From mboxrd@z Thu Jan  1 00:00:00 1970
From: Ingo Molnar <mingo@kernel.org>
Subject: Re: [PATCH v2 00/10] evacuate struct page from the block layer,
 introduce __pfn_t
Date: Thu, 7 May 2015 21:53:13 +0200
Message-ID: <20150507195313.GA23597@gmail.com>
References: <CA+55aFwL321_xxcXwj5+G7q=rc8HmqMBpot5dK_AW2=-7xOn0g@mail.gmail.com>
 <CAPcyv4iY+2SKqSxNijhzgwBMDAbcd01kzWrVnW06yqtOinoSQg@mail.gmail.com>
 <CA+55aFyK==+bMHMdfaP4EHjUNseoTWu1-pp2M2WyLVN73ig+_g@mail.gmail.com>
 <CAPcyv4i5957h0PW5JL0rv4iurL=c7ppEFj3cDrrrMc4MQmJQGQ@mail.gmail.com>
 <20150507173641.GA21781@gmail.com>
 <CAPcyv4hHwW4U8x1_VLGj2Q4a3HxWgK4F4n9qXg8009_n7sxkmg@mail.gmail.com>
 <554BA748.9030804@linux.intel.com>
 <20150507191107.GB22952@gmail.com>
 <20150507193635.GC5966@gmail.com>
 <20150507194832.GB23511@gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Dave Hansen <dave.hansen@linux.intel.com>,
	Dan Williams <dan.j.williams@intel.com>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Boaz Harrosh <boaz@plexistor.com>, Jan Kara <jack@suse.cz>,
	Mike Snitzer <snitzer@redhat.com>, Neil Brown <neilb@suse.de>,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>,
	Heiko Carstens <heiko.carstens@de.ibm.com>,
	Chris Mason <clm@fb.com>, Paul Mackerras <paulus@samba.org>,
	"H. Peter Anvin" <hpa@zytor.com>, Christoph Hellwig <hch@lst.de>,
	Alasdair Kergon <agk@redhat.com>,
	"linux-nvdimm@lists.01.org" <linux-nvdimm@lists.01.org>,
	Mel Gorman <mgorman@suse.de>,
	Matthew Wilcox <willy@linux.intel.com>,
	Ross Zwisler <ross.zwisler@linux.intel.com>,
	Rik van Riel <riel@redhat.com>,
	Martin Schwidefsky <schwidefsky@de.ibm.com>,
	Jens Axboe <axboe@kernel.dk>, Theodore Ts'o <tytso@mit.edu>,
To: Jerome Glisse <j.glisse@gmail.com>
Return-path: <linux-kernel-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <20150507194832.GB23511@gmail.com>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org


* Ingo Molnar <mingo@kernel.org> wrote:

> > Is handling kernel pagefault on the vmemmap completely out of the 
> > picture ? So we would carveout a chunck of kernel address space 
> > for those pfn and use it for vmemmap and handle pagefault on it.
> 
> That's pretty clever. The page fault doesn't even have to do remote 
> TLB shootdown, because it only establishes mappings - so it's pretty 
> atomic, a bit like the minor vmalloc() area faults we are doing.
> 
> Some sort of LRA (least recently allocated) scheme could unmap the 
> area in chunks if it's beyond a certain size, to keep a limit on 
> size. Done from the same context and would use remote TLB shootdown.
> 
> The only limitation I can see is that such faults would have to be 
> able to sleep, to do the allocation. So pfn_to_page() could not be 
> used in arbitrary contexts.

So another complication would be that we cannot just unmap such pages 
when we want to recycle them, because the struct page in them might be 
in use - so all struct page uses would have to refcount the underlying 
page. We don't really do that today: code just looks up struct pages 
and assumes they never go away.

Thanks,

	Ingo