Re: + prctl-pr_set_mm-introduce-pr_set_mm_map-operation-v3.patch added to -mm tree

[Cyrill Gorcunov <gorcunov@gmail.com>]

On Fri, Aug 22, 2014 at 09:22:41PM +0200, Oleg Nesterov wrote:
> Hi Cyrill,
> 
> I think the patch is fine but I can't understand the usage of mmap_sem
> and alloc_lock,
> 
> > +	stack_vma = find_vma(mm, (unsigned long)prctl_map->start_stack);
> 
> OK, find_vma() needs mmap_sem. But otherwise, why this should be called
> under down_read(&mm->mmap_sem) ? What this lock tries to protect?

It should protect from allocation/devetion/mergin of another vma. IOW when
I lookup for vma I need to be sure it exist and won't disappear at least
while I validate it.

> > +	if (prctl_map.auxv_size) {
> > +		up_read(&mm->mmap_sem);
> > +		memset(user_auxv, 0, sizeof(user_auxv));
> > +		error = copy_from_user(user_auxv,
> > +				       (const void __user *)prctl_map.auxv,
> > +				       prctl_map.auxv_size);
> > +		down_read(&mm->mmap_sem);
> 
> And if we actually need this lock, why it is safe to drop it temporary?
> And why we can't move this copy_from_user() up before down_read) in any
> case?

Good point, thanks! I agreed that better to move this copying of user_auxv
before taking read-lock (once Andrew answer me my question about copying of
offsets in prctl_mm_map structure, I'll update on top).

> > +	if (prctl_map.auxv_size) {
> > +		/* Last entry must be AT_NULL as specification requires */
> > +		user_auxv[AT_VECTOR_SIZE - 2] = AT_NULL;
> > +		user_auxv[AT_VECTOR_SIZE - 1] = AT_NULL;
> > +
> > +		task_lock(current);
> > +		memcpy(mm->saved_auxv, user_auxv, sizeof(user_auxv));
> > +		task_unlock(current);
> 
> Again, could you explain this task_lock() ?

It is used for serialization access to saved_auxv, ie when we fill it
with new data the other reader (via procfs interface) should wait until
we finish.

Thanks for comments, Oleg!