[gentoo-amd64] Kernel 2.6.21 problem

[gentoo-amd64] Kernel 2.6.21 problem

[Peter Humphrey]

On syncing today I was offered gentoo-sources-2.6.21. On attempting to 
compile it I get:

drivers/built-in.o: In function `acpi_rtc_init':
glue.c:(.init.text+0x34c2): undefined reference to `pnp_bus_type'
drivers/built-in.o: In function `acpi_bus_generate_event':
(.text.acpi_bus_generate_event+0x25): undefined reference to `event_is_open'
drivers/built-in.o: In function `acpi_bus_get_power':
(.text.acpi_bus_get_power+0x97): undefined reference to 
`acpi_power_get_inferred_state'
drivers/built-in.o: In function `acpi_bus_set_power':
(.text.acpi_bus_set_power+0xf3): undefined reference to 
`acpi_power_transition'
drivers/built-in.o: In function `acpi_bus_set_power':
(.text.acpi_bus_set_power+0x165): undefined reference to 
`acpi_power_transition'

Does this imply that I'm required to include a full ACPI setup in my 
kernels? I'm trying to prevent all attempts to switch my monitor off, but 
without a lot of success so far. (If I haven't told the machine to do 
something, it shouldn't do it; and I want control over my peripherals 
thank-you-very-much.)

My kernel-compilation method is as follows:
1.	Emerge -u gentoo-sources
2.	Set /usr/src/linux manually (I know there is an automatic method, but 
that kills any package that's recompiled with every kernel version)
3.	Copy the previous .config from its store on /boot
4.	Make oldconfig and set any new options, making a note of any that need 
further exploration
5.	If any such notes, make menuconfig and read help; change any settings 
that need it
6.	Make && make modules_install
7.	Copy new .config and new kernel to /boot
8.	Change /boot/grub/grub.conf to choose the new kernel by default
9.	Reboot no-X and remerge nvidia-drivers and vmware-server
10.	Reboot into KDE and check the system over
11.	Celebrate with a pint.

-- 
Rgds
Peter Humphrey
Linux Counter 5290, Aug 93
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Kernel 2.6.21 problem

[Hemmann, Volker Armin]

On Sonntag, 29. April 2007, Peter Humphrey wrote:
> On syncing today I was offered gentoo-sources-2.6.21. On attempting to
> compile it I get:
>
> drivers/built-in.o: In function `acpi_rtc_init':
> glue.c:(.init.text+0x34c2): undefined reference to `pnp_bus_type'
> drivers/built-in.o: In function `acpi_bus_generate_event':
> (.text.acpi_bus_generate_event+0x25): undefined reference to
> `event_is_open' drivers/built-in.o: In function `acpi_bus_get_power':
> (.text.acpi_bus_get_power+0x97): undefined reference to
> `acpi_power_get_inferred_state'
> drivers/built-in.o: In function `acpi_bus_set_power':
> (.text.acpi_bus_set_power+0xf3): undefined reference to
> `acpi_power_transition'
> drivers/built-in.o: In function `acpi_bus_set_power':
> (.text.acpi_bus_set_power+0x165): undefined reference to
> `acpi_power_transition'
>
> Does this imply that I'm required to include a full ACPI setup in my
> kernels? I'm trying to prevent all attempts to switch my monitor off, but
> without a lot of success so far. (If I haven't told the machine to do
> something, it shouldn't do it; and I want control over my peripherals
> thank-you-very-much.)

ACPI and switching the monitor off are two completly unrelated problems. ACPI 
is not evil - and some machines even need ACPI to boot. There is no reason to 
not have ACPI in the kernel. And monitor - that is something for your 
xorg.conf and xset.
-- 
gentoo-amd64@gentoo.org mailing list

[gentoo-amd64] Re: Kernel 2.6.21 problem

[Duncan]

"Hemmann, Volker Armin" <volker.armin.hemmann@tu-clausthal.de> posted
200704291349.08002.volker.armin.hemmann@tu-clausthal.de, excerpted below,
on  Sun, 29 Apr 2007 13:49:07 +0200:

> ACPI and switching the monitor off are two completly unrelated problems.
> ACPI is not evil - and some machines even need ACPI to boot. There is no
> reason to not have ACPI in the kernel. And monitor - that is something
> for your xorg.conf and xset.

