Re: [PATCH] power: disable hibernation if module loading is disabled

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Friday, March 04, 2011, Vasiliy Kulikov wrote:
> If /proc/sys/kernel/modules_disabled is set to 1, then nobody (even full
> root) may not read/write arbitrary kernel memory.  In spite of it,
> hibernation allows anyone with an access to either /dev/snapshot or
> /sys/power/ make the full snapshot of the system.  This snapshot may be
> freely changed and uploaded back.
> 
> Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>

That "everyone" is actually the "full root" (in the case of /sys/power/state)
or someone having CAP_SYS_ADMIN in the /dev/snapshot case, right?

So the changelog is misleading and please fix it.

Second, there's _zero_ relationship between /proc/sys/kernel/modules_disabled
and the hibernation interface, so please find a different way to solve the
problem (if there is any).

Thanks,
Rafael


> ---
>  kernel/power/hibernate.c |    6 ++++++
>  1 files changed, 6 insertions(+), 0 deletions(-)
> 
> diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c
> index 1832bd2..1ac9eee 100644
> --- a/kernel/power/hibernate.c
> +++ b/kernel/power/hibernate.c
> @@ -328,6 +328,9 @@ int hibernation_snapshot(int platform_mode)
>  {
>  	int error;
>  
> +	if (modules_disabled)
> +		return -EPERM;
> +
>  	error = platform_begin(platform_mode);
>  	if (error)
>  		goto Close;
> @@ -385,6 +388,9 @@ static int resume_target_kernel(bool platform_mode)
>  {
>  	int error;
>  
> +	if (modules_disabled)
> +		return -EPERM;
> +
>  	error = dpm_suspend_noirq(PMSG_QUIESCE);
>  	if (error) {
>  		printk(KERN_ERR "PM: Some devices failed to power down, "
>