sleep: measure the remaining time against a clock, not the kernel's remainder - #642

…emainder

(sleep n) returns late, and on a machine that collects often enough it does
not return at all, when another thread allocates.  This is issue Clozure#639.

%nanosleep calls #_nanosleep and, when a signal interrupts the call, sleeps
again for the remaining time the kernel wrote back.  Each GC suspends the
sleeping thread with a signal, and suspend_resume_handler blocks inside the
handler until the collection is over.  The kernel computes the remaining time
before it runs the handler, so the time the thread spends blocked in the
handler is not subtracted.  Every world stop adds its own duration to the
sleep, and the error grows with the collection rate.  On two 2-vCPU burstable
instances a 10 s sleep took 31.8 s on linuxx8664 and 170.1 s on linuxarm64.
Earlier runs on the same arm64 cell were killed at 90 s without returning, so
the overrun has no bound that I have established.

Darwin does not have the defect: its nanosleep works the remaining time out
after the signal handler returns, so the parked time is counted.  FreeBSD and
illumos behave like Linux.  Those three measurements are not mine; they come
from the maintainer, who ran the reproducer on platforms I cannot reach.

The fix computes the deadline once and, on each EINTR, sleeps again for
(stop - now).  The kernel's write-back is no longer read at all, so
#_nanosleep now gets a NULL rem pointer, and the Darwin check for a negative
zero-extended remainder goes with the code it guarded.

The clock is get-internal-real-time, which is monotonic on every platform
since 1e88afc.  The loop is shaped like the one in
%timed-wait-on-semaphore-ptr, which solves the same problem for a semaphore
wait: keep a stop in internal-time-units, re-read the clock on each wakeup,
and floor the difference back into the units the call wants.
clock_nanosleep with TIMER_ABSTIME was considered and not used: Darwin does
not have it, so it would need a second loop with a different error convention
for one target.  This form is one loop for every target.

Cost: two clock reads per %nanosleep call and one more per interruption.
process-wait polls through %nanosleep once per tick, so this is noise there.

Measured red then green with ccl-tests extended-tests/threads/sleep-vs-alloc.lisp
on linuxx8664 and linuxarm64.  The pull request carries the cell.  Not built on
a 32-bit port, and not built on Darwin, FreeBSD or illumos.