Re: [PATCH v5] mm/ksm: introduce ksm_force for each process

[Ammar Faizi]

On 5/11/22 3:30 AM, Andrew Morton wrote:
> On Wed, 11 May 2022 03:10:31 +0700 Ammar Faizi <ammarfaizi2@xxxxxxxxxxx> wrote:
>
> > On 5/8/22 4:27 PM, cgel.zte@xxxxxxxxx wrote:
> > > +static ssize_t ksm_force_write(struct file *file, const char __user *buf,
> > > +				size_t count, loff_t *ppos)
> > > +{
> > > +	struct task_struct *task;
> > > +	struct mm_struct *mm;
> > > +	char buffer[PROC_NUMBUF];
> > > +	int force;
> > > +	int err = 0;
> > > +
> > > +	memset(buffer, 0, sizeof(buffer));
> > > +	if (count > sizeof(buffer) - 1)
> > > +		count = sizeof(buffer) - 1;
> > > +	if (copy_from_user(buffer, buf, count)) {
> > > +		err = -EFAULT;
> > > +		goto out_return;
> > > +	}
> >
> > This one looks like over-zeroing to me. You don't need to zero
> > all elements in the array. You're going to overwrite it with
> > `copy_from_user()` anyway.
> >
> > Just zero the last potentially useful element by using @count
> > as the index. It can be like this:
> >
> > ```
> > 	char buffer[PROC_NUMBUF];
> >
> > 	if (count > sizeof(buffer) - 1)
> > 		count = sizeof(buffer) - 1;
> > 	if (copy_from_user(buffer, buf, count))
> > 		return -EFAULT;
> > 	buffer[count] = '\0';
> > ```
>
> Use strncpy_from_user()?

Sounds better.

> Can this code use proc_dointvec_minmax() or similar?

Not familiar with that API at all. Leaving it to other participants...

--
Ammar Faizi