From: Pankaj Gupta <pagupta@redhat.com>
To: Luiz Capitulino <lcapitulino@redhat.com>
Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
qemu-devel@nongnu.org, linux-nvdimm@ml01.01.org, jack@suse.cz,
stefanha@redhat.com, dan j williams <dan.j.williams@intel.com>,
riel@surriel.com, nilal@redhat.com, kwolf@redhat.com,
pbonzini@redhat.com, ross zwisler <ross.zwisler@intel.com>,
david@redhat.com,
xiaoguangrong eric <xiaoguangrong.eric@gmail.com>,
hch@infradead.org, mst@redhat.com, niteshnarayanlal@hotmail.com,
imammedo@redhat.com, eblake@redhat.com
Subject: Re: [RFC v3 1/2] libnvdimm: Add flush callback for virtio pmem
Date: Mon, 16 Jul 2018 04:13:47 -0400 (EDT) [thread overview]
Message-ID: <9008103.50980761.1531728827436.JavaMail.zimbra@redhat.com> (raw)
In-Reply-To: <20180713163559.692aca30@doriath>
Hi Luiz,
>
> > This patch adds functionality to perform flush from guest to host
> > over VIRTIO. We are registering a callback based on 'nd_region' type.
> > As virtio_pmem driver requires this special flush interface, for rest
> > of the region types we are registering existing flush function.
> > Also report the error returned by virtio flush interface.
>
> This patch doesn't apply against latest upstream. A few more comments
> below.
My bad, I tested it with 4.17-rc1. Will rebase it.
>
> >
> > Signed-off-by: Pankaj Gupta <pagupta@redhat.com>
> > ---
> > drivers/nvdimm/nd.h | 1 +
> > drivers/nvdimm/pmem.c | 4 ++--
> > drivers/nvdimm/region_devs.c | 24 ++++++++++++++++++------
> > include/linux/libnvdimm.h | 5 ++++-
> > 4 files changed, 25 insertions(+), 9 deletions(-)
> >
> > diff --git a/drivers/nvdimm/nd.h b/drivers/nvdimm/nd.h
> > index 32e0364..1b62f79 100644
> > --- a/drivers/nvdimm/nd.h
> > +++ b/drivers/nvdimm/nd.h
> > @@ -159,6 +159,7 @@ struct nd_region {
> > struct badblocks bb;
> > struct nd_interleave_set *nd_set;
> > struct nd_percpu_lane __percpu *lane;
> > + int (*flush)(struct device *dev);
> > struct nd_mapping mapping[0];
> > };
> >
> > diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c
> > index 9d71492..29fd2cd 100644
> > --- a/drivers/nvdimm/pmem.c
> > +++ b/drivers/nvdimm/pmem.c
> > @@ -180,7 +180,7 @@ static blk_qc_t pmem_make_request(struct request_queue
> > *q, struct bio *bio)
> > struct nd_region *nd_region = to_region(pmem);
> >
> > if (bio->bi_opf & REQ_FLUSH)
> > - nvdimm_flush(nd_region);
> > + bio->bi_status = nvdimm_flush(nd_region);
> >
> > do_acct = nd_iostat_start(bio, &start);
> > bio_for_each_segment(bvec, bio, iter) {
> > @@ -196,7 +196,7 @@ static blk_qc_t pmem_make_request(struct request_queue
> > *q, struct bio *bio)
> > nd_iostat_end(bio, start);
> >
> > if (bio->bi_opf & REQ_FUA)
> > - nvdimm_flush(nd_region);
> > + bio->bi_status = nvdimm_flush(nd_region);
> >
> > bio_endio(bio);
> > return BLK_QC_T_NONE;
> > diff --git a/drivers/nvdimm/region_devs.c b/drivers/nvdimm/region_devs.c
> > index a612be6..124aae7 100644
> > --- a/drivers/nvdimm/region_devs.c
> > +++ b/drivers/nvdimm/region_devs.c
> > @@ -1025,6 +1025,7 @@ static struct nd_region *nd_region_create(struct
> > nvdimm_bus *nvdimm_bus,
> > dev->of_node = ndr_desc->of_node;
> > nd_region->ndr_size = resource_size(ndr_desc->res);
> > nd_region->ndr_start = ndr_desc->res->start;
> > + nd_region->flush = ndr_desc->flush;
> > nd_device_register(dev);
> >
> > return nd_region;
> > @@ -1065,13 +1066,10 @@ struct nd_region
> > *nvdimm_volatile_region_create(struct nvdimm_bus *nvdimm_bus,
> > }
> > EXPORT_SYMBOL_GPL(nvdimm_volatile_region_create);
> >
> > -/**
> > - * nvdimm_flush - flush any posted write queues between the cpu and pmem
> > media
> > - * @nd_region: blk or interleaved pmem region
> > - */
> > -void nvdimm_flush(struct nd_region *nd_region)
> > +void pmem_flush(struct device *dev)
> > {
> > - struct nd_region_data *ndrd = dev_get_drvdata(&nd_region->dev);
> > + struct nd_region_data *ndrd = dev_get_drvdata(dev);
> > + struct nd_region *nd_region = to_nd_region(dev);
> > int i, idx;
> >
> > /*
> > @@ -1094,6 +1092,20 @@ void nvdimm_flush(struct nd_region *nd_region)
> > writeq(1, ndrd_get_flush_wpq(ndrd, i, idx));
> > wmb();
> > }
> > +
> > +/**
> > + * nvdimm_flush - flush any posted write queues between the cpu and pmem
> > media
> > + * @nd_region: blk or interleaved pmem region
> > + */
> > +int nvdimm_flush(struct nd_region *nd_region)
> > +{
> > + if (nd_region->flush)
> > + return(nd_region->flush(&nd_region->dev));
> > +
> > + pmem_flush(&nd_region->dev);
>
> IMHO, a better way of doing this would be to allow nvdimm_flush() to
> be overridden. That is, in nd_region_create() you set nd_region->flush
> to the original nvdimm_flush() if ndr_desc->flush is NULL. And then
> always call nd_region->flush() where nvdimm_flush() is called today.
