2008-03-13 13:27:17

by Benny Halevy

[permalink] [raw]
Subject: [PATCH] nfs: fix encode_fsinfo_maxsz

From: Fred Isaman <[email protected]>

The previous value was not taking into account space for bitmap array size.

Signed-off-by: Fred Isaman <[email protected]>
Signed-off-by: Benny Halevy <[email protected]>
---
fs/nfs/nfs4xdr.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 51dd380..0277821 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -110,7 +110,7 @@ static int nfs4_stat_to_errno(int);
#define decode_savefh_maxsz (op_decode_hdr_maxsz)
#define encode_restorefh_maxsz (op_encode_hdr_maxsz)
#define decode_restorefh_maxsz (op_decode_hdr_maxsz)
-#define encode_fsinfo_maxsz (op_encode_hdr_maxsz + 2)
+#define encode_fsinfo_maxsz (encode_getattr_maxsz)
#define decode_fsinfo_maxsz (op_decode_hdr_maxsz + 11)
#define encode_renew_maxsz (op_encode_hdr_maxsz + 3)
#define decode_renew_maxsz (op_decode_hdr_maxsz)
--
1.5.3.3



2008-03-14 01:00:23

by Rajesham Gajjela

[permalink] [raw]
Subject: Re: nfs over ipv6 support


Thanks Chuck for the quick reply.
- this support is not there even in latest
kernels (2.6.24 and such) ?
- Whats is your plan on this ? Do you think,
it can be done any time soon ?

Rajesham

--- Chuck Lever <[email protected]> wrote:

