Re: [PATCH 0/5] mm: rework hmm to use devm_memremap_pages

From: Christoph Hellwig <hch@lst.de>
To: Dan Williams <dan.j.williams@intel.com>
Cc: Jerome Glisse <jglisse@redhat.com>,
	Dave Airlie <airlied@gmail.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Logan Gunthorpe <logang@deltatee.com>,
	Christoph Hellwig <hch@lst.de>, Michal Hocko <mhocko@suse.com>,
	Linux MM <linux-mm@kvack.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 0/5] mm: rework hmm to use devm_memremap_pages
Date: Wed, 6 Jun 2018 09:14:18 +0200	[thread overview]
Message-ID: <20180606071418.GA7660@lst.de> (raw)
In-Reply-To: <CAPcyv4gsS4xDXahZdOggURBHS2y-oJ5tPG9vXPDdY2p6jPufxA@mail.gmail.com>

On Tue, Jun 05, 2018 at 06:33:04PM -0700, Dan Williams wrote:
> Unless the nouveau patches are using the entirety of what is already
> upstream for HMM, we should look to pare HMM back.
> 
> There is plenty of precedent of building a large capability
> out-of-tree and piecemeal merging it later, so I do not buy the
> "chicken-egg" argument. The change in the export is to make sure we
> don't repeat this backward "merge first, ask questions later" mistake
> in the future as devm_memremap_pages() is continuing to find new users
> like peer-to-peer DMA support and Linux is better off if that
> development is upstream. From a purely technical standpoint
> devm_memremap_pages() is EXPORT_SYMBOL_GPL because it hacks around
> several implementation details in the core kernel to achieve its goal,
> and it leaks new assumptions all over the kernel. It is strictly not a
> self contained interface.

Agreed with all of that.  And remember EXPORT_SYMBOL_GPL really just is
a clear expression of the authors they think these are internals.
The lack of it doesn't make it any less a derived work, we just remove
a very clear hint to users that they are poking very deeply into internals.

And with HMM they very clearly do.