Re: [PATCH] initscripts: added save-rtc to runlevel S

[Felipe Ferreri Tonello <eu@felipetonello.com>]

Hi Chen,

Thank you for your answer.

On 01/09/2013 12:16 AM, ChenQi wrote:
 > On 01/09/2013 01:35 PM, Felipe Ferreri Tonello wrote:
 >> Hi Chen,
 >>
 >> On 01/08/2013 06:16 PM, ChenQi wrote:
 >>> On 01/09/2013 08:18 AM, eu@felipetonello.com wrote:
 >>>> From: "Felipe F. Tonello" <ftonello@cercacor.com>
 >>>>
 >>>> It is necessary to add save-rtc.sh to runlevel S so the system is
 >>>> updated when
 >>>> it boots up.
 >>> Hi ftonello,
 >>> What do you mean by "system is updated"?
 >> I meant system clock.
 >>
 >> What is happening now is that when you turn off the device, without
 >> system halt, the next time the device is booted up the system clock is
 >> not in sync with the rtc.
 > Hi Felipe,
 >
 > I'm sorry, but I really don't see why this patch works.
 > Below is my understanding for the system clock, hardware clock and
 > /etc/timestamp.
 > (The file name 'save-rtc.sh' is somewhat misleading, 'save-timestamp.sh'
 > would be a more reasonable one.)
 >
 > /etc/timestamp is used to provide a reasonable reference for system time.
 > The initial contents in this file is the building time of the image.
 >
 > The system clock should always be in sync with the rtc as long as the
 > /etc/init.d/hwclock.sh is present, whose main purpose is to sync system
 > clock and hardware clock.
 > No matter whether the system is shutdown normally or crashes, the system
 > clock is according to the hardware clock by hwclock.sh.

Is there anything special to enable hwclock.sh?

I understand what you say, but here that patch was the only way to have 
system clock (down to minutes, also, or even seconds, don't recall) sync 
up with the rtc.

Btw, this patch I did few months ago. So I don't remember exactly what 
was happening.

Do you know by the top of your mind how is this process? Because, as I 
remember, save-rtc.sh was never been called. I'm not sure.

Felipe