kernel compilation fails

[Ben Hsu]

Hello

In the kernel source package for severn beta2 (2.4.22-1.2061) compiling
the kernel with the following commands--

  552  cp configs/kernel-2.4.22-i686.config ./.config
  553  make oldconfig
  554  make dep bzImage

-- fails with the following error message-- 
kernel/kernel.o(.text+0xf9f): In function `schedule':
: undefined reference to `active_load_balance'

[Robert P. J. Day]

On Sat, 27 Sep 2003, Ben Hsu wrote:

>  Hello
>  
>  In the kernel source package for severn beta2 (2.4.22-1.2061) compiling
>  the kernel with the following commands--
>  
>    552  cp configs/kernel-2.4.22-i686.config ./.config
>    553  make oldconfig
>    554  make dep bzImage
>  
>  -- fails with the following error message-- 
>  kernel/kernel.o(.text+0xf9f): In function `schedule':
>  : undefined reference to `active_load_balance' 
unlikely to fix it, but did you start with a "make mrproper", just
to play it safe?

rday

[Ben Hsu]

On Sat, 2003-09-27 at 08:49, Robert P. J. Day wrote:

>  unlikely to fix it, but did you start with a "make
> mrproper", just
>  to play it safe?
>   
Yes.

It looks like the problem is that in kernel.c the definition of
"active_load_balance" is surrounded by #if CONFIG_SHARE_RUNQUEUE, but
the actual call to that function on line 1712 is not in such an #if.

Also, some weirdness when debugging this:

cp configs/kernel-2.4.22-i686.config ./.config
grep CONFIG_SHARE_RUNQUEUE .config #nothing
make oldconfig
grep CONFIG_SHARE_RUNQUEUE .config #set to "y"
make xconfig
<save without making any changes>
grep CONFIG_SHARE_RUNQUEUE .config #nothing again!

The same happens for "make menuconfig"

[Ben Hsu]

On Sat, 2003-09-27 at 10:12, Ben Hsu wrote:

>  
>  It looks like the problem is that in kernel.c the definition of
>  "active_load_balance" is surrounded by #if CONFIG_SHARE_RUNQUEUE, but
>  the actual call to that function on line 1712 is not in such an #if.
>   
I tracked down the problem that the definition of the
active_load_rebalance on line 1321 is inside of an #ifdef CONFIG_SMP on
1087

I compiled with SMP enabled and it worked, is it safe for uniprocessor
machines to run with SMP enabled?

[Michael K. Johnson]

On Sat, Sep 27, 2003 at 06:02:25PM -0700, Ben Hsu wrote:
>  On Sat, 2003-09-27 at 10:12, Ben Hsu wrote:
>  
>  > 
>  > It looks like the problem is that in kernel.c the definition of
>  > "active_load_balance" is surrounded by #if CONFIG_SHARE_RUNQUEUE, but
>  > the actual call to that function on line 1712 is not in such an #if.
>  > 
>  I tracked down the problem that the definition of the
>  active_load_rebalance on line 1321 is inside of an #ifdef CONFIG_SMP on
>  1087
>  
>  I compiled with SMP enabled and it worked, is it safe for uniprocessor
>  machines to run with SMP enabled? 
Sometimes, even usually.  :-)  (If it boots, it should be OK; there are
some UP machines that won't boot the SMP kernel)

Just a trifle slower...

What's the bugzilla number for this one?

michaelkjohnson

 "He that composes himself is wiser than he that composes a book."
 Linux Application Development                     -- Ben Franklin
 http://people.redhat.com/johnsonm/lad/

[Ben Hsu]

On Mon, 2003-09-29 at 17:39, Michael K. Johnson wrote:

>  What's the bugzilla number for this one?
>   
105978