Re: [PATCH] x86/mtrr: Remove unnecessary strlen() in mtrr_write()

[Ingo Molnar <mingo@kernel.org>]

* Thorsten Blum <thorsten.blum@linux.dev> wrote:

> On 25. Feb 2025, at 14:35, Ingo Molnar wrote:
> > * Thorsten Blum <thorsten.blum@linux.dev> wrote:
> > 
> >> The local variable length already holds the string length after calling
> >> strncpy_from_user(). Using another local variable linlen and calling
> >> strlen() is therefore unnecessary and can be removed. Remove linlen
> >> and strlen() and use length instead.
> >> 
> >> Compile-tested only.
> >> 
> >> Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
> >> ---
> > 
> > I see no corner-case analysis in the changelog about what may happen if 
> > the copy fails partially.
> 
> Hi Ingo,
> 
> I'm not sure what you mean. Why would I describe something I didn't
> change? This patch only removes an unnecessary string (line) length
> calculation and shouldn't affect the logic in any way.
> 
> If strncpy_from_user() fails, we immediately return length from
> mtrr_write(), but my patch doesn't change this - unless I'm missing
> something?

My bad: I was fixated on this 'some data may have been copied' language 
in the strncpy_from_user() documentation:

 * If access to userspace fails, returns -EFAULT (some data may have been
 * copied).

which would make the strlen() result ambiguous. But this doesn't matter 
here - if there was a fault then we get an unconditional -EFAULT from 
strncpy_from_user(), so your patch should be equivalent to the previous 
logic.

Thanks,

	Ingo