Re: Some Doubts

2006-01-11 Thread Post, Mark K
If the RPM name has noarch in it, it means it does not contain any
binary files specific to a particular architecture.  If it has i386,
s390x, or anything like that, it is almost guaranteed to not run on any
architecture other than the one specified.  You might get lucky if
someone made a mistake and didn't mark something as noarch when they
should have, but I wouldn't count it working.

If you don't use RPM to install something, RPM will *not* know about it.
So, no  surprise there.  If you want to install something that SUSE
doesn't provide, create a SRPM for it, and build a "binary" RPM from it.


Mark Post

-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Yu
Safin
Sent: Wednesday, January 11, 2006 6:39 PM
To: LINUX-390@VM.MARIST.EDU
Subject: Some Doubts


Some doubts: (cross posted with the OpenSuSE forum)
1) I am not clear as to  why some rpm's are named i386 and some noarch.
eg)
perl-libwww-perl-5.801-8.noarch.rpm
 perl-GDTextUtil-0.86-8.i386.rpm

I suspect the i386 means Intel but when I install a perl i386 rpm on
my Mainframe z890 it works (under Linux).
2) when I install some perl RPM's on my SuSe 9.3 from rpmpam I had no
problems.  However, when I go into CPAN (perl -MCPAN -e shell), I find
that the rpm's for the perl modules are not showing up when I do an
"i" on the name (e.g. perl-AnyData-xxx.rpm was installed and then
under CPAN, "i AnyData").
is this normal?

I am trying to avoid using CPAN because I have been burned before when
I install and module that I need to remove later on.  CPAN won't allow
me.
Another problem with CPAN is that I need to do it with me present, I
can't just automate the installation of a bunch of RPM's for new
servers.

Any suggestions will be appreciated.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Some Doubts

2006-01-11 Thread Yu Safin
Some doubts: (cross posted with the OpenSuSE forum)
1) I am not clear as to  why some rpm's are named i386 and some noarch.
eg)
perl-libwww-perl-5.801-8.noarch.rpm
 perl-GDTextUtil-0.86-8.i386.rpm

I suspect the i386 means Intel but when I install a perl i386 rpm on
my Mainframe z890 it works (under Linux).
2) when I install some perl RPM's on my SuSe 9.3 from rpmpam I had no
problems.  However, when I go into CPAN (perl -MCPAN -e shell), I find
that the rpm's for the perl modules are not showing up when I do an
"i" on the name (e.g. perl-AnyData-xxx.rpm was installed and then
under CPAN, "i AnyData").
is this normal?

I am trying to avoid using CPAN because I have been burned before when
I install and module that I need to remove later on.  CPAN won't allow
me.
Another problem with CPAN is that I need to do it with me present, I
can't just automate the installation of a bunch of RPM's for new
servers.

Any suggestions will be appreciated.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: S390 Suse FTP installation - IP address question.

2006-01-11 Thread Rob van der Heij
On 1/11/06, MOEUR TIM C <[EMAIL PROTECTED]> wrote:

> Here's how I did this.

The reason Mike did his mksles9root script to prepare the file tree is
that this way you install the latest versions from the service pack
rather than the older base levels. This way you avoid creating a
system that does not run because of missing fixes, save yourself a lot
of time because avoiding the upgrade, avoid trouble because of massive
upgrade with YOU.

That's why I recommend following that route (there's still enough
service beyond SP2 that will show you how the upgrade works).

Rob

--
Rob van der Heij  rvdheij @ gmail.com
Velocity Software, Inc

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: [2.6 patch] arch/s390/Makefile: remove -finline-limit=10000

2006-01-11 Thread David S. Miller
From: Martin Schwidefsky <[EMAIL PROTECTED]>
Date: Wed, 11 Jan 2006 10:21:20 +0100

> On Tue, 2006-01-10 at 21:57 +0100, Adrian Bunk wrote:
> > -finline-limit might have been required for older compilers, but
> > nowadays it does no longer make sense.
>
> I didn't check the effects of reverting to the default inline-limit, did
> you find any negative impacts? I'm thinking about the critical code
> paths e.g. minor faults. There better should not be an additional
> function call that would have been inlined with the bigger inline limit,
> since function calls are quite expensive on s390.

You need to be careful now that -Os is specified by default
in 2.6.x

The inline-limit GCC option is interpreted differently in
gcc-4.x when -Os is given vs. when it is not.

On Sparc this caused schedule() to be inlined (I'm not kidding)
which caused all kinds of troubles.

