From: Dan Williams <dan.j.williams@intel.com> To: Ross Zwisler <ross.zwisler@linux.intel.com>, Dave Chinner <david@fromorbit.com>, Dan Williams <dan.j.williams@intel.com>, Brian Foster <bfoster@redhat.com>, Jan Kara <jack@suse.cz>, xfs@oss.sgi.com, linux-fsdevel <linux-fsdevel@vger.kernel.org>, "linux-nvdimm@lists.01.org" <linux-nvdimm@lists.01.org> Subject: Re: [PATCH 3/6] xfs: Don't use unwritten extents for DAX Date: Tue, 3 Nov 2015 17:02:34 -0800 [thread overview] Message-ID: <CAPcyv4hbrM4+P-8=SXU8BFP8tr1Dw9u1zSU9o0M=wzYjzsp8rw@mail.gmail.com> (raw) In-Reply-To: <20151104005056.GA24710@linux.intel.com> On Tue, Nov 3, 2015 at 4:50 PM, Ross Zwisler <ross.zwisler@linux.intel.com> wrote: > On Tue, Nov 03, 2015 at 04:04:13PM +1100, Dave Chinner wrote: >> On Mon, Nov 02, 2015 at 07:53:27PM -0800, Dan Williams wrote: >> > On Mon, Nov 2, 2015 at 1:44 PM, Dave Chinner <david@fromorbit.com> wrote: > <> >> > > This comes back to the comments I made w.r.t. the pmem driver >> > > implementation doing synchronous IO by immediately forcing CPU cache >> > > flushes and barriers. it's obviously correct, but it looks like >> > > there's going to be a major performance penalty associated with it. >> > > This is why I recently suggested that a pmem driver that doesn't do >> > > CPU cache writeback during IO but does it on REQ_FLUSH is an >> > > architecture we'll likely have to support. >> > > >> > >> > The only thing we can realistically delay is wmb_pmem() i.e. the final >> > sync waiting for data that has *left* the cpu cache. Unless/until we >> > get a architecturally guaranteed method to write-back the entire >> > cache, or flush the cache by physical-cache-way we're stuck with >> > either non-temporal cycles or looping on potentially huge virtual >> > address ranges. >> >> I'm missing something: why won't flushing the address range returned >> by bdev_direct_access() during a fsync operation work? i.e. we're >> working with exactly the same address as dax_clear_blocks() and >> dax_do_io() use, so why can't we look up that address and flush it >> from fsync? > > I could be wrong, but I don't see a reason why DAX can't use the strategy of > writing data and marking it dirty in one step and then flushing later in > response to fsync/msync. I think this could be used everywhere we write or > zero data - dax_clear_blocks(), dax_io() etc. (I believe that lots of the > block zeroing code will go away once we have the XFS and ext4 patches in that > guarantee we will only get written and zeroed extents from the filesystem in > response to get_block().) I think the PMEM driver, lacking the ability to > mark things as dirty in the radix tree, etc, will need to keep doing things > synchronously. Not without numbers showing the relative performance of dirtying cache followed by flushing vs non-temporal + pcommit. > Hmm...if we go this path, though, is that an argument against moving the > zeroing from DAX down into the driver? True, with BRD it makes things nice > and efficient because you can zero and never flush, and the driver knows > there's nothing else to do. > > For PMEM, though, you lose the ability to zero the data and then queue the > flushing for later, as you would be able to do if you left the zeroing code in > DAX. The benefit of this is that if you are going to immediately re-write the > newly zeroed data (which seems common), PMEM will end up doing an extra cache > flush of the zeroes, only to have them overwritten and marked as dirty by DAX. > If we leave the zeroing to DAX we can mark it dirty once, zero it once, write > it once, and flush it once. Why do we lose the ability to flush later if the driver supports blkdev_issue_zeroout? > This would make us lose the ability to do hardware-assisted flushing in the > future that requires driver specific knowledge, though I don't think that > exists yet. ioatdma has supported memset() for a while now, but I would prioritize a non-temporal SIMD implementation first. > Perhaps we should leave the zeroing in DAX for now to take > advantage of the single flush, and then move it down if a driver can improve > performance with hardware assisted PMEM zeroing? Not convinced. I think we should implement the driver zeroing solution and take a look at performance.
WARNING: multiple messages have this Message-ID (diff)
From: Dan Williams <dan.j.williams@intel.com> To: Ross Zwisler <ross.zwisler@linux.intel.com>, Dave Chinner <david@fromorbit.com>, Dan Williams <dan.j.williams@intel.com>, Brian Foster <bfoster@redhat.com>, Jan Kara <jack@suse.cz>, xfs@oss.sgi.com, linux-fsdevel <linux-fsdevel@vger.kernel.org>, "linux-nvdimm@lists.01.org" <linux-nvdimm@lists.01.org> Subject: Re: [PATCH 3/6] xfs: Don't use unwritten extents for DAX Date: Tue, 3 Nov 2015 17:02:34 -0800 [thread overview] Message-ID: <CAPcyv4hbrM4+P-8=SXU8BFP8tr1Dw9u1zSU9o0M=wzYjzsp8rw@mail.gmail.com> (raw) In-Reply-To: <20151104005056.GA24710@linux.intel.com> On Tue, Nov 3, 2015 at 4:50 PM, Ross Zwisler <ross.zwisler@linux.intel.com> wrote: > On Tue, Nov 03, 2015 at 04:04:13PM +1100, Dave Chinner wrote: >> On Mon, Nov 02, 2015 at 07:53:27PM -0800, Dan Williams wrote: >> > On Mon, Nov 2, 2015 at 1:44 PM, Dave Chinner <david@fromorbit.com> wrote: > <> >> > > This comes back to the comments I made w.r.t. the pmem driver >> > > implementation doing synchronous IO by immediately forcing CPU cache >> > > flushes and barriers. it's obviously correct, but it looks like >> > > there's going to be a major performance penalty associated with it. >> > > This is why I recently suggested that a pmem driver that doesn't do >> > > CPU cache writeback during IO but does it on REQ_FLUSH is an >> > > architecture we'll likely have to support. >> > > >> > >> > The only thing we can realistically delay is wmb_pmem() i.e. the final >> > sync waiting for data that has *left* the cpu cache. Unless/until we >> > get a architecturally guaranteed method to write-back the entire >> > cache, or flush the cache by physical-cache-way we're stuck with >> > either non-temporal cycles or looping on potentially huge virtual >> > address ranges. >> >> I'm missing something: why won't flushing the address range returned >> by bdev_direct_access() during a fsync operation work? i.e. we're >> working with exactly the same address as dax_clear_blocks() and >> dax_do_io() use, so why can't we look up that address and flush it >> from fsync? > > I could be wrong, but I don't see a reason why DAX can't use the strategy of > writing data and marking it dirty in one step and then flushing later in > response to fsync/msync. I think this could be used everywhere we write or > zero data - dax_clear_blocks(), dax_io() etc. (I believe that lots of the > block zeroing code will go away once we have the XFS and ext4 patches in that > guarantee we will only get written and zeroed extents from the filesystem in > response to get_block().) I think the PMEM driver, lacking the ability to > mark things as dirty in the radix tree, etc, will need to keep doing things > synchronously. Not without numbers showing the relative performance of dirtying cache followed by flushing vs non-temporal + pcommit. > Hmm...if we go this path, though, is that an argument against moving the > zeroing from DAX down into the driver? True, with BRD it makes things nice > and efficient because you can zero and never flush, and the driver knows > there's nothing else to do. > > For PMEM, though, you lose the ability to zero the data and then queue the > flushing for later, as you would be able to do if you left the zeroing code in > DAX. The benefit of this is that if you are going to immediately re-write the > newly zeroed data (which seems common), PMEM will end up doing an