From: Peter Zijlstra <peterz@infradead.org>
To: Matthew Wilcox <willy@infradead.org>
Cc: SeongJae Park <sj38.park@gmail.com>,
Michal Hocko <mhocko@suse.com>, Vlastimil Babka <vbabka@suse.cz>,
"Kirill A. Shutemov" <kirill@shutemov.name>,
linux-mm@kvack.org
Subject: Re: Re: Splitting the mmap_sem
Date: Thu, 6 Feb 2020 21:55:29 +0100 [thread overview]
Message-ID: <20200206205529.GZ14914@hirez.programming.kicks-ass.net> (raw)
In-Reply-To: <20200206201536.GX8731@bombadil.infradead.org>
On Thu, Feb 06, 2020 at 12:15:36PM -0800, Matthew Wilcox wrote:
> On Thu, Feb 06, 2020 at 02:59:20PM +0100, Peter Zijlstra wrote:
> > > The proposal consists of three phases. In phase 1, we convert the
> > > rbtree to the maple tree, and leave the locking alone. In phase 2,
> > > we change the locking to a per-VMA refcount, looked up under RCU.
> > >
> > > This problem arises during phase 3 where we attempt to handle page
> > > faults entirely under the RCU read lock. If we encounter problems,
> > > we can fall back to acquiring the VMA refcount, but we need the
> > > page allocation to fail rather than sleep (or magically drop the
> > > RCU lock and return an indication that it has done so, but that
> > > doesn't seem to be an approach that would find any favour).
> >
> > So why not use SRCU? You can do full blocking faults under SRCU and
> > don't need no 'stinkin' refcounts ;-)
>
> I have to say, SRCU is not in my mental toolbox of "how to solve a
> problem", so it simply hadn't occurred to me. Thanks.
>
> So, we'd DEFINE_SRCU(vma_srcu); in mm/memory.c
>
> then, at the beginning of a page fault call srcu_read_lock(&vma_srcu);
> walk the tree as we do now, allocate memory for PTEs, sleep waiting for
> pages to arrive back from disc, etc, etc, then at the end of the fault,
> call srcu_read_unlock(&vma_srcu).
So far so good,...
> munmap() would consist of removing the
> VMA from the tree, then calling synchronize_srcu() to wait for all faults
> to finish, then putting the backing file, etc, etc and freeing the VMA.
call_srcu(), and the (s)rcu callback will then fput() and such things
more.
synchronize_srcu() (like synchronize_rcu()) is stupid slow and would
make munmap()/exit()/etc.. unusable.
> This seems pretty reasonable, and investigation could actually proceed
> before the Maple tree work lands. Today, that would be:
>
> srcu_read_lock(&vmas_srcu);
> down_read(&mm->mmap_sem);
> find_vma(mm, address);
> up_read(&mm->mmap_sem);
> ... rest of fault handler path ...
> srcu_read_unlock(&vmas_srcu);
>
> Kind of a pain because we still call find_vma() in the per-arch page
> fault handler, but for prototyping, we'd only have to do one or two
> architectures.
If you look at the earlier speculative page-fault patches by Laurent,
which were based on my still earlier patches, you'll find most of this
there.
The tricky bit was validating everything on the second page-table walk,
so see if nothing had fundamentally changed, specifically the VMA,
before installing the PTE. If you do this without mmap_sem, you need to
hold ptlock to pin stuff while validating everything you did earlier.
next prev parent reply other threads:[~2020-02-06 20:55 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-12-03 22:21 Splitting the mmap_sem Matthew Wilcox
2019-12-05 17:21 ` Jerome Glisse
2019-12-06 5:13 ` Matthew Wilcox
2019-12-06 17:30 ` Jerome Glisse
2019-12-09 3:33 ` Matthew Wilcox
2019-12-09 14:17 ` Jerome Glisse
2019-12-10 15:26 ` Vlastimil Babka
2019-12-10 16:07 ` Jerome Glisse
2019-12-10 18:09 ` Vlastimil Babka
2019-12-12 14:24 ` Kirill A. Shutemov
2019-12-12 15:40 ` Matthew Wilcox
2019-12-12 15:46 ` Kirill A. Shutemov
2019-12-13 14:33 ` Matthew Wilcox
2019-12-13 18:06 ` Kirill A. Shutemov
2019-12-13 18:21 ` Matthew Wilcox
2020-01-06 22:09 ` Matthew Wilcox
2020-01-07 12:34 ` Kirill A. Shutemov
2020-01-07 13:54 ` Matthew Wilcox
2020-01-07 14:27 ` Kirill A. Shutemov
2020-01-09 13:56 ` Vlastimil Babka
2020-01-09 17:03 ` Michal Hocko
2020-01-09 17:07 ` Michal Hocko
2020-01-09 17:32 ` SeongJae Park
2020-01-09 20:13 ` Matthew Wilcox
2020-02-06 13:59 ` Peter Zijlstra
2020-02-06 20:15 ` Matthew Wilcox
2020-02-06 20:55 ` Peter Zijlstra [this message]
2020-02-06 21:20 ` Matthew Wilcox
2020-02-07 8:52 ` Peter Zijlstra
2020-02-10 22:00 ` Matthew Wilcox
2020-02-19 17:14 ` Laurent Dufour
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=20200206205529.GZ14914@hirez.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=kirill@shutemov.name \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=sj38.park@gmail.com \
--cc=vbabka@suse.cz \
--cc=willy@infradead.org \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).