I wanted to do minimal changes for actual 'nvdimm_flush' function because it
does not return an error or return status for fsync. So, I needed to differentiate
between 'fake DAX' & 'NVDIMM' at the time of calling 'flush', otherwise I need to
change 'nvdimm_flush' to return zero for all the calls.
Looks like I am already doing this, will change as suggested.
>
> > +
> > + return 0;
> > +}
> > EXPORT_SYMBOL_GPL(nvdimm_flush);
> >
> > /**
> > diff --git a/include/linux/libnvdimm.h b/include/linux/libnvdimm.h
> > index 097072c..33b617f 100644
> > --- a/include/linux/libnvdimm.h
> > +++ b/include/linux/libnvdimm.h
> > @@ -126,6 +126,7 @@ struct nd_region_desc {
> > int numa_node;
> > unsigned long flags;
> > struct device_node *of_node;
> > + int (*flush)(struct device *dev);
> > };
> >
> > struct device;
> > @@ -201,7 +202,9 @@ unsigned long nd_blk_memremap_flags(struct
> > nd_blk_region *ndbr);
> > unsigned int nd_region_acquire_lane(struct nd_region *nd_region);
> > void nd_region_release_lane(struct nd_region *nd_region, unsigned int
> > lane);
> > u64 nd_fletcher64(void *addr, size_t len, bool le);
> > -void nvdimm_flush(struct nd_region *nd_region);
> > +int nvdimm_flush(struct nd_region *nd_region);
> > +void pmem_set_flush(struct nd_region *nd_region, void (*flush)
> > + (struct device *));
>
> It seems pmem_set_flush() doesn't exist.
Sorry! will remove it.
>
> > int nvdimm_has_flush(struct nd_region *nd_region);
> > int nvdimm_has_cache(struct nd_region *nd_region);
> >
>
>
Thanks,
Pankaj
next prev parent reply other threads:[~2018-07-16 8:13 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-07-13 7:52 [RFC v3 0/2] kvm "fake DAX" device flushing Pankaj Gupta
2018-07-13 7:52 ` [RFC v3 1/2] libnvdimm: Add flush callback for virtio pmem Pankaj Gupta
2018-07-13 20:35 ` Luiz Capitulino
2018-07-16 8:13 ` Pankaj Gupta [this message]
2018-07-13 7:52 ` [RFC v3 2/2] virtio-pmem: Add virtio pmem driver Pankaj Gupta
2018-07-13 20:38 ` Luiz Capitulino
2018-07-16 11:46 ` [Qemu-devel] " Pankaj Gupta
2018-07-16 14:03 ` Luiz Capitulino
2018-07-16 15:11 ` Pankaj Gupta
2018-07-17 13:11 ` Stefan Hajnoczi
2018-07-18 7:05 ` Pankaj Gupta
2018-07-13 7:52 ` [RFC v3] qemu: Add virtio pmem device Pankaj Gupta
2018-07-18 12:55 ` Luiz Capitulino
2018-07-19 5:48 ` [Qemu-devel] " Pankaj Gupta
2018-07-19 12:16 ` Stefan Hajnoczi
2018-07-19 12:48 ` Luiz Capitulino
2018-07-19 12:57 ` Luiz Capitulino
2018-07-20 13:04 ` Pankaj Gupta
2018-07-19 13:58 ` David Hildenbrand
2018-07-19 15:48 ` Luiz Capitulino
2018-07-20 13:02 ` Pankaj Gupta
2018-07-19 12:39 ` Luiz Capitulino
2018-07-24 16:13 ` Eric Blake
2018-07-25 5:01 ` [Qemu-devel] " Pankaj Gupta
2018-07-25 12:19 ` Eric Blake
2018-07-25 12:47 ` Pankaj Gupta
2018-08-28 12:13 ` [RFC v3 0/2] kvm "fake DAX" device flushing David Hildenbrand
2018-08-28 12:39 ` [Qemu-devel] " Pankaj Gupta
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=9008103.50980761.1531728827436.JavaMail.zimbra@redhat.com \
--to=pagupta@redhat.com \
--cc=dan.j.williams@intel.com \
--cc=david@redhat.com \
--cc=eblake@redhat.com \
--cc=hch@infradead.org \
--cc=imammedo@redhat.com \
--cc=jack@suse.cz \
--cc=kvm@vger.kernel.org \
--cc=kwolf@redhat.com \
--cc=lcapitulino@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nvdimm@ml01.01.org \
--cc=mst@redhat.com \
--cc=nilal@redhat.com \
--cc=niteshnarayanlal@hotmail.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=riel@surriel.com \
--cc=ross.zwisler@intel.com \
--cc=stefanha@redhat.com \
--cc=xiaoguangrong.eric@gmail.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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).