Subject: kern/31275: mclpool hardlimit problem
To: None <kern-bug-people@netbsd.org, gnats-admin@netbsd.org,>
From: None <yamt@mwd.biglobe.ne.jp>
List: netbsd-bugs
Date: 09/09/2005 08:12:00
>Number: 31275
>Category: kern
>Synopsis: mclpool hardlimit problem
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: kern-bug-people
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Fri Sep 09 08:12:00 +0000 2005
>Originator: YAMAMOTO Takashi <yamt@mwd.biglobe.ne.jp>
>Release: NetBSD 3.99.8
>Organization:
>Environment:
System: NetBSD kaeru 3.99.8 NetBSD 3.99.8 (build.kaeru.xen.nodebug) #10: Thu Aug 25 20:00:08 JST 2005 takashi@kaeru:/home/takashi/work/kernel/build.kaeru.xen.nodebug i386
Architecture: i386
Machine: i386
>Description:
mclpool assumes that there is always enough kva to allocate items
up to hardlimit.
after "kmapent" changes, it isn't likely true.
>How-To-Repeat:
>Fix:
one of the followings?
1. make PR_LIMITFAIL fail on kva starvation as well.
2. adjust the size of the submap, considering overhead of kmapent.
3. revive kmem_map-backed mapent pool.
>Unformatted: