[SOLVED] rtc_cmos slow boot

[Haggard]

Hi,

I´ve just installed Mageia2 on an ASUS X42N laptop. It takes minutes to boot, waiting at

rtc_cmos 00:05 rtccore: registered rtc_cmos as rtc0

for ages before finishing the process. I´ve tried to google an answer, but I can´t find anything I can understand.

It happens about 90% of the time. Occasionally it will boot quickly.

I appreciate any help you can give me.

Haggard

[wintpe]

the output from dmesg after a boot gives times for each event that is involved in the boot up.

for example mine around the rtc_cmos is as follows.

Code: Select all

[    0.505152] rtc_cmos 00:06: RTC can wake from S4
[    0.505275] rtc_cmos 00:06: rtc core: registered rtc_cmos as rtc0
[    0.505303] rtc0: alarms up to one month, y3k, 242 bytes nvram, hpet irqs
[    0.505312] cpuidle: using governor ladder
[    0.505313] cpuidle: using governor menu
[    0.505408] TCP: cubic registered
[    0.505410] Registering the dns_resolver key type
[    0.505528] PM: Checking hibernation image partition UUID=d39e3d98-6989-42e3-9573-cf4c1b44d712
[    0.538740] input: AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input0
[    0.623495] PM: Hibernation image not present or could not be loaded.
[    0.623521] registered taskstats version 1
[    0.623902]   Magic number: 13:138:21
[    0.624091] rtc_cmos 00:06: setting system clock to 2013-01-02 10:02:55 UTC (1357120975)


is yours doing anything in parrallel or is it simply held up by this one event, look at the times, maybe this will indicate what its hanging on.

regards peter

[Haggard]

Hi wintpe,

I appreciate the reply.

Here is a chunk of dmesg from around the pause. It waits at the 2nd last line before carrying on. I don´t understand enough about what I´m looking for to comment on whether it is doing anything in parrallel or not.

Code: Select all

[    0.480243] GHES: HEST is not enabled!
[    0.480707] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
[    0.481813] Linux agpgart interface v0.103
[    0.483362] brd: module loaded
[    0.483807] i8042: PNP: PS/2 Controller [PNP0303:PS2K,PNP0f03:PS2M] at 0x60,0x64 irq 1,12
[    0.489007] i8042: Detected active multiplexing controller, rev 1.1
[    0.492221] serio: i8042 KBD port at 0x60,0x64 irq 1
[    0.492763] serio: i8042 AUX0 port at 0x60,0x64 irq 12
[    0.493382] serio: i8042 AUX1 port at 0x60,0x64 irq 12
[    0.493970] serio: i8042 AUX2 port at 0x60,0x64 irq 12
[    0.494598] serio: i8042 AUX3 port at 0x60,0x64 irq 12
[    0.511963] mousedev: PS/2 mouse device common for all mice
[    0.529523] rtc_cmos 00:05: RTC can wake from S4
[    0.546923] rtc_cmos 00:05: rtc core: registered rtc_cmos as rtc0
[    0.565129] hpet1: lost 117879626 rtc interrupts


What should I be looking for?
Thanks again.
Haggard

[wintpe]

[ 0.565129] hpet1: lost 117879626 rtc interrupts is the issue

thats 117 million interupts lost

theres a thread on google , several in fact that deal with this.

one

http://old.nabble.com/Please-help-with- ... 35062.html

that suggests this

Fixed. All I had to do was to disable ACPI. booting with acpi=off
removed the hpet/rtc errors listed in the log.


otherwise if thats not it i need to see the rest of your dmesg up till the line that says

[ 0.624091] rtc_cmos 00:06: setting system clock to ............................

hence why i included it in my example

regards peter

[doktor5000]

http://old.nabble.com/Please-help-with- ... 35062.html

that suggests this

Fixed. All I had to do was to disable ACPI. booting with acpi=off
removed the hpet/rtc errors listed in the log.

Completely disabling acpi seems a rather hard cut and may have serious consequences.
Maybe you want to try first with boot option clocksource=jiffies

[Haggard]

Thanks wintpe,

Here´s the dmesg block between the pause and setting sytem clock.

Code: Select all

[    0.529523] rtc_cmos 00:05: RTC can wake from S4
[    0.546923] rtc_cmos 00:05: rtc core: registered rtc_cmos as rtc0
[    0.565129] hpet1: lost 117879626 rtc interrupts
[  149.821261] Switching to clocksource tsc
[  149.838098] rtc0: alarms up to one month, y3k, 114 bytes nvram, hpet irqs
[  149.855628] cpuidle: using governor ladder
[  149.872867] cpuidle: using governor menu
[  149.889485] TCP cubic registered
[  149.905720] Registering the dns_resolver key type
[  149.922379] input: AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input0
[  149.939810] PM: Checking hibernation image partition UUID=549ae68c-ed61-4145-83c5-2de3e743cdec
[  149.939820] PM: Hibernation image not present or could not be loaded.
[  149.939832] registered taskstats version 1
[  149.957571]   Magic number: 13:792:189
[  149.974519] acpi device:1b: hash matches
[  149.991597] rtc_cmos 00:05: setting system clock to 2013-01-04 18:12:43 UTC (1357323163)

What is the boot process doing here? I think(?) it is trying to set the time from hpet1, getting no response and eventually changing to tsc as a clocksource.

Earlier in dmseg I´ve found

Code: Select all

[    0.111503] RTC time: 18:10:13, date: 01/04/13

. Is this information even slightly relevent? Why is it trying to set the time again?

thanks again, I appreciate your help.

[wintpe]

the problem is the [ 0.565129] hpet1: lost 117879626 rtc interrupts

its not setting the time twice, thats normal for RTC, and its not relevant

rtc registers first then later the time gets set, it is not the symptom of your problem..

[ 0.565129] hpet1: lost 117879626 rtc interrupts is taking 149 to finish because of the 117 million or so lost interupts.

thats the problem.

hpet is the high precision event timer.

also check this out this is exactly your problem, http://ubuntuforums.org/showthread.php?t=2097993
they suggest "hpet=disable" option to grub config file
try it, and try doctors suggestion also if that does not work.
we can see if it makes a difference then try something else.

regards peter

[Haggard]

OK,

I added hpet=disable to grub, and it seems to boot fine now. woohoo.

Code: Select all

[    0.513152] serio: i8042 AUX3 port at 0x60,0x64 irq 12
[    0.530516] mousedev: PS/2 mouse device common for all mice
[    0.548104] rtc_cmos 00:05: RTC can wake from S4
[    0.565445] rtc_cmos 00:05: rtc core: registered rtc_cmos as rtc0
[    0.582501] rtc0: alarms up to one month, y3k, 114 bytes nvram


Thanks heaps for the assistance. I suppose this can be marked as solved (I don´t know how to make that happen).

So now my system isn´t using hpet. Do this have any implications? Is there anything I can read to get a better understanding of the different options?

[doktor5000]

Thanks heaps for the assistance. I suppose this can be marked as solved (I don´t know how to make that happen).

By editing the topic of the first post and prefix it by [SOLVED], thanks.

So now my system isn´t using hpet. Do this have any implications? Is there anything I can read to get a better understanding of the different options?

You may want to read http://www.kernel.org/doc/Documentation/timers/hpet.txt or http://en.wikipedia.org/wiki/High_Precision_Event_Timer
or http://www.intel.com/content/www/us/en/ ... -1-0a.html
or http://elinux.org/System_Tap_Timestamp_Notes or http://blog.fpmurphy.com/2009/07/linux- ... pport.html
As an enduser there are no implications for you.