| Age | Commit message (Collapse) | Author |
|---|
|
* lukas/erts/non-blocking-shell:
erts: Fix big-endian issue with printf_putc
|
|
On little-endian machines, doing &int will give the smallest
byte which is what we want to give to write. But on big-endian
it will give the highest byte, which will always be \000 here
which results in write never doing any writes.
So we have to cast c to an unsigned char before passing it to
write.
|
|
* lukas/erts/non-blocking-shell:
erts: Make sure IOV_MAX is enforced for writev
erts: Make erts_printf work with non-blocking
|
|
Now that stdout can be set as non-blocking by the ttsl driver
we can no longer use putc and friends for the stdout/stderr
streams. Therefore we fallback on write.
OTP-12239
|
|
Code incorrectly relied on undefined C behavior. Clang and gcc 4.9 do not behave
as code expected. This affected cores and processors detection on FreeBSD.
|
|
The commit adb5dc0090bc419e2c4c1250653badbddeb6263b (ETHR_FORCE_INLINE)
broke some platforms without adequate thread support.
|
|
* sverk/ethr_aux-cache-align-bug:
erts: Fix cache line alignment bug for ts_event_pool
|
|
|
|
Some win32 specific code does not compile with gcc (mingw-w64) since
'__forceinline' is not supported by gcc. This can be avoided by
defining a new macro ETHR_FORCE_INLINE similar to ETHR_INLINE.
|
|
|
|
|
|
This port has support for both non-smp and smp.
It contains a new way to do io checking in which erts_poll_wait
receives the payload of the polled entity. This has implications
for all linked-in drivers.
|
|
|
|
* sverk/win-64-pointer-fix:
erts: Correct term type for printf %T
erts: Correct internal printf integer type for win64
erts: Correct some printf type formatting
erts: Fix type bug in get_proc_affinity for windows
OTP-10887
Forgot this ticket for sverk/erlang_pid-revert:
OTP-10885
|
|
|
|
|
|
|
|
|
|
This reverts commit 624751813b30a367c8255e86f26f7342aeb33cbc.
Overzealous removal of vxworks includes
|
|
The following are deliberately left, as I have only a list of compiler
warnings and no system to test on:
hipe/hipe_x86_signal.c:264:5: warning: no previous prototype for function '_sigaction' [-Wmissing-prototypes]
int __SIGACTION(int signum, const struct sigaction *act, struct sigaction *oldact)
^
hipe/hipe_x86_signal.c:222:21: note: expanded from macro '__SIGACTION'
^
1 warning generated.
sys/unix/sys_float.c:835:16: warning: declaration of 'struct exception' will not be visible outside of this function [-Wvisibility]
matherr(struct exception *exc)
^
sys/unix/sys_float.c:835:1: warning: no previous prototype for function 'matherr' [-Wmissing-prototypes]
matherr(struct exception *exc)
^
2 warnings generated.
drivers/unix/unix_efile.c:1504:11: warning: implicit declaration of function 'sendfile' [-Wimplicit-function-declaration]
retval = sendfile(in_fd, out_fd, *offset, SENDFILE_CHUNK_SIZE,
^
1 warning generated.
|
|
On architectures where cpu topology was unavailable,
warnings erupted due to unused static functions.
Those warnings are now silenced.
|
|
|
|
|
|
Valgrind complains "Conditional jump or move depends on uninitialised value"
when strlen steps past given string maxlen (precision).
|
|
|
|
|
|
|
|
Add probes to the virtual machine, except (mostly) the efile_drv.c
driver and other file I/O-related source files.
|
|
If the number of processors actually found while reading sysfs
is lower than the configured value, we realloc() the cpuinfo array
to the smaller size, but we then iterate it using the original
configured size, thus corrupting memory beyond the allocated
block.
|
|
* rickard/barriers/OTP-9922:
Remove newly introduced excessive barrier in erts_thr_progress_later()
Fix implementation with ddrb barriers
|
|
|
|
Fallback on using /sys/devices/system/cpu to get cpu topology,
as is already done when node directory doesn't exist.
|
|
|
|
- Document barrier semantics
- Introduce ddrb suffix on atomic ops
- Barrier macros for both non-SMP and SMP case
- Make the thread progress API a bit more intuitive
|
|
|
|
|
|
* rickard/bad-rwmtx-assert:
Fix bad assertion in rwmtx implementation
|
|
Windows native critical sections are now used internally in the
runtime system as mutex implementation. This since they perform
better under extreme contention than our own implementation.
|
|
|
|
Conflicts:
erts/aclocal.m4
erts/emulator/beam/erl_db.c
erts/emulator/sys/win32/sys.c
erts/include/internal/ethread_header_config.h.in
|
|
|
|
* pan/erts_printf_term/OTP-9435:
Correct return values from write-functions in erl_printf
|
|
* hw/detect-cpus-on-irix:
Detect the available CPUs on IRIX
OTP-9531
|
|
* cg/fix-constant-logical-operand:
Fix use of logical operator && with constant operand instead of bitwise &.
OTP-9454
|
|
* pg/fix-freebsd-dualcore-detection:
Fix bug in FreeBSD topology detection code
OTP-9453
|
|
|
|
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.
|
|
Add support for querying the number of configured and online processors
on SGI systems running IRIX.
|
|
|
|
|