I highly recommed you don't specify it and let the compiler
make the decisions, and add inline tags to places where you
think it is hyper-important for inlining to occur.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: S390 Suse FTP installation - IP address question.

2006-01-11 Thread MOEUR TIM C
Ted,

Here's how I did this.



Instead of 3 (FTP) I used 1 NFS. BUT FIRST, you need to do some stuff.


- Before you choose NFS find a nearby Linux desktop machine that is on
your network.
- Make a copy of CD #1 (or you can take it out of the HMC, but two
copies are easier).
- Put CD #1 in the drive on the desktop machine.  
- Configure your desktop machine to use NFS, and export the cd
directory.  It was /media/cdrom in my case.  I had have to read up on
NFS, exportfs to get that all straight.  
- Back on the HMC, now select 1) NFS
- The installation dialog will then ask for the nfs server address and
directory where it can find the installation disk.  Answer with your
desktop machine and /media/cdrom

You still have to then log on to the installation Linux machine and
start YaST, but hopefully this will get you past the unusual "where's
the installation media?" problem.



Here's how I did this 

-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of
Ted Keesey
Sent: Wednesday, January 11, 2006 1:22 PM
To: LINUX-390@VM.MARIST.EDU
Subject: S390 Suse FTP installation - IP address question.

Hello,

I have begun installing s390 Suse Linux into a Mainframe LPAR.
I am installing the boot code from CD #1 (from the HMC).
I have successfully set up my network parameters.
I am now at the point of choosing the installation source.  Here is the
message
:
Please specify the installation source:
1) NFS
2) SAMBA
3) FTP
4) Abort

I chose option 3 (FTP)

The response is the following:
Please enter IP-Number of the host providing the installation

It is here that I am stuck.  I have been unable to find the FTP IP
address for Novell's FTP site for S390.  Does anyone know how I should
proceed from here?  Are there mirror sites?
If it does exist, is it anonymous?

Thanks for any help you can provide.

Ted.

--
For LINUX-390 subscribe / signoff / archive access instructions, send
email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or
visit http://www.marist.edu/htbin/wlvindex?LINUX-390

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: S390 Suse FTP installation - IP address question.

2006-01-11 Thread Eddie Chen
 The quickest  way is to have a FTP server... The question I like to know ,
do   you have a LINUX  machine running else where and allow FTP.
  You might need this server  to  ssh

  You can then create a directory and account(userid)  on that server...
   i.e.

. mkdir  /SuSE9CD
. useradd  suse9inst-d   /SuSE9CD
. passwd   <- set the password


. cp -dpR /media/cdrom /SuSE9CD,<-- do these two step for all 6 CDs

 . eject
 .  chown suse9inst:users -R /SuSE9CD
  . ifconfig <- get the IP address of the server

Go back to HMC and  you need the IP address of the FTP server, the
userid and the password.

the directory path is "../SuSE9CD

You will then need to "ssh" into Mainframe(your mainframe IP address)
then  "yast"





 Ted Keesey
 <[EMAIL PROTECTED]
 an.com>To
 Sent by: Linux on LINUX-390@vm.marist.edu
 390 Port   cc
 <[EMAIL PROTECTED]
 ist.edu>  Subject
   S390 Suse FTP installation - IP
   address question.
 01/11/2006 03:21
 PM


 Please respond to
 Linux on 390 Port
 <[EMAIL PROTECTED]
 ist.edu>






Hello,

I have begun installing s390 Suse Linux into a Mainframe LPAR.
I am installing the boot code from CD #1 (from the HMC).
I have successfully set up my network parameters.
I am now at the point of choosing the installation source.  Here
is the message
:
Please specify the installation source:
1) NFS
2) SAMBA
3) FTP
4) Abort

I chose option 3 (FTP)

The response is the following:
Please enter IP-Number of the host providing the installation

It is here that I am stuck.  I have been unable to find the
FTP IP address for Novell's FTP site for S390.  Does anyone
know how I should proceed from here?  Are there mirror sites?
If it does exist, is it anonymous?

Thanks for any help you can provide.

Ted.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or
visit
http://www.marist.edu/htbin/wlvindex?LINUX-390



-
This message and its attachments may contain  privileged and
confidential information.  If you are not the intended recipient(s),
you are prohibited from printing, forwarding, saving or copying this
email.  If you have received this e-mail in error, please immediately
notify the sender and delete this e-mail and its attachments from your
computer.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: S390 Suse FTP installation - IP address question.

