Re: [PATCH] microblaze: remove CONFIG_SET_FS

[Arnd Bergmann]

On Thu, Feb 10, 2022 at 3:21 PM David Laight <David.Laight@xxxxxxxxxx> wrote:
> From: Arnd Bergmann Sent: 10 February 2022 13:30
>
> > static inline int __range_ok(unsigned long addr, unsigned long size)
> > {
> >         return size <= TASK_SIZE && addr <= (TASK_SIZE - size);
> > }
> >
> > since 'size' is usually constant, so this turns into a single comparison
> > against a compile-time constant.
>
> Hmmm... maybe there should be a comment that it is the same as
> the more obvious:
>         (addr <= TASK_SIZE && addr <= TASK_SIZE - size)
> but is better for constant size.
> (Provided TASK_SIZE is a constant.)
>
> I'm sure Linus was 'unhappy' about checking against 2^63 for
> 32bit processes on a 64bit kernel.
>
> Hmmm compat code that has 32bit addr/size needn't even call
> access_ok() - it can never access kernel memory at all.

I suppose the generic function should compare against
TASK_SIZE_MAX or user_addr_max() then, to make it a
constant while TASK_SIZE potentially depends on compat
mode.

        Arnd