Re: [PATCH 3/6] bitmap_parselist: rework input string parser

[Mike Travis]

On 3/26/2019 2:09 PM, Yuri Norov wrote:
> + Mike Travis <travis@xxxxxxx>
> + Thomas Gleixner <tglx@xxxxxxxxxxxxx>
>
> ----------------------------------------------------------------------
> On Tue, Mar 26, 2019 at 12:07:45AM +0300, Yury Norov wrote:
> > > The requirement for this rework is to keep the __bitmap_parselist()
> > > copy-less and single-pass but make it more readable and maintainable by
> > > splitting into logical parts and removing explicit nested cycles and
> > > opaque local variables.
> > >
> > > __bitmap_parselist() can parse userspace inputs and therefore we cannot
> > > use simple_strtoul() to parse numbers.
>
> > So, all above depends to what memory we access kernel / user space.
> > Perhaps we can get copy of memory of a given size and then parse it in kernel space always?
>
> > --
> > With Best Regards,
> > Andy Shevchenko
>
> What I missed during rework is that we have only one caller of *parselist_user  -
> it's write_irq_affinity() introduced by Mike Travis in kernel/irq/proc.c. It doesn't look
> like a hot path as it's file operations handler. If no objections from Mike or Thomas,

No objections from me as long as you can still change irq affinity using 
a cpulist instead of a cpumask.

Thanks,
Mike Travis


> I think it would make sense to copy_from_user() the userspace data at the beginning
> in sake of simplicity of __bitmap_parselist(), as you suggested above.
>
> Yury