Well, xorg.conf/xset don't help in console mode.  At least here, 
something does monitor blanking (it doesn't actually turn off) in console 
mode if I leave it sit.  It's not /that/ big a deal here, but it'd be 
nice to at least know what's responsible and where the config for it is.

You're right about ACPI, however.  IIRC I had it compiled out of my 
kernel entirely, back in the 2.4 era, but that's been awhile, and a lot 
of machines, particularly laptops I've read, need it to function properly.

That said, if ACPI is in the kernel, IIRC it's possible to disable ACPI 
from the kernel command line if necessary.  See ($KERNELDIR)/
Documentation/x86_64/boot-options.txt for a whole list of interesting and 
potentially useful options.

@PH:  I'm running .21 and have run the -rcs and some gits before that, 
and have only partial ACPI enabled here, without issue.  However, my 
machine being a dual socket-940 Opteron server targeted system, only a 
limited subset of the various ACPI components (sleep-statess, processor, 
button, *NUMA*, which I'm using for CPU detection, and power management 
timer, that's about it) are even supported on my hardware.  I don't have 
the stuff like fan and thermal zone enabled as they aren't supported 
anyway, but it compiled and is running fine.

Anyway, I'd suggest filing a bug, as if symbols are missing due to unmet 
dependencies, the menu config needs to be changed such that it doesn't 
allow that option, either turning on the dependencies along with what 
depends on them or disabling the depended option unless the dependencies 
are met.  Alternatively, if it's a mistake due to code refactoring, it 
needs to be corrected, and the way to bring it to their attention is via 
bug.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Re: Kernel 2.6.21 problem

[Wil Reichert]

> Well, xorg.conf/xset don't help in console mode.  At least here,
> something does monitor blanking (it doesn't actually turn off) in console
> mode if I leave it sit.  It's not /that/ big a deal here, but it'd be
> nice to at least know what's responsible and where the config for it is.

setterm?

So far as I can tell, having ACPI enabled these days is generally a
good thing.  If theres a problem with screens blanking when you don't
want them to then its most likely something else thats mis-configured.
 At the same time, the kernel should still compile without ACPI
support.

Wil
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Kernel 2.6.21 problem

[Peter Humphrey]

On Sunday 29 April 2007 12:49:07 Hemmann, Volker Armin wrote:

> ACPI and switching the monitor off are two completly unrelated problems.
> ACPI is not evil - and some machines even need ACPI to boot. There is no
> reason to not have ACPI in the kernel.

I find it all somewhat confusing. APM, ACPI, others...

> And monitor - that is something for your xorg.conf and xset.

Before I took dpms out of xorg.conf the monitor was being blanked after 
about five minutes. If kdm was still showing its login screen at that time 
the machine would lock solid. Having removed dpms the period is now much 
longer, but I still get a blank screen eventually.

-- 
Rgds
Peter Humphrey
Linux Counter 5290, Aug 93
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Kernel 2.6.21 problem

[Hemmann, Volker Armin]

On Montag, 30. April 2007, Peter Humphrey wrote:
> On Sunday 29 April 2007 12:49:07 Hemmann, Volker Armin wrote:
> > ACPI and switching the monitor off are two completly unrelated problems.
> > ACPI is not evil - and some machines even need ACPI to boot. There is no
> > reason to not have ACPI in the kernel.
>
> I find it all somewhat confusing. APM, ACPI, others...

there are no 'others' and APM is completly obsolete for years. ACPI on the 
other hand is needed for a lot more than powersaving.

