lkml.org 
[lkml]   [2020]   [Apr]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 1/2] mm/mempolicy: Allow lookup_node() to handle fatal signal
Any opinion on this Linus? Should I just repost the patch?

On Tue 14-04-20 13:04:32, Michal Hocko wrote:
> On Thu 09-04-20 09:42:20, Linus Torvalds wrote:
> > On Thu, Apr 9, 2020 at 12:03 AM Michal Hocko <mhocko@kernel.org> wrote:
> > >
> > > This patch however doesn't go all the way to revert it because 0 return
> > > value is impossible.
> >
> > I'm not convinced it's impossible.
>
> __get_user_pages is documented as
> * -- If nr_pages is 0, returns 0.
> * -- If nr_pages is >0, but no pages were pinned, returns -errno.
> * -- If nr_pages is >0, and some pages were pinned, returns the number of
> * pages pinned. Again, this may be less than nr_pages.
>
> but let me double check the actual code... There seem to be only one
> exception the above rule AFAICS. faultin_page returning EBUSY will
> be overriden to either 0 for the first page or return the number of
> already pinned pages. So nr_pages > 0 && ret = 0 is indeed possible
> from __get_user_pages :/ That will be the case only for VM_FAULT_RETRY,
> thoug.
>
> Now __get_user_pages_locked behaves differently. It keeps retrying the
> fault until it succeeds unless FOLL_NOWAIT is specified. Then it would
> return 0. Why we need to return 0 is not really clear to me but it
> seem to be a long term behavior. I believe we need to document it.
>
> > And if it is, then the current code is harmless.
>
> Yes from the above it seems that the check is indeed harmless becasue
> this path doesn't use FOLL_NOWAIT and so it will never see 0 return.
> I find a reference to EINTR confusing so I would still love to change
> that.
>
> > Now, I do agree that we probably should go through and clarify the
> > whole range of different get_user_pages() cases of returning zero (or
> > not doing so), but right now it's so confusing that I'd prefer to keep
> > that (possibly unnecessary) belt-and-suspenders check for zero in
> > there.
> >
> > If/when somebody actually does a real audit and the result is "these
> > functions cannot return zero" and it's documented, then we can remove
> > those checks.
>
> Would you mind this patch instead?
>
> commit bc6c0fa7c7fb5eb54963dca65ae4a62ba04c9efa
> Author: Michal Hocko <mhocko@suse.com>
> Date: Thu Apr 9 08:26:57 2020 +0200
>
> mm, mempolicy: fix up gup usage in lookup_node
>
> ba841078cd05 ("mm/mempolicy: Allow lookup_node() to handle fatal signal") has
> added a special casing for 0 return value because that was a possible
> gup return value when interrupted by fatal signal. This has been fixed
> by ae46d2aa6a7f ("mm/gup: Let __get_user_pages_locked() return -EINTR
> for fatal signal") in the mean time so ba841078cd05 can be reverted.
>
> This patch however doesn't go all the way to revert it because the check
> for 0 is wrong and confusing here. Firstly it is inherently unsafe to
> access the page when get_user_pages_locked returns 0 (aka no page
> returned).
> Fortunatelly this will not happen because get_user_pages_locked will not
> return 0 when nr_pages > 0 unless FOLL_NOWAIT is specified which is not
> the case here. Document this potential error code in gup code while we
> are at it.
>
> Signed-off-by: Michal Hocko <mhocko@suse.com>
>
> diff --git a/mm/gup.c b/mm/gup.c
> index 50681f0286de..a8575b880baf 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -980,6 +980,7 @@ static int check_vma_flags(struct vm_area_struct *vma, unsigned long gup_flags)
> * -- If nr_pages is >0, but no pages were pinned, returns -errno.
> * -- If nr_pages is >0, and some pages were pinned, returns the number of
> * pages pinned. Again, this may be less than nr_pages.
> + * -- 0 return value is possible when the fault would need to be retried.
> *
> * The caller is responsible for releasing returned @pages, via put_page().
> *
> @@ -1247,6 +1248,10 @@ int fixup_user_fault(struct task_struct *tsk, struct mm_struct *mm,
> }
> EXPORT_SYMBOL_GPL(fixup_user_fault);
>
> +/*
> + * Please note that this function, unlike __get_user_pages will not
> + * return 0 for nr_pages > 0 without FOLL_NOWAIT
> + */
> static __always_inline long __get_user_pages_locked(struct task_struct *tsk,
> struct mm_struct *mm,
> unsigned long start,
> diff --git a/mm/mempolicy.c b/mm/mempolicy.c
> index 48ba9729062e..1965e2681877 100644
> --- a/mm/mempolicy.c
> +++ b/mm/mempolicy.c
> @@ -927,10 +927,7 @@ static int lookup_node(struct mm_struct *mm, unsigned long addr)
>
> int locked = 1;
> err = get_user_pages_locked(addr & PAGE_MASK, 1, 0, &p, &locked);
> - if (err == 0) {
> - /* E.g. GUP interrupted by fatal signal */
> - err = -EFAULT;
> - } else if (err > 0) {
> + if (err > 0) {
> err = page_to_nid(p);
> put_page(p);
> }
> --
> Michal Hocko
> SUSE Labs

--
Michal Hocko
SUSE Labs

\
 
 \ /
  Last update: 2020-04-20 14:49    [W:0.109 / U:2.040 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site