[bit_flipper]

Code: Select all

Pid: 2641, comm: mkdir Not tainted (2.6.26.8.tex4 #1)
EIP: 0060:[<f923d181>] EFLAGS: 00010046 CPU: 1
EIP is at cmos_pnp_driver+0x21/0xfffff23b [rtc_cmos]
EAX: c0f923c3 EBX: c02819ca ECX: 01b83000 EDX: c0457a40
ESI: f923d17c EDI: f751ce54 EBP: f53cce40 ESP: f51cbf28
DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068

Process mkdir (pid: 2641, ti=f51ca000 task=f4c257f0 task.ti=f51ca000)
Stack: c0112d82 b7f83ff4 f53cce78 c0118b80 c0104610 b7f83ff4 0000007b b7f62765 f53cc378 c0118b80 f53cc340 f51cbfb8 0000007b 0000007b 000000d8 ffffff04 c0118bd8 00000060 00000206 f510a580 00000000 f77e1680 f510a580

Call Trace:
[<c0112d825>] smp_call_function_interrupt0x41/0x80
[<c0118b80>] do_page_fault+0x0/0x680
[<c0104610>] call_function_interrupt+0x28/0x30
[<c0118b80>] do_page_fault+0x0/0x680
[<c0118bd8>] do_page_fault+0x58/0x680
[<c017e319>] filp_close+0x49/0x80
[<c0118b80>] do_page_fault+0x0/0x680
[<c033fa42>] error_code+0x72/0x78
Code: 00 00 00 00 00 00 00 00 00 00 b8 c3 23 f9 20 c4 23 f9 01 00 00 00 20 c3 23 f9 10 c3 23 f9 00 b2 23 f9 f0 b0 23 f9 b8 c3 23 f9 c0 <fe> 3e c0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EIP: [<f923d181>] cmos_pnp_driver+0x21/0xfffff23b [rtc_cmos] SS:ESP 0068:f51cbf28
Kernel panic - not syncing:  Fatal exception in interrupt


I get the above kernel panic with exactly the same error when:

a) After doing a Media Check that yields a PASS of a remaster of my system and I strike the ENTER key to continue booting. Doing a hard reset of the system results in a successful boot from the LiveCD with no problems. Booting directly from a LiveCD remaster (without doing a Media Check detour) results in a successful boot with no problems.

b) When I try to boot the new ZEN-mini 2009.2 LiveCD. The previous version, ZEN-mini 2009.1 LiveCD, boots with no problems whatsoever.

c) When I make a LiveUSB and try to boot from it, BUT have no problems when I make a LiveUSB using MiniMe 2008.

At my site, on multiple machines, there appears to be a pattern of regression for remasters:
* LiveUSB remasters don't work on PCLOS versions newer than 2008...
* Zen-mini remasters don't work on PCLOS versions newer than 2009.1...
* PCLOS-Gnome remasters used to boot without problems after a Media Check and now they don't...
* PCLOS-Gnome remasters using upcoming kernel 2.6.27.31 don't boot at all as described here: http://linuxgator.org/forums/viewtopic.php?f=3&t=2139

Being able to make a remaster is part of my backup strategy and this distro is starting to make me nervous.

[siamer]

cmos_pnp_driver

can You make update Your bios version ? Maybe something is wrong with bios... ? <-- just sugestion, not sure if that is a problem... i didn't see that error before

what mainboard do You have in that computer ? is it laptop or desktop ?

[bit_flipper]

what mainboard do You have in that computer ? is it laptop or desktop ?

Desktop machines. Assorted ASUS motherboards with assorted AMD/64 Athlon X2 processors. All BIOSs are up to date. As to how many machines? Let's see: Each machine comes with a loud squawking non-technical user attached and my office holds about a dozen squawking users when there is standing room only, so my guess is that I have more machines having problems with PCLOS-Gnome than I can count on my fingers...

I've spent the past three days testing Ubuntu 9.10 and One Mandriva 2010 (Gnome). Both of these distros run flawlessly just like PCLOS-Gnome 2008 did on these machines, so I am quite certain that this is not a hardware problem.

