> Actually, the CMOS has always been designed to make the clock
> slow when the battery's weak to warn the user.
this is true, to a point.
it is not intended _by_design_, it is due to nature of the clock
oscillator circuit used.
I'm sceptical, too (about that belief). I'm inclined to believe that
story has been conjured up by someone to explain things, and everyone
else has simply gone along with it.
If I were designing a BIOS to automatically deal with low batteries, I'd
be far more obvious than to cause an obscure error. I'd use the beeper,
and display LOW BATTERY on the bootscreen, perhaps halting bootup, or at
least delaying it (which would allow unattended PCs to work). I'd make
the low battery warning public data so that OS builders can monitor it
on a booted system without any guesswork, encouraging OS writers to
actually make use of the data.
I'd also, not prevent systems from working when they're connected to the
mains i.e. BIOS would be battery-backed, not solely battery powered.
There's really little need for BIOS settings on systems where the clock
is automatically maintained by the OS, the hardware is automatically
discovered and configured, etc. Other computer systems manage it (such
as my old Amigas, they cold boot in 13 seconds, every boot does
auto-hardware configuration in that 13 seconds, for all the hardware;
and Macintoshes don't have the PC-type BIOS nightmare, either).
I've got a dedicated video editing PC, here, a Panasonic AU-A950 edit
controller. It's designed to control industrial/broadcast VTRs over
RS422 serial ports. It's now inoperable because its BIOS battery
failed, and cannot be replaced. It's buried inside a clock chip,
soldered onto the board. If it were just the clock that was wrong, that
wouldn't really matter. But it's also lost the parameters for the hard
disk drive, and I have no details for recreating them. And, of course,
to make things more difficult, I have no floppy disks to reset the
system (it's second hand). This is XT-era vintage equipment, with all
those heads/cylinder settings that need manually setting according to
the specific drive.
That kind of malarkey could be avoided if BIOS writers put more thought
into what they designed.
i have never seen a cmos _lose_days_ during a power cycle.
I've only seen such radical clock changes with major crashes. I think
that if you've had a crash so severe that the hardware clock has gone
random (or seriously back to the past), you really need to go into the
BIOS, check all you settings, and save them.
[tim@localhost ~]$ uname -rsvp
Linux 3.9.10-100.fc17.x86_64 #1 SMP Sun Jul 14 01:31:27 UTC 2013 x86_64
Boilerplate: All mail to my mailbox is automatically deleted, there is
no point trying to privately email me, I only get to see the messages
posted to the mailing list.
George Orwell's '1984' was supposed to be a warning against tyranny, not
a set of instructions for supposedly democratic governments.