| Age | Commit message (Collapse) | Author |
|---|
|
__lll_trylock_elision sets the adapt_count variable too
aggressively, and incorrectly on persistent aborts. Taking
a cue from s390, adapt_count is only updated if the lock
is locked, or a persistent failure occurs.
In addition, the abort codes have been renumbered and
refactored for clarity. As it stands, glibc only cares
if the abort is persistent or not.
All aborts are now persistent, excepting a busy lock. This
includes changing _ABORT_NESTED_TRYLOCK into a persistent
abort.
* sysdeps/unix/sysv/linux/powerpc/elision-trylock.c
(__lll_trylock_elision): Fix setting of adapt_count.
* sysdeps/unix/sysv/linux/powerpc/htm.h
(_ABORT_PERSISTENT): Define to clarify persistent aborts.
(_ABORT_NESTED_TRYLOCK): Renumber, and make persistent.
(_ABORT_SYSCALL): Renumber, and clarify definition.
(_ABORT_LOCK_BUSY): Renumber, make non-persistent.
(cherry picked from commit 86b4939846caf2bb072bba6057e1dc3ad187c2c2)
Conflicts:
sysdeps/unix/sysv/linux/powerpc/elision-trylock.c
|
|
Some extraneous semicolons were included in a
recent patch which causes a build failure with
newer compilers.
(cherry picked from commit af8ea0f449af7d3847351a4a5bafcd435a22ac31)
|
|
Work around a GCC behavior with hardware transactional memory built-ins.
GCC doesn't treat the PowerPC transactional built-ins as compiler
barriers, moving instructions past the transaction boundaries and
altering their atomicity.
(cherry picked from commit 42bf1c897170ff951c7fd0ee9da25f97ff787396)
Conflicts:
sysdeps/unix/sysv/linux/powerpc/elision-trylock.c
|
|
This patch fix the TABORT encoding for toolchains with no support for
HTM builtins.
|
|
This patch fixes the copyright dates from files created by commit
8d2c0a5, 4b45943, and 56cf276.
|
|
This patch adds support for lock elision using ISA 2.07 hardware
transactional memory instructions for pthread_mutex primitives.
Similar to s390 version, the for elision logic defined in
'force-elision.h' is only enabled if ENABLE_LOCK_ELISION is defined.
Also, the lock elision code should be able to be built even with
a compiler that does not provide HTM support with builtins.
However I have noted the performance is sub-optimal due scheduling
pressures.
|
