devil-linux-develop

[Devil-linux-develop] [0.6Beta1] One more bug to report...

[Dean N. <di...@ti...>]

I have one more bug to report on the latest beta release (08-08-i686-SMP).

When I go to shutdown the environment, I get the following error right after
"Stopping LVM [OK]":
flushing ide devices: hda hdb ide_cdrom_cleanup: hdb: failed to
ide_unregister_subdriver

If it is meaningful, hdb is my cdrom drive.

I am also getting a message from grsecurity about a resource over
allocation - but it scrolls off the screen before I can catch it.  (And
other than knowing it is just before or about the time of the DHCP update
for eth0, but before the WLAN stuff - I can't pin it down much.

Dean Nedelman
TimeLord Consulting

Re: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Heiko Z. <he...@zu...>]

Dean Nedelman wrote:
> I have one more bug to report on the latest beta release (08-08-i686-SMP).
> 
> When I go to shutdown the environment, I get the following error right after
> "Stopping LVM [OK]":
> flushing ide devices: hda hdb ide_cdrom_cleanup: hdb: failed to
> ide_unregister_subdriver

I saw this message, too. It must be something with the Kernel.
Kernel 2.4.22 seems to come out anyway soon, so just ignore it.


> If it is meaningful, hdb is my cdrom drive.
> 
> I am also getting a message from grsecurity about a resource over
> allocation - but it scrolls off the screen before I can catch it.  (And
> other than knowing it is just before or about the time of the DHCP update
> for eth0, but before the WLAN stuff - I can't pin it down much.

Try to stop the output by pressing SCROLL LOCK, it's bad when grsecurity 
reports a resource overstep. You need to know which program it is, 
otherwise you could run into problems.

Heiko

RE: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Dean N. <di...@ti...>]

>>
>> I am also getting a message from grsecurity about a resource over
>> allocation - but it scrolls off the screen before I can catch it.  (And
>> other than knowing it is just before or about the time of the DHCP update
>> for eth0, but before the WLAN stuff - I can't pin it down much.
>
>Try to stop the output by pressing SCROLL LOCK, it's bad when grsecurity
>reports a resource overstep. You need to know which program it is,
>otherwise you could run into problems.

Starting DHCP client daemon on interface eth0
grsec: attempted resource overstep by requesting 76591104 for RLIMIT_STACK
against limit 8388608 by (pidof:333) UID(0) EUID(0), parent (S05network:321)
UID(0) EUID(0)
grsec: attempted resource overstep by requesting 203358208 for RLIMIT_STACK
against limit 8388608 by (pidof:333) UID(0) EUID(0), parent (S05network:321)
UID(0) EUID(0)

BTW:  RedHat (and I presume others) automatically turns off the machine for
me when I shutdown.  I notice that Devil-Linux just gives me the message to
go ahead and power down.  Might be something that should be added.....

Dean

RE: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Bruce S. <bw...@ar...>]

> BTW:  RedHat (and I presume others) automatically turns off the machine for
> me when I shutdown.  I notice that Devil-Linux just gives me the message to
> go ahead and power down.  Might be something that should be added.....

Agree, it should be added, IF it's not already there. (it may be)

Keep in mind that even Redhat (and others) will NOT power off machines
for certain types of hardware.  i.e. SMP boxes will not power off.
I'm not sure what happens if you're running a SMP kernel on a UP PC.

 - BS

Re: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Heiko Z. <he...@zu...>]

Bruce Smith wrote:
>>BTW:  RedHat (and I presume others) automatically turns off the machine for
>>me when I shutdown.  I notice that Devil-Linux just gives me the message to
>>go ahead and power down.  Might be something that should be added.....
> 
> 
> Agree, it should be added, IF it's not already there. (it may be)
> 
> Keep in mind that even Redhat (and others) will NOT power off machines
> for certain types of hardware.  i.e. SMP boxes will not power off.
> I'm not sure what happens if you're running a SMP kernel on a UP PC.

I was thinking for a while if we should enable the power management stuff.
Since you both agree with it, let's do it.

Heiko

Re: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Bruce S. <bw...@ar...>]

> > Agree, it should be added, IF it's not already there. (it may be)
> > 
> > Keep in mind that even Redhat (and others) will NOT power off machines
> > for certain types of hardware.  i.e. SMP boxes will not power off.
> > I'm not sure what happens if you're running a SMP kernel on a UP PC.
> 
> I was thinking for a while if we should enable the power management stuff.
> Since you both agree with it, let's do it.

