Re: [PATCH 02/36] usercopy: Include offset in overflow report

[Kees Cook]

On Wed, Jan 10, 2018 at 7:25 AM, Christopher Lameter <cl@xxxxxxxxx> wrote:
> On Tue, 9 Jan 2018, Kees Cook wrote:
>
>> -static void report_usercopy(unsigned long len, bool to_user, const char *type)
>> +int report_usercopy(const char *name, const char *detail, bool to_user,
>> +                 unsigned long offset, unsigned long len)
>>  {
>> -     pr_emerg("kernel memory %s attempt detected %s '%s' (%lu bytes)\n",
>> +     pr_emerg("kernel memory %s attempt detected %s %s%s%s%s (offset %lu, size %lu)\n",
>>               to_user ? "exposure" : "overwrite",
>> -             to_user ? "from" : "to", type ? : "unknown", len);
>> +             to_user ? "from" : "to",
>> +             name ? : "unknown?!",
>> +             detail ? " '" : "", detail ? : "", detail ? "'" : "",
>> +             offset, len);
>>       /*
>>        * For greater effect, it would be nice to do do_group_exit(),
>>        * but BUG() actually hooks all the lock-breaking and per-arch
>>        * Oops code, so that is used here instead.
>>        */
>>       BUG();
>
> Should this be a WARN() or so? Or some configuration that changes
> BUG() behavior? Otherwise

This BUG() is the existing behavior, with the new behavior taking the
WARN() route in a following patch.

>> +
>> +     return -1;
>
> This return code will never be returned.
>
> Why a return code at all? Maybe I will see that in the following patches?

I was trying to simplify the callers, but I agree, the result is
rather ugly. I'll see if I can fix this up.

-Kees

-- 
Kees Cook
Pixel Security