Thanks Nick. will try out your suggestions.

On Tuesday, October 13, 2015 at 1:54:47 PM UTC-5, Nick Kralevich wrote:
>
>
> This was covered on the seandroid mailing list. Please see 
> https://www.mail-archive.com/[email protected]/msg02418.html
>
> A few things you can do:
>
> 1) Patch the kernel to not attempt module loading for alternate execution 
> environments. This code never worked in the kernel anyway, so no harm in 
> pulling it. Something like:
>
> diff --git a/kernel/exec_domain.c b/kernel/exec_domain.c
> index 0dbeae3..a66ea19 100644
> --- a/kernel/exec_domain.c
> +++ b/kernel/exec_domain.c
> @@ -68,7 +68,8 @@
>                                 goto out;
>         }
>  
> -#ifdef CONFIG_MODULES
> +
> +#if 0
>         read_unlock(&exec_domains_lock);
>         request_module("personality-%d", pers);
>         read_lock(&exec_domains_lock);
>
> 2) Add a dontaudit SELinux rule to suppress these:
>
> Add the following to your device specific policy:
>
>   dontaudit domain kernel:system module_request;
>
> However, please read 
> https://www.mail-archive.com/[email protected]/msg02433.html 
> before doing so.
>
> 3) Backport the patches from 
> https://www.mail-archive.com/[email protected]/msg02418.html
>
> -- Nick
>
> On Monday, October 12, 2015 at 8:55:28 PM UTC-7, Vishal Mahaveer wrote:
>>
>> Hi,
>>
>>
>> In Android-M, below change in bionic was introduced
>>
>> https://android-review.googlesource.com/122131
>>
>>  
>>
>> Which caused lot of SELinux denials to be logged regarding module_request 
>> (similar to below one) for 32bit systems, There is quite a lot of 
>> module_request denials when any 32bit executable is executed.
>>
>>  
>>
>> type=1400 audit(946685139.719:12): avc: denied { module_request } for 
>> pid=1970 comm="tzdatacheck" kmod="personality-8" 
>> scontext=u:r:tzdatacheck:s0 tcontext=u:r:kernel:s0 tclass=system 
>> permissive=1
>>
>>  
>>
>>
>> The issue is discussed here
>>
>> http://marc.info/?t=144190478400004&r=1&w=2
>>
>>
>>
>> For kernel version < 4.0, is there any resolution available for resolving 
>> these denials with out having to backport the removal of execution domain 
>> patches from kernel 4.0?
>>
>> Is adding a dontaudit for kernel module_request good option?
>>
>>
>> In our system disabling modules (CONFIG_MODULES=n) is currently not an 
>> option.
>>
>>
>>  
>>
>> Regards,
>>
>> Vishal
>>
>

-- 
-- 
unsubscribe: [email protected]
website: http://groups.google.com/group/android-porting

--- 
You received this message because you are subscribed to the Google Groups 
"android-porting" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to