Re: modules, "modules" and CONFIG_LIST_SORT

[Randy Dunlap <randy.dunlap@oracle.com>]

On 03/07/10 03:23, Linus Torvalds wrote:
> 
> 
> On Sun, 7 Mar 2010, Alexey Dobriyan wrote:
>>
>> Unpleasant side effect of the change is that some modules stop being
>> true modules, i. e. admin is unable to start using them without reboot
>> if kernel was compiled without that tiny amount of core kernel.
>>
>> Having used this feature several times, I think it'd be correct
>> to preserve this behaviour, at least not regress for those modules
>> which benefitted from it. For modules which were always "modules" (ipv6)
>> it's fine to continue.
>>
>> Can we declare some policy about it?
>>
>> And revert LIST_SORT commit if yes.
> 
> Yeah, I think that in cases like this, you have a very good argument: 
> LIST_SORT enables code that isn't that large, and is clearly very generic.
> 
> And changing the config later and trying to compile and install a module 
> is rather sane. And if that new module needs LIST_SORT, you're screwed 
> because it didn't get compiled in originally.
> 
> Honestly, personally I'd rather have a real library that modules can link 
> to _before_ even loading into kernel space, but that's not how we've 
> traditionally done things. So I guess we should just revert that commit.

xfs also needs "select LIST_SORT".  I posted a patch for that a few days
ago and now Christoph Hellwig has asked me to send the patch directly to Linus,
but if Linus is going to revert the 'config LIST_SORT' patch, I'll skip it.

-- 
~Randy