Shmuel Metz , Seymour J. wrote:
In <[EMAIL PROTECTED]>, on 04/21/2006
at 10:48 PM, "John S. Giltner, Jr." <[EMAIL PROTECTED]> said:
The OS can't support hardware that isn't there. And BTW, the TOD clock
is only part of the timing facilities on zSeries.
My original statement was:
"Any OS can support the hardware clock for the platform it is running on
to be set to local time or UTC."
Where did I imply that a OS can support hardware that is not there? I
said that it can support what the platform supplies. I guess somebody
could infer that I was implying all platforms have hardware clocks.
Which I did not mean to do. I simply meant that a OS can support what
the platform supplies. If there is one there, it can use it, if there
is not one there, will obviously (at least to me) it can't use it.
Because the various computers don't have the same timing facilities.
Again, I stated that OS can support what the platform supplies. I never
stated, nor did I mean to imply, that all platforms have the same functions.
>>Again, logically, aren't they are providing the same function.
No.
So a TOD clock does not supply the time of day? I did not mean to imply
that the TOD clock is a single function piece of hardware. The original
question was specifically about time and date and how mainframes handle
UTC vs. local.
Telling the OS what time it is?
No, that's one function, and not the most important one.
I did not mean to imply that its only function is telling the time of
day, nor did I mean to imply that that was even the most important
function. However that is one of the functions it does do and that
function can be compared with hardware clocks on other platforms (at
least those platforms that supply hardware clocks). I did go and read
the POP and it seems to imply that the TOD clock only provides the time
of day. That there are two other timers that do the clock comparator
and CPU synchronization and that these three clocks together provide the
timer functions for z/Series hardware.
No. An OS is limited by the hardware at hand. Deciding not to exploit
hardware is an option, but when the hardware is missing then deciding
to exploit it is not an option.
Again, my original statement was basically any OS can support what the
platform supplies.
The Devil is in the details. I don't believe that it is possible to
look at it from a logical point of view without first understanding it
from a technical point of view.
I some instances I would agree and in some I disagree. I don't believe
the original question was worried about how they actually work or what
all functions they may provide. I felt they were asking more about UTC
vs. local and which one TOD clock were normally set to and which
timestamps (UTC or local) was used in mainframe logs.
A TOD clock is a clock, just like a RTC, and
just like the watch on my wrist.
No.
Or am I wrong, does a TOD clock provide some other purpose that
providing the time?
Yes. It also provides notification and synchronization facilities.
Please note your word "also". A TOD clock does provide time of day,
granted along with other functions.
I believe that he was looking for a *correct* comparison. The timing
facilities on the PC are fundamentally different from those on
zSeries, and an OS like Linux has to abstract away those differences
and provide portable services rather than access to the underlying
hardware.
I interpreted his question dealing more date and time, not timing. That
is, which time MF's clocks are set to, UTC or local? Does it use UTC or
local used in logs?
Again, according to the POP, some of the other functions you mention are
not provided by the TOD clock, but by other clocks all considered part
of the mainframe timer facilities.
Is the TOD clock like a RTC/CMOS clock? IMHO from a very simplistic
point of view, yes, the RTC/CMOS clock provides the time of day, so the
the TOD clock, as one of its funtions.
Is the RTC/CMOS clock like the TOD clock? No, as the TOD clock does more.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
