From: Christoph Hellwig <hch@lst.de>
To: Dan Williams <dan.j.williams@intel.com>
Cc: Michal Hocko <mhocko@suse.com>, Jan Kara <jack@suse.cz>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Dave Hansen <dave.hansen@linux.intel.com>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
"J. Bruce Fields" <bfields@fieldses.org>,
linux-mm@kvack.org, Paul Mackerras <paulus@samba.org>,
Jeff Layton <jlayton@poochiereds.net>,
hch@lst.de, Matthew Wilcox <mawilcox@microsoft.com>,
linux-rdma@vger.kernel.org, Michael Ellerman <mpe@ellerman.id.au>,
Jason Gunthorpe <jgunthorpe@obsidianresearch.com>,
Doug Ledford <dledford@redhat.com>,
Hal Rosenstock <hal.rosenstock@gmail.com>,
Sean Hefty <sean.hefty@intel.com>,
Dave Chinner <david@fromorbit.com>,
linux-fsdevel@vger.kernel.org,
Alexander Viro <viro@zeniv.linux.org.uk>,
Gerald Schaefer <gerald.schaefer@de.ibm.com>,
linux-nvdimm@lists.01.org, linux-kernel@vger.kernel.org,
linux-xfs@vger.kernel.org,
Martin Schwidefsky <schwidefsky@de.ibm.com>,
akpm@linux-foundation.org,
"Darrick J. Wong" <darrick.wong@oracle.com>,
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Subject: Re: [PATCH v3 00/13] dax: fix dma vs truncate and remove 'page-less' support
Date: Fri, 20 Oct 2017 09:47:50 +0200 [thread overview]
Message-ID: <20171020074750.GA13568@lst.de> (raw)
In-Reply-To: <150846713528.24336.4459262264611579791.stgit@dwillia2-desk3.amr.corp.intel.com>
> The solution presented is not pretty. It creates a stream of leases, one
> for each get_user_pages() invocation, and polls page reference counts
> until DMA stops. We're missing a reliable way to not only trap the
> DMA-idle event, but also block new references being taken on pages while
> truncate is allowed to progress. "[PATCH v3 12/13] dax: handle truncate of
> dma-busy pages" presents other options considered, and notes that this
> solution can only be viewed as a stop-gap.
I'd like to brainstorm how we can do something better.
How about:
If we hit a page with an elevated refcount in truncate / hole puch
etc for a DAX file system we do not free the blocks in the file system,
but add it to the extent busy list. We mark the page as delayed
free (e.g. page flag?) so that when it finally hits refcount zero we
call back into the file system to remove it from the busy list.
_______________________________________________
Linux-nvdimm mailing list
Linux-nvdimm@lists.01.org
https://lists.01.org/mailman/listinfo/linux-nvdimm
WARNING: multiple messages have this Message-ID (diff)
From: Christoph Hellwig <hch@lst.de>
To: Dan Williams <dan.j.williams@intel.com>
Cc: akpm@linux-foundation.org, Michal Hocko <mhocko@suse.com>,
Jan Kara <jack@suse.cz>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Dave Hansen <dave.hansen@linux.intel.com>,
Dave Chinner <david@fromorbit.com>,
"J. Bruce Fields" <bfields@fieldses.org>,
linux-mm@kvack.org, Paul Mackerras <paulus@samba.org>,
Sean Hefty <sean.hefty@intel.com>,
Jeff Layton <jlayton@poochiereds.net>,
Matthew Wilcox <mawilcox@microsoft.com>,
linux-rdma@vger.kernel.org, Michael Ellerman <mpe@ellerman.id.au>,
Jeff Moyer <jmoyer@redhat.com>,
hch@lst.de, Jason Gunthorpe <jgunthorpe@obsidianresearch.com>,
Doug Ledford <dledford@redhat.com>,
Ross Zwisler <ross.zwisler@linux.intel.com>,
Hal Rosenstock <hal.rosenstock@gmail.com>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
linux-nvdimm@lists.01.org,
Alexander Viro <viro@zeniv.linux.org.uk>,
Gerald Schaefer <gerald.schaefer@de.ibm.com>,
Darri
Subject: Re: [PATCH v3 00/13] dax: fix dma vs truncate and remove 'page-less' support
Date: Fri, 20 Oct 2017 09:47:50 +0200 [thread overview]
Message-ID: <20171020074750.GA13568@lst.de> (raw)
In-Reply-To: <150846713528.24336.4459262264611579791.stgit@dwillia2-desk3.amr.corp.intel.com>
> The solution presented is not pretty. It creates a stream of leases, one
> for each get_user_pages() invocation, and polls page reference counts
> until DMA stops. We're missing a reliable way to not only trap the
> DMA-idle event, but also block new references being taken on pages while
> truncate is allowed to progress. "[PATCH v3 12/13] dax: handle truncate of
> dma-busy pages" presents other options considered, and notes that this
> solution can only be viewed as a stop-gap.
