From: Bruce Fields <bfields@fieldses.org>
To: Chuck Lever <chuck.lever@oracle.com>
Cc: Bill Baker <Bill.Baker@oracle.com>,
Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH v2 26/27] NFSD: Add tracepoints in the NFS dispatcher
Date: Fri, 25 Sep 2020 10:47:14 -0400 [thread overview]
Message-ID: <20200925144714.GE1096@fieldses.org> (raw)
In-Reply-To: <801F3A94-4668-4DF6-9CAF-27171EEBA17A@oracle.com>
On Fri, Sep 25, 2020 at 09:59:54AM -0400, Chuck Lever wrote:
>
>
> > On Sep 24, 2020, at 7:45 PM, J. Bruce Fields <bfields@fieldses.org> wrote:
> >
> > On Mon, Sep 21, 2020 at 02:13:07PM -0400, Chuck Lever wrote:
> >> This is follow-on work to the tracepoints added in the NFS server's
> >> duplicate reply cache. Here, tracepoints are introduced that report
> >> replies from cache as well as encoding and decoding errors.
> >>
> >> The NFSv2, v3, and v4 dispatcher requirements have diverged over
> >> time, leaving us with a little bit of technical debt. In addition,
> >> I wanted to add a tracepoint for NFSv2 and NFSv3 similar to the
> >> nfsd4_compound/compoundstatus tracepoints. Lastly, removing some
> >> conditional branches from this hot path helps optimize CPU
> >> utilization. So, I've duplicated the nfsd_dispatcher function.
> >
> > Comparing current nfsd_dispatch to the nfsv2/v3 nfsd_dispatch: the only
> > thing I spotted removed from the v2/v3-specific dispatch is the
> > rq_lease_breaker = NULL. (I think that's not correct, actually. We
> > could remove the need for that to be set in the v2/v3 case, but with the
> > current code it does need to be set.)
>
> Noted with thanks.
>
>
> > Comparing current nfsd_dispatch to the nfsv4 nfsd4_dispatch, the
> > v4-specific dispatch does away with nfs_request_too_big() and the
> > v2-specific shortcut in the error encoding case.
> >
> > So these still look *very* similar. I don't feel like we're getting a
> > lot of benefit out of splitting these out.
>
> I don't disagree with that at all. At this point I'm just noodling
> to see what's possible. I'm now toying with other ways to add high-
> value tracing in the legacy ULPs. In the end I might end up avoiding
> significant changes in the dispatchers in order to add tracing.
OK.
> However, a few thoughts I had while learning how the dispatcher
> code works.
>
> There are some opportunities for reducing instruction path length
> and the number of conditional branches in here. It's a hot path,
> so I think we should consider some careful micro-optimizations
> even if they don't add significant new features or do add some
> code duplication.
>
> In user space, the library (iirc) assumes each ULP provides it's
> own dispatcher function. I'd consider duplicating and removing
> svc_generic_dispatcher() to simplify the pasta in svc_process(),
> again as a micro-optimization and for better code legibility.
Not sure you even have to duplicate it, just export the generic
dispatcher and let individual programs point to it, right?
> lockd's pc_func returns an RPC accept_stat, but the NFSD pc_func
> methods return an NFS status. The latter feels like a layering
> violation for the sake of reducing a small amount of code
> duplication. I'd rather see encoding of the NFS status handled in
> the NFS Reply encoders, since that is an XDR function, and because
> that logic seems slightly different for NFSv2, support for which
> we'd like to deprecate at some point.
>
> Note also that *statp in nfsd_dispatch is never explicitly set to
> rpc_success in the normal execution flow. It relies on the
> equivalence of rpc_success and nfs_ok, which is convenient, but
> confusing to read. It might be cleaner if *statp was made an enum
> to make it explicit what set of values go in that return variable.
OK.
--b.
next prev parent reply other threads:[~2020-09-25 14:47 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-21 18:10 [PATCH v2 00/27] NFSD operation monitoring tracepoints Chuck Lever
2020-09-21 18:10 ` [PATCH v2 01/27] NFS: Move generic FS show macros to global header Chuck Lever
2020-09-21 18:11 ` [PATCH v2 02/27] NFS: Move NFS protocol display " Chuck Lever
2020-09-21 18:11 ` [PATCH v2 03/27] NFSD: Add SPDX header for fs/nfsd/trace.c Chuck Lever
2020-09-21 18:11 ` [PATCH v2 04/27] SUNRPC: Move the svc_xdr_recvfrom() tracepoint Chuck Lever
2020-09-21 18:11 ` [PATCH v2 05/27] SUNRPC: Add svc_xdr_authenticate tracepoint Chuck Lever
2020-09-21 18:11 ` [PATCH v2 06/27] lockd: Replace PROC() macro with open code Chuck Lever
2020-09-21 18:11 ` [PATCH v2 07/27] NFSACL: " Chuck Lever
2020-09-21 18:11 ` [PATCH v2 08/27] SUNRPC: Make trace_svc_process() display the RPC procedure symbolically Chuck Lever
2020-09-21 18:11 ` [PATCH v2 09/27] NFSD: Clean up the show_nf_may macro Chuck Lever
2020-09-21 18:11 ` [PATCH v2 10/27] NFSD: Remove extra "0x" in tracepoint format specifier Chuck Lever
2020-09-21 18:11 ` [PATCH v2 11/27] NFSD: Constify @fh argument of knfsd_fh_hash() Chuck Lever
2020-09-21 18:11 ` [PATCH v2 12/27] NFSD: Add tracepoint in nfsd_setattr() Chuck Lever
2020-09-21 18:11 ` [PATCH v2 13/27] NFSD: Add tracepoint for nfsd_access() Chuck Lever
2020-09-21 18:12 ` [PATCH v2 14/27] NFSD: nfsd_compound_status tracepoint should record XID Chuck Lever
2020-09-21 18:12 ` [PATCH v2 15/27] NFSD: Add client ID lifetime tracepoints Chuck Lever
2020-09-21 18:12 ` [PATCH v2 16/27] NFSD: Add tracepoints to report NFSv4 session state Chuck Lever
2020-09-21 18:12 ` [PATCH v2 17/27] NFSD: Add a tracepoint to report the current filehandle Chuck Lever
2020-09-21 18:12 ` [PATCH v2 18/27] NFSD: Add GETATTR tracepoint Chuck Lever
2020-09-21 18:12 ` [PATCH v2 19/27] NFSD: Add tracepoint in nfsd4_stateid_preprocess() Chuck Lever
2020-09-21 18:12 ` [PATCH v2 20/27] NFSD: Add tracepoint to report arguments to NFSv4 OPEN Chuck Lever
2020-09-21 18:12 ` [PATCH v2 21/27] NFSD: Add a tracepoint for DELEGRETURN Chuck Lever
2020-09-21 18:12 ` [PATCH v2 22/27] NFSD: Add a lookup tracepoint Chuck Lever
2020-09-21 18:12 ` [PATCH v2 23/27] NFSD: Add lock and locku tracepoints Chuck Lever
2020-09-21 18:12 ` [PATCH v2 24/27] NFSD: Add tracepoints to record the result of TEST_STATEID and FREE_STATEID Chuck Lever
2020-09-21 18:13 ` [PATCH v2 25/27] NFSD: Rename nfsd_ tracepoints to nfsd4_ Chuck Lever
2020-09-21 18:13 ` [PATCH v2 26/27] NFSD: Add tracepoints in the NFS dispatcher Chuck Lever
2020-09-24 23:45 ` J. Bruce Fields
2020-09-25 13:59 ` Chuck Lever
2020-09-25 14:17 ` Bruce Fields
2020-09-25 14:21 ` Chuck Lever
2020-09-25 14:18 ` Chuck Lever
2020-09-25 14:47 ` Bruce Fields [this message]
2020-09-21 18:13 ` [PATCH v2 27/27] NFSD: Replace dprintk callsites in fs/nfsd/nfsfh.c Chuck Lever
2020-09-24 21:36 ` [PATCH v2 00/27] NFSD operation monitoring tracepoints J. Bruce Fields
2020-09-25 13:59 ` Chuck Lever
2020-09-25 14:32 ` Bruce Fields
2020-09-25 14:36 ` Chuck Lever
2020-09-25 15:00 ` Bruce Fields
2020-09-25 15:05 ` Chuck Lever
2020-09-25 17:04 ` Chuck Lever
2020-09-25 18:37 ` Bruce Fields
2020-09-25 18:41 ` Chuck Lever
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=20200925144714.GE1096@fieldses.org \
--to=bfields@fieldses.org \
--cc=Bill.Baker@oracle.com \
--cc=chuck.lever@oracle.com \
--cc=linux-nfs@vger.kernel.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 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).