headers
Bryce Harrington | 1 Mar 2006 01:26

Re: licensing for libgssapi and librpcsecgss

On Fri, Feb 24, 2006 at 04:44:17PM -0500, J. Bruce Fields wrote:
> On Fri, Feb 24, 2006 at 01:06:43PM -0800, Bryce Harrington wrote:
> > For reference, below are the licenses in that directory:
> 
> Aie, OK, I'm not reading through all those--I'll let you figure it out.
> 
> But I can add more verbose COPYING files with copies of the various
> licenses.  I've committed those to CVS, so they'll show up on the
> webpage some day.  For now, enough details from them appended to make
> you sorry you asked.
> 
> (The executive summary: anything we do is under a modified-BSD/X11-like
> (non-copyleft) license that should be compatible with just about
> anything.  We're also using code from Sun and MIT that's under pretty
> similar licenses.)

Okay, I've submitted the packaging for gentoo for these two libraries.  

  http://bugs.gentoo.org/show_bug.cgi?id=124473
  http://bugs.gentoo.org/show_bug.cgi?id=124474

I couldn't find existing licenses that match what these need, so
submitted new ones, based on what you had in debian/copyright.  

Bryce
Permalink | Reply | 0 comments |
headers
J. Bruce Fields | 1 Mar 2006 01:23

linux-2.6.16-rc4-CITI_NFS4_ALL-4

http://www.citi.umich.edu/projects/nfsv4/linux/kernel-patches/2.6.16-rc4-4
Changes since 2.6.16-rc4-CITI_NFS4_ALL-3
	* fix a compile error--dump mistake on my part when applying
	  Fred's patches

Sorry!  (Isn't there a good mailbox-import script for stgit somewhere?  That
would make this kind of mistake less likely.  The standard stg import doesn't
quite do the job.)

--b.
Permalink | Reply | 5 comments |
headers
Trond Myklebust | 1 Mar 2006 08:29
Picon
Picon

Re: linux-2.6.16-rc4-CITI_NFS4_ALL-4

On Tue, 2006-02-28 at 19:23 -0500, J. Bruce Fields wrote:
> http://www.citi.umich.edu/projects/nfsv4/linux/kernel-patches/2.6.16-rc4-4
> Changes since 2.6.16-rc4-CITI_NFS4_ALL-3
> 	* fix a compile error--dump mistake on my part when applying
> 	  Fred's patches
> 
> Sorry!  (Isn't there a good mailbox-import script for stgit somewhere?  That
> would make this kind of mistake less likely.  The standard stg import doesn't
> quite do the job.)

Here's mine...

Cheers,
  Trond
Attachment (mbox_to_patches): application/x-shellscript, 2141 bytes
_______________________________________________
NFSv4 mailing list
NFSv4 <at> linux-nfs.org
http://linux-nfs.org/cgi-bin/mailman/listinfo/nfsv4
Permalink | Reply | 0 comments |
headers
Vincent Roqueta | 1 Mar 2006 12:13
Picon

2.6.16-rc4-CITI_NFS4_ALL-4 server : oops

Hello Bruce,

Here is the traces I have got using last 2.6.16-rc4-CITI_NFS4_ALL-4.
That have happened some time after an IOZone test with krb5p enabled. I am not 
sure that is related with the test.

Server's load with krb5p is still very low. I send an other mail for that.

<1>Unable to handle kernel paging request at virtual address 7374656c
<1> printing eip:
<4>c01d14ad
<1>*pde = 00000000
<0>Oops: 0002 [#1]
<0>PREEMPT SMP
<4>Modules linked in: oprofile video ipv6
<0>CPU:    0
<4>EIP:    0060:[<c01d14ad>]    Not tainted VLI
<4>EFLAGS: 00010206   (2.6.16-rc4-CITI_NFS4_ALL-4 #2)
<0>EIP is at svc_export_put+0x57/0xac
<0>eax: f7f5e2a0   ebx: 7374656c   ecx: c040d500   edx: c040d500
<0>esi: c03e7200   edi: f7fddc00   ebp: 00000213   esp: c216ef18
<0>ds: 007b   es: 007b   ss: 0068
<0>Process events/0 (pid: 8, threadinfo=c216e000 task=c2167030)
<0>Stack: <0>f7fddc00 c03e7200 00000001 c0351872 f7fddc00 c03e7200 00000005 
c040d724
<0>       c21422c0 c03518aa c040d720 c0126bbd 00000000 00000000 c035189f 
c216e000
<0>       c216ef88 c21422c0 c21422c0 c0126cf3 c21422c0 ffffffff ffffffff 
00000001
<0>Call Trace:
(Continue reading)

