Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commita948e49

Browse files
committed
Use nanosleep() to implement pg_usleep().
The previous coding based on select() had commentary about historicalportability concerns. Use POSIX nanosleep() instead.This has independently been suggested a couple of times before, butnever managed to stick. Since recent and proposed work removes otheruses of select(), and associated code and comments relating to itsnon-portable interaction with signals, it seems like a good time to tidyup this case, too.Also modernize the explanation of why WaitLatch() is a better way towait.Reviewed-by: Nathan Bossart <nathandbossart@gmail.com>Suggested-by: Paul Guo <paulguo@gmail.com>Suggested-by: Tom Lane <tgl@sss.pgh.pa.us>Discussion:https://postgr.es/m/CAAKRu_b-q0hXCBUCAATh0Z4Zi6UkiC0k2DFgoD3nC-r3SkR3tg%40mail.gmail.comDiscussion:https://postgr.es/m/CABQrizfxpBLZT5mZeE0js5oCh1tqEWvcGF3vMRCv5P-RwUY5dQ@mail.gmail.comDiscussion:https://postgr.es/m/4902.1552349020@sss.pgh.pa.us
1 parente4da2a4 commita948e49

File tree

1 file changed

+10
-15
lines changed

1 file changed

+10
-15
lines changed

‎src/port/pgsleep.c

Lines changed: 10 additions & 15 deletions
Original file line numberDiff line numberDiff line change
@@ -12,9 +12,7 @@
1212
*/
1313
#include"c.h"
1414

15-
#include<unistd.h>
16-
#include<sys/select.h>
17-
#include<sys/time.h>
15+
#include<time.h>
1816

1917
/*
2018
* In a Windows backend, we don't use this implementation, but rather
@@ -32,27 +30,24 @@
3230
*
3331
* On machines where "long" is 32 bits, the maximum delay is ~2000 seconds.
3432
*
35-
* CAUTION: the behavior when a signal arrives during the sleep is platform
36-
* dependent. On most Unix-ish platforms, a signal does not terminate the
37-
* sleep; but on some, it will (the Windows implementation also allows signals
38-
* to terminate pg_usleep). And there are platforms where not only does a
39-
* signal not terminate the sleep, but it actually resets the timeout counter
40-
* so that the sleep effectively starts over! It is therefore rather hazardous
41-
* to use this for long sleeps; a continuing stream of signal events could
42-
* prevent the sleep from ever terminating. Better practice for long sleeps
43-
* is to use WaitLatch() with a timeout.
33+
* CAUTION: It's not a good idea to use long sleeps in the backend. They will
34+
* silently return early if a signal is caught, but that doesn't include
35+
* latches being set on most OSes, and even signal handlers that set MyLatch
36+
* might happen to run before the sleep begins, allowing the full delay.
37+
* Better practice is to use WaitLatch() with a timeout, so that backends
38+
* respond to latches and signals promptly.
4439
*/
4540
void
4641
pg_usleep(longmicrosec)
4742
{
4843
if (microsec>0)
4944
{
5045
#ifndefWIN32
51-
structtimevaldelay;
46+
structtimespecdelay;
5247

5348
delay.tv_sec=microsec /1000000L;
54-
delay.tv_usec=microsec %1000000L;
55-
(void)select(0,NULL,NULL,NULL,&delay);
49+
delay.tv_nsec=(microsec %1000000L)*1000;
50+
(void)nanosleep(&delay,NULL);
5651
#else
5752
SleepEx((microsec<500 ?1 : (microsec+500) /1000), FALSE);
5853
#endif

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp