[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
clock_getcpuclockid versus clock-time
From: |
Andy Wingo |
Subject: |
clock_getcpuclockid versus clock-time |
Date: |
Fri, 17 Jun 2011 11:15:53 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/23.3 (gnu/linux) |
Hello gnulibbers,
I have a curious issue. I would like to use clock_gettime with
CLOCK_PROCESS_CPUTIME_ID. clock_gettime(2) notes some portability
issues with this timer, suggesting that the user call
clock_getcpuclockid(0) to verify that the timer is available.
I was using the clock-time module, but it turns out that on uclibc,
clock_gettime is available in libc, but clock_getcpuclockid is only
available in librt. Thus the clock-time check doesn't add -lrt.
Here's a note from a user. The quoted text is mine.
> Is it perhaps that clock_getcpuclockid is in -lrt but clock_gettime is
> in your libc?
clock_gettime is in both of them:
/lib # readelf -s librt.so.0 | grep clock
42: 00001cb4 68 FUNC GLOBAL DEFAULT 6 clock_gettime
44: 00001a78 172 FUNC GLOBAL DEFAULT 6 clock_nanosleep
64: 00001b74 88 FUNC GLOBAL DEFAULT 6 clock_getcpuclockid
/lib # readelf -s libc.so.0 | grep clock
166: 0000fcf4 72 FUNC GLOBAL DEFAULT 6 clock_gettime
514: 00011868 72 FUNC GLOBAL DEFAULT 6 clock_settime
537: 0000fa84 72 FUNC GLOBAL DEFAULT 6 clock_getres
762: 0002a588 56 FUNC GLOBAL DEFAULT 6 clock
but getcpuclockid is only in librt.
I "solved" this issue with a patch like the attached (to a locally
included gl_CLOCK_TIME). What do you think is the right thing to do
here?
Thanks,
Andy
0001-check-for-clock_getcpuclockid-in-gl_CLOCK_TIME.patch
Description: Text Data
--
http://wingolog.org/
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- clock_getcpuclockid versus clock-time,
Andy Wingo <=