From: Dan Williams <dan.j.williams@intel.com>
To: Jason Gunthorpe <jgg@nvidia.com>,
Dan Williams <dan.j.williams@intel.com>
Cc: <akpm@linux-foundation.org>, Matthew Wilcox <willy@infradead.org>,
"Jan Kara" <jack@suse.cz>, "Darrick J. Wong" <djwong@kernel.org>,
"Christoph Hellwig" <hch@lst.de>,
John Hubbard <jhubbard@nvidia.com>,
<linux-fsdevel@vger.kernel.org>, <nvdimm@lists.linux.dev>,
<linux-xfs@vger.kernel.org>, <linux-mm@kvack.org>,
<linux-ext4@vger.kernel.org>
Subject: Re: [PATCH v2 10/18] fsdax: Manage pgmap references at entry insertion and deletion
Date: Wed, 21 Sep 2022 08:18:38 -0700 [thread overview]
Message-ID: <632b2b4edd803_66d1a2941a@dwillia2-xfh.jf.intel.com.notmuch> (raw)
In-Reply-To: <YysZrdF/BSQhjWZs@nvidia.com>
Jason Gunthorpe wrote:
> On Thu, Sep 15, 2022 at 08:36:07PM -0700, Dan Williams wrote:
> > The percpu_ref in 'struct dev_pagemap' is used to coordinate active
> > mappings of device-memory with the device-removal / unbind path. It
> > enables the semantic that initiating device-removal (or
> > device-driver-unbind) blocks new mapping and DMA attempts, and waits for
> > mapping revocation or inflight DMA to complete.
>
> This seems strange to me
>
> The pagemap should be ref'd as long as the filesystem is mounted over
> the dax. The ref should be incrd when the filesystem is mounted and
> decrd when it is unmounted.
>
> When the filesystem unmounts it should zap all the mappings (actually
> I don't think you can even unmount a filesystem while mappings are
> open) and wait for all page references to go to zero, then put the
> final pagemap back.
>
> The rule is nothing can touch page->pgmap while page->refcount == 0,
> and if page->refcount != 0 then page->pgmap must be valid, without any
> refcounting on the page map itself.
>
> So, why do we need pgmap refcounting all over the place? It seems like
> it only existed before because of the abuse of the page->refcount?
Recall that this percpu_ref is mirroring the same function as
blk_queue_enter() whereby every new request is checking to make sure the
device is still alive, or whether it has started exiting.
So pgmap 'live' reference taking in fs/dax.c allows the core to start
failing fault requests once device teardown has started. It is a 'block
new, and drain old' semantic.
next prev parent reply other threads:[~2022-09-21 15:18 UTC|newest]
Thread overview: 84+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-16 3:35 [PATCH v2 00/18] Fix the DAX-gup mistake Dan Williams
2022-09-16 3:35 ` [PATCH v2 01/18] fsdax: Wait on @page not @page->_refcount Dan Williams
2022-09-20 14:30 ` Jason Gunthorpe
2022-09-16 3:35 ` [PATCH v2 02/18] fsdax: Use dax_page_idle() to document DAX busy page checking Dan Williams
2022-09-20 14:31 ` Jason Gunthorpe
2022-09-16 3:35 ` [PATCH v2 03/18] fsdax: Include unmapped inodes for page-idle detection Dan Williams
2022-09-16 3:35 ` [PATCH v2 04/18] ext4: Add ext4_break_layouts() to the inode eviction path Dan Williams
2022-09-16 3:35 ` [PATCH v2 05/18] xfs: Add xfs_break_layouts() " Dan Williams
2022-09-18 22:57 ` Dave Chinner
2022-09-19 16:11 ` Dan Williams
2022-09-19 21:29 ` Dave Chinner
2022-09-20 16:44 ` Dan Williams
2022-09-21 22:14 ` Dave Chinner
2022-09-21 22:28 ` Jason Gunthorpe
2022-09-23 0:18 ` Dave Chinner
2022-09-23 0:41 ` Dan Williams
2022-09-23 2:10 ` Dave Chinner
2022-09-23 9:38 ` Jan Kara
2022-09-23 23:06 ` Dan Williams
2022-09-25 23:54 ` Dave Chinner
2022-09-26 14:10 ` Jan Kara
2022-09-29 23:33 ` Dan Williams
