Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 25 May 2023 00:59:39 -0700
From:   Christoph Hellwig <hch@infradead.org>
To:     Dave Chinner <david@fromorbit.com>
Cc:     Uladzislau Rezki <urezki@gmail.com>,
        Hyeonggon Yoo <42.hyeyoo@gmail.com>, linux-mm@kvack.org,
        Andrew Morton <akpm@linux-foundation.org>,
        LKML <linux-kernel@vger.kernel.org>, Baoquan He <bhe@redhat.com>,
        Lorenzo Stoakes <lstoakes@gmail.com>,
        Christoph Hellwig <hch@infradead.org>,
        Matthew Wilcox <willy@infradead.org>,
        "Liam R . Howlett" <Liam.Howlett@oracle.com>,
        "Paul E . McKenney" <paulmck@kernel.org>,
        Joel Fernandes <joel@joelfernandes.org>,
        Oleksiy Avramchenko <oleksiy.avramchenko@sony.com>,
        linux-xfs@vger.kernel.org
Subject: Re: [PATCH 0/9] Mitigate a vmap lock contention
Message-ID: <ZG8Va0hc67IC6Cuf@infradead.org>
References: <20230522110849.2921-1-urezki@gmail.com>
 <ZGyqiaRnMJPFhxR6@debian-BULLSEYE-live-builder-AMD64>
 <ZGzX3vRMlGHIcYCe@pc636>
 <ZG0AE9mjHkRZIGmr@debian-BULLSEYE-live-builder-AMD64>
 <ZG3d1FUXiCk3QL3D@pc636>
 <ZG6IKE7yNEkJhge+@dread.disaster.area>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <ZG6IKE7yNEkJhge+@dread.disaster.area>
Precedence: bulk

On Thu, May 25, 2023 at 07:56:56AM +1000, Dave Chinner wrote:
> > It is up to community to decide. As i see XFS needs it also. Maybe in
> > the future it can be removed(who knows). If the vmalloc code itself can
> > deliver such performance as vm_map* APIs.
> 
> vm_map* APIs cannot be replaced with vmalloc, they cover a very
> different use case.  i.e.  vmalloc allocates mapped memory,
> vm_map_ram() maps allocated memory....
> 
> > vm_map_ram() and friends interface was added because of vmalloc drawbacks.
> 
> No. vm_map*() were scalability improvements added in 2009 to replace
> on vmap() and vunmap() to avoid global lock contention in the vmap
> allocator that XFS had been working around for years with it's own
> internal vmap cache....

All of that is true.  At the same time XFS could very much switch to
vmalloc for !XBF_UNMAPPED && size > PAGE_SIZE buffers IFF that provided
an advantage.  The need for vmap and then vm_map_* initially came from
the fact that we were using the page cache to back the xfs_buf (or
page_buf back then).  With your work on getting rid of the pagecache
usage we could just use vmalloc now if we wanted to and it improves
performance.  Or at some point we could even look into just using large
folios for that with the infrastructure for that improving a lot (but
I suspect we're not quite there yet).

But ther are other uses of vm_map_* that can't do that, and they will
benefit from the additional scalability as well even IFF just using
vmalloc was fine for XFS now (I don't know, I haven't actually looked
into it).