LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Jerome Glisse <jglisse@redhat.com>
To: Ralph Campbell <rcampbell@nvidia.com>
Cc: Souptick Joarder <jrdr.linux@gmail.com>,
	Linux-MM <linux-mm@kvack.org>,
	linux-kernel@vger.kernel.org, John Hubbard <jhubbard@nvidia.com>,
	Ira Weiny <ira.weiny@intel.com>,
	Dan Williams <dan.j.williams@intel.com>,
	Arnd Bergmann <arnd@arndb.de>,
	Balbir Singh <bsingharora@gmail.com>,
	Dan Carpenter <dan.carpenter@oracle.com>,
	Matthew Wilcox <willy@infradead.org>,
	Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH 4/5] mm/hmm: hmm_vma_fault() doesn't always call hmm_range_unregister()
Date: Sun, 12 May 2019 11:28:15 -0400	[thread overview]
Message-ID: <20190512152814.GC4238@redhat.com> (raw)
In-Reply-To: <20190512150724.GA4238@redhat.com>

On Sun, May 12, 2019 at 11:07:24AM -0400, Jerome Glisse wrote:
> On Tue, May 07, 2019 at 11:12:14AM -0700, Ralph Campbell wrote:
> > 
> > On 5/7/19 6:15 AM, Souptick Joarder wrote:
> > > On Tue, May 7, 2019 at 5:00 AM <rcampbell@nvidia.com> wrote:
> > > > 
> > > > From: Ralph Campbell <rcampbell@nvidia.com>
> > > > 
> > > > The helper function hmm_vma_fault() calls hmm_range_register() but is
> > > > missing a call to hmm_range_unregister() in one of the error paths.
> > > > This leads to a reference count leak and ultimately a memory leak on
> > > > struct hmm.
> > > > 
> > > > Always call hmm_range_unregister() if hmm_range_register() succeeded.
> > > 
> > > How about * Call hmm_range_unregister() in error path if
> > > hmm_range_register() succeeded* ?
> > 
> > Sure, sounds good.
> > I'll include that in v2.
> 
> NAK for the patch see below why
> 
> > 
> > > > 
> > > > Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
> > > > Cc: John Hubbard <jhubbard@nvidia.com>
> > > > Cc: Ira Weiny <ira.weiny@intel.com>
> > > > Cc: Dan Williams <dan.j.williams@intel.com>
> > > > Cc: Arnd Bergmann <arnd@arndb.de>
> > > > Cc: Balbir Singh <bsingharora@gmail.com>
> > > > Cc: Dan Carpenter <dan.carpenter@oracle.com>
> > > > Cc: Matthew Wilcox <willy@infradead.org>
> > > > Cc: Souptick Joarder <jrdr.linux@gmail.com>
> > > > Cc: Andrew Morton <akpm@linux-foundation.org>
> > > > ---
> > > >   include/linux/hmm.h | 3 ++-
> > > >   1 file changed, 2 insertions(+), 1 deletion(-)
> > > > 
> > > > diff --git a/include/linux/hmm.h b/include/linux/hmm.h
> > > > index 35a429621e1e..fa0671d67269 100644
> > > > --- a/include/linux/hmm.h
> > > > +++ b/include/linux/hmm.h
> > > > @@ -559,6 +559,7 @@ static inline int hmm_vma_fault(struct hmm_range *range, bool block)
> > > >                  return (int)ret;
> > > > 
> > > >          if (!hmm_range_wait_until_valid(range, HMM_RANGE_DEFAULT_TIMEOUT)) {
> > > > +               hmm_range_unregister(range);
> > > >                  /*
> > > >                   * The mmap_sem was taken by driver we release it here and
> > > >                   * returns -EAGAIN which correspond to mmap_sem have been
> > > > @@ -570,13 +571,13 @@ static inline int hmm_vma_fault(struct hmm_range *range, bool block)
> > > > 
> > > >          ret = hmm_range_fault(range, block);
> > > >          if (ret <= 0) {
> > > > +               hmm_range_unregister(range);
> > > 
> > > what is the reason to moved it up ?
> > 
> > I moved it up because the normal calling pattern is:
> >     down_read(&mm->mmap_sem)
> >     hmm_vma_fault()
> >         hmm_range_register()
> >         hmm_range_fault()
> >         hmm_range_unregister()
> >     up_read(&mm->mmap_sem)
> > 
> > I don't think it is a bug to unlock mmap_sem and then unregister,
> > it is just more consistent nesting.
> 
> So this is not the usage pattern with HMM usage pattern is:
> 
> hmm_range_register()
> hmm_range_fault()
> hmm_range_unregister()
> 
> The hmm_vma_fault() is gonne so this patch here break thing.
> 
> See https://cgit.freedesktop.org/~glisse/linux/log/?h=hmm-5.2-v3

