Subject: Re: system clock mess on SPARC64 1.6ZK (and 1.6.2)
To: Timo Schoeler <email@example.com>
From: Andrey Petrov <firstname.lastname@example.org>
Date: 03/06/2004 20:16:52
On Sat, Mar 06, 2004 at 05:12:16PM +0100, Timo Schoeler wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> hi list,
> since i pulled the creator (FFB2+) graphics card out of my U10 (dmesg
> attached below) the system clock loses approx. five (5) minutes per
> hour (!) -- it appears both running 1.6.2 and 1.6ZK.
> when i set the correct time via ntpdate and then restart ntpd (4.2 from
> pkgsrc) ntpd can't catch up and refuses to sync due to the large
> difference the system clock 'creates'.
I saw quite noticable time drift (well, delay) when openprom was
used for console output, which is practically always the case unless
you configure raster console. If your console display changed from ffb
to aty then it would be a cause. Say, openprom works better for ffb.
And you have to have pretty console output, btw, if you do, then
stop using console and check that time stays correct.
Here is intercept from u5 config for raster console, colors are funny
(pink on white:-), due to some aty problem) but it works faster and without
options RASTERCONSOLE # fast rasterop console
options FONT_GALLANT12x22 # the console font
# wscons options
# builtin terminal emulations
options WSEMUL_SUN # sun terminal emulation
#options WSEMUL_VT100 # VT100 / VT220 emulation
# different kernel output - see dev/wscons/wsdisplayvar.h
# compatibility to other console drivers
#options WSDISPLAY_COMPAT_PCVT # emulate some ioctls
options WSDISPLAY_COMPAT_SYSCONS # emulate some ioctls
#options WSDISPLAY_COMPAT_USL # VT handling
options WSDISPLAY_COMPAT_RAWKBD # can get raw scancodes
# see dev/pckbc/wskbdmap_mfii.c for implemented layouts
##options PCKBD_LAYOUT="(KB_DE | KB_NODEAD)"
# allocate a number of virtual screens at autoconfiguration time
# use a large software cursor that doesn't blink
# modify the screen type of the console; defaults to "80x25"
# work around a hardware bug that loaded fonts don't work; found on ATI cards
# the following enables some functions to get mouse console support.
# if you want a really secure system, it may be better not to enable them,
# see wsmoused(8), section SECURITY CONSIDERATIONS for more info.
#options WSDISPLAY_CHARFUNCS # mouse console support
# enable VGA raster mode capable of displaying multilingual text on console
machfb* at pci? dev ? function ?
wsdisplay* at machfb?
wskbd* at kbd? # console ?
wsmouse* at ms?
> i wonder whether it is known behavior (on SPARC64) or my rtc is
> wracked...? as soon as build.sh is finished, i'll check up that machine
> with Sun's VTS tools to evaluate the status of the machines rtc.
Alaverdy, we don't use rtc on sparc64 anymore, it always tick register,
I don't is that bad or good.