[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <274f7ceff9a840023c299ba750ff7ccc0d00897c.camel@hammerspace.com>
Date: Thu, 12 Sep 2019 13:14:24 +0000
From: Trond Myklebust <trondmy@...merspace.com>
To: "bcodding@...hat.com" <bcodding@...hat.com>
CC: "tibbs@...h.uh.edu" <tibbs@...h.uh.edu>,
"bfields@...ldses.org" <bfields@...ldses.org>,
"linux@...m.de" <linux@...m.de>,
"linux-nfs@...r.kernel.org" <linux-nfs@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"km@...all.com" <km@...all.com>,
"chuck.lever@...cle.com" <chuck.lever@...cle.com>
Subject: Re: Regression in 5.1.20: Reading long directory fails
On Thu, 2019-09-12 at 09:08 -0400, Benjamin Coddington wrote:
> On 12 Sep 2019, at 8:53, Trond Myklebust wrote:
> > Let's please just scrap this function and rewrite it as a generic
> > function for reading the MIC. It clearly is not a generic function
> > for
> > reading arbitrary netobjs, and modifications like the above just
> > make
> > the misnomer painfully obvious.
> >
> > Let's rewrite it as xdr_buf_read_mic() so that we can simplify it
> > where
> > possible.
>
> Ok. I want to assume the mic will not land in the head, but I am not
> sure..
> Is there a scenario where the mic might land in the head, or is that
> bit of
> the current function left over from other uses?
It is likely to land in the head for most RPC calls that have short
replies. Particularly in error cases.
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@...merspace.com
Powered by blists - more mailing lists