I'd like to brainstorm how we can do something better.
How about:
If we hit a page with an elevated refcount in truncate / hole puch
etc for a DAX file system we do not free the blocks in the file system,
but add it to the extent busy list. We mark the page as delayed
free (e.g. page flag?) so that when it finally hits refcount zero we
call back into the file system to remove it from the busy list.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
WARNING: multiple messages have this Message-ID (diff)
From: Christoph Hellwig <hch@lst.de>
To: Dan Williams <dan.j.williams@intel.com>
Cc: akpm@linux-foundation.org, Michal Hocko <mhocko@suse.com>,
Jan Kara <jack@suse.cz>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Dave Hansen <dave.hansen@linux.intel.com>,
Dave Chinner <david@fromorbit.com>,
"J. Bruce Fields" <bfields@fieldses.org>,
linux-mm@kvack.org, Paul Mackerras <paulus@samba.org>,
Sean Hefty <sean.hefty@intel.com>,
Jeff Layton <jlayton@poochiereds.net>,
Matthew Wilcox <mawilcox@microsoft.com>,
linux-rdma@vger.kernel.org, Michael Ellerman <mpe@ellerman.id.au>,
Jeff Moyer <jmoyer@redhat.com>,
hch@lst.de, Jason Gunthorpe <jgunthorpe@obsidianresearch.com>,
Doug Ledford <dledford@redhat.com>,
Ross Zwisler <ross.zwisler@linux.intel.com>,
Hal Rosenstock <hal.rosenstock@gmail.com>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
linux-nvdimm@lists.01.org,
Alexander Viro <viro@zeniv.linux.org.uk>,
Gerald Schaefer <gerald.schaefer@de.ibm.com>,
"Darrick J. Wong" <darrick.wong@oracle.com>,
linux-kernel@vger.kernel.org, linux-xfs@vger.kernel.org,
Martin Schwidefsky <schwidefsky@de.ibm.com>,
linux-fsdevel@vger.kernel.org,
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Subject: Re: [PATCH v3 00/13] dax: fix dma vs truncate and remove 'page-less' support
Date: Fri, 20 Oct 2017 09:47:50 +0200 [thread overview]
Message-ID: <20171020074750.GA13568@lst.de> (raw)
In-Reply-To: <150846713528.24336.4459262264611579791.stgit@dwillia2-desk3.amr.corp.intel.com>
> The solution presented is not pretty. It creates a stream of leases, one
> for each get_user_pages() invocation, and polls page reference counts
> until DMA stops. We're missing a reliable way to not only trap the
> DMA-idle event, but also block new references being taken on pages while
> truncate is allowed to progress. "[PATCH v3 12/13] dax: handle truncate of
> dma-busy pages" presents other options considered, and notes that this
> solution can only be viewed as a stop-gap.
I'd like to brainstorm how we can do something better.
How about:
If we hit a page with an elevated refcount in truncate / hole puch
etc for a DAX file system we do not free the blocks in the file system,
but add it to the extent busy list. We mark the page as delayed
free (e.g. page flag?) so that when it finally hits refcount zero we
call back into the file system to remove it from the busy list.
WARNING: multiple messages have this Message-ID (diff)
From: Christoph Hellwig <hch@lst.de>
To: Dan Williams <dan.j.williams@intel.com>
Cc: akpm@linux-foundation.org, Michal Hocko <mhocko@suse.com>,
Jan Kara <jack@suse.cz>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Dave Hansen <dave.hansen@linux.intel.com>,
Dave Chinner <david@fromorbit.com>,
"J. Bruce Fields" <bfields@fieldses.org>,
linux-mm@kvack.org, Paul Mackerras <paulus@samba.org>,
Sean Hefty <sean.hefty@intel.com>,
Jeff Layton <jlayton@poochiereds.net>,
Matthew Wilcox <mawilcox@microsoft.com>,
linux-rdma@vger.kernel.org, Michael Ellerman <mpe@ellerman.id.au>,
Jeff Moyer <jmoyer@redhat.com>,
hch@lst.de, Jason Gunthorpe <jgunthorpe@obsidianresearch.com>,
Doug Ledford <dledford@redhat.com>,
Ross Zwisler <ross.zwisler@linux.intel.com>,
Hal Rosenstock <hal.rosenstock@gmail.com>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
linux-nvdimm@lists.01.org,
Alexander Viro <viro@zeniv.linux.org.uk>,
Gerald Schaefer <gerald.schaefer@de.ibm.com>,
"Darrick J. Wong" <darrick.wong@oracle.com>,
linux-kernel@vger.kernel.org, linux-xfs@vger.kernel.org,
Martin Schwidefsky <schwidefsky@de.ibm.com>,
linux-fsdevel@vger.kernel.org,
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Subject: Re: [PATCH v3 00/13] dax: fix dma vs truncate and remove 'page-less' support
Date: Fri, 20 Oct 2017 09:47:50 +0200 [thread overview]
Message-ID: <20171020074750.GA13568@lst.de> (raw)
In-Reply-To: <150846713528.24336.4459262264611579791.stgit@dwillia2-desk3.amr.corp.intel.com>
> The solution presented is not pretty. It creates a stream of leases, one
> for each get_user_pages() invocation, and polls page reference counts
> until DMA stops. We're missing a reliable way to not only trap the
> DMA-idle event, but also block new references being taken on pages while
> truncate is allowed to progress. "[PATCH v3 12/13] dax: handle truncate of
> dma-busy pages" presents other options considered, and notes that this
> solution can only be viewed as a stop-gap.