2006-01-11 Thread Rob van der Heij
On 1/11/06, Ted Keesey <[EMAIL PROTECTED]> wrote:

> It is here that I am stuck.  I have been unable to find the
> FTP IP address for Novell's FTP site for S390.  Does anyone
> know how I should proceed from here?  Are there mirror sites?
> If it does exist, is it anonymous?

You'll have to provide your own ftp server loaded with stuff from the
CD's you got from Novell. Recommended to use Mike MacIsaac's
mksles9root program to build the proper structures.
I believe you can even boot the first CD on a PC to get a working
Linux with ftp server.

Rob
--
Rob van der Heij  rvdheij @ gmail.com

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


S390 Suse FTP installation - IP address question.

2006-01-11 Thread Ted Keesey
Hello,

I have begun installing s390 Suse Linux into a Mainframe LPAR.
I am installing the boot code from CD #1 (from the HMC).
I have successfully set up my network parameters.
I am now at the point of choosing the installation source.  Here
is the message
:
Please specify the installation source:
1) NFS
2) SAMBA
3) FTP
4) Abort

I chose option 3 (FTP)

The response is the following:
Please enter IP-Number of the host providing the installation

It is here that I am stuck.  I have been unable to find the
FTP IP address for Novell's FTP site for S390.  Does anyone
know how I should proceed from here?  Are there mirror sites?
If it does exist, is it anonymous?

Thanks for any help you can provide.

Ted.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Very interesting article!

2006-01-11 Thread Ranga Nathan

People often make the mistake of comparing Linux to Windows. I know one
is a serious operating system  - multi-user, multi-tasking and the other
is not. I still can not run multiple isolated sessions or login as
different users concurrently.

There is a place for Windows and a place for Linux. I do not see one
dislodging the other. In fact they co-exist very well thanks to the
standards. Good job EFF and IETF and other standards bodies. Without you
we would have had no choice.

Linux is the greatest fruit anarchy has yielded. The organized anarchy
is an amazing phenomenon of the last two decades.

McKown, John wrote:


"Linux is Not Windows"

Granted, more oriented towards the desktop crowd, but hopefully on-topic
enough. If not, I apologize. The only thing that I see wrong with it is
that the author does not get "libre" != "gratis". He states that FOSS is
free. He states that there is no monetary benefit to the developer with
FOSS as there is with proprietary software. I don't think this is
entirely true. But I guess it is more so than with proprietary (even
shareware).

http://linux.oneandoneis2.org/LNW.htm

--
John McKown
Senior Systems Programmer
UICI Insurance Center
Information Technology

This message (including any attachments) contains confidential
information intended for a specific individual and purpose, and its
content is protected by law.  If you are not the intended recipient, you
should delete this message and are hereby notified that any disclosure,
copying, or distribution of this transmission, or taking any action
based on it, is strictly prohibited.


--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390





--
__
Ranga Nathan
Work: 714-442-7591


--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
begin:vcard
fn:Ranga Nathan
n:Nathan;Ranga
email;internet:[EMAIL PROTECTED]
tel;work:714-442-7591
version:2.1
end:vcard



Re: glibc compilation error

2006-01-11 Thread Post, Mark K
I added the include for  to
linux-2.6.14.3/include/asm-s390/sigcontext.h and that did resolve the
undefined errors.  Do you think we can get that fix sent upstream?


Thanks,

Mark Post

-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of
Peter 1 Oberparleiter
Sent: Wednesday, January 11, 2006 8:15 AM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: glibc compilation error


Mark Post <[EMAIL PROTECTED]> wrote on 11.01.2006 08:09:58:
> I'm trying to compile glibc 2.3.6 using linux 2.6.14.3 headers.  I'm
getting
> this error:
[...]

> Apparently, __user isn't being defined anywhere in the chain of header
files
> being pulled in.  Does anyone know where it _should_ be defined, and 
> how ../nptl/sysdeps/s390/tcb-offsets.sym should be modified to include

> that file?

I recently came across a possibly related bug report on the SUSE
bugzilla database that suggested that "asm/sigcontext.h" is missing an
"#include " statement which leads to "__user" being
undefined. Maybe try adding that statement to the file and retry.


