linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Steve Dickson <SteveD@redhat.com>
To: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH 1/2] configure.ac: Added server 4.1 support configuration option
Date: Mon, 19 Sep 2011 10:20:38 -0400	[thread overview]
Message-ID: <4E774FB6.6040707@RedHat.com> (raw)
In-Reply-To: <20110919133845.GA1106@fieldses.org>



On 09/19/2011 09:38 AM, J. Bruce Fields wrote:
> On Mon, Sep 19, 2011 at 05:51:51AM -0400, Steve Dickson wrote:
>>
>>
>> On 09/17/2011 02:36 PM, J. Bruce Fields wrote:
>>> On Sat, Sep 17, 2011 at 11:17:01AM -0400, Steve Dickson wrote:
>>>> Separate the v4.1 support in the client and server
>>>> by creating the --enable-nfsdv4 configuration option.
>>>
>>> I really dislike --enable-nfsdv4: people shouldn't have to rebuild
>>> nfs-utils in order to turn on server 4.1 support.  They should be able
>>> to do that at run time.
>>>
>>> It makes sense for something like the block client, because there we
>>> actually have to decide whether to *build* something or not, and if
>>> someone knows they're never going to use that daemon, maybe they don't
>>> want to bother building and installing it.
>>>
>>> For server 4.1 support, we're making them rebuild just to change the
>>> value of a single constant.  Why?
>> If I remember correctly, that's the way you wanted... We disabled 
>> 4.1 server supported this way, because the code did not conform to 
>> the RFC. This was done back in Fed 2010.
>>
>> All this patch is doing is reorganizing things so there is a
>> configuration option to enable/disable the 4.1 client.   
>>
>> Now, I agree. Enabling/disabling the 4.1 server in a more 
>> dynamic way would be good... Is the code ready for that?
>> Should this constraint be remove completely? 
> 
> We still don't want it on by default.  But that's no reason to force
> early adopters to recompile nfs-utils just to turn it on.
Fine...

> 
> So how about something like this?  (Untested.)
> 
> --b.
> 
> commit 3df526795eed38b27fecb51775c039b7eaeed088
> Author: J. Bruce Fields <bfields@citi.umich.edu>
> Date:   Thu Dec 17 17:11:54 2009 -0500
> 
>     nfsd: allow choosing server 41 support at runtime
>     
>     Provide an enable-41 option to rpc.nfsd to turn on experimental 4.1
>     support in the server.  (Current kernels have 41 off by default until
>     basic 41 support is finished, so this option is intended only for people
>     that know what they're doing.)
>     
>     Note once 4.1 defaults to on, this option will become unnecessary (and
>     can probably just be dropped from nfs-utils), and only the -N 4.1 option
>     will be necessary, for those that want to turn 4.1 off for some reason.
>     
>     Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
> 
> diff --git a/configure.ac b/configure.ac
> index 1a28f8a..7000324 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -73,19 +73,6 @@ AC_ARG_ENABLE(nfsv4,
>  	AC_SUBST(enable_nfsv4)
>  	AM_CONDITIONAL(CONFIG_NFSV4, [test "$enable_nfsv4" = "yes"])
>  
> -AC_ARG_ENABLE(nfsv41,
> -	[AC_HELP_STRING([--enable-nfsv41],
> -                        [enable support for NFSv41 @<:@default=no@:>@])],
> -	enable_nfsv41=$enableval,
> -	enable_nfsv41=no)
> -	if test "$enable_nfsv41" = yes; then
> -		AC_DEFINE(NFS41_SUPPORTED, 1, [Define this if you want NFSv41 support compiled in])
> -	else
> -		enable_nfsv41=
> -	fi
> -	AC_SUBST(enable_nfsv41)
> -	AM_CONDITIONAL(CONFIG_NFSV41, [test "$enable_nfsv41" = "yes"])
> -
I think we still need this for the 4.1 client code.

>  AC_ARG_ENABLE(gss,
>  	[AC_HELP_STRING([--enable-gss],
>                          [enable support for rpcsec_gss @<:@default=yes@:>@])],
> diff --git a/support/include/nfs/nfs.h b/support/include/nfs/nfs.h
> index c939d78..320880e 100644
> --- a/support/include/nfs/nfs.h
> +++ b/support/include/nfs/nfs.h
> @@ -15,13 +15,6 @@
>  #define NFSD_MINVERS 2
>  #define NFSD_MAXVERS 4
>  
> -#define NFSD_MINMINORVERS4 1
> -#ifdef  NFS41_SUPPORTED
> -#define NFSD_MAXMINORVERS4 1
> -#else
> -#define NFSD_MAXMINORVERS4 0
> -#endif
> -
>  struct nfs_fh_len {
>  	int		fh_size;
>  	u_int8_t	fh_handle[NFS3_FHSIZE];
> diff --git a/utils/nfsd/nfsd.c b/utils/nfsd/nfsd.c
> index e7e1470..fa9c429 100644
> --- a/utils/nfsd/nfsd.c
> +++ b/utils/nfsd/nfsd.c
> @@ -40,6 +40,7 @@ static struct option longopts[] =
>  	{ "port", 1, 0, 'p' },
>  	{ "debug", 0, 0, 'd' },
>  	{ "syslog", 0, 0, 's' },
> +	{ "enable-41", 0, 0, 'X' },
>  	{ NULL, 0, 0, 0 }
>  };
Introducing temporary, undocumented command line arguments
is generally not a good thing... IMHO... I was think more
of an echo to /proc/fs/nfsd/versions to enable the support.

steved.