> Rajesham-
>
> IPv6 is not supported for NFS. The code you are
> looking at is a
> development snapshot.
>
> On Mar 13, 2008, at 10:31 AM, Rajesham Gajjela
> wrote:
> >> Mar 13 19:32:20 DXi5500-12345 kernel: RPC:
> transport
> >> (2/17) not supported
> >
> > After looking at the code
> (include/linux/socket.h):
> > #define AF_INET 2 /* Internet IP
> > Protocol */
> >
> > include/linux/in.h:
> > IPPROTO_UDP = 17, /* User Datagram
> > Protocol */
> >
> > Why 2/17 (which is AF_INET/IPPROTO_UDP) not
> > registered ?
> >
> > Looking at the register code:
> net/sunrpc/xprtsock.c:
> > static int __init init_xs_transport(void)
> > {
> > xprt_register(&xs_udp4_type);
> > #if defined(CONFIG_IPV6) ||
> > defined(CONFIG_IPV6_MODULE)
> > xprt_register(&xs_udp6_type);
> > #endif
> > xprt_register(&xs_tcp4_type);
> > #if defined(CONFIG_IPV6) ||
> > defined(CONFIG_IPV6_MODULE)
> > xprt_register(&xs_tcp6_type);
> > }
> >
> > The above code must add to "xprt_list". Note that,
> > I configured my kernel with CONFIG_IPV6.
> >
> >
> >
> > --- Rajesham Gajjela <[email protected]> wrote:
> >
> >>
> >> Did any body test nfs (v2 and v3) over ipv6 ?
> >> do you guys claim "this feature suppose to work"
> ?
> >>
> >> It did not work for me. Here are the details of
> >> experiments I did (for me server and client is
> >> same):
> >>
> >> 1) Kernel side : Linux-2.6.12-5 +
> >> linux-2.6.12-rc6-CITI_NFS4_ALL-1.dif +
> >> Chuck Lever Transport Patch
> >>
> >
>
(http://oss.oracle.com/~cel/linux-2.6/2.6.12/release-notes.html)
> >>
> >> 2) User land (tools) side: libtirpc-0.1.7-1 +
> >> rpcbind-0.1.4 + nfs-utils-1.1.1 +
> quota-3.12-5
> >> (Ref:
> >>
> >
>
http://nfsv4.bullopensource.org/patches/nfsv4_ipv6-2)
> >>
> >> 3) rpcbind suppose to recognize ipv6/ipv4 tcp/udp
> >> protocols (after modifying /etc/netconfig). I
> did
> >>
> >> softlink of /usr/bin/rpcbind to
> /sbin/portmap).
> >> Then did "service portmap restart".
> >> Then "service nfs restart". NFS services
> failed.
> >>
> >> syslog messages:
> >>
> >> Mar 13 19:32:20 DXi5500-12345 nfs: Starting NFS
> >> services: succeeded
> >> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.rquotad
> >> startup
> >> succeeded
> >> Mar 13 19:32:20 DXi5500-12345 kernel: RPC:
> transport
> >> (2/17) not supported
> >> Mar 13 19:32:20 DXi5500-12345 last message
> repeated
> >> 2
> >> times
> >> Mar 13 19:32:20 DXi5500-12345 nfsd[12453]:
> nfssvc:
> >> Input/output error
> >> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.nfsd
> startup
> >> failed
> >> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.mountd
> >> startup
> >> succeeded
> >> Mar 13 19:32:20 DXi5500-12345 rpc.idmapd:
> >> nfsdreopen:
> >> Opening '/proc/net/rpc/nfs4.idtoname/channel'
> >> failed:
> >> errno 2 (No such file or directory)
> >> Mar 13 19:32:20 DXi5500-12345 rpc.idmapd:
> >> nfsdreopen:
> >> Opening '/proc/net/rpc/nfs4.nametoid/channel'
> >> failed:
> >> errno 2 (No such file or directory)
> >> Mar 13 19:32:20 DXi5500-12345 rpcidmapd:
> rpc.idmapd
> >> -SIGHUP succeeded
> >>
> >>
> >>
> >>
> >>
> >
>
______________________________________________________________________
>
> > ______________
> >> Never miss a thing. Make Yahoo your home page.
> >> http://www.yahoo.com/r/hs
> >>
> >
> >
> >
> >
> >
>
______________________________________________________________________
>
> > ______________
> > Never miss a thing. Make Yahoo your home page.
> > http://www.yahoo.com/r/hs
> > --
> > To unsubscribe from this list: send the line
> "unsubscribe linux-
> > nfs" in
> > the body of a message to [email protected]
> > More majordomo info at
> http://vger.kernel.org/majordomo-info.html
>
> --
> Chuck Lever
> chuck[dot]lever[at]oracle[dot]com
>
>
>
> --
> To unsubscribe from this list: send the line
> "unsubscribe linux-nfs" in
> the body of a message to [email protected]
> More majordomo info at
> http://vger.kernel.org/majordomo-info.html
>



____________________________________________________________________________________
Never miss a thing. Make Yahoo your home page.
http://www.yahoo.com/r/hs

2008-03-13 14:15:10

by Rajesham Gajjela

[permalink] [raw]
Subject: nfs over ipv6 support


Did any body test nfs (v2 and v3) over ipv6 ?
do you guys claim "this feature suppose to work" ?

It did not work for me. Here are the details of
experiments I did (for me server and client is
same):

1) Kernel side : Linux-2.6.12-5 +
linux-2.6.12-rc6-CITI_NFS4_ALL-1.dif +
Chuck Lever Transport Patch
(http://oss.oracle.com/~cel/linux-2.6/2.6.12/release-notes.html)

2) User land (tools) side: libtirpc-0.1.7-1 +
rpcbind-0.1.4 + nfs-utils-1.1.1 + quota-3.12-5
(Ref:
http://nfsv4.bullopensource.org/patches/nfsv4_ipv6-2)

3) rpcbind suppose to recognize ipv6/ipv4 tcp/udp
protocols (after modifying /etc/netconfig). I did
softlink of /usr/bin/rpcbind to /sbin/portmap).
Then did "service portmap restart".
Then "service nfs restart". NFS services failed.

syslog messages:

