On Wed, 2014-01-22 at 14:33 -0500, Paul Gortmaker wrote: > On 14-01-22 02:20 PM, James Bottomley wrote: > > On Wed, 2014-01-22 at 13:19 -0500, Paul Gortmaker wrote: > >> The pdc_cons.c code is always built in. It will never be modular, > >> so using module_init as an alias for __initcall is rather > >> misleading. > >> > >> Fix this up now, so that we can relocate module_init from > >> init.h into module.h in the future. If we don't do this, we'd > >> have to add module.h to obviously non-modular code, and that > >> would be a worse thing. > > > > I don't buy this. We've already had an argument about using MODULE_ > > tags in non-modular code here: > > > > http://marc.info/?t=138947344500007 > > > > The consensus was that we'd continue to do so, so that would seem to > > invalidate the rationale for doing this patch set. > > > > Without the rationale, this is churn for no gain, so I'd rather not do > > it. > > Hi James, > > Thanks for the link. Here is another one. > > https://lkml.org/lkml/2014/1/21/434 > > In there I explain the rationale for doing this and what value add we > get from it. Perhaps you'll say those things aren't important, and > then I guess we'll have to agree to disagree then... but I hope not. Well, OK, so 1 is covered by the link I sent. 3 is an EDONTCARE because we don't care when the console is initialised as long as it is. Your reason 2 I'm not sure I understand. module_exit() functions are automatically discarded by the linker scripts ... if we have a problem with reference to code in them it will cause a link failure. I'm grateful for dumping all the spurious section mismatches caused by HOTPLUG, I really am, but I'm seeing less benefit to this code churn. James -- To unsubscribe from this list: send the line "unsubscribe linux-parisc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
