From: Jeff Layton <jlayton@poochiereds.net>
To: "J. Bruce Fields" <bfields@fieldses.org>
Cc: anna.schumaker@netapp.com, trondmy@primarydata.com,
tigran.mkrtchyan@desy.de, thomas.haynes@primarydata.com,
linux-nfs@vger.kernel.org
Subject: Re: [PATCH v2 3/3] pnfs: add a new mechanism to select a layout driver according to an ordered list
Date: Mon, 13 Jun 2016 21:46:47 -0400 [thread overview]
Message-ID: <1465868807.12825.7.camel@poochiereds.net> (raw)
In-Reply-To: <20160613200525.GD19825@fieldses.org>
On Mon, 2016-06-13 at 16:05 -0400, J. Bruce Fields wrote:
> On Thu, Jun 09, 2016 at 08:21:05AM -0400, Jeff Layton wrote:
> > Currently, the layout driver selection code always chooses the first one
> > from the list. That's not really ideal however, as the server can send
> > the list of layout types in any order that it likes. It's up to the
> > client to select the best one for its needs.
> >
> > This patch adds an ordered list of preferred driver types and has the
> > selection code sort the list of avaiable layout drivers according to it.
> > Any unrecognized layout type is sorted to the end of the list.
> >
> > For now, the order of preference is hardcoded, but it should be possible
> > to make this configurable in the future.
> >
> > Signed-off-by: Jeff Layton <jeff.layton@primarydata.com>
> > ---
> > fs/nfs/pnfs.c | 57 ++++++++++++++++++++++++++++++++++++++++++++++++++-------
> > 1 file changed, 50 insertions(+), 7 deletions(-)
> >
> > diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
> > index 2fc1b9354345..c31d30f1e5f2 100644
> > --- a/fs/nfs/pnfs.c
> > +++ b/fs/nfs/pnfs.c
> > @@ -99,6 +99,21 @@ unset_pnfs_layoutdriver(struct nfs_server *nfss)
> > }
> >
> > /*
> > + * When the server sends a list of layout types, we choose one in the order
> > + * given in the list below.
> > + *
> > + * FIXME: should this list be configurable via module_param or mount option?
> > + */
> > +static const u32 ld_prefs[] = {
> > + LAYOUT_SCSI,
> > + LAYOUT_BLOCK_VOLUME,
> > + LAYOUT_OSD2_OBJECTS,
> > + LAYOUT_FLEX_FILES,
> > + LAYOUT_NFSV4_1_FILES,
> > + 0
> > +};
> > +
> > +/*
> > * Try to set the server's pnfs module to the pnfs layout type specified by id.
> > * Currently only one pNFS layout driver per filesystem is supported.
> > *
> > @@ -110,6 +125,8 @@ set_pnfs_layoutdriver(struct nfs_server *server, const struct nfs_fh *mntfh,
> > {
> > struct pnfs_layoutdriver_type *ld_type = NULL;
> > u32 id;
> > + int i, j;
> > + bool swapped;
> >
> > if (!(server->nfs_client->cl_exchange_flags &
> > (EXCHGID4_FLAG_USE_NON_PNFS | EXCHGID4_FLAG_USE_PNFS_MDS))) {
> > @@ -118,18 +135,44 @@ set_pnfs_layoutdriver(struct nfs_server *server, const struct nfs_fh *mntfh,
> > goto out_no_driver;
> > }
> >
> > - id = ids[0];
> > - if (!id)
> > - goto out_no_driver;
> > + /* bubble sort into ld_prefs order */
> > + do {
> > + swapped = false;
> > + for (i = 0; i < (NFS_MAX_LAYOUT_TYPES - 1); i++) {
> >
> > - ld_type = find_pnfs_driver(id);
> > - if (!ld_type) {
> > - request_module("%s-%u", LAYOUT_NFSV4_1_MODULE_PREFIX, id);
> > + /* If either is zero then we're done with this pass */
> > + if (ids[i] == 0 || ids[i + 1] == 0)
> > + break;
> > +
> > + for (j = 0; ld_prefs[j] != 0; j++) {
> > + /* If we match left entry first, no swap */
> > + if (ids[i] == ld_prefs[j])
> > + break;
> > +
> > + /* if we match the right, we need to swap */
> > + if (ids[i + 1] == ld_prefs[j]) {
> > + swap(ids[i], ids[i + 1]);
> > + swapped = true;
> > + break;
> > + }
> > + }
> > + }
> > + } while (swapped);
>
> Maybe I'm strange, but, yes, to me this is easier to read as on a first
> pass I can just take your word that the above is doing a sort and see
> much more quickly what's happening....
>
Yeah, it is for me too, actually. Good call on asking for it. ;)
That said, I did notice the code in lib/sort.c after I sent this
series. It might be better to use that instead of this hand-rolled
bubble sort (if only for conciseness -- I doubt it matters for
efficiency).
> Anyway, ACK to the series from me.
>
Thanks!
> --b.
>
> > +
> > + for (i = 0; i < NFS_MAX_LAYOUT_TYPES && ids[i] != 0; i++) {
> > + id = ids[i];
> > ld_type = find_pnfs_driver(id);
> > + if (!ld_type) {
> > + request_module("%s-%u", LAYOUT_NFSV4_1_MODULE_PREFIX,
> > + id);
> > + ld_type = find_pnfs_driver(id);
> > + }
> > + if (ld_type)
> > + break;
> > }
> >
> > if (!ld_type) {
> > - dprintk("%s: No pNFS module found for %u.\n", __func__, id);
> > + dprintk("%s: No pNFS module found!\n", __func__);
> > goto out_no_driver;
> > }
> >
> > --
> > 2.5.5
--
Jeff Layton <jlayton@poochiereds.net>
next prev parent reply other threads:[~2016-06-14 1:46 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-06-09 12:21 [PATCH v2 0/3] pnfs/nfsd: have client and server support multiple layout types Jeff Layton
2016-06-09 12:21 ` [PATCH v2 1/3] nfsd: allow nfsd to advertise " Jeff Layton
2016-06-09 12:21 ` [PATCH v2 2/3] pnfs: track multiple layout types in fsinfo structure Jeff Layton
2016-06-09 12:21 ` [PATCH v2 3/3] pnfs: add a new mechanism to select a layout driver according to an ordered list Jeff Layton
2016-06-13 20:05 ` J. Bruce Fields
2016-06-14 1:46 ` Jeff Layton [this message]
2016-06-15 15:18 ` J. Bruce Fields
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=1465868807.12825.7.camel@poochiereds.net \
--to=jlayton@poochiereds.net \
--cc=anna.schumaker@netapp.com \
--cc=bfields@fieldses.org \
--cc=linux-nfs@vger.kernel.org \
--cc=thomas.haynes@primarydata.com \
--cc=tigran.mkrtchyan@desy.de \
--cc=trondmy@primarydata.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).