Re: [PATCH 28/28] Blackfin SPI: init early

[Grant Likely <grant.likely-s3s/WqlpOiPyB63q8FvJNQ@public.gmane.org>]

On Sun, Oct 17, 2010 at 06:59:41PM -0400, Mike Frysinger wrote:
> From: Michael Hennerich <michael.hennerich-OyLXuOCK7orQT0dZR+AlfA@public.gmane.org>
> 
> Some systems using this bus sometimes have very basic devices on them
> such as regulators.  So we need to be loaded even earlier in case the
> devices are used before userspace is loaded.  Therefore register in
> subsys_initcall().

This shouldn't have anything to do with userspace since regardless of
the initcall, all of this setup should be complete before userspace
starts.  Rather I expect it has to do with an earlier initcall wanting
to have the spi bus up earlier than module_initcall.  The description
of this patch seems wrong.

On that note, this is something that needs to be revisited.  Needing
to play these kind of games highlights a failure of the device model,
or the way that it is being used.  (The patch is okay; but it does
highlight some ugliness in the current kernel)

g.

> 
> Signed-off-by: Michael Hennerich <michael.hennerich-OyLXuOCK7orQT0dZR+AlfA@public.gmane.org>
> Signed-off-by: Mike Frysinger <vapier-aBrp7R+bbdUdnm+yROfE0A@public.gmane.org>
> ---
>  drivers/spi/spi_bfin5xx.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/spi/spi_bfin5xx.c b/drivers/spi/spi_bfin5xx.c
> index bd18214..d17ab37 100644
> --- a/drivers/spi/spi_bfin5xx.c
> +++ b/drivers/spi/spi_bfin5xx.c
> @@ -1466,7 +1466,7 @@ static int __init bfin_spi_init(void)
>  {
>  	return platform_driver_probe(&bfin_spi_driver, bfin_spi_probe);
>  }
> -module_init(bfin_spi_init);
> +subsys_initcall(bfin_spi_init);
>  
>  static void __exit bfin_spi_exit(void)
>  {
> -- 
> 1.7.3.1
>