[OpenAFS] OpenAFS 1.3.78 on Linux 2.6

Mikkel Kruse Johnsen mikkel@linet.dk
Sun, 06 Feb 2005 12:12:08 +0100


--=-0UKJWwgT51vM0sVEsSTR
Content-Type: multipart/alternative; boundary="=-IIea0RFvl/kJeJ3thNjT"


--=-IIea0RFvl/kJeJ3thNjT
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Hi

You say that you don't have to export the syscall table. But when I run
the configure it still say it can't find the syscall and assuming it is
not exportet:

[mkj@frodo openafs-1.3.78]$ ./configure --enable-bos-new-config --
enable-largefile-fileserver --enable-supergroups --enable-fast-restart
--enable-bitmap-later --with-linux-kernel-
headers=3D/lib/modules/2.6.10-1.12_FC2smp/build/ --enable-transarc-paths
--enable-debug --enable-debug-lwp --with-afs-sysname=3Di386_linux26

...
checking for linux/syscall.h in kernel... no
checking for SELinux kernel... yes
checking for 5th argument in sock_create found in some SELinux
kernels... no
checking for page_follow_link_light vs page_follow_link... no
checking for redhat kernel configuration... no
checking for recalc_sigpending arg type... yes
checking for parent in struct task_struct... yes
checking for real_parent in struct task_struct... yes
checking for sig in struct task_struct... no
checking for sighand in struct task_struct... yes
checking for sigmask_lock in struct task_struct... no
checking for rlim in struct task_struct... no
checking for signal->rlim in struct task_struct... yes
checking for exit_state in struct task_struct... yes
checking if kernel uses MODVERSIONS... yes
checking which kernel modules to build... MP
configure: warning: Cannot determine sys_call_table status. assuming it
isn't exported
...

This is Fedora Core 2 kernel 2.6.10-1.12_FC2smp

What should I do to make it use the syscall table ?

/Mikkel

On Fri, 2005-02-04 at 09:26 -0500, Jason McCormick wrote:

>   First off, the 1.3.78 release is working VERY well for us.  It's nice n=
ot
> to have to re-spin the kernel RPMs to export the syscall table.  OpenAFS
> 1.3.78 also seems to have fixed a number of issues:
>=20
>   * Our cache consistency problems seem to have been solved
>=20
>   * Kernel panics under high I/O no longer occur
>  =20
>   * Kernel oopsing occasionally using memcache no longer occur.
>=20
>   The only major problem I can still see is the inability to unmount the
> filesystem containing the cache (bug #17451) and the problem I posted to
> -info about the default configuration not allowing AFS to start on system=
s
> with no separate /usr/vice/cache partition.
>=20
>=20
>   Looking great!
>=20

Med Venlig Hilsen

Linet
Tlf:
21287793
Mikkel Kruse Johnsen
Direkte:=20
21287793
=C3=98rholmgade 6 st. tv
email:
mikkel@linet.dk
DK-2200 K=C3=B8benhavn N
web:
http://www.linet.dk



