Strange clock behaviour on powerpc-smp

[Dean Loros]

I'm just a beta user--noticed a "problem" with the 15-12 & 15-13 smp ppc kernels--I've tried to use them & both have locked the desktop up almost tight--The times I've been able to have a "sort of" working desktop, I notice that clock functions seem to "bounce" around (Gkrellm aLLtraX-Clock goes "crazy"). The last time I used 15-13, I opened the time & date to change to the correct setting & saw the date go from current to a date around 2031(is this the 31-04 distro?)---I'd love to be around then, but not quite so soon--No unusual information in boot log (other than system clock can't be set--occurs in 15-11 also)--If requested, I'll reboot with 15-13 & recover any logs requested--just E me--autocrosser(at)macdialup(dot)com

[Dean Loros (autocrosser)]

Thank you Ben!

For some reason I couldn't get Malone to do it--

Ben Collins wrote:
> Public bug report changed:
> https://launchpad.net/malone/bugs/29420
>
> Task: ubuntu linux-source-2.6.12
> Sourcepackagename: linux-source-2.6.12 => linux-source-2.6.15
>
>

[Dean Loros (autocrosser)]

I'm just a beta user--noticed a "problem" with the 15-12 & 15-13 smp ppc kernels--I've tried to use them & both have locked the desktop up almost tight--The times I've been able to have a "sort of" working desktop, I notice that clock functions seem to "bounce" around (Gkrellm aLLtraX-Clock goes "crazy"). The last time I used 15-13, I opened the time & date to change to the correct setting & saw the date go from current to a date around 2031(is this the 31-04 distro?)---I'd love to be around then, but not quite so soon--No unusual information in boot log (other than system clock can't be set--occurs in 15-11 also)--If requested, I'll reboot with 15-13 & recover any logs requested--just E me--autocrosser(at)macdialup(dot)com

[Dean Loros (autocrosser)]

Hello Matt--

I have reported on this problem several times over the last few
months--every kernel after 15-11-smp has had this strange clock/timing
issue--I am running a very stock Dual MDD 1.25 Mac with 2 gig of
memory--I re-installed Dapper about a month ago in the hope that a fresh
install would fix the problem--the only change is that the desktop is
slightly more stable--I still can not use a smp kernel for my daily
use--Other than this issue, Dapper has been MUCH better than Breezy.

Thank you for your time.

Dean Loros
Performance by Design Ltd.

Matt Zimmerman wrote:
> ** Summary changed:
>
> - Problem with development 2.6.15-12 & 13 powerpc-smp kernel
> + Strange clock behaviour on powerpc-smp
>

[Dean Loros (autocrosser)]

I have tried the new kernel & still have the problems listed in bug
29420. Time was rapidly changing from current date/time to either May 20
at 6:20 am or July 17 at 7:10am--Gnome-panel "worked" with this for
about 15 minutes & then stopped responding--System "froze" within
seconds of panel non-operation & I had to hard reboot.

[Dean Loros (autocrosser)]

Just tried the "new" 15-21 SMP kernel--no change--still having the same issues as before--crashed Gnome-panel several times--screensaver starts at random--rebooted four times, each time the problem got worse--date/time changes became more pronounced. The single processor kernel with the same number is stable.

[Dean Loros (autocrosser)]

2.6.15-22 SMP still having the same problems--just tried & am now back to one processor--:/

[jbauman (jonas-abgensets)]

I have the same machine with the same problem. I looked through the smp kernel config file and noticed that the CONFIG_IRQ_ALL_CPUS option is set to y, which means interrupts are distributed across all cpus. I seem to remember in the back of my mind that some dual processor Power Macs had a problem with this and it needed to be set to route interrupts to the first cpu only. Another thing I noticed is that the nvram is not accessible with the smp kernel, CONFIG_NVRAM is set to m in the config and when I try to load the nvram module it fails. I did a quick google search on the error message and found this: http://patchwork.ozlabs.org/linuxppc/patch?id=5419
which seems to indicate that the nvram module should be compiled into the kernel instead of as a separate module. Perhaps someone can re-compile the kernel with these options corrected and see if it fixes the problem?

[jbauman (jonas-abgensets)]

I have tried recompiling the kernel as per previous post with no improvements :-(

[Dean Loros (autocrosser)]

I tried to compile a 2.6.16 kernel based off the .config from the latest -SMP kernel & changed the two areas as per the posts--no luck :(

[Niels Kristian Bech Jensen (nkbjensen)]

This patch from 2.6.16 might be a fix for this bug. I don't have a SMP system to test it.

[Ben Collins (ben-collins)]

Patch applied.

[Niels Kristian Bech Jensen (nkbjensen)]

Ben: Beware that I have not tested this fix. On the other hand it is so simple it should not break anything. ;-)

[robT (rob-terhaar)]

i'm a bit of a noob, so if someone could tell me the ubuntu way to test this out.. or give me a .deb file- i'll give this patch a test :)

[Ben Collins (ben-collins)]

2.6.15-25 is now available via dapper-security. Please update to it and retest. Thanks

[Samuel Fogh (gaffa)]

I experience this with Gutsy. My clock is running too fast. Which is a serious problem when using sudo.

[Launchpad Janitor (janitor)]

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.