Mar 13 19:32:20 DXi5500-12345 nfs: Starting NFS
services: succeeded
Mar 13 19:32:20 DXi5500-12345 nfs: rpc.rquotad startup
succeeded
Mar 13 19:32:20 DXi5500-12345 kernel: RPC: transport
(2/17) not supported
Mar 13 19:32:20 DXi5500-12345 last message repeated 2
times
Mar 13 19:32:20 DXi5500-12345 nfsd[12453]: nfssvc:
Input/output error
Mar 13 19:32:20 DXi5500-12345 nfs: rpc.nfsd startup
failed
Mar 13 19:32:20 DXi5500-12345 nfs: rpc.mountd startup
succeeded
Mar 13 19:32:20 DXi5500-12345 rpc.idmapd: nfsdreopen:
Opening '/proc/net/rpc/nfs4.idtoname/channel' failed:
errno 2 (No such file or directory)
Mar 13 19:32:20 DXi5500-12345 rpc.idmapd: nfsdreopen:
Opening '/proc/net/rpc/nfs4.nametoid/channel' failed:
errno 2 (No such file or directory)
Mar 13 19:32:20 DXi5500-12345 rpcidmapd: rpc.idmapd
-SIGHUP succeeded



____________________________________________________________________________________
Never miss a thing. Make Yahoo your home page.
http://www.yahoo.com/r/hs

2008-03-13 14:31:25

by Rajesham Gajjela

[permalink] [raw]
Subject: Re: nfs over ipv6 support



> Mar 13 19:32:20 DXi5500-12345 kernel: RPC: transport
> (2/17) not supported

After looking at the code (include/linux/socket.h):
#define AF_INET 2 /* Internet IP
Protocol */

include/linux/in.h:
IPPROTO_UDP = 17, /* User Datagram
Protocol */

Why 2/17 (which is AF_INET/IPPROTO_UDP) not
registered ?

Looking at the register code: net/sunrpc/xprtsock.c:
static int __init init_xs_transport(void)
{
xprt_register(&xs_udp4_type);
#if defined(CONFIG_IPV6) ||
defined(CONFIG_IPV6_MODULE)
xprt_register(&xs_udp6_type);
#endif
xprt_register(&xs_tcp4_type);
#if defined(CONFIG_IPV6) ||
defined(CONFIG_IPV6_MODULE)
xprt_register(&xs_tcp6_type);
}

The above code must add to "xprt_list". Note that,
I configured my kernel with CONFIG_IPV6.



--- Rajesham Gajjela <[email protected]> wrote:

>
> Did any body test nfs (v2 and v3) over ipv6 ?
> do you guys claim "this feature suppose to work" ?
>
> It did not work for me. Here are the details of
> experiments I did (for me server and client is
> same):
>
> 1) Kernel side : Linux-2.6.12-5 +
> linux-2.6.12-rc6-CITI_NFS4_ALL-1.dif +
> Chuck Lever Transport Patch
>
(http://oss.oracle.com/~cel/linux-2.6/2.6.12/release-notes.html)
>
> 2) User land (tools) side: libtirpc-0.1.7-1 +
> rpcbind-0.1.4 + nfs-utils-1.1.1 + quota-3.12-5
> (Ref:
>
http://nfsv4.bullopensource.org/patches/nfsv4_ipv6-2)
>
> 3) rpcbind suppose to recognize ipv6/ipv4 tcp/udp
> protocols (after modifying /etc/netconfig). I did
>
> softlink of /usr/bin/rpcbind to /sbin/portmap).
> Then did "service portmap restart".
> Then "service nfs restart". NFS services failed.
>
> syslog messages:
>
> Mar 13 19:32:20 DXi5500-12345 nfs: Starting NFS
> services: succeeded
> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.rquotad
> startup
> succeeded
> Mar 13 19:32:20 DXi5500-12345 kernel: RPC: transport
> (2/17) not supported
> Mar 13 19:32:20 DXi5500-12345 last message repeated
> 2
> times
> Mar 13 19:32:20 DXi5500-12345 nfsd[12453]: nfssvc:
> Input/output error
> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.nfsd startup
> failed
> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.mountd
> startup
> succeeded
> Mar 13 19:32:20 DXi5500-12345 rpc.idmapd:
> nfsdreopen:
> Opening '/proc/net/rpc/nfs4.idtoname/channel'
> failed:
> errno 2 (No such file or directory)
> Mar 13 19:32:20 DXi5500-12345 rpc.idmapd:
> nfsdreopen:
> Opening '/proc/net/rpc/nfs4.nametoid/channel'
> failed:
> errno 2 (No such file or directory)
> Mar 13 19:32:20 DXi5500-12345 rpcidmapd: rpc.idmapd
> -SIGHUP succeeded
>
>
>
>
>
____________________________________________________________________________________
> Never miss a thing. Make Yahoo your home page.
> http://www.yahoo.com/r/hs
>