--=-IIea0RFvl/kJeJ3thNjT
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
  <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; CHARSET=3DUTF-8">
  <META NAME=3D"GENERATOR" CONTENT=3D"GtkHTML/3.3.2">
</HEAD>
<BODY>
Hi<BR>
<BR>
You say that you don't have to export the syscall table. But when I run the=
 configure it still say it can't find the syscall and assuming it is not ex=
portet:<BR>
<BR>
[mkj@frodo openafs-1.3.78]$ ./configure --enable-bos-new-config --enable-la=
rgefile-fileserver --enable-supergroups --enable-fast-restart --enable-bitm=
ap-later --with-linux-kernel-headers=3D/lib/modules/2.6.10-1.12_FC2smp/buil=
d/ --enable-transarc-paths --enable-debug --enable-debug-lwp --with-afs-sys=
name=3Di386_linux26<BR>
<BR>
...<BR>
checking for linux/syscall.h in kernel... no<BR>
checking for SELinux kernel... yes<BR>
checking for 5th argument in sock_create found in some SELinux kernels... n=
o<BR>
checking for page_follow_link_light vs page_follow_link... no<BR>
checking for redhat kernel configuration... no<BR>
checking for recalc_sigpending arg type... yes<BR>
checking for parent in struct task_struct... yes<BR>
checking for real_parent in struct task_struct... yes<BR>
checking for sig in struct task_struct... no<BR>
checking for sighand in struct task_struct... yes<BR>
checking for sigmask_lock in struct task_struct... no<BR>
checking for rlim in struct task_struct... no<BR>
checking for signal-&gt;rlim in struct task_struct... yes<BR>
checking for exit_state in struct task_struct... yes<BR>
checking if kernel uses MODVERSIONS... yes<BR>
checking which kernel modules to build... MP<BR>
configure: warning: Cannot determine sys_call_table status. assuming it isn=
't exported<BR>
...<BR>
<BR>
This is Fedora Core 2 kernel 2.6.10-1.12_FC2smp<BR>
<BR>
What should I do to make it use the syscall table ?<BR>
<BR>
/Mikkel<BR>
<BR>
On Fri, 2005-02-04 at 09:26 -0500, Jason McCormick wrote:
<BLOCKQUOTE TYPE=3DCITE>
<PRE>
<FONT COLOR=3D"#000000">  First off, the 1.3.78 release is working VERY wel=
l for us.  It's nice not</FONT>
<FONT COLOR=3D"#000000">to have to re-spin the kernel RPMs to export the sy=
scall table.  OpenAFS</FONT>
<FONT COLOR=3D"#000000">1.3.78 also seems to have fixed a number of issues:=
</FONT>

<FONT COLOR=3D"#000000">  * Our cache consistency problems seem to have bee=
n solved</FONT>

<FONT COLOR=3D"#000000">  * Kernel panics under high I/O no longer occur</F=
ONT>
<FONT COLOR=3D"#000000">  </FONT>
<FONT COLOR=3D"#000000">  * Kernel oopsing occasionally using memcache no l=
onger occur.</FONT>

<FONT COLOR=3D"#000000">  The only major problem I can still see is the ina=
bility to unmount the</FONT>
<FONT COLOR=3D"#000000">filesystem containing the cache (bug #17451) and th=
e problem I posted to</FONT>
<FONT COLOR=3D"#000000">-info about the default configuration not allowing =
AFS to start on systems</FONT>
<FONT COLOR=3D"#000000">with no separate /usr/vice/cache partition.</FONT>


<FONT COLOR=3D"#000000">  Looking great!</FONT>

</PRE>
</BLOCKQUOTE>
<TABLE CELLSPACING=3D"0" CELLPADDING=3D"0" WIDTH=3D"100%">
<TR>
<TD>
Med Venlig Hilsen<BR>
<BR>
<TABLE CELLSPACING=3D"1">
<TR>
<TD WIDTH=3D"250">
Linet
</TD>
<TD>
Tlf:
</TD>
<TD>
21287793
</TD>
</TR>
<TR>
<TD WIDTH=3D"100%">
Mikkel Kruse Johnsen
</TD>
<TD>
Direkte:=20
</TD>
<TD>
21287793
</TD>
</TR>
<TR>
<TD>
&#216;rholmgade 6 st. tv
</TD>
<TD>
email:
</TD>
<TD>
mikkel@linet.dk
</TD>
</TR>
<TR>
<TD WIDTH=3D"100%">
DK-2200 K&#248;benhavn N
</TD>
<TD>
web:
</TD>
<TD>
http://www.linet.dk
</TD>
</TR>
</TABLE>
<BR>
<BR>
</TD>
</TR>
</TABLE>
</BODY>
</HTML>

--=-IIea0RFvl/kJeJ3thNjT--

--=-0UKJWwgT51vM0sVEsSTR
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQBCBfuIs/DIef2Kq/IRArx2AKCmX5yI+S6s8nk1JcHdPRqM7Gtx+QCgmIKk
b/saYpgJ+quc2kTmkRrLu+g=
=5xMN
-----END PGP SIGNATURE-----

--=-0UKJWwgT51vM0sVEsSTR--