| Message ID | 1390414784-6472-1-git-send-email-paul.gortmaker@windriver.com (mailing list archive) |
|---|---|
| State | Awaiting Upstream |
| Headers | show |
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. Thanks, James -- To unsubscribe from this list: send the line "unsubscribe linux-parisc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
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. Thanks, Paul. -- > > Thanks, > > James > > -- To unsubscribe from this list: send the line "unsubscribe linux-parisc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
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@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/arch/parisc/kernel/pdc_cons.c b/arch/parisc/kernel/pdc_cons.c index d5cae55195ec..10a5ae9553fd 100644 --- a/arch/parisc/kernel/pdc_cons.c +++ b/arch/parisc/kernel/pdc_cons.c @@ -207,8 +207,7 @@ static int __init pdc_console_tty_driver_init(void) return 0; } - -module_init(pdc_console_tty_driver_init); +device_initcall(pdc_console_tty_driver_init); static struct tty_driver * pdc_console_device (struct console *c, int *index) {
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. Direct use of __initcall is discouraged, vs prioritized ones. Use of device_initcall is consistent with what __initcall maps onto, and hence does not change the init order, making the impact of this change zero. Should someone with real hardware for boot testing want to change it later to arch_initcall or something different, they can do that at a later date. Reported-by: kbuild test robot <fengguang.wu@intel.com> Cc: "James E.J. Bottomley" <jejb@parisc-linux.org> Cc: Helge Deller <deller@gmx.de> Cc: linux-parisc@vger.kernel.org Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> --- [patch will be added to init cleanup series: http://git.kernel.org/cgit/linux/kernel/git/paulg/init.git/ ] arch/parisc/kernel/pdc_cons.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)