Permalink | Reply | 4 comments |
headers
Vincent Roqueta | 1 Mar 2006 13:06
Picon

Re: 2.6.16-rc4-CITI_NFS4_ALL-4 server : oops

Hello Trond,

After a server oops, I get a client oops at mount time:
___________________________________________________
general protection fault: 0000 [#1]
PREEMPT SMP
Modules linked in: video ipv6
CPU:    2
EIP:    0060:[<c014de24>]    Not tainted VLI
EFLAGS: 00010297   (2.6.16-rc4-CITI_NFS4_ALL-4 #2)
EIP is at free_percpu+0x29/0x62
eax: 0000000f   ebx: 00000000   ecx: f58f85e0   edx: 00000000
esi: ffffffff   edi: f62f1000   ebp: 00000000   esp: f584ceac
ds: 007b   es: 007b   ss: 0068
Process mount (pid: 3604, threadinfo=f584c000 task=f784b550)
Stack: <0>f5be4000 f72e2800 c01b35e4 00000000 f517d000 00000000 c219e440 
c03e57e0
       c0156572 c03e57e0 00000000 f517d000 f62f1000 f6493000 f60b2000 00000000
       f60b2004 00000000 c0169254 f60b2000 00000000 f517d000 f62f1000 00000000
Call Trace:
 [<c01b35e4>] nfs4_get_sb+0x200/0x210
 [<c0156572>] do_kern_mount+0x85/0x12d
 [<c0169254>] do_new_mount+0x49/0x6f
 [<c01697d7>] do_mount+0x172/0x18a
 [<c0169618>] copy_mount_options+0x4c/0x99
 [<c0169ab7>] sys_mount+0x79/0xb3
 [<c010282f>] sysenter_past_esp+0x54/0x75
Code: 5d c3 56 a1 a0 70 4c c0 ba 20 00 00 00 53 85 c0 8b 74 24 0c f7 d6 74 03 
0f bc d0 83 fa 09 bb 08 00 00 00 0f 4c da 83 fb 07 7f 2f <ff> 34 9e e8 78 ff 
ff ff 8d 43 01 89 04 24 6a 08 68 a0 70 4c c0
(Continue reading)

Permalink | Reply | 3 comments |
headers
Timo Aaltonen | 1 Mar 2006 13:48
Picon
Picon

Re: util-linux-patch breaks nfs-user-server

On Sat, 25 Feb 2006, Trond Myklebust wrote:

> I doubt this is the reason for the failure, though. The standard Linux
> portmapper works fine with both. I suspect, rather, that nfs-user-server
> is screwed up w.r.t. some of the other assumptions that we make when we
> probe it for tcp support.
>
> Can you get us a full binary tcpdump of what is going on? 'tcpdump -s
> 9000 -w /tmp/dump' should be fine.

Here goes:

http://users.tkk.fi/~tjaalton/tmp/dump.client

t
Permalink | Reply | 12 comments |
headers
Gopal Santhanam | 1 Mar 2006 16:18

Compile error linux-2.6.16-rc4-CITI_NFS4_ALL-4.diff

Hi,

I am trying to compile the vanilla linux-2.6.16-rc4 with the patch
linux-2.6.16-rc4-CITI_NFS4_ALL-4.diff.

I get the following compile error!

CC [M]  fs/nfsd/export.o
fs/nfsd/export.c: In function 'svc_export_put':
fs/nfsd/export.c:284: error: 'struct svc_export' has no member named 'ex_fslocs'
fs/nfsd/export.c: In function 'svc_export_parse':
fs/nfsd/export.c:424: error: 'struct svc_export' has no member named 'ex_fslocs'
fs/nfsd/export.c: In function 'svc_export_update':
fs/nfsd/export.c:567: error: 'struct svc_export' has no member named 'ex_fslocs'
fs/nfsd/export.c:567: error: 'struct svc_export' has no member named 'ex_fslocs'

I did not see this problem with linux-2.6.16-rc4-CITI_NFS4_ALL-1.diff.

Any assistance would be appreciated.