____________________________________________________________________________________
Never miss a thing. Make Yahoo your home page.
http://www.yahoo.com/r/hs

2008-03-13 16:21:08

by Trond Myklebust

[permalink] [raw]
Subject: Re: [PATCH] nfs: fix encode_fsinfo_maxsz


On Thu, 2008-03-13 at 15:26 +0200, Benny Halevy wrote:
> From: Fred Isaman <[email protected]>
>
> The previous value was not taking into account space for bitmap array size.
>
> Signed-off-by: Fred Isaman <[email protected]>
> Signed-off-by: Benny Halevy <[email protected]>
> ---
> fs/nfs/nfs4xdr.c | 2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
> index 51dd380..0277821 100644
> --- a/fs/nfs/nfs4xdr.c
> +++ b/fs/nfs/nfs4xdr.c
> @@ -110,7 +110,7 @@ static int nfs4_stat_to_errno(int);
> #define decode_savefh_maxsz (op_decode_hdr_maxsz)
> #define encode_restorefh_maxsz (op_encode_hdr_maxsz)
> #define decode_restorefh_maxsz (op_decode_hdr_maxsz)
> -#define encode_fsinfo_maxsz (op_encode_hdr_maxsz + 2)
> +#define encode_fsinfo_maxsz (encode_getattr_maxsz)
> #define decode_fsinfo_maxsz (op_decode_hdr_maxsz + 11)
> #define encode_renew_maxsz (op_encode_hdr_maxsz + 3)
> #define decode_renew_maxsz (op_decode_hdr_maxsz)

Thanks!

Has this ever been observed to cause problems in the mainline NFSv4?
Having never seen any such problems myself, I'm a bit inclined to delay
this until the 2.6.26 merge window. That said, if you can point to an
instance where it caused a bug, then I'll push it to Linus sooner.

Cheers
Trond


2008-03-13 16:28:01

by Benny Halevy

[permalink] [raw]
Subject: Re: [PATCH] nfs: fix encode_fsinfo_maxsz

Trond Myklebust wrote:
> On Thu, 2008-03-13 at 15:26 +0200, Benny Halevy wrote:
>> From: Fred Isaman <[email protected]>
>>
>> The previous value was not taking into account space for bitmap array size.
>>
>> Signed-off-by: Fred Isaman <[email protected]>
>> Signed-off-by: Benny Halevy <[email protected]>
>> ---
>> fs/nfs/nfs4xdr.c | 2 +-
>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
>> index 51dd380..0277821 100644
>> --- a/fs/nfs/nfs4xdr.c
>> +++ b/fs/nfs/nfs4xdr.c
>> @@ -110,7 +110,7 @@ static int nfs4_stat_to_errno(int);
>> #define decode_savefh_maxsz (op_decode_hdr_maxsz)
>> #define encode_restorefh_maxsz (op_encode_hdr_maxsz)
>> #define decode_restorefh_maxsz (op_decode_hdr_maxsz)
>> -#define encode_fsinfo_maxsz (op_encode_hdr_maxsz + 2)
>> +#define encode_fsinfo_maxsz (encode_getattr_maxsz)
>> #define decode_fsinfo_maxsz (op_decode_hdr_maxsz + 11)
>> #define encode_renew_maxsz (op_encode_hdr_maxsz + 3)
>> #define decode_renew_maxsz (op_decode_hdr_maxsz)
>
> Thanks!
>
> Has this ever been observed to cause problems in the mainline NFSv4?
> Having never seen any such problems myself, I'm a bit inclined to delay
> this until the 2.6.26 merge window. That said, if you can point to an
> instance where it caused a bug, then I'll push it to Linus sooner.

I can't point to such an instance.
I *think* it was found by code inspection only.
Maybe Fred can...

Benny

>
> Cheers
> Trond
>


2008-03-13 16:29:08

by Fred Isaman

[permalink] [raw]
Subject: Re: [PATCH] nfs: fix encode_fsinfo_maxsz



On Thu, 13 Mar 2008, Trond Myklebust wrote:

>
> On Thu, 2008-03-13 at 15:26 +0200, Benny Halevy wrote:
>> From: Fred Isaman <[email protected]>
>>
>> The previous value was not taking into account space for bitmap array size.
>>
>> Signed-off-by: Fred Isaman <[email protected]>
>> Signed-off-by: Benny Halevy <[email protected]>
>> ---
>> fs/nfs/nfs4xdr.c | 2 +-
>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
>> index 51dd380..0277821 100644
>> --- a/fs/nfs/nfs4xdr.c
>> +++ b/fs/nfs/nfs4xdr.c
>> @@ -110,7 +110,7 @@ static int nfs4_stat_to_errno(int);
>> #define decode_savefh_maxsz (op_decode_hdr_maxsz)
>> #define encode_restorefh_maxsz (op_encode_hdr_maxsz)
>> #define decode_restorefh_maxsz (op_decode_hdr_maxsz)
>> -#define encode_fsinfo_maxsz (op_encode_hdr_maxsz + 2)
>> +#define encode_fsinfo_maxsz (encode_getattr_maxsz)
>> #define decode_fsinfo_maxsz (op_decode_hdr_maxsz + 11)
>> #define encode_renew_maxsz (op_encode_hdr_maxsz + 3)
>> #define decode_renew_maxsz (op_decode_hdr_maxsz)
>
> Thanks!
>
> Has this ever been observed to cause problems in the mainline NFSv4?
> Having never seen any such problems myself, I'm a bit inclined to delay
> this until the 2.6.26 merge window. That said, if you can point to an
> instance where it caused a bug, then I'll push it to Linus sooner.
>
> Cheers
> Trond
>
>

I just noticed it in the process of adding some attributes to fsinfo, as
opposed to searching down a bug.

Fred


2008-03-13 16:55:18

by Trond Myklebust

[permalink] [raw]
Subject: Re: [PATCH] nfs: fix encode_fsinfo_maxsz


On Thu, 2008-03-13 at 12:29 -0400, Fredric Isaman wrote:
>
> On Thu, 13 Mar 2008, Trond Myklebust wrote:
> > Has this ever been observed to cause problems in the mainline NFSv4?
> > Having never seen any such problems myself, I'm a bit inclined to delay
> > this until the 2.6.26 merge window. That said, if you can point to an
> > instance where it caused a bug, then I'll push it to Linus sooner.
> >
> > Cheers
> > Trond
> >
> >
>
> I just noticed it in the process of adding some attributes to fsinfo, as
> opposed to searching down a bug.

OK. I'll queue it up for 2.6.26, then.

Again thanks!

Trond


2008-03-13 18:16:10

by Chuck Lever III

[permalink] [raw]
Subject: Re: nfs over ipv6 support

Rajesham-

IPv6 is not supported for NFS. The code you are looking at is a
development snapshot.

