Re: [PATCH -mm2] net: Fix compiler-error on dgrs.c when !CONFIG_PCI

[Richard Knutsson <ricknu-0@student.ltu.se>]

Herbert Xu wrote:

>On Mon, Nov 21, 2005 at 01:52:49PM +0100, Richard Knutsson wrote:
>  
>
>>This patch requirer the 
>>"net-fix-compiler-error-on-dgrsc-when-config_pci.patch" (added to the 
>>-mm tree after 2.6.15-rc1-mm2):
>>
>>--- 
>>devel/drivers/net/dgrs.c~net-fix-compiler-error-on-dgrsc-when-config_pci 
>>2005-11-19 18:00:34.000000000 -0800
>>+++ devel-akpm/drivers/net/dgrs.c	2005-11-19 18:00:34.000000000 -0800
>>@@ -1458,6 +1458,8 @@ static struct pci_driver dgrs_pci_driver
>>	.probe = dgrs_pci_probe,
>>	.remove = __devexit_p(dgrs_pci_remove),
>>};
>>+#else
>>+static struct pci_driver dgrs_pci_driver = {};
>>#endif
>>    
>>
>
>I don't see the point.  We shouldn't have this structure at all
>if CONFIG_PCI is not set.
>
>Cheers,
>  
>
Opps, misread your mail. Sorry.
But in that case, why shall we have any pci_*-function in the first 
place when !CONFIG_PCI? As it was before, they were contained within 
#ifdef CONFIG_PCI's.
You said your patch were easier to read, please elaborate. Yes, in the 
dgrs_init_module() (no arguments there), but you introduces new 
functions (who need to be checked out if you read the code for the first 
time) and is really #ifdef's a good idea to change function behavior? 
Isn't better to change the input? (I know, linux/pci.h does it, but at 
least that is in a .h-file with inline functions containing at most 
"return 0;").

Btw, you said Jeff should decide. Why not Rich Richardson who is the author?

Apologize for the ranting.
As I said before, thinking of deleting CONFIG_PCI's containing 
pci_*-functions and if so, need a valid plan for it (because of the 
pci_driver-struct). If there is any no-no in "my" way, please point it 
out to spare me/you/lkml the patches. If not for this, I would've let 
this rest and leave it to Jeff (or Rich ;) ).

cu,
Richard