Regards,
  Peter Oberparleiter

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


[PATCH] arch/s390/kernel/setup.c: fix compilation on UP

2006-01-11 Thread Alexey Dobriyan
  CC  arch/s390/kernel/setup.o
arch/s390/kernel/setup.c: In function `do_machine_restart_nonsmp':
arch/s390/kernel/setup.c:271: error: too few arguments to function `__cpcmd'
arch/s390/kernel/setup.c: In function `do_machine_halt_nonsmp':
arch/s390/kernel/setup.c:279: error: too few arguments to function `__cpcmd'
arch/s390/kernel/setup.c: In function `do_machine_power_off_nonsmp':
arch/s390/kernel/setup.c:286: error: too few arguments to function `__cpcmd'

Signed-off-by: Alexey Dobriyan <[EMAIL PROTECTED]>
---

 arch/s390/kernel/setup.c |6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -268,7 +268,7 @@ static void do_machine_restart_nonsmp(ch
reipl_diag();

if (MACHINE_IS_VM)
-   cpcmd ("IPL", NULL, 0);
+   cpcmd ("IPL", NULL, 0, NULL);
else
reipl (0x1 | S390_lowcore.ipl_device);
 }
@@ -276,14 +276,14 @@ static void do_machine_restart_nonsmp(ch
 static void do_machine_halt_nonsmp(void)
 {
 if (MACHINE_IS_VM && strlen(vmhalt_cmd) > 0)
-cpcmd(vmhalt_cmd, NULL, 0);
+cpcmd(vmhalt_cmd, NULL, 0, NULL);
 signal_processor(smp_processor_id(), sigp_stop_and_store_status);
 }

 static void do_machine_power_off_nonsmp(void)
 {
 if (MACHINE_IS_VM && strlen(vmpoff_cmd) > 0)
-cpcmd(vmpoff_cmd, NULL, 0);
+cpcmd(vmpoff_cmd, NULL, 0, NULL);
 signal_processor(smp_processor_id(), sigp_stop_and_store_status);
 }

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: What size memory footprint does the kernel take

2006-01-11 Thread Rob van der Heij
My current SLES9 definition would look like this:
0154 SUS90244 NSS  016M 0  00011   EW  A  0   OMITTED   NO
00100  003E1   SR
00400  0045E   EW
00800  00985   EW

That means potentiall 737 pages to be shared. If you're talking about
Linux virtual machines of 2GB, then being able to share 2MB of it is
not an obvious winner. When you run a lot of small virtual machines it
becomes more appealing. Some benefit is in having the kernel pages
shared because page faults on those typically block the entire virtual
machine. And I believe there still is the side effect of CP being
reluctant to page out the EW portion (static kernel variables, so also
good to keep resident - but less fun for the initrd at 00800).

The documentation on the VM pages for creating the NSS are not correct
anymore. The current SuSE kernel comes with the option enabled by
default, so you should be able to use an NSS without having to compile
it yourself (and lose support). There's some options to simplify this
process with the latest version of zipl. When I can't find useful
documentation of that I will post the intructions one of these days.

We're currently in the process of repeating  the measurements on Linux
memory usage and the possible savings of NSS and DCSS facilities for
current releases. Watch this space on how your favorite performance
monitor will guide you in this.

Rob
--
Rob van der Heij  rvdheij @ gmail.com
Velocity Software, Inc

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


What size memory footprint does the kernel take

2006-01-11 Thread James Melin
In an active running system, using the default SuSE kernel, what is the 
footprint in memory?

I am being directed to put the Linux kernel in a name-save-segment in VM to 
share it. While I agree that this is goodness from a memory standpoint, I
would have to have at least 2 of them (prod and test) and I've only got mebbe 4 
prod systems and 6 test systems to begin with?

Secondly, is there a comprehensive document on sharing the kernel among 
systems? And lastly what are the best candidates to take advantage of the
execute in place file system?

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: glibc compilation error

2006-01-11 Thread Peter 1 Oberparleiter
Mark Post <[EMAIL PROTECTED]> wrote on 11.01.2006 08:09:58:
> I'm trying to compile glibc 2.3.6 using linux 2.6.14.3 headers.  I'm
getting
> this error:
[...]

> Apparently, __user isn't being defined anywhere in the chain of header
files
> being pulled in.  Does anyone know where it _should_ be defined, and how
> ../nptl/sysdeps/s390/tcb-offsets.sym should be modified to include that
> file?

I recently came across a possibly related bug report on the SUSE bugzilla
database that suggested that "asm/sigcontext.h" is missing an "#include
" statement which leads to "__user" being undefined.
Maybe try adding that statement to the file and retry.


Regards,
  Peter Oberparleiter

--
Peter Oberparleiter
Linux on zSeries Development
IBM Development Lab, Boeblingen/Germany

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Secure file transfers: thoughts on zLinux as server for MVS sysplex?