On Mar 13, 2008, at 10:31 AM, Rajesham Gajjela wrote:
>> Mar 13 19:32:20 DXi5500-12345 kernel: RPC: transport
>> (2/17) not supported
>
> After looking at the code (include/linux/socket.h):
> #define AF_INET 2 /* Internet IP
> Protocol */
>
> include/linux/in.h:
> IPPROTO_UDP = 17, /* User Datagram
> Protocol */
>
> Why 2/17 (which is AF_INET/IPPROTO_UDP) not
> registered ?
>
> Looking at the register code: net/sunrpc/xprtsock.c:
> static int __init init_xs_transport(void)
> {
> xprt_register(&xs_udp4_type);
> #if defined(CONFIG_IPV6) ||
> defined(CONFIG_IPV6_MODULE)
> xprt_register(&xs_udp6_type);
> #endif
> xprt_register(&xs_tcp4_type);
> #if defined(CONFIG_IPV6) ||
> defined(CONFIG_IPV6_MODULE)
> xprt_register(&xs_tcp6_type);
> }
>
> The above code must add to "xprt_list". Note that,
> I configured my kernel with CONFIG_IPV6.
>
>
>
> --- Rajesham Gajjela <[email protected]> wrote:
>
>>
>> Did any body test nfs (v2 and v3) over ipv6 ?
>> do you guys claim "this feature suppose to work" ?
>>
>> It did not work for me. Here are the details of
>> experiments I did (for me server and client is
>> same):
>>
>> 1) Kernel side : Linux-2.6.12-5 +
>> linux-2.6.12-rc6-CITI_NFS4_ALL-1.dif +
>> Chuck Lever Transport Patch
>>
> (http://oss.oracle.com/~cel/linux-2.6/2.6.12/release-notes.html)
>>
>> 2) User land (tools) side: libtirpc-0.1.7-1 +
>> rpcbind-0.1.4 + nfs-utils-1.1.1 + quota-3.12-5
>> (Ref:
>>
> http://nfsv4.bullopensource.org/patches/nfsv4_ipv6-2)
>>
>> 3) rpcbind suppose to recognize ipv6/ipv4 tcp/udp
>> protocols (after modifying /etc/netconfig). I did
>>
>> softlink of /usr/bin/rpcbind to /sbin/portmap).
>> Then did "service portmap restart".
>> Then "service nfs restart". NFS services failed.
>>
>> syslog messages:
>>
>> Mar 13 19:32:20 DXi5500-12345 nfs: Starting NFS
>> services: succeeded
>> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.rquotad
>> startup
>> succeeded
>> Mar 13 19:32:20 DXi5500-12345 kernel: RPC: transport
>> (2/17) not supported
>> Mar 13 19:32:20 DXi5500-12345 last message repeated
>> 2
>> times
>> Mar 13 19:32:20 DXi5500-12345 nfsd[12453]: nfssvc:
>> Input/output error
>> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.nfsd startup
>> failed
>> Mar 13 19:32:20 DXi5500-12345 nfs: rpc.mountd
>> startup
>> succeeded
>> Mar 13 19:32:20 DXi5500-12345 rpc.idmapd:
>> nfsdreopen:
>> Opening '/proc/net/rpc/nfs4.idtoname/channel'
>> failed:
>> errno 2 (No such file or directory)
>> Mar 13 19:32:20 DXi5500-12345 rpc.idmapd:
>> nfsdreopen:
>> Opening '/proc/net/rpc/nfs4.nametoid/channel'
>> failed:
>> errno 2 (No such file or directory)
>> Mar 13 19:32:20 DXi5500-12345 rpcidmapd: rpc.idmapd
>> -SIGHUP succeeded
>>
>>
>>
>>
>>
> ______________________________________________________________________
> ______________
>> Never miss a thing. Make Yahoo your home page.
>> http://www.yahoo.com/r/hs
>>
>
>
>
>
> ______________________________________________________________________
> ______________
> Never miss a thing. Make Yahoo your home page.
> http://www.yahoo.com/r/hs
> --
> To unsubscribe from this list: send the line "unsubscribe linux-
> nfs" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html

--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com