Re: [PSEUDOPATCH] rename is_compat_task

[Andy Lutomirski]

On Mon, Dec 7, 2015 at 9:15 PM, Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote:
> On Tue, Dec 08, 2015 at 06:01:48AM +0100, Ingo Molnar wrote:
>
>> Hm, so if Sparc has no notion of compat-ness of the system call then how does it
>> implement runtime compat checks, such as AUDIT_ARCH et al?
>
> Badly.  Things like compat_sys_ioctl() vs. ioctl() work (we use different
> arrays of function pointers in 32bit and 64bit traps), but anything
> dynamic assumes that things match the task.  Not that we had a lot of
> such dynamic checks, actually...

I wouldn't take x86 as a shining example of how to do this, but it
does mostly work.  In 4.4, it's not even all that messy, modulo a
bunch of checks that check the wrong condition (hence, in part, this
proposal).

--Andy
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/