2006-01-11 Thread Brian France

We use FDR/Upstream here to move data between z/OS and MainFrameLinux using
hipersockets. On the MFL image, we then encrypt the data using PGP and FTP
it. The reverse works quite nicely too. What we like about Upstream is that
we use CA-7 to schedule the job on the z/OS image and Upstreams agent
monitors the job on both the z/OS and MFL image.

At 03:47 PM 1/9/2006, you wrote:

MVS access: that's where the question of practicality arises.  One can ftp
the file to the Linux image and if it heads across a hipersocket it can't
be sniffed therefore it needs not be encrypted.  That leaves the
authentication issue to be dealt with, but I admit I haven't given that a
great deal of thought.  Once the file is there a scheduler agent or
something similar can initiate the transfer from Linux to the final
location.  One can reverse the flow and the situation is changed very
little, the agent lets the MVS system know to retrieve the file once it
arrives at the Linux image.

OpenSSH on MVS uses OpenSSL software encryption routines and can consume a
lot of cycles.   That workload would likely be less expensive out of the
sysplex where it would not inflate all the related yet totally uninvolved
software license fees.




 [EMAIL PROTECTED]
 Sent by:
 [EMAIL PROTECTED]  To
 ST.EDULINUX-390@VM.MARIST.EDU
cc

 01/09/2006 15:10  Subject
   Re: Secure file transfers: thoughts
   on zLinux as server for MVS
 Please respond to sysplex?
 [EMAIL PROTECTED]
  ST.EDU








I don't understand how installing a Linux/390 system running SSH is
going to allow your MVS systems to access the data.  Besides, if running
SSH on your Linux system is good enough, why isn't running SSH on your
MVS systems good enough?


Mark Post

-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of
Tom Ambros
Sent: Monday, January 09, 2006 3:01 PM
To: LINUX-390@VM.MARIST.EDU
Subject: Secure file transfers: thoughts on zLinux as server for MVS
sysplex?


We are looking to eliminate password authentication and, probably,
encrypt all production file transfers on our internal network.

Our Unix engineers are loathe to install SSL enabled ftp clients but
instead wish to exploit OpenSSH.  As a matter of fact, all our new Unix
machines will be built with ftp disabled.  OpenSSH is not the best
solution for our MVS installation, so we look at alternatives.

One thought is to implement a zLinux image and use it as an ssh server
for the MVS sysplex.  Is this a practical idea?  How are others
approaching this issue, if one is unable to influence the Unix engineers
to install an SSL enabled ftp client?

Thanks...

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or
visit
http://www.marist.edu/htbin/wlvindex?LINUX-390





***
This communication may contain privileged and/or confidential information. It
is intended solely for the use of the addressee. If you are not the intended
recipient, you are strictly prohibited from disclosing, copying, distributing
or using any of this information. If you received this communication in error,
please contact the sender immediately and destroy the material in its
entirety,
whether electronic or hard copy. This communication may contain nonpublic
information about individuals and businesses subject to the restrictions
of the
Gramm-Leach-Bliley Act. You may not directly or indirectly reuse or redisclose
such information for any purpose other than to provide the services for which
you are receiving the information.

127 Public Square, Cleveland, OH 44114
***


If you prefer not to receive future e-mail offers for products or services
from
Key send an e-mail to [EMAIL PROTECTED] with 'No Promotional E-mails' in the
SUBJECT line.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390



Brian W. France
Systems Administrator (Mainframe)
Pennsylvania State University
Administrative Information Services - Infrastructure/Sysarc
Rm 25 Shields Bldg., University Park, Pa. 16802
814-863-4739
[EMAIL PROTECTED]

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECT

Re: [2.6 patch] arch/s390/Makefile: remove -finline-limit=10000

2006-01-11 Thread Martin Schwidefsky
On Tue, 2006-01-10 at 21:57 +0100, Adrian Bunk wrote:
> -finline-limit might have been required for older compilers, but
> nowadays it does no longer make sense.

I didn't check the effects of reverting to the default inline-limit, did
you find any negative impacts? I'm thinking about the critical code
paths e.g. minor faults. There better should not be an additional
function call that would have been inlined with the bigger inline limit,
since function calls are quite expensive on s390.

