[Linux-ia64] heads up: ucd-snmp 64-bit problem

[Linux-ia64] heads up: ucd-snmp 64-bit problem

[David Mosberger]

Just a quick note: I discovered that there are some ugly 64-bit
dirtiness issues in the UCD snmp package.  It seems that Bill has
fixed some of them already, but even with the "rawhide" version of
ucd-snmp-utils, setting an IP address doesn't work.  For example, a
command along the lines of:

 $ snmpset aws110 public .enterprises.410.1.1.2.1.0 a 192.168.1.10

would fail because "snmpset" tries to write the IP address as an eight
byte quantity.  I hope someone who's more familiar with SNMP can fix
this.

	--david

RE: [Linux-ia64] heads up: ucd-snmp 64-bit problem

[Matt_Domsch]

> I hope someone who's more familiar with SNMP can fix
> this.

One of our server management guys has done a great deal of patching on this
code.  In many many many places in the code, it's assumed that an IP address
is a long, and it's not a simple change.  You've basically got to track down
everywhere an IP address value is used and fix it.  Given this is SNMP,
that's a lot of places.  I'll forward this to him and see what can be done.

Thanks for the info!
-Matt

-- 
Matt Domsch
Sr. Software Engineer
Dell Linux Solutions
www.dell.com/linux
#2 Linux Server provider with 17% in the US and 14% Worldwide (IDC)!
#3 Unix provider with 18% in the US (Dataquest)!

RE: [Linux-ia64] heads up: ucd-snmp 64-bit problem

[David Mosberger]

>>>>> On Tue, 24 Jul 2001 07:50:41 -0500, Matt_Domsch@Dell.com said:

  >> I hope someone who's more familiar with SNMP can fix this.

  Matt> One of our server management guys has done a great deal of
  Matt> patching on this code.  In many many many places in the code,
  Matt> it's assumed that an IP address is a long, and it's not a
  Matt> simple change.

Yes, I have to say I was a bit shocked when I looked at the code.
Looks like straight from the seventies.  The code also ought to use
inet_aton() instead of inet_addr().

  Matt> You've basically got to track down everywhere an IP address
  Matt> value is used and fix it.  Given this is SNMP, that's a lot of
  Matt> places.  I'll forward this to him and see what can be done.

If he does work on this, could you ask him to cc any patches to
linux-ia64 so everybody can see them?

BTW: I think snmptrapd is suffering from similar problems.

Thanks,

	--david