Let's think about this a little harder.  Are other power management
features going to cause us problems?  Causing a PC to go into sleep mode
would be bad.  We need to be really careful on what features we select.

Can we turn on just the power off feature?

 - BS

Re: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Heiko Z. <he...@zu...>]

Bruce Smith wrote:
>>>Agree, it should be added, IF it's not already there. (it may be)
>>>
>>>Keep in mind that even Redhat (and others) will NOT power off machines
>>>for certain types of hardware.  i.e. SMP boxes will not power off.
>>>I'm not sure what happens if you're running a SMP kernel on a UP PC.
>>
>>I was thinking for a while if we should enable the power management stuff.
>>Since you both agree with it, let's do it.
> 
> 
> Let's think about this a little harder.  Are other power management
> features going to cause us problems?  Causing a PC to go into sleep mode
> would be bad.  We need to be really careful on what features we select.
> 
> Can we turn on just the power off feature?

I definitely don't want any harddisk spin-downs, or sleep modes.
Don't you need a special daemon which puts the PC into sleep mode?
Lets see if there's a program where we can turn that stuff off during 
system bootup.

Heiko

Re: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Bruce S. <bw...@ar...>]

> > Let's think about this a little harder.  Are other power management
> > features going to cause us problems?  Causing a PC to go into sleep mode
> > would be bad.  We need to be really careful on what features we select.
> > 
> > Can we turn on just the power off feature?
> 
> I definitely don't want any harddisk spin-downs, or sleep modes.

Right.

> Don't you need a special daemon which puts the PC into sleep mode?

I don't think so.  I think it's done with the kernel & the BIOS.

On my Redhat system, where HDD spin-down work, I have a process:
root         3     1  0 Aug08 ?        00:00:00 [kapmd]

But there is no binary for kapmd, so I think it's a kernel process.

> Lets see if there's a program where we can turn that stuff off during 
> system bootup.

I haven't looked for a long time, but there used to be a lot of
selections you could make in the kernel config about what worked.

I'm starting to wonder if this is a bad idea to turn on.
Especially since APM is being replaced with ACPI, and Linux 
currently has poor support for ACPI.

 - BS

Re: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Heiko Z. <he...@zu...>]

Bruce Smith wrote:
>>>Let's think about this a little harder.  Are other power management
>>>features going to cause us problems?  Causing a PC to go into sleep mode
>>>would be bad.  We need to be really careful on what features we select.
>>>
>>>Can we turn on just the power off feature?
>>
>>I definitely don't want any harddisk spin-downs, or sleep modes.
> 
> 
> Right.
> 
> 
>>Don't you need a special daemon which puts the PC into sleep mode?
> 
> 
> I don't think so.  I think it's done with the kernel & the BIOS.
> 
> On my Redhat system, where HDD spin-down work, I have a process:
> root         3     1  0 Aug08 ?        00:00:00 [kapmd]
> 
> But there is no binary for kapmd, so I think it's a kernel process.

They write in the Kernel config, that nothing like this happens.
E.g. for HDD spindown, you have to specify this via the hdparm program.
On the other hand, when somebody has enabled it in the BIOS -> disable it.


>>Lets see if there's a program where we can turn that stuff off during 
>>system bootup.
> 
> 
> I haven't looked for a long time, but there used to be a lot of
> selections you could make in the kernel config about what worked.
> 
> I'm starting to wonder if this is a bad idea to turn on.
> Especially since APM is being replaced with ACPI, and Linux 
> currently has poor support for ACPI.

I turned on only APM, ACPI is not well enough supported yet.

cya
   Heiko

Re: [Devil-linux-develop] [0.6Beta1] One more bug to report...

[Heiko Z. <he...@zu...>]

Bruce Smith wrote:
>>>Agree, it should be added, IF it's not already there. (it may be)
>>>
>>>Keep in mind that even Redhat (and others) will NOT power off machines
>>>for certain types of hardware.  i.e. SMP boxes will not power off.
>>>I'm not sure what happens if you're running a SMP kernel on a UP PC.
>>
>>I was thinking for a while if we should enable the power management stuff.
>>Since you both agree with it, let's do it.
> 
> 
> Let's think about this a little harder.  Are other power management
> features going to cause us problems?  Causing a PC to go into sleep mode
> would be bad.  We need to be really careful on what features we select.
> 
> Can we turn on just the power off feature?

I love Linux !
I did some research on the web and of course Linux doesn't do anything 
without being told so. We just need to enable the APM feature in the 
Kernel, then we should be set. There won't be any spindowns or so.

Heiko