--
blue skies,
   Martin

Martin Schwidefsky
Linux for zSeries Development & Services
IBM Deutschland Entwicklung GmbH

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


glibc compilation error

2006-01-11 Thread Mark Post
I'm trying to compile glibc 2.3.6 using linux 2.6.14.3 headers.  I'm getting
this error:
make[2]: Entering directory `/tmp/glibc-2.3.6/csu'
gawk -f ../scripts/gen-as-const.awk ../nptl/sysdeps/s390/tcb-offsets.sym \
| s390x-slackware-linux-gcc -S -o
/tmp/glibc-2.3.6/build-glibc-2.3.6/tcb-offsets.hT3 -std=gnu99 -O3 -Wall -Win
line -Wstrict-prototypes -Wwrite-strings -g  -I../include -I. -I/tmp/gli
bc-2.3.6/build-glibc-2.3.6/csu -I.. -I../libio -I../nptl -I/tmp/glibc-2.3.6/
build-glibc-2.3.6 -I../sysdeps/s390/s390-64/elf -I../libidn/sysdeps/unix -I.
./nptl/sysdeps/unix/sysv/linux/s390/s390-64 -I../nptl/sysdeps/unix/sysv/linu
x/s390 -I../nptl/sysdeps/unix/sysv/linux -I../nptl/sysdeps/pthread -I../sysd
eps/pthread -I../nptl/sysdeps/unix/sysv -I../nptl/sysdeps/unix -I../nptl/sys
deps/s390 -I../sysdeps/unix/sysv/linux/s390/s390-64 -I../sysdeps/unix/sysv/l
inux/wordsize-64 -I../sysdeps/unix/sysv/linux/s390 -I../sysdeps/unix/sysv/li
nux -I../sysdeps/gnu -I../sysdeps/unix/common -I../sysdeps/unix/mman -I../sy
sdeps/unix/inet -I../sysdeps/unix/sysv -I../sysdeps/unix -I../sysdeps/posix
-I../sysdeps/s390/s390-64 -I../sysdeps/wordsize-64 -I../sysdeps/s390/fpu -I.
./sysdeps/s390 -I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754/flt-32 -I../
sysdeps/ieee754 -I../sysdeps/generic/elf -I../sysdeps/generic -nostdinc -isy
stem /usr/lib/gcc-lib/s390x-slackware-linux/3.3.4/include -isystem
/usr/src/linux-2.6.14.3/include -D_LIBC_REENTRANT -D_LIBC_REENTRANT -include
../include/libc-symbols.h   -DHAVE_INITFINI -x c - \
-MD -MP -MF /tmp/glibc-2.3.6/build-glibc-2.3.6/tcb-offsets.h.dT -MT
'/tmp/glibc-2.3.6/build-glibc-2.3.6/tcb-offsets.h.d
/tmp/glibc-2.3.6/build-glibc-2.3.6/tcb-offsets.h'
In file included from ../sysdeps/unix/sysv/linux/bits/sigcontext.h:28,
 from ../signal/signal.h:333,
 from ../include/signal.h:5,
 from ../sysdeps/unix/sysv/linux/s390/sys/ucontext.h:26,
 from ../sysdeps/unix/sysv/linux/s390/sys/procfs.h:35,
 from ../nptl_db/thread_db.h:29,
 from ../nptl/descr.h:33,
 from ../nptl/sysdeps/s390/tls.h:77,
 from ../include/tls.h:6,
 from :2:
/usr/src/linux-2.6.14.3/include/asm/sigcontext.h:64: error: parse error
before '*' token
make[2]: *** [/tmp/glibc-2.3.6/build-glibc-2.3.6/tcb-offsets.h] Error 1
make[2]: Leaving directory `/tmp/glibc-2.3.6/csu'


The offending definition is this:
struct sigcontext
{
unsigned long   oldmask[_SIGCONTEXT_NSIG_WORDS];
_sigregs__user *sregs;
};


Apparently, __user isn't being defined anywhere in the chain of header files
being pulled in.  Does anyone know where it _should_ be defined, and how
../nptl/sysdeps/s390/tcb-offsets.sym should be modified to include that
file?

The output of the gawk command is this:
#include 
#include 
void dummy(void) {
asm ("@@@name@@@MULTIPLE_THREADS_OFFSET@@@value@@@%0@@@end@@@" : : "i"
(offsetof (tcbhead_t, multiple_threads)));
asm ("@@@name@@@PID@@@value@@@%0@@@end@@@" : : "i" (offsetof (struct
pthread, pid)));
asm ("@@@name@@@TID@@@value@@@%0@@@end@@@" : : "i" (offsetof (struct
pthread, tid)));
}


Thanks,

Mark Post

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390