[KJ] This was the last of my pci __devinitdata patches!

[KJ] This was the last of my pci __devinitdata patches!

[Henne]

So, this was the first set of my patches. :)
Things I learned:
-	The default linelength of thunderbird (72 chars) are not enough for patches and configuring the mailclient can save a lot of time.

-	There are many pci-drivers which are not conform in all points to
	<kernel>/Documentation/pci.txt
  	e.g. : if the remove funcion of the driver is marked as __devexit,
		the __devexit_p() is forgotten in struct pci_driver,
		which can leave a pointer to a function which doesn't exist. (evil)

-	Bringing more ifdefs into the kernel is evil too.

-	Typing 286 (what a number) times "__devinitdata" could have benn much more boring.

Questions I have:

-	What is the prefered way to make powemanagement-stuff depend on CONFIG_PM?
	ifdef the .suspend and .resume in struct pci_driver and gather these functions into ONE
	#ifdef CONFIG_PM
	#endif
	?
	There are lots of drivers who have there PM-functions compiled, even if no CONFIG_PM is set.

-	Will these patches be put into the mainline-kernel by you or must every janitor send them
	again to the subsys-maintainer?

Greets,
Henne
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] This was the last of my pci __devinitdata patches!

[Randy.Dunlap]

[-- Attachment #1: Type: text/plain, Size: 1127 bytes --]

On Wed, 26 Apr 2006 22:18:54 +0200 Henne wrote:

> So, this was the first set of my patches. :)
> Things I learned:
> -	The default linelength of thunderbird (72 chars) are not enough for patches and configuring the mailclient can save a lot of time.

Uh, yes, that's well-known.  :(
I have written email about how I can use tbird successfully, or
you can try this web page, although our recommendations are different:

http://mbligh.org/linuxdocs/Email/Clients/Thunderbird

> -	There are many pci-drivers which are not conform in all points to
> 	<kernel>/Documentation/pci.txt
>   	e.g. : if the remove funcion of the driver is marked as __devexit,
> 		the __devexit_p() is forgotten in struct pci_driver,
> 		which can leave a pointer to a function which doesn't exist. (evil)

Make a .config file with CONFIG_HOTPLUG=n (tricky[1]) and then build
a kernel.  The modpost phase will check/catch/warn about these.
I have sent patches for quite a few of them recently.


> -	Bringing more ifdefs into the kernel is evil too.
> 
> -	Typing 286 (what a number) times "__devinitdata" could have benn much more boring.


---
~Randy

[-- Attachment #2: Type: text/plain, Size: 168 bytes --]

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors