Date: Thu, 25 Apr 2019 21:12:43 +0000 From: Rick Macklem <rmacklem@uoguelph.ca> To: Adam <amvandemore@gmail.com>, Kurt Jaeger <pi@freebsd.org> Cc: freebsd-fs <freebsd-fs@freebsd.org> Subject: Re: NFS 4.2 plans ? Message-ID: <YQBPR0101MB0852BC9B8EDFC40C9C1728DCDD3D0@YQBPR0101MB0852.CANPRD01.PROD.OUTLOOK.COM> In-Reply-To: <CA%2BtpaK39AKs4Y0q5bpymbD21oQK0B7ODzet9Y1Kgzparva_nRg@mail.gmail.com> References: <20190425191224.GI9159@fc.opsec.eu>, <CA%2BtpaK39AKs4Y0q5bpymbD21oQK0B7ODzet9Y1Kgzparva_nRg@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Adam wrote: >> >> I've read that NFS 4.2 was specified 2016 in >> >> https://tools.ietf.org/html/rfc7862 >> >> so who knows if someone is working on this ? >> > > >https://lists.freebsd.org/pipermail/freebsd-current/2017-December/067755.h= tml The NFSv4.2 stuff lives in the FreeBSD svn under: projects/nfsv42 What is there works, but doesn't have many of the new features (which are a= ll optional) implemented in it. I do have some work I've done on it during the winter, which I will be sync= 'ing into the above projects branch sometime soon. To be honest, I don't think most of it will be particularily useful. Exampl= es of what it will are: - Some support for the posix_fadvise POSIX_FADV_WILLNEED POSIX_FADV_DONTNEED - Some support for the posix_fallocate() operation. - pNFS LayoutError, which allows the client to report an I/O error without = returning the layout (for NFSv4.1, the client must report the I/O error via a Layou= tReturn). - I don't think this is useful for the FreeBSD client, which will always = do a LayoutReturn The big one that I intend to do for the is the intra-server file copy. I do plan to get this into FreeBSD13, but I can't be sure. It will require = a fair amount of "negotiation" for things like adding a system call to do the file copy. (I will start with whatever Linux has, since Solaris is no longer mai= nstream imho.) I am also not sure if it possible to do per-file cloning within ZFS. (ZFS h= as per-fs cloning, but the NFSv4.2 spec supports a similar operation, but for a singl= e file. Basically, cloning of a file allows the clones to be copied on write after = the clone is created. Unless it can easily be done within ZFS, it won't be in the NFS= v4.2 implementation. If you are interested, check the commit logs for projects/nfsv42 accessible= via svnweb.freebsd.org. rick
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?YQBPR0101MB0852BC9B8EDFC40C9C1728DCDD3D0>