chrisa@beldar.ma30.bull.com (Chris Anderson) (12/14/90)
Is it possible to provide an external clock for either or both of the serial ports on the personal Iris? The manuals indicate that ioctl can set an external clock mode, but the signal is undocumented. Is the clock signal a spare pin on the D connector, and if so, what should drive it? Thanks for any info. ---Chris A. -- ______________________________. /_____________________________/| | Chris Anderson || | chrisa@beldar.ma30.bull.com || |_____________________________|/
xiaoyan@ecf.toronto.edu (Yan Xiao) (12/15/90)
In article <CHRISA.90Dec13145507@beldar.ma30.bull.com> chrisa@beldar.ma30.bull.com (Chris Anderson) writes: > >Is it possible to provide an external clock for either or both of the >serial ports on the personal Iris? > >The manuals indicate that ioctl can set an external clock mode, >but the signal is undocumented. Is the clock signal a spare pin on the >D connector, and if so, what should drive it? > >Thanks for any info. > >---Chris A. I am wondering why you need an external clock? gettimeofday() provides time in micro-second. xiao
vjs@rhyolite.wpd.sgi.com (Vernon Schryver) (12/17/90)
In article <CHRISA.90Dec13145507@beldar.ma30.bull.com>, chrisa@beldar.ma30.bull.com (Chris Anderson) writes: > > Is it possible to provide an external clock for either or both of the > serial ports on the personal Iris? No. > The manuals indicate that ioctl can set an external clock mode, > but the signal is undocumented. Is the clock signal a spare pin on the > D connector, and if so, what should drive it? Which manual is this? It should not say that, since it is wrong. Have the "EXTA" and "EXTB" speeds of /usr/include/termio.h and termio(7) been misinterpreted? In some distant and ancient universe, they may have referred to external clocks. Now they are the same as 19.2 and 38.4. As chapter 10 of the "IXIX System Administrator's Guide" says, the only signals that do anything interesting on the standard SGI DB-9 are 2,3,4,5,7,8,and 9/20. (or RD,TD,RTS,CTS,SG,DCD,DTR) Vernon Schryver, vjs@sgi.com
vjs@rhyolite.wpd.sgi.com (Vernon Schryver) (12/18/90)
> I am wondering why you need an external clock? gettimeofday() provides > time in micro-second. While the resolution of gettimeofday() is in microseconds, its accuracy is only that of the system clock, which is usually 10 milliseconds but can be set to about 1 millisecond. That is, while getttimeofday() may return an odd number of microseconds, that number will change only every 1 or 10 milliseconds. Perhaps the originial question concerned an alternative to timed, timeslave, or NTP. If so, an external time-of-day clock can be connected via a built-in serial port to synchronize the system clock with something external. For example, the timeslave command (see timeslave(1M)) can talk to a "Precision Standard Time" WWV receiver. PST went out of business, but rumor has it that the new owners, "Traconex", can be reached at: Robert Young 1140 N. Kramer Blvd. Unit G Aneheim, CA 92806 (714) 632-9510 There are rumors that GEOS "satelite clocks" have been connected as well. Vernon Schryver, vjs@sgi.com
dale@lamont.ldgo.columbia.edu (dale chayes) (12/19/90)
In article <78445@sgi.sgi.com>, vjs@rhyolite.wpd.sgi.com (Vernon Schryver) writes: > > > I am wondering why you need an external clock? gettimeofday() provides > > time in micro-second. > . lines deleted . > > > Perhaps the originial question concerned an alternative to timed, Actually, here is the original posting: From: chrisa@beldar.ma30.bull.com (Chris Anderson) ... Subject: External clock for P-I serial port? Message-ID: <CHRISA.90Dec13145507@beldar.ma30.bull.com> Date: 13 Dec 90 19:55:07 GMT Is it possible to provide an external clock for either or both of the serial ports on the personal Iris? The manuals indicate that ioctl can set an external clock mode, but the signal is undocumented. Is the clock signal a spare pin on the D connector, and if so, what should drive it? Thanks for any info. ---Chris A. It has nothing to do with time-of-day. I read this as a requst for information about SYNCHRONOUS serial I/O through the supported serial ports. In fact, Vernon, you answered the original question on December 16th with a clear NO, and a reminder that the IOCTL EXTA/B are for two additional baud rates. I will be the first to thank you for helping us (the net community interested in SIG issues) but converting a discussion about serial (async or not) i/o into a timed/ntp/GOES time-of-day discussion to be a bit of a devolution. Dale ============= -- Dale Chayes Lamont-Doherty Geological Observatory of Columbia University Route 9W, Palisades, N.Y. 10964 dale@lamont.ldgo.columbia.edu voice: (914) 359-2900 extension 434 fax: (914) 359-6817