Date: Thu, 9 Sep 2004 12:29:56 -0500 (CDT) From: "Paul Seniura" <pdseniura@techie.com> To: <FreeBSD-gnats-submit@freebsd.org> Cc: freebsd-emulation@freebsd.org Subject: emulators/rtc: kernel msg "rtc: [number] > kern.hz: Timing will be inaccurate, please increase hz." issued even when [number] == kern.hz Message-ID: <20040909172956.4A6095C13@techpc04.okladot.state.ok.us>
next in thread | raw e-mail | index | archive | help
>Submitter-Id: current-users
>Originator: "Paul Seniura" <pdseniura@techie.com>
>Organization: State of Okla. D.O.T.
>Confidential: no
>Synopsis: emulators/rtc: kernel msg "rtc: [number] > kern.hz: Timing will be inaccurate, please increase hz." issued even when [number] == kern.hz
>Severity: non-critical
>Priority: medium
>Category: ports
>Class: sw-bug
>Release: FreeBSD 6.0-CURRENT i386
>Environment:
System: FreeBSD techpc04.okladot.state.ok.us 6.0-CURRENT FreeBSD 6.0-CURRENT #0: Wed Sep 8 13:47:43 CDT 2004 root@techpc04.okladot.state.ok.us:/usr/obj/src/sys/IBM300SY_4BSD_Os i386
>Description:
We see the kernel msg:
rtc: [number] > kern.hz: Timing will be inaccurate, please increase hz.
on the console as well as in /var/log/messages (maybe dmesg)
even when [number] and kern.hz are the same value
or close to each other.
>How-To-Repeat:
(1) Choose a suitable kern.hz value that your machine can
handle. For the "puny p2" I'm using here at work, I
set this to 512 for testing.
(2) Add a line in your /boot/loader.conf to set kern.hz
to the same value e.g. for me: kern.hz="512" .
(3) Compile the port multimedia/mplayer with WITH_RTC=yes
and WITH_KERN_HZ=[number] set in your /etc/make.conf
e.g. for me: WITH_KERN_HZ=512 .
(4) Compile & install emulators/rtc normally. Load rtc.ko
and ensure /dev/rtc is ready to be used etc.
(5) Test the freshly-built mplayer with any supported sound
file i.e. something simple. I test this on a local
ttyv session with no X running at all.
(6) Look at your console and/or /var/log/messages (maybe dmesg)
to see the 'Timing will be inaccurate' msg is logged.
>Fix:
I don't know much about this, but it sure seems rtc.c is
comparing apples with soccerballs ;) . I have been using the
following patch for many many weeks so far:
===cut-here===
diff -ur files_orig/rtc.c files/rtc.c
--- files_orig/rtc.c Thu Jun 24 08:54:46 2004
+++ files/rtc.c Wed Sep 8 17:11:56 2004
@@ -268,7 +268,7 @@
break;
}
sc->var.freq = freq;
- if (sc->var.freq * 9 > hz * 8) {
+ if (sc->var.freq > hz) {
sc->var.freq = hz;
printf("rtc: %d > kern.hz: Timing will be inaccurate, please increase hz.\n", sc->var.freq);
}
===cut-here===
FWIW I will usually use 500 for kern.hz et al. as it is
more evenly-divisable for other related system parms.
Also FWIW this "puny p2" machine really slows down if
kern.hz gets close to or over 1000 for kern.hz; this
slowdown really shows while running KDE or Gnome. <sigh>
At any rate, I'm not so interested in atomic-level
accuracy here, but rather have a cleaner log, since
this particular msg is issued every single time mplayer
is invoked. ;) Besides, wouldn't the accuracy be
'best' if both were set to exactly the same value?
-- thx, Paul Seniura
System Specialist
State of Okla. D.O.T.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040909172956.4A6095C13>