Re: [PATCH] mm: Fix dirty_bytes/dirty_background_bytes sysctls on 64bit arches

[Peter Zijlstra <peterz@infradead.org>]

On Sun, 2009-02-01 at 02:22 +0100, Sven Wegener wrote:
> We need to pass an unsigned long as the minimum, because it gets casted
> to an unsigned long in the sysctl handler. If we pass an int, we'll
> access four more bytes on 64bit arches, resulting in a random minimum
> value.

If that's so, how can any of those other limit values still be good?

> Signed-off-by: Sven Wegener <sven.wegener@stealer.net>
> ---
>  kernel/sysctl.c |    5 +++--
>  1 files changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> index 790f9d7..c5ef44f 100644
> --- a/kernel/sysctl.c
> +++ b/kernel/sysctl.c
> @@ -101,6 +101,7 @@ static int two = 2;
>  
>  static int zero;
>  static int one = 1;
> +static unsigned long one_ul = 1;
>  static int one_hundred = 100;
>  
>  /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
> @@ -974,7 +975,7 @@ static struct ctl_table vm_table[] = {
>  		.mode		= 0644,
>  		.proc_handler	= &dirty_background_bytes_handler,
>  		.strategy	= &sysctl_intvec,
> -		.extra1		= &one,
> +		.extra1		= &one_ul,
>  	},
>  	{
>  		.ctl_name	= VM_DIRTY_RATIO,
> @@ -995,7 +996,7 @@ static struct ctl_table vm_table[] = {
>  		.mode		= 0644,
>  		.proc_handler	= &dirty_bytes_handler,
>  		.strategy	= &sysctl_intvec,
> -		.extra1		= &one,
> +		.extra1		= &one_ul,
>  	},
>  	{
>  		.procname	= "dirty_writeback_centisecs",