Messages in this thread | ![/](/images/icornerl.gif) | | Date | Tue, 30 Nov 2010 16:05:27 +0000 | From | Russell King - ARM Linux <> | Subject | Re: [PATCH] ARM: Gemini: Add support for PCI Bus |
| |
On Mon, Nov 29, 2010 at 04:57:53PM +0100, Arnd Bergmann wrote: > On Monday 29 November 2010, Hans Ulli Kroll wrote: > > @@ -1164,7 +1169,7 @@ config ISA_DMA_API > > bool > > > > config PCI > > - bool "PCI support" if ARCH_INTEGRATOR_AP || ARCH_VERSATILE_PB || > > ARCH_IXP4XX || ARCH_KS8695 || MACH_ARMCORE || ARCH_CNS3XXX > > + bool "PCI support" if MIGHT_HAVE_PCI > > help > > Find out whether you have a PCI motherboard. PCI is the name of > > a > > bus system, i.e. the way the CPU talks to the other stuff inside > > This does not solve the problem that Russell mentioned: existing platforms > select PCI unconditionally, e.g. Iop13XX, some IXP, Orion, Shark > and more. At the very least, these would need to also select MIGHT_HAVE_PCI > to avoid the warning.
No, the above is correct. What I was talking about was the difference between these:
config PCI bool "PCI support" if MIGHT_HAVE_PCI
and
config PCI bool "PCI support" depends on MIGHT_HAVE_PCI
In the first instance, PCI itself does not depend on MIGHT_HAVE_PCI. MIGHT_HAVE_PCI controls whether the user is offered the "PCI support" option.
In the second instance, PCI depends on MIGHT_HAVE_PCI, which must be set to 'y' to offer the option _and_ also if PCI is selected.
We want the first behaviour. Platforms which must have PCI support can continue to select PCI as they currently do, and leave the MIGHT_HAVE_PCI option alone.
Platforms which may optionally have PCI support should select MIGHT_HAVE_PCI instead.
| ![\](/images/icornerr.gif) |