Re: [PATCH v2 1/4] syscalls: Restore address limit after a syscall

[Thomas Garnier <thgarnie@google.com>]

On Thu, Mar 9, 2017 at 9:27 AM, Andy Lutomirski <luto@amacapital.net> wrote:
> On Thu, Mar 9, 2017 at 12:42 AM, Borislav Petkov <bp@alien8.de> wrote:
>> On Wed, Mar 08, 2017 at 05:24:53PM -0800, Thomas Garnier wrote:
>>> @@ -191,6 +191,22 @@ extern struct trace_event_functions exit_syscall_print_funcs;
>>>       SYSCALL_METADATA(sname, x, __VA_ARGS__)                 \
>>>       __SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
>>>
>>> +asmlinkage void verify_pre_usermode_state(void);
>>> +
>>> +#ifndef CONFIG_ARCH_NO_SYSCALL_VERIFY_PRE_USERMODE_STATE
>>> +static inline bool has_user_ds(void) {
>>> +     bool ret = segment_eq(get_fs(), USER_DS);
>>> +     // Prevent re-ordering the call
>>
>> This is not the kernel comments style. Use /* */ instead.
>>
>>> +     barrier();
>>> +     return ret;
>>> +}
>>> +#else
>>> +static inline bool has_user_ds(void) {
>>> +     return false;
>>> +}
>>> +#endif
>>
>> ... and then you could slim down the ifdeffery a bit:
>>
>> static inline bool has_user_ds(void) {
>>         bool ret = false;
>>
>> #ifndef CONFIG_ARCH_NO_SYSCALL_VERIFY_PRE_USERMODE_STATE
>>         ret = segment_eq(get_fs(), USER_DS);
>>         /* Prevent re-ordering the call. */
>>         barrier();
>> #endif
>>
>>         return ret;
>> }
>
> I don't like having any kernel configuration in which has_user_ds()
> unconditionally returns false.  Can we put the ifdeffery in the caller
> instead?

I don't like it either to be honest. We could have something like:

#ifndef CONFIG_ARCH_NO_SYSCALL_VERIFY_PRE_USERMODE_STATE
#define CHECK_USER_CALLER(_x) bool _x = segment_eq(get_fs(), USER_DS)
#else
#define CHECK_USER_CALLER(_x) bool _x = false
#endif

// In the syscall macro:
CHECK_CALLED_BY_USER(user_caller);

>
> --Andy



-- 
Thomas