While it may seem pointless, it the rouding mode is set to round towards
minus infinity then acos(-1) calculates atan2(sqrt(1.0 - 1.0), -1) the subtract
generates -0.0 which sqrt() preserves.
atan2(-0, -1) is -pi, but acos(-1) is expected to be +pi.
This might 'fix' the test failures seen in some environments, but they
are not failing due to an obvously incorrent x87 rounding mode.
- convert to using ${EXTERNAL_GCC_SUBDIR}
- define base-external-gpl3-gcc* subdir as GCC_SUBDIR
- use <bsd.init.mk> over <bsd.own.mk> for a bunch of places; mostly
because it arranges for ../Makefile.inc to be included earlier, and
don't bother including the latter if the former is already included.
- move all .PATH: settings after <bsd.{own,lib}.mk> so that all
valid variables are set before it is evaluated
- rename mknative-gcc* to match their subdir name.
XXX the relationship between the Makefile.inc/Makefile.gcc_path files
is kind of sketchy, it would be great if this was fixed.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.
to sync with their default gcc configuration.
We could relax this again later, after a filing a defect report with the
ISO C comitee and getting an answer contradicting the gcc teams
interpretation (and then fix the gcc target configurations instead).
alignment by default. Override this to 8byte alignment for alpha and arm
eabi.
Someone (tm) can review this change once the standards and compiler(s)
have sorted themselves out.
non-interlocked CAS in the fast unlock path -- it is unsafe to test for
the waiters-bit while the owner thread is running, we have to spin for
the owner or its state change to be sure about the presence of the bit.
Split off the logic into the pthread__mutex_setwaiters() routine.
This is a partial fix to the named lockup problem (also see PR/44756).
It seems there is another race which can be reproduced on faster CPUs.