>
> > And monitor - that is something for your xorg.conf and xset.
>
> Before I took dpms out of xorg.conf the monitor was being blanked after
> about five minutes. If kdm was still showing its login screen at that time
> the machine would lock solid. Having removed dpms the period is now much
> longer, but I still get a blank screen eventually.

xset is your friend.
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Kernel 2.6.21 problem

[Mark Knecht]

On 4/29/07, Peter Humphrey <prh@gotadsl.co.uk> wrote:
> On syncing today I was offered gentoo-sources-2.6.21. On attempting to
> compile it I get:
>
> drivers/built-in.o: In function `acpi_rtc_init':
> glue.c:(.init.text+0x34c2): undefined reference to `pnp_bus_type'
> drivers/built-in.o: In function `acpi_bus_generate_event':
> (.text.acpi_bus_generate_event+0x25): undefined reference to `event_is_open'
> drivers/built-in.o: In function `acpi_bus_get_power':
> (.text.acpi_bus_get_power+0x97): undefined reference to
> `acpi_power_get_inferred_state'
> drivers/built-in.o: In function `acpi_bus_set_power':
> (.text.acpi_bus_set_power+0xf3): undefined reference to
> `acpi_power_transition'
> drivers/built-in.o: In function `acpi_bus_set_power':
> (.text.acpi_bus_set_power+0x165): undefined reference to
> `acpi_power_transition'
>
> Does this imply that I'm required to include a full ACPI setup in my
> kernels? I'm trying to prevent all attempts to switch my monitor off, but
> without a lot of success so far. (If I haven't told the machine to do
> something, it shouldn't do it; and I want control over my peripherals
> thank-you-very-much.)
>
> My kernel-compilation method is as follows:
> 1.      Emerge -u gentoo-sources
> 2.      Set /usr/src/linux manually (I know there is an automatic method, but
> that kills any package that's recompiled with every kernel version)
> 3.      Copy the previous .config from its store on /boot
> 4.      Make oldconfig and set any new options, making a note of any that need
> further exploration
> 5.      If any such notes, make menuconfig and read help; change any settings
> that need it
> 6.      Make && make modules_install
> 7.      Copy new .config and new kernel to /boot
> 8.      Change /boot/grub/grub.conf to choose the new kernel by default
> 9.      Reboot no-X and remerge nvidia-drivers and vmware-server
> 10.     Reboot into KDE and check the system over
> 11.     Celebrate with a pint.
>
> --
> Rgds
> Peter Humphrey
> Linux Counter 5290, Aug 93

Hi Peter,
   I've been running rt-sources-2.6.21-rc5-rt12 for a few weeks with
no problems but likely I didn't have the same config options as you.
rt-sources-2.6.21-r1 is out now so I should really upgrade.

   Our build processes are pretty much identical except I would build
the video driver without the reboot. seems to work as long as the
/usr/src/linux link is in place. Maybe I shouldn't be doing that?

   Before filing a bug report I'd suggest building 2.6.21 from
kernel.org instead of gentoo-sources to see if it's a kernel problem
or something introduced by our developers.

Good luck,
Mark
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Kernel 2.6.21 problem

[William Tetrault]

Mark Knecht wrote:
> On 4/29/07, Peter Humphrey <prh@gotadsl.co.uk> wrote:
>> On syncing today I was offered gentoo-sources-2.6.21. On attempting to
>> compile it I get:

snip

>> Rgds
>> Peter Humphrey
>> Linux Counter 5290, Aug 93
> 
> Hi Peter,

snip

>   Before filing a bug report I'd suggest building 2.6.21 from
> kernel.org instead of gentoo-sources to see if it's a kernel problem
> or something introduced by our developers.
> 
> Good luck,
> Mark

Use vanilla-sources rather than current from kernel.org.  It's the same
thing and allows you to use portage.

Bill
-- 
gentoo-amd64@gentoo.org mailing list