>  
> @@ -94,7 +95,7 @@ main(int argc, char **argv)
>  	char *p, *progname, *port;
>  	char *haddr = NULL;
>  	int	socket_up = 0;
> -	int minorvers4 = NFSD_MAXMINORVERS4;	/* nfsv4 minor version */
> +	int minorvers41 = 0;	/* nfsv4 minor version */
>  	unsigned int versbits = NFSCTL_ALLBITS;
>  	unsigned int protobits = NFSCTL_ALLBITS;
>  	unsigned int proto4 = 0;
> @@ -154,7 +155,12 @@ main(int argc, char **argv)
>  			switch((c = strtol(optarg, &p, 0))) {
>  			case 4:
>  				if (*p == '.') {
> -					minorvers4 = -atoi(p + 1);
> +					int i = atoi(p+1);
> +					if (i != 1) {
> +						fprintf(stderr, "%s: unsupported minor version\n", optarg);
> +						exit(1);
> +					}
> +					minorvers41 = -1;
>  					break;
>  				}
>  			case 3:
> @@ -176,6 +182,9 @@ main(int argc, char **argv)
>  		case 'U':
>  			NFSCTL_UDPUNSET(protobits);
>  			break;
> +		case 'X':
> +			minorvers41 = 1;
> +			break;
>  		default:
>  			fprintf(stderr, "Invalid argument: '%c'\n", c);
>  		case 'h':
> @@ -251,7 +260,7 @@ main(int argc, char **argv)
>  	 * registered with rpcbind. Note that on older kernels w/o the right
>  	 * interfaces, these are a no-op.
>  	 */
> -	nfssvc_setvers(versbits, minorvers4);
> +	nfssvc_setvers(versbits, minorvers41);
>   
>  	error = nfssvc_set_sockets(AF_INET, proto4, haddr, port);
>  	if (!error)
> @@ -303,7 +312,7 @@ static void
>  usage(const char *prog)
>  {
>  	fprintf(stderr, "Usage:\n"
> -		"%s [-d|--debug] [-H hostname] [-p|-P|--port port] [-N|--no-nfs-version version ] [-s|--syslog] [-T|--no-tcp] [-U|--no-udp] nrservs\n", 
> +		"%s [-d|--debug] [-H hostname] [-p|-P|--port port] [-N|--no-nfs-version version ] [-s|--syslog] [-T|--no-tcp] [-U|--no-udp] [--enable-41] nrservs\n", 
>  		prog);
>  	exit(2);
>  }
> diff --git a/utils/nfsd/nfssvc.c b/utils/nfsd/nfssvc.c
> index f607214..683008e 100644
> --- a/utils/nfsd/nfssvc.c
> +++ b/utils/nfsd/nfssvc.c
> @@ -269,7 +269,7 @@ nfssvc_set_sockets(const int family, const unsigned int protobits,
>  }
>  
>  void
> -nfssvc_setvers(unsigned int ctlbits, int minorvers4)
> +nfssvc_setvers(unsigned int ctlbits, int minorvers41)
>  {
>  	int fd, n, off;
>  	char *ptr;
> @@ -280,11 +280,9 @@ nfssvc_setvers(unsigned int ctlbits, int minorvers4)
>  	if (fd < 0)
>  		return;
>  
> -	n = minorvers4 >= 0 ? minorvers4 : -minorvers4;
> -	if (n >= NFSD_MINMINORVERS4 && n <= NFSD_MAXMINORVERS4)
> -		    off += snprintf(ptr+off, sizeof(buf) - off, "%c4.%d ",
> -				    minorvers4 > 0 ? '+' : '-',
> -				    n);
> +	if (minorvers41)
> +		off += snprintf(ptr+off, sizeof(buf) - off, "%c4.1",
> +				minorvers41 > 0 ? '+' : '-');
>  	for (n = NFSD_MINVERS; n <= NFSD_MAXVERS; n++) {
>  		if (NFSCTL_VERISSET(ctlbits, n))
>  		    off += snprintf(ptr+off, sizeof(buf) - off, "+%d ", n);

  reply	other threads:[~2011-09-19 14:20 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-09-17 15:17 [PATCH 1/2] configure.ac: Added server 4.1 support configuration option Steve Dickson
2011-09-17 15:17 ` [PATCH 2/2] configure.ac: Remove unused NFS3_SUPPORTED define Steve Dickson
2011-09-17 18:36 ` [PATCH 1/2] configure.ac: Added server 4.1 support configuration option J. Bruce Fields
2011-09-19  9:51   ` Steve Dickson
2011-09-19 13:38     ` J. Bruce Fields
2011-09-19 14:20       ` Steve Dickson [this message]
2011-09-19 14:35         ` J. Bruce Fields
2011-09-19 15:16           ` Steve Dickson
2011-09-19 15:27             ` J. Bruce Fields
2011-09-19 18:15               ` J. Bruce Fields
2011-09-20 11:43                 ` Steve Dickson
     [not found]                   ` <4E787C6E.8080104-AfCzQyP5zfLQT0dZR+AlfA@public.gmane.org>
2011-09-20 11:47                     ` 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=4E774FB6.6040707@RedHat.com \
    --to=steved@redhat.com \
    --cc=bfields@fieldses.org \
    --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).