Subject: port-alpha/13600: 1.5.1 install hangs on an Alpha PC164 with 2 3Com EtherLinkXL cards
To: None <gnats-bugs@gnats.netbsd.org>
From: None <mhitch@montana.edu>
List: netbsd-bugs
Date: 07/30/2001 11:59:39
>Number:         13600
>Category:       port-alpha
>Synopsis:       mclpool limit reached; hangs on a PC164 with 2 3Com EtherLinkXL cards
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    port-alpha-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Jul 30 10:56:00 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     Michael L. Hitch
>Release:        NetBSD 1.5.1
>Organization:
Montana State University
	
>Environment:
	
System: NetBSD alpha.msu.montana.edu 1.5.1 NetBSD 1.5.1 (PC164) #2: Thu Jul 26 04:26:36 GMT 2001 mhitch@alpha.msu.montana.edu:/usr/src/sys/arch/alpha/compile/PC164 alpha


>Description:
	Attempt to boot 1.5.1 installation kernel on an Alpha with 2 3Com
	EtherLinkXL type adapters.  Systems hangs after the message:
	WARNING: mclpool limit reached; increase NMBCLUSTERS

	This is due to an increase in the number of receive mbuf clusters allocated
	by the elinkxl driver since the 1.5 release.  I think this will probably
	also happen in -current as well.  The elinkxl driver now allocates 128
	mbuf clusters for receive buffers, and two interfaces will use all of the
	default 256.
>How-To-Repeat:
	Boot the 1.5.1 installation kernel (or even a plain old GENERIC kernel) on
	an Alpha system with 2 3Com EtherLinkXL type network cards.
>Fix:
	The default NMBCLUSTERS should be increased in alpha/include/param.h, or
	in the kernel configuration files for at least GENERIC and INSTALL.
>Release-Note:
>Audit-Trail:
>Unformatted: