Hi guys!
Could someone please tell me where to get the openglsdk for linux? im using mandrake8.2 …
wfg codi
The document at http://developer.nvidia.com/docs/IO/1462/SUPP/AutoupdateGuide.htm gives good instructions on how to download the latest version of the SDK from CVS. If you manage to get it built, I would love to know what you did. I’m also running Mandrake 8.2 and I’m having huge problems with the compilation.
That sounds very difficult, i don`t understand the documentation on how to download via cvs, could you explain?
Could not ssh into NVIDIA’s server… what’s up?
I’m running redhat 7.2, open SSH 2.9p2, and CVS 1.11.
ssh just hangs…
ssh -v dumps this:
debug1: Reading configuration data /etc/ssh/ssh_config
debug1: Applying options for *
debug1: Seeding random number generator
debug1: Rhosts Authentication disabled, originating port will not be trusted.
debug1: restore_uid
debug1: ssh_connect: getuid 1083 geteuid 0 anon 1
debug1: Connecting to cvs1.nvidia.com [209.213.198.12] port 22.
debug1: temporarily_use_uid: 1083/40 (e=0)
debug1: restore_uid
debug1: temporarily_use_uid: 1083/40 (e=0)
Brainleacher,
Ssh is supposed to just “hang” when logging into our server directly, this is the correct behavior as its waiting for CVS commands. You will not get a console prompt since we restricted access so you can only execute CVS commands.
[This message has been edited by jra101 (edited 05-29-2002).]
Hey Jra101,
Thanks for the quick reply.
Actually, the cvs co also hangs.
After about 3 mins or so, it times out
error:
cvs1.nvidia.com: Connection timed out
cvs [checkout aborted]: end of file from server (consult above messages if any)
and no files show up…
Here’s my cvs co command:
cvs co -kb nvsdk/Common/media
everything is set up exactly according to the web page and the firewall settings allow SSH out across port 22…
I don’t get it…
- Brainleacher
nope… the IP resolves to 209.213.198.12
and the traceroute goes to 209.249.170.25
and then to 10.0.0.10.
I just figured that they had their ping daemon off…
can you ping them from where you are?
Yes, I can ping them from work (inside our network though). I will try from home later but I might not be able to do so since my internet has been broken for 4-5 days 
Edit: Also, that is the correct IP so its not a DNS problem.
[This message has been edited by jra101 (edited 05-29-2002).]
Yup, I can ping it fine from outside the NVIDIA network. Looks like the problem is on your end
Hi!
I hae not so much experience with linux, could somone tell how CVS works? r is there a graphical front-end?
wfg codi
Yes there are gfx frontends like pharamacy using GTK+ for instance.
Speaking of the devil.
Nvidia messed the ‘linux protable’ SDK up.
First off it only compiles ( but not whole ) when using gcc and g++ 3.0.x ( nameless unions and stuff inside the math stuff ).
And even when using the GNU compiler suit 3.0.x can’t seem to compile the stuff propalry under my SuSE 7.3 profesional.
The guys should make the linux port with use of autoconf , automake and autoheader.
Those makefiles are so messed up.
Hey Jra101,
So I’ve found the problem. The problem is that our router/firewall supports IPSEC, and therefore requires packet truncation, your router at 10.0.0.10 does not allow packet truncation. Which probably means that you also do not support VPN/IPSEC, and is a problem for anyone who’s behind a firewall that does, such as us over here at centropolis effects. Do you think that your routers can be reconfigured any time soon or do you think there’s any other way of getting the Linux SDK? Thanks for the follow-ups
- Brainleacher
What do i have to compile?
I have downloaded the SDK, but what do i have to compile there?
wfg codi
CopyCat,
Can you email me specifically what problems you are having?
Brainleacher,
I will ask our network guys and see if thats feasable.
Codemasta,
go to the OpenGL/src directory and run ‘make’ and that should compile everything in the correct order.
I have the same problem. I don’t know if the analysis Brainleacher gave is correct, but I too am behind a firewall (though I don’t know if we use IPSEC, or even what it is…)
Several weeks ago there was a ‘time window’ when cvs could work, I think it was right at the time you moved from SSH1 to SSH2
But I no longer can ‘cvs update’ my tree…
When I put my pc outside of the firewall, I could download the SDK w/o any problems. I’m sure that it’s not because the outgoing SSH or CVS ports are blocked cuz I checked our firewall rules and could also ssh and cvs to other servers. So I’m sure it’s the interaction between our firewall NVidias… and I really believe it’s the packet truncation feature that the NVidia firewalls lack…
- Brainleacher
Don’t forget to export the CVS_RSH and CVSROOT variables otherwise the checkout will fail. An interesting one is when CVSROOT is exported (because you use CVS elsewhere) and CVS_RSH is not (because it is seldom used). The checkout will proceed but you will use rsh instead of the script with the ssh call. The effect is that CVS will try to connect at port 514 and after the second connect attempt, the Nvidia’s firewall will classify your IP as an attacker. You will be banished (nothing works, including pings) for some time. The banishment is less than an hour, seems to be about 15 minutes but I did not actually time it.
Don’t forget to set and export the CVS_RSH variable before doing an update. CVSROOT will be handled for you (stored in the ./CVS/Root file of each directory). If you forget, you will trigger the firewall’s security and will again be locked out.
Maybe the web page should mention exporting the variables. Many that are new to Unix will not know about this requirement.
enjoy,
scott
Originally posted by jra101:
[b]Moshe Nissim,
Can you ping cvs1.nvidia.com at all? Are you using SSH2 and the new SSH2 key?[/b]
I cannot ping cvs1.nvidia.com
From home I can. I don’t know if your firewall blocks incoming ICMP responses.
And yes, I am using the new ssh2 key, and the one time which I could actually cvs-download everything (from work) was with the ssh2 key