time.
There are now functions implementing each atomic operation with the
following implied memory barrier semantics: none, read, write,
acquire, release, and full. Some of the operation-barrier
combinations aren't especially useful. But instead of filtering
useful ones out, and potentially miss a useful one, we implement
them all.
A much smaller set of functionality for native atomics are required
to be implemented than before. More or less only cmpxchg and a
membar macro are required to be implemented for each atomic size.
Other functions will automatically be constructed from these. It is,
of course, often wise to implement more that this if possible from a
performance perspective.
|
|
|
|
Atomic operations with specified barriers have specified barrier semantics.
Set and read operations have undefined barrier semantics. All other atomic
operations implied full memory barriers, except when using the libatomic_ops
library and the tilera atomics api.
Some code in the runtime system assumed that all operations used (except for
set, read and specified) implied full memory barriers. The use of the
libatomic_ops library and the tilera atomics api have therefore been modified
to behave as the other implementations.
Some atomic operations with specified barrier semantics on sparc32 have also
been been relaxed in this commit.
|
|
|
|
|
|
The atomic memory operations interface used the 'long' type and assumed that
it was of the same size as 'void *'. This is true on most platforms, however,
not on Windows 64.
|
|
* rickard/configure-help/OTP-8859:
Cleanup configure help texts
Conflicts:
erts/aclocal.m4
|
|
* rickard/pre-pentium4-compat/OTP-8847:
Allow usage of libatomic_ops when other compilers than gcc are used
Add pre pentium 4 compatibilty configure test
Fix --enable-ethread-pre-pentium4-compatibility
|
|
|
|
|
|
|
|
The `configure' command line argument
`--enable-ethread-pre-pentium4-compatibility'
had no effect.
|
|
|
|
Large parts of the ethread library have been rewritten. The
ethread library is an Erlang runtime system internal, portable
thread library used by the runtime system itself.
Most notable improvement is a reader optimized rwlock
implementation which dramatically improve the performance of
read-lock/read-unlock operations on multi processor systems by
avoiding ping-ponging of the rwlock cache lines. The reader
optimized rwlock implementation is used by miscellaneous
rwlocks in the runtime system that are known to be read-locked
frequently, and can be enabled on ETS tables by passing the
`{read_concurrency, true}' option upon table creation. See the
documentation of `ets:new/2' for more information.
The ethread library can now also use the libatomic_ops library
for atomic memory accesses. This makes it possible for the
Erlang runtime system to utilize optimized atomic operations
on more platforms than before. Use the
`--with-libatomic_ops=PATH' configure command line argument
when specifying where the libatomic_ops installation is
located. The libatomic_ops library can be downloaded from:
http://www.hpl.hp.com/research/linux/atomic_ops/
The changed API of the ethread library has also caused
modifications in the Erlang runtime system. Preparations for
the to come "delayed deallocation" feature has also been done
since it depends on the ethread library.
Note: When building for x86, the ethread library will now use
instructions that first appeared on the pentium 4 processor. If
you want the runtime system to be compatible with older
processors (back to 486) you need to pass the
`--enable-ethread-pre-pentium4-compatibility' configure command
line argument when configuring the system.
|
|
Writer preferred pthread read/write locks has been enabled on Linux.
|
|
Support for using gcc's built-in functions for atomic memory access has
been added. This functionallity will be used if available and no other
native atomic implementation in ERTS is available.
|
|
|
|
improvements.
Most notable:
Lots of cross compilation improvements. The old cross compilation
support was more or less non-existing as well as broken. Please,
note that the cross compilation support should still be
considered as experimental. Also note that old cross compilation
configurations cannot be used without modifications. For more
information on cross compiling Erlang/OTP see the
$ERL_TOP/xcomp/README file.
Support for staged install using <url
href="http://www.gnu.org/prep/standards/html_node/DESTDIR.html">D
ESTDIR</url>. The old broken INSTALL_PREFIX has also been fixed.
For more information see the $ERL_TOP/README file.
Documentation of the release target of the top Makefile. For more
information see the $ERL_TOP/README file.
make install now by default creates relative symbolic links
instead of absolute ones. For more information see the
$ERL_TOP/README file.
$ERL_TOP/configure --help=recursive now works and prints help for
all applications with configure scripts.
Doing make install, or make release directly after make all no
longer triggers miscellaneous rebuilds.
Existing bootstrap system is now used when doing make install, or
make release without a preceding make all.
The crypto and ssl applications use the same runtime library path
when dynamically linking against libssl.so and libcrypto.so. The
runtime library search path has also been extended.
The configure scripts of erl_interface and odbc now search for
thread libraries and thread library quirks the same way as erts
do.
The configure script of the odbc application now also looks for
odbc libraries in lib64 and lib/64 directories when building on a
64-bit system.
The config.h.in file in the erl_interface application is now
automatically generated in instead of statically updated which
reduces the risk of configure tests without any effect.
|
|
improvements.
Most notable:
Lots of cross compilation improvements. The old cross compilation
support was more or less non-existing as well as broken. Please,
note that the cross compilation support should still be
considered as experimental. Also note that old cross compilation
configurations cannot be used without modifications. For more
information on cross compiling Erlang/OTP see the
$ERL_TOP/xcomp/README file.
Support for staged install using <url
href="http://www.gnu.org/prep/standards/html_node/DESTDIR.html">D
ESTDIR</url>. The old broken INSTALL_PREFIX has also been fixed.
For more information see the $ERL_TOP/README file.
Documentation of the release target of the top Makefile. For more
information see the $ERL_TOP/README file.
make install now by default creates relative symbolic links
instead of absolute ones. For more information see the
$ERL_TOP/README file.
$ERL_TOP/configure --help=recursive now works and prints help for
all applications with configure scripts.
Doing make install, or make release directly after make all no
longer triggers miscellaneous rebuilds.
Existing bootstrap system is now used when doing make install, or
make release without a preceding make all.
The crypto and ssl applications use the same runtime library path
when dynamically linking against libssl.so and libcrypto.so. The
runtime library search path has also been extended.
The configure scripts of erl_interface and odbc now search for
thread libraries and thread library quirks the same way as erts
do.
The configure script of the odbc application now also looks for
odbc libraries in lib64 and lib/64 directories when building on a
64-bit system.
The config.h.in file in the erl_interface application is now
automatically generated in instead of statically updated which
reduces the risk of configure tests without any effect.
|
|
new mandatory configuration variable that replaces all other
previously mandatory configuration variables, and better
documentation. Note that old cross compilation configurations
cannot be used without modifications.
For more information see the $ERL_TOP/xcomp/README file.
|
|
|