As I've indicated in a previous message, PCLOS-Gnome 2008 ran fine. My problems started with PCLOS-Gnome 2009 and these problems seem to be escalating rather than abating as time progresses. I see two possible explanations for this:

a) A fundamental change in packaging methodology for PCLOS-Gnome may have occurred somewhere along the line. Coincident with the release of PCLOS-Gnome 2009, was the departure of a significant group of talented developers. Perhaps, whoever took over the task of packaging is now using different packaging procedures, different kernel compile options and/or different kernel patches for use with PCLOS-Gnome. If this is indeed the case, it appears that these changes in packaging methodology have now propagated to ZEN-mini 2009.2, which would explain why ZEN-mini 2009.1 works for me and ZEN-mini 2009.2 does not -- a repeat of what I experienced between PCLOS-Gnome 2008 and PCLOS-Gnome 2009; or

b) There may be some fundamental flaw with both the 2.6.26.8 kernel I am currently using and the 2.6.27.31 kernel in the repos, totally unrelated to kernel compile options selected and/or kernel patches applied, which would explain why I am having problems with PCLOS-Gnome and no problems with both Ubuntu 9.10 and One Mandriva 2010 (Gnome), both of which use the 2.6.31 series kernel.

It is not clear how one would go about sorting out whether the cause of the problem is a) above, b) above, or a combiniation of both. If a ZEN-mini with a 2.6.31.5 kernel were available, I would try it on my hardware to see if it works. Knowing that One Mandriva 2010 works with this kernel, this kind of test would quickly establish whether the cause of my PCLOS-Gnome problem is a) or b) above, but alas, a ZEN-mini with such a kernel is not available.

Meanwhile, in order to reclaim my office from a bunch of squawking users, I've come to the conclusion that, in my case, at this point in time, PCLOS-Gnome is best enjoyed from a distance on a machine at someone else's premises... I've converted several of my machines having problems with PCLOS-Gnome to running One Mandriva 2010 (Gnome) and so far, everything just works, just like it did with PCLOS-Gnome 2008, including the screen saver, samba networking, Nautilus, and all the other niggling bits that seem to have crept into PCLOS-Gnome lately. I look forward to soon being the only one in my office.

[don_crissti]

Off-topic:

a) A fundamental change in packaging methodology for PCLOS-Gnome may have occurred somewhere along the line.

There is no specific packaging methodology for PCLOS-Gnome. The rules are the same for all PCLOS packages no matter where they belong (core, KDE, Gnome etc) or who is working on those packages. Same goes for any other distro, be it RPM-based or not.
On-topic:
While working on the first ZEN-mini I had to hack into the mklivecd linuxrc file to make it boot on SATA desktop machines:
http://linuxgator.org/forums//viewtopic.php?f=18&t=550&st=0&sk=t&sd=a&start=15
To me it seems this issue is still related to how mklivecd builds the initrd/ISO during remaster. I might be wrong though...

[bit_flipper]

Off-topic:

a) A fundamental change in packaging methodology for PCLOS-Gnome may have occurred somewhere along the line.

Correction:
This is very much on topic for those who Google for: ISO IEC 90003

There is no specific packaging methodology for PCLOS-Gnome.

Correction:
There is always a "methodology" or "process". Thoughtlessly kicking your dirty underwear under your bed is just as much a "methodology" or "process" for handling soiled clothing as consciously putting your dirty underwear into a laundry hamper. The question is whether or not you understand that you have a "methodology" or "process". Those who document their "methodology" or "process" are in a postion to reflect upon it and improve it.

While working on the first ZEN-mini I had to hack into the mklivecd linuxrc file to make it boot on SATA desktop machines

Let me understand this: ZEN-mini 2009.1 works because you hacked that file and ZEN-mini 2009.2 does not work because the individual who took over ZEN-mini from you did not hack that file, or did not know he had to hack that file? This kind of looks like some kind of change in the "methodology" or "process" that produces a ZEN-mini distribution, don't you think?