Sorry not enough coffee on sunday morning, so yeah this patch
looks good except that you do not need to move it up.

Note that hmm_vma_fault() is a gonner once ODP to HMM is upstream
and i converted nouveau/amd to new API then we can remove that
one.

Cheers,
Jérôme

  reply	other threads:[~2019-05-12 15:28 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-06 23:29 [PATCH 0/5] mm/hmm: HMM documentation updates and code fixes rcampbell
2019-05-06 23:29 ` [PATCH 1/5] mm/hmm: Update HMM documentation rcampbell
2019-06-06 14:02   ` Jason Gunthorpe
2019-06-06 18:50     ` Ralph Campbell
2019-06-06 19:32       ` Jason Gunthorpe
2019-05-06 23:29 ` [PATCH 2/5] mm/hmm: Clean up some coding style and comments rcampbell
2019-06-06 14:16   ` Jason Gunthorpe
2019-06-06 14:27     ` Jerome Glisse
2019-06-06 15:41       ` Jason Gunthorpe
2019-06-06 15:52         ` Jerome Glisse
2019-06-06 16:55           ` Joe Perches
2019-06-06 18:54           ` Jason Gunthorpe
2019-06-06 15:57   ` Jason Gunthorpe
2019-06-07  0:44     ` Ralph Campbell
2019-05-06 23:29 ` [PATCH 3/5] mm/hmm: Use mm_get_hmm() in hmm_range_register() rcampbell
2019-05-23 12:51   ` Jason Gunthorpe
2019-05-23 18:19     ` Ralph Campbell
2019-05-23 18:46     ` John Hubbard
2019-05-06 23:29 ` [PATCH 4/5] mm/hmm: hmm_vma_fault() doesn't always call hmm_range_unregister() rcampbell
2019-05-07 13:15   ` Souptick Joarder
2019-05-07 18:12     ` Ralph Campbell
2019-05-09  4:42       ` Souptick Joarder
2019-05-12 15:07       ` Jerome Glisse
2019-05-12 15:28         ` Jerome Glisse [this message]
2019-05-13 17:23         ` Ralph Campbell
2019-06-06 14:50   ` Jason Gunthorpe
2019-06-06 19:44     ` Ralph Campbell
2019-06-06 19:54       ` Jason Gunthorpe
2019-06-06 21:08         ` Ralph Campbell
2019-05-12 15:08 ` [PATCH 0/5] mm/hmm: HMM documentation updates and code fixes Jerome Glisse
2019-05-13 17:26   ` Ralph Campbell
2019-05-13 18:10     ` Jerome Glisse
2019-06-06 14:53 ` Jason Gunthorpe

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190512152814.GC4238@redhat.com \
    --to=jglisse@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=arnd@arndb.de \
    --cc=bsingharora@gmail.com \
    --cc=dan.carpenter@oracle.com \
    --cc=dan.j.williams@intel.com \
    --cc=ira.weiny@intel.com \
    --cc=jhubbard@nvidia.com \
    --cc=jrdr.linux@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=rcampbell@nvidia.com \
    --cc=willy@infradead.org \
    --subject='Re: [PATCH 4/5] mm/hmm: hmm_vma_fault() doesn'\''t always call hmm_range_unregister()' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).