2022-09-30 13:41 ` Jan Kara
2022-09-30 17:56 ` Dan Williams
2022-09-30 18:06 ` Jason Gunthorpe
2022-09-30 18:46 ` Dan Williams
2022-10-03 7:55 ` Jan Kara
2022-09-23 12:39 ` Jason Gunthorpe
2022-09-26 0:34 ` Dave Chinner
2022-09-26 13:04 ` Jason Gunthorpe
2022-09-22 0:02 ` Dan Williams
2022-09-22 0:10 ` Jason Gunthorpe
2022-09-16 3:35 ` [PATCH v2 06/18] fsdax: Rework dax_layout_busy_page() to dax_zap_mappings() Dan Williams
2022-09-16 3:35 ` [PATCH v2 07/18] fsdax: Update dax_insert_entry() calling convention to return an error Dan Williams
2022-09-16 3:35 ` [PATCH v2 08/18] fsdax: Cleanup dax_associate_entry() Dan Williams
2022-09-16 3:36 ` [PATCH v2 09/18] fsdax: Rework dax_insert_entry() calling convention Dan Williams
2022-09-16 3:36 ` [PATCH v2 10/18] fsdax: Manage pgmap references at entry insertion and deletion Dan Williams
2022-09-21 14:03 ` Jason Gunthorpe
2022-09-21 15:18 ` Dan Williams [this message]
2022-09-21 21:38 ` Dan Williams
2022-09-21 22:07 ` Jason Gunthorpe
2022-09-22 0:14 ` Dan Williams
2022-09-22 0:25 ` Jason Gunthorpe
2022-09-22 2:17 ` Dan Williams
2022-09-22 17:55 ` Jason Gunthorpe
2022-09-22 21:54 ` Dan Williams
2022-09-23 1:36 ` Dave Chinner
2022-09-23 2:01 ` Dan Williams
2022-09-23 13:24 ` Jason Gunthorpe
2022-09-23 16:29 ` Dan Williams
2022-09-23 17:42 ` Jason Gunthorpe
2022-09-23 19:03 ` Dan Williams
2022-09-23 19:23 ` Jason Gunthorpe
2022-09-27 6:07 ` Alistair Popple
2022-09-27 12:56 ` Jason Gunthorpe
2022-09-16 3:36 ` [PATCH v2 11/18] devdax: Minor warning fixups Dan Williams
2022-09-16 3:36 ` [PATCH v2 12/18] devdax: Move address_space helpers to the DAX core Dan Williams
2022-09-27 6:20 ` Alistair Popple
2022-09-29 22:38 ` Dan Williams
2022-09-16 3:36 ` [PATCH v2 13/18] dax: Prep mapping helpers for compound pages Dan Williams
2022-09-21 14:06 ` Jason Gunthorpe
2022-09-21 15:19 ` Dan Williams
2022-09-16 3:36 ` [PATCH v2 14/18] devdax: add PUD support to the DAX mapping infrastructure Dan Williams
2022-09-16 3:36 ` [PATCH v2 15/18] devdax: Use dax_insert_entry() + dax_delete_mapping_entry() Dan Williams
2022-09-21 14:10 ` Jason Gunthorpe
2022-09-21 15:48 ` Dan Williams
2022-09-21 22:23 ` Jason Gunthorpe
2022-09-22 0:15 ` Dan Williams
2022-09-16 3:36 ` [PATCH v2 16/18] mm/memremap_pages: Support initializing pages to a zero reference count Dan Williams
2022-09-21 15:24 ` Jason Gunthorpe
2022-09-21 23:45 ` Dan Williams
2022-09-22 0:03 ` Alistair Popple
2022-09-22 0:04 ` Jason Gunthorpe
2022-09-22 0:34 ` Dan Williams
2022-09-22 1:36 ` Alistair Popple
2022-09-22 2:34 ` Dan Williams
2022-09-26 6:17 ` Alistair Popple
2022-09-22 0:13 ` John Hubbard
2022-09-16 3:36 ` [PATCH v2 17/18] fsdax: Delete put_devmap_managed_page_refs() Dan Williams
2022-09-16 3:36 ` [PATCH v2 18/18] mm/gup: Drop DAX pgmap accounting Dan Williams
2022-09-20 14:29 ` [PATCH v2 00/18] Fix the DAX-gup mistake Jason Gunthorpe
2022-09-20 16:50 ` Dan Williams
2022-11-09 0:20 ` Andrew Morton
2022-11-09 11:38 ` Jan Kara
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=632b2b4edd803_66d1a2941a@dwillia2-xfh.jf.intel.com.notmuch \
--to=dan.j.williams@intel.com \
--cc=akpm@linux-foundation.org \
--cc=djwong@kernel.org \
--cc=hch@lst.de \
--cc=jack@suse.cz \
--cc=jgg@nvidia.com \
--cc=jhubbard@nvidia.com \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-xfs@vger.kernel.org \
--cc=nvdimm@lists.linux.dev \
--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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.