I'd like to brainstorm how we can do something better.
How about:
If we hit a page with an elevated refcount in truncate / hole puch
etc for a DAX file system we do not free the blocks in the file system,
but add it to the extent busy list. We mark the page as delayed
free (e.g. page flag?) so that when it finally hits refcount zero we
call back into the file system to remove it from the busy list.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2017-10-20 7:44 UTC|newest]
Thread overview: 143+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-20 2:38 [PATCH v3 00/13] dax: fix dma vs truncate and remove 'page-less' support Dan Williams
2017-10-20 2:38 ` Dan Williams
2017-10-20 2:38 ` Dan Williams
2017-10-20 2:38 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 01/13] dax: quiet bdev_dax_supported() Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 02/13] dax: require 'struct page' for filesystem dax Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 7:57 ` Christoph Hellwig
2017-10-20 7:57 ` Christoph Hellwig
2017-10-20 15:23 ` Dan Williams
2017-10-20 15:23 ` Dan Williams
2017-10-20 15:23 ` Dan Williams
2017-10-20 16:29 ` Christoph Hellwig
2017-10-20 16:29 ` Christoph Hellwig
2017-10-20 16:29 ` Christoph Hellwig
2017-10-20 16:29 ` Christoph Hellwig
2017-10-20 22:29 ` Dan Williams
2017-10-20 22:29 ` Dan Williams
2017-10-20 22:29 ` Dan Williams
2017-10-21 3:20 ` Matthew Wilcox
2017-10-21 3:20 ` Matthew Wilcox
2017-10-21 3:20 ` Matthew Wilcox
2017-10-21 4:16 ` Dan Williams
2017-10-21 4:16 ` Dan Williams
2017-10-21 4:16 ` Dan Williams
2017-10-21 8:15 ` Christoph Hellwig
2017-10-21 8:15 ` Christoph Hellwig
2017-10-21 8:15 ` Christoph Hellwig
2017-10-23 5:18 ` Martin Schwidefsky
2017-10-23 5:18 ` Martin Schwidefsky
2017-10-23 5:18 ` Martin Schwidefsky
2017-10-23 8:55 ` Dan Williams
2017-10-23 8:55 ` Dan Williams
2017-10-23 10:44 ` Martin Schwidefsky
2017-10-23 10:44 ` Martin Schwidefsky
2017-10-23 10:44 ` Martin Schwidefsky
2017-10-23 11:20 ` Dan Williams
2017-10-23 11:20 ` Dan Williams
2017-10-23 11:20 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 03/13] dax: stop using VM_MIXEDMAP for dax Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 04/13] dax: stop using VM_HUGEPAGE " Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 05/13] dax: stop requiring a live device for dax_flush() Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 06/13] dax: store pfns in the radix Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 07/13] dax: warn if dma collides with truncate Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 08/13] tools/testing/nvdimm: add 'bio_delay' mechanism Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 09/13] IB/core: disable memory registration of fileystem-dax vmas Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 10/13] mm: disable get_user_pages_fast() for dax Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` [PATCH v3 11/13] fs: use smp_load_acquire in break_{layout,lease} Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 2:39 ` Dan Williams
2017-10-20 12:39 ` Jeffrey Layton
2017-10-20 12:39 ` Jeffrey Layton
2017-10-20 12:39 ` Jeffrey Layton
2017-10-20 12:39 ` Jeffrey Layton
2017-10-20 2:40 ` [PATCH v3 12/13] dax: handle truncate of dma-busy pages Dan Williams
2017-10-20 2:40 ` Dan Williams
