Re: [PATCH 2/2 resend] mm: various cleanups in get_user_pages()

[Eugene Teo]

Hi Nick,

Thanks for the review.

<quote sender="Nick Piggin">
> On Wednesday 13 February 2008 00:10, Eugene Teo wrote:
[...]
> > diff --git a/mm/memory.c b/mm/memory.c
> > index 54f951b..c7e0610 100644
> > --- a/mm/memory.c
> > +++ b/mm/memory.c
> > @@ -1003,7 +1003,9 @@ int get_user_pages(struct task_struct *tsk, struct
> > mm_struct *mm, unsigned int foll_flags;
> >
> >  		vma = find_extend_vma(mm, start);
> > -		if (!vma && in_gate_area(tsk, start)) {
> > +		if (!vma)
> > +			goto finish_or_fault;
> > +		if (in_gate_area(tsk, start)) {
> >  			unsigned long pg = start & PAGE_MASK;
> >  			struct vm_area_struct *gate_vma = get_gate_vma(tsk);
> >  			pgd_t *pgd;
> 
> Doesn't this break the logic?
> 
> If you don't have a vma, but you are in the gate area, then you
> should use the gate vma. With your patch, gate area will fault.

Yes, you are right. I also relooked at the patch, and actually vma is
validated after if (... in_gate_area(tsk, start)) { ... }, so my patch
is not correct.

> > @@ -1011,7 +1013,7 @@ int get_user_pages(struct task_struct *tsk, struct
> > mm_struct *mm, pmd_t *pmd;
> >  			pte_t *pte;
> >  			if (write) /* user gate pages are read-only */
> > -				return i ? : -EFAULT;
> > +				goto finish_or_fault;
> 
> I don't know if this is exactly a cleanup or not... I guess gcc
> probably isn't smart enough to fold them all together, so it should
> use a little less code in the unlikely branches. Does it?

Agree.

Eugene
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/