extra cache > flush of the zeroes, only to have them overwritten and marked as dirty by DAX. > If we leave the zeroing to DAX we can mark it dirty once, zero it once, write > it once, and flush it once. Why do we lose the ability to flush later if the driver supports blkdev_issue_zeroout? > This would make us lose the ability to do hardware-assisted flushing in the > future that requires driver specific knowledge, though I don't think that > exists yet. ioatdma has supported memset() for a while now, but I would prioritize a non-temporal SIMD implementation first. > Perhaps we should leave the zeroing in DAX for now to take > advantage of the single flush, and then move it down if a driver can improve > performance with hardware assisted PMEM zeroing? Not convinced. I think we should implement the driver zeroing solution and take a look at performance. _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs
next prev parent reply other threads:[~2015-11-04 1:02 UTC|newest] Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top 2015-10-19 3:27 [PATCH 0/6 V2] xfs: upfront block zeroing for DAX Dave Chinner 2015-10-19 3:27 ` [PATCH 1/6] xfs: fix inode size update overflow in xfs_map_direct() Dave Chinner 2015-10-29 14:27 ` Brian Foster 2015-10-19 3:27 ` [PATCH 2/6] xfs: introduce BMAPI_ZERO for allocating zeroed extents Dave Chinner 2015-10-29 14:27 ` Brian Foster 2015-10-29 23:35 ` Dave Chinner 2015-10-30 12:36 ` Brian Foster 2015-11-02 1:21 ` Dave Chinner 2015-10-19 3:27 ` [PATCH 3/6] xfs: Don't use unwritten extents for DAX Dave Chinner 2015-10-29 14:29 ` Brian Foster 2015-10-29 23:37 ` Dave Chinner 2015-10-30 12:36 ` Brian Foster 2015-11-02 1:14 ` Dave Chinner 2015-11-02 14:15 ` Brian Foster 2015-11-02 21:44 ` Dave Chinner 2015-11-02 21:44 ` Dave Chinner 2015-11-02 21:44 ` Dave Chinner 2015-11-03 3:53 ` Dan Williams 2015-11-03 3:53 ` Dan Williams 2015-11-03 3:53 ` Dan Williams 2015-11-03 5:04 ` Dave Chinner 2015-11-03 5:04 ` Dave Chinner 2015-11-04 0:50 ` Ross Zwisler 2015-11-04 0:50 ` Ross Zwisler 2015-11-04 1:02 ` Dan Williams [this message] 2015-11-04 1:02 ` Dan Williams 2015-11-04 4:46 ` Ross Zwisler 2015-11-04 4:46 ` Ross Zwisler 2015-11-04 9:06 ` Jan Kara 2015-11-04 9:06 ` Jan Kara 2015-11-04 15:35 ` Ross Zwisler 2015-11-04 15:35 ` Ross Zwisler 2015-11-04 17:21 ` Jan Kara 2015-11-04 17:21 ` Jan Kara 2015-11-03 9:16 ` Jan Kara 2015-11-03 9:16 ` Jan Kara 2015-10-19 3:27 ` [PATCH 4/6] xfs: DAX does not use IO completion callbacks Dave Chinner 2015-10-29 14:29 ` Brian Foster 2015-10-29 23:39 ` Dave Chinner 2015-10-30 12:37 ` Brian Foster 2015-10-19 3:27 ` [PATCH 5/6] xfs: add ->pfn_mkwrite support for DAX Dave Chinner 2015-10-29 14:30 ` Brian Foster 2015-10-19 3:27 ` [PATCH 6/6] xfs: xfs_filemap_pmd_fault treats read faults as write faults Dave Chinner 2015-10-29 14:30 ` Brian Foster 2015-11-05 23:48 ` [PATCH 0/6 V2] xfs: upfront block zeroing for DAX Ross Zwisler 2015-11-06 22:32 ` Dave Chinner 2015-11-06 18:12 ` Boylston, Brian
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='CAPcyv4hbrM4+P-8=SXU8BFP8tr1Dw9u1zSU9o0M=wzYjzsp8rw@mail.gmail.com' \ --to=dan.j.williams@intel.com \ --cc=bfoster@redhat.com \ --cc=david@fromorbit.com \ --cc=jack@suse.cz \ --cc=linux-fsdevel@vger.kernel.org \ --cc=linux-nvdimm@lists.01.org \ --cc=ross.zwisler@linux.intel.com \ --cc=xfs@oss.sgi.com \ /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: linkBe 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.