why is CONFIG_PCI unset ? [was: build #350 issue for v2.6.25-rc1 in drivers/i2c/busses/i2c-i801.c]

At Friday 22 February 2008 Sam Ravnborg wrote :

On Fri, Feb 22, 2008 at 09:43:36PM +0100, Toralf Förster wrote:

quoted

Hello,

pls see below for an issue of KConfig which happens for isntance if randconfig selects
CONFIG_INPUT_APANEL.

Which select a symbol with dependencies => boom!

Select is evil. We need 'requires' that could be used by the user to say "I want this".

[Comment from Randy] 

quoted

Yes, I see this one often.  I wish someone were willing to try
to fix it.  I've looked at the source code, but I'm not willing to
dive deeply into it.

I hope Roman suddenly fixes it.

	Sam

Hello,

a _lot_ of build tests of my (compile) test script failed with:

...
  LD      drivers/i2c/algos/built-in.o
  CC      drivers/i2c/busses/i2c-i801.o
drivers/i2c/busses/i2c-i801.c: In function 'i801_probe':
drivers/i2c/busses/i2c-i801.c:642: error: implicit declaration of function 'pci_request_region'
drivers/i2c/busses/i2c-i801.c:682: error: implicit declaration of function 'pci_release_region'
make[3]: *** [drivers/i2c/busses/i2c-i801.o] Error 1
make[2]: *** [drivers/i2c/busses] Error 2
make[1]: *** [drivers/i2c] Error 2
make: *** [drivers] Error 2

I hard coded my script to prepend these 2 lines
CONFIG_PCI=y
CONFIG_I2C=y
at top of the each .config file before I run "make oldconfig" however the
resulting config CONFIG_PCI is very often unset.

I'm wondering which kernel config I've to set in addition to the 2 lines to
circumvent the current situation until kbuild handles such case correctly. Most
of the reported build failures results from the issue seen above which makes
the current test runs nearly useless :-(
 

I attached a config file as an example.


-- 
MfG/Sincerely

Toralf Förster
pgp finger print: 7B1A 07F4 EC82 0F90 D4C2 8936 872A E508 7DB6 9DA3