[uClinux-dev] nios2: apparent I2C_GPIO-related patch error

[Robert P. J. Day]

  i'm almost at the point where i can build a kernel with:

1) latest CVS checkout of uClinux-dist
2) latest git pull of kernel source tree
3) minimal nios2-related patchset

  the only glitch is the following error when running "make
vendor_hwselect ..."

...
drivers/i2c/Kconfig:6:error: found recursive dependency: I2C -> I2C_GPIO -> I2C
...

  this *appears* to be because the most recent uclinux patch (the big
patch for 2.6.23) includes the following patch snippet:

diff -Naurp linux-2.6.23/arch/nios2nommu/drivers/Kconfig linux-2.6.23-uc0/arch/nios2nommu/drivers/Kconfig
--- linux-2.6.23/arch/nios2nommu/drivers/Kconfig        1970-01-01 10:00:00.000000000 +1000
+++ linux-2.6.23-uc0/arch/nios2nommu/drivers/Kconfig    2007-10-19 11:24:52.000000000 +1000
@@ -0,0 +1,45 @@
...
+config I2C_GPIO
+       tristate "GPIO-Based I2C Interface"
+       default N
+       select I2C
+       select I2C_ALGOBIT
+       help
+         Say Y here if you use GPIO lines for an I2C bus.
+
...

  however, the Kconfig variable I2C_GPIO is *already* defined in the
stock kernel file drivers/i2c/busses/Kconfig:

config I2C_GPIO
        tristate "GPIO-based bitbanging I2C"
        depends on GENERIC_GPIO
        select I2C_ALGOBIT
        help
          This is a very simple bitbanging I2C driver utilizing the
          arch-neutral GPIO API to control the SCL and SDA lines.

  so it's not clear why it's being *redefined* by the uclinux patch
when it already exists, which is apparently what's causing the
circular definition error.

  thoughts?  is there something subtle here i'm not appreciating?

rday
-- 
========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry
Waterloo, Ontario, CANADA

http://crashcourse.ca
========================================================================