Skip to content

Commit e4da2a4

Browse files
committed
Update obsolete comment about pg_usleep() accuracy.
There are still some systems that use traditional tick-based sleep timing, but many including Linux, FreeBSD and macOS started using high resolution timer hardware more directly a decade or two ago. Update our comment about that. Also highlight that Windows is like the older Unixen in that respect. Author: Nathan Bossart <nathandbossart@gmail.com> Discussion: https://postgr.es/m/CA%2BhUKG%2BogAon8_V223Ldv6taPR2uKH3X_UJ_A7LJAf3-VRARPA%40mail.gmail.com
1 parent 805b821 commit e4da2a4

File tree

1 file changed

+4
-3
lines changed

1 file changed

+4
-3
lines changed

src/port/pgsleep.c

Lines changed: 4 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -25,9 +25,10 @@
2525
/*
2626
* pg_usleep --- delay the specified number of microseconds.
2727
*
28-
* NOTE: although the delay is specified in microseconds, the effective
29-
* resolution is only 1/HZ, or 10 milliseconds, on most Unixen. Expect
30-
* the requested delay to be rounded up to the next resolution boundary.
28+
* NOTE: Although the delay is specified in microseconds, older Unixen and
29+
* Windows use periodic kernel ticks to wake up, which might increase the delay
30+
* time significantly. We've observed delay increases as large as 20
31+
* milliseconds on supported platforms.
3132
*
3233
* On machines where "long" is 32 bits, the maximum delay is ~2000 seconds.
3334
*

0 commit comments

Comments
 (0)