While working on the first ZEN-mini I had to hack into the mklivecd linuxrc file to make it boot on SATA desktop machines:
/viewtopic.php?f=18&t=550&st=0&sk=t&sd=a&start=15

I find it interesting that you chose to reveal this tidbit of information after learning I was moving my organization away from PCLOS-Gnome while choosing to remain silent here:
http://linuxgator.org/forums/viewtopic.php?f=3&t=2139
and here:
http://linuxgator.org/forums/viewtopic.php?f=25&t=1997#p11211
when I was highly motivated to find a solution to the problem. You see, this too, illustrates a systemic "methodology" or "process" in action. Whether or not you are aware that these methodologies and processes you are engaged in exist, let alone understand them, is a moot point but, since I've brought this one to your attention, I'll let you finger out what it means for the future of PCLOS-Gnome.

[don_crissti]

bit_flipper,

I think you misunderstood my whole post.
1.

There is always a "methodology" or "process". Thoughtlessly kicking your dirty underwear under your bed is just as much a "methodology" or "process" for handling soiled clothing as consciously putting your dirty underwear into a laundry hamper. The question is whether or not you understand that you have a "methodology" or "process". Those who document their "methodology" or "process" are in a postion to reflect upon it and improve it.

Thanks for the explanation... The question is whether or not you understand my words. Let me rephrase:
There is no packaging methodology specific to PCLOS-Gnome. All PCLOS devs use the same techniques, be it RPM packaging or ISO building.
If we are talking about RPMS, there's only one "methodology": the PCLOS way. There is nothing specific to Gnome or KDE or LXDE etc... For some people the KDE ISO works and the Gnome one doesn't. For some others it's quite the opposite. That doesn't mean packages on a certain PCLOS ISO are better built than the packages on a different PCLOS flavor.
As to ISO creation, everyone's using the same tool, mklivecd.
So yes, there is a "process" and there is a "methodology" but they are not specific to PCLOS-Gnome. Maybe this time I was clear enough for you.
2.

Let me understand this: ZEN-mini 2009.1 works because you hacked that file and ZEN-mini 2009.2 does not work because the individual who took over ZEN-mini from you did not hack that file, or did not know he had to hack that file?

NO. I was talking about the first ZEN-mini 2008.1 not 2009.1 which worked OK with the file available from upstream. I was trying to explain a known fact: that various PCLOS ISOs have trouble booting on some hardware and that IMO the problem is, as I said, the way mklivecd builds the initrd/ISO during remaster. What I did for the 2008.1 was a hack not a change in methodology. That hack worked for the 2008 release and wasn't needed anymore for the next release. It was just an example.
As to the rest of your post... I'll save my breath as apparently, you're too frustrated and therefore too biased for my taste.

[bit_flipper]

The question is whether or not you understand my words.

When there is a question, I never hesitate to provide an answer: My entire professional day is spent "not understanding" things. It is part of a managment process that allows people to step right up, take charge, express themselves and to clarify where exactly they might fit on an organization chart. Around here, getting a promotion and a pay increase is as easy as that.

To illustrate how this works, you were presented with an opportunity to respond either from a software engineering/project management perspective (I even gave you a clue by mentioning ISO IEC 90003) or to respond from a program coder's perspective. You chose to respond from a program coder's perspective. There is nothing wrong with responding from the perspective you chose, but it is a data point suggesting that your "fit" is that of "Programmer" not "Project Manager".

Furthermore, the rules of the jungle dictate that in order to be a leader, one must have followers. Followers are quite selective when choosing a leader and I have strong feeling in my bones (or, maybe it is just my rheumatism acting up) that there would be a problem in gaining or retaining followers if a "leader", like for example, the President of the U.S.A., ran around expressing sentiments like this:

I'll save my breath as apparently, you're too frustrated and therefore too biased for my taste.

There is nothing wrong with not being a leader and not having any followers, but it is a data point suggesting that your name might not fit into any box on an organization chart that has the word "Leader" in its label.

All these data points in and of themselves don't mean much but, when aggregated from multiple individuals they, like individual pixels on a computer monitor, begin to paint an image that highlights strengths and deficiencies of an organization, including that of my own.

I hope you don't take this as something personal. It's just an answer to your question. Now, it is time for me to move on. I don't think I will be visiting this board very frequently, if at all.

[ClareOldie]

Now, it is time for me to move on. I don't think I will be visiting this board very frequently, if at all.

Sounds like a good idea ...

[ongoto]

Whew !