Gopal
Permalink | Reply | 7 comments |
headers
Vincent Roqueta | 1 Mar 2006 16:30
Picon

Re: Compile error linux-2.6.16-rc4-CITI_NFS4_ALL-4.diff

Le Mercredi 1 Mars 2006 16:18, Gopal Santhanam a écrit :
> Hi,
>
> I am trying to compile the vanilla linux-2.6.16-rc4 with the patch
> linux-2.6.16-rc4-CITI_NFS4_ALL-4.diff.
>
> I get the following compile error!
>
> CC [M]  fs/nfsd/export.o
> fs/nfsd/export.c: In function 'svc_export_put':
> fs/nfsd/export.c:284: error: 'struct svc_export' has no member named
> 'ex_fslocs' fs/nfsd/export.c: In function 'svc_export_parse':
> fs/nfsd/export.c:424: error: 'struct svc_export' has no member named
> 'ex_fslocs' fs/nfsd/export.c: In function 'svc_export_update':
> fs/nfsd/export.c:567: error: 'struct svc_export' has no member named
> 'ex_fslocs' fs/nfsd/export.c:567: error: 'struct svc_export' has no member
> named 'ex_fslocs'
>
>
> I did not see this problem with linux-2.6.16-rc4-CITI_NFS4_ALL-1.diff.

I am surprised, because I haven't seen anything so.
The problem remains if you set nfsd built-in (and not as a module) ?

Vincent
Permalink | Reply | 6 comments |
headers
Gopal Santhanam | 1 Mar 2006 16:42

Re: Compile error linux-2.6.16-rc4-CITI_NFS4_ALL-4.diff

No, it does not work when compiling as built-in:

CC      fs/nfsd/export.o
fs/nfsd/export.c: In function 'svc_export_put':
fs/nfsd/export.c:284: error: 'struct svc_export' has no member named 'ex_fslocs'
fs/nfsd/export.c: In function 'svc_export_parse':
fs/nfsd/export.c:424: error: 'struct svc_export' has no member named 'ex_fslocs'
fs/nfsd/export.c: In function 'svc_export_update':
fs/nfsd/export.c:567: error: 'struct svc_export' has no member named 'ex_fslocs'
fs/nfsd/export.c:567: error: 'struct svc_export' has no member named 'ex_fslocs'
make[2]: *** [fs/nfsd/export.o] Error 1
make[1]: *** [fs/nfsd] Error 2
make: *** [fs] Error 2

Gopal

On Wednesday 01 March 2006 07:30, Vincent Roqueta wrote:
> Le Mercredi 1 Mars 2006 16:18, Gopal Santhanam a écrit :
> > Hi,
> >
> > I am trying to compile the vanilla linux-2.6.16-rc4 with the
> > patch linux-2.6.16-rc4-CITI_NFS4_ALL-4.diff.
> >
> > I get the following compile error!
> >
> > CC [M]  fs/nfsd/export.o
> > fs/nfsd/export.c: In function 'svc_export_put':
> > fs/nfsd/export.c:284: error: 'struct svc_export' has no
> > member named 'ex_fslocs' fs/nfsd/export.c: In function
> > 'svc_export_parse': fs/nfsd/export.c:424: error: 'struct
(Continue reading)

Permalink | Reply | 4 comments |
headers
Gopal Santhanam | 1 Mar 2006 16:52

NFSv3 + Kerberos + x86_64 (Fedora Core 4)

Hi all,

I have been using FC4's kernel-2.6.15-1.1831 on both
i686 and x86_64 architecture machines.

I want to use NFSv3 with Kerberos.  This works fine
on i686 with the aforementioned FC4 kernel.  I am
able to do the following:

1) Setup NFS server with kerberos

2) Mount on NFS client with -o sec=krb5.

Now, if I have (1) on i686 and do (2) on x86_64, things just break:

{root <at> x86_64client:3:~} mount -o sec=krb5 server:/export/test /mnt
{root <at> x86_64client:4:~} ls -al /mnt
ls: /mnt: Broken pipe

The clients are setup exactly the same
(both have /etc/krb5.keytab, /etc/idmapd.conf). 
They are both running rpc.idmapd and rpc.gssd.

An important thing to mention is that rpc.gssd is still running
after the mount command but then has vanished (crashed?) after the ls.

This has got to be a known problem.  How do I fix it?

Thanks,
Gopal
(Continue reading)

Permalink | Reply | 2 comments |

Gmane