2017-10-20 2:40 ` Dan Williams
2017-10-20 13:05 ` Jeff Layton
2017-10-20 13:05 ` Jeff Layton
2017-10-20 13:05 ` Jeff Layton
2017-10-20 15:42 ` Dan Williams
2017-10-20 15:42 ` Dan Williams
2017-10-20 15:42 ` Dan Williams
2017-10-20 16:32 ` Christoph Hellwig
2017-10-20 16:32 ` Christoph Hellwig
2017-10-20 16:32 ` Christoph Hellwig
2017-10-20 17:27 ` Dan Williams
2017-10-20 17:27 ` Dan Williams
2017-10-20 17:27 ` Dan Williams
2017-10-20 20:36 ` Brian Foster
2017-10-20 20:36 ` Brian Foster
2017-10-20 20:36 ` Brian Foster
2017-10-21 8:11 ` Christoph Hellwig
2017-10-21 8:11 ` Christoph Hellwig
2017-10-20 2:40 ` [PATCH v3 13/13] xfs: wire up FL_ALLOCATED support Dan Williams
2017-10-20 2:40 ` Dan Williams
2017-10-20 2:40 ` Dan Williams
2017-10-20 7:47 ` Christoph Hellwig [this message]
2017-10-20 7:47 ` [PATCH v3 00/13] dax: fix dma vs truncate and remove 'page-less' support Christoph Hellwig
2017-10-20 7:47 ` Christoph Hellwig
2017-10-20 7:47 ` Christoph Hellwig
2017-10-20 9:31 ` Christoph Hellwig
2017-10-20 9:31 ` Christoph Hellwig
2017-10-20 9:31 ` Christoph Hellwig
2017-10-26 10:58 ` Jan Kara
2017-10-26 10:58 ` Jan Kara
2017-10-26 10:58 ` Jan Kara
2017-10-26 10:58 ` Jan Kara
2017-10-26 23:51 ` Williams, Dan J
2017-10-26 23:51 ` Williams, Dan J
2017-10-26 23:51 ` Williams, Dan J
2017-10-26 23:51 ` Williams, Dan J
2017-10-27 6:48 ` Dave Chinner
2017-10-27 6:48 ` Dave Chinner
2017-10-27 6:48 ` Dave Chinner
2017-10-27 6:48 ` Dave Chinner
2017-10-27 6:48 ` Dave Chinner
2017-10-27 11:42 ` Dan Williams
2017-10-27 11:42 ` Dan Williams
2017-10-27 11:42 ` Dan Williams
2017-10-29 21:52 ` Dave Chinner
2017-10-29 21:52 ` Dave Chinner
2017-10-29 21:52 ` Dave Chinner
2017-10-27 6:45 ` Christoph Hellwig
2017-10-27 6:45 ` Christoph Hellwig
2017-10-27 6:45 ` Christoph Hellwig
2017-10-29 23:46 ` Dan Williams
2017-10-29 23:46 ` Dan Williams
2017-10-29 23:46 ` Dan Williams
2017-10-30 2:00 ` Dave Chinner
2017-10-30 2:00 ` Dave Chinner
2017-10-30 2:00 ` Dave Chinner
2017-10-30 2:00 ` Dave Chinner
2017-10-30 8:38 ` Jan Kara
2017-10-30 8:38 ` Jan Kara
2017-10-30 8:38 ` Jan Kara
2017-10-30 11:20 ` Dave Chinner
2017-10-30 11:20 ` Dave Chinner
2017-10-30 11:20 ` Dave Chinner
2017-10-30 11:20 ` Dave Chinner
2017-10-30 17:51 ` Dan Williams
2017-10-30 17:51 ` Dan Williams
2017-10-30 17:51 ` Dan Williams
2017-10-30 17:51 ` Dan Williams
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=20171020074750.GA13568@lst.de \
--to=hch@lst.de \
--cc=akpm@linux-foundation.org \
--cc=benh@kernel.crashing.org \
--cc=bfields@fieldses.org \
--cc=dan.j.williams@intel.com \
--cc=darrick.wong@oracle.com \
--cc=dave.hansen@linux.intel.com \
--cc=david@fromorbit.com \
--cc=dledford@redhat.com \
--cc=gerald.schaefer@de.ibm.com \
--cc=hal.rosenstock@gmail.com \
--cc=heiko.carstens@de.ibm.com \
--cc=jack@suse.cz \
--cc=jgunthorpe@obsidianresearch.com \
--cc=jlayton@poochiereds.net \
--cc=kirill.shutemov@linux.intel.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-nvdimm@lists.01.org \
--cc=linux-rdma@vger.kernel.org \
--cc=linux-xfs@vger.kernel.org \
--cc=mawilcox@microsoft.com \
--cc=mhocko@suse.com \
--cc=mpe@ellerman.id.au \
--cc=paulus@samba.org \
--cc=schwidefsky@de.ibm.com \
--cc=sean.hefty@intel.com \
--cc=viro@zeniv.linux.org.uk \
/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.