Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 8 Oct 2010 16:12:10 -0700
From:      Steve Kargl <sgk@troutmask.apl.washington.edu>
To:        Kostik Belousov <kostikbel@gmail.com>
Cc:        freebsd-amd64@freebsd.org
Subject:   Re: linux32 sysctl limits?
Message-ID:  <20101008231210.GA90183@troutmask.apl.washington.edu>
In-Reply-To: <20101008222013.GT2392@deviant.kiev.zoral.com.ua>
References:  <20101008221701.GA89977@troutmask.apl.washington.edu> <20101008222013.GT2392@deviant.kiev.zoral.com.ua>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Oct 09, 2010 at 01:20:13AM +0300, Kostik Belousov wrote:
> On Fri, Oct 08, 2010 at 03:17:01PM -0700, Steve Kargl wrote:
> > 'man linux' does not list or describe the following sysctl
> > variables.
> > 
> > compat.linux32.maxvmem: 0
> > compat.linux32.maxssiz: 67108864
> > compat.linux32.maxdsiz: 536870912
> > 
> > Are there any limitations with setting these to
> > say 1G for stack and 4G for data on a system with
> > 16G of memory?
> 
> The linux ABI processes are 32bit, so you cannot reallistically assign
> 4Gb to maxdsiz. The space between data (bss) end and stack bottom is
> used for mappings, so it should be kept large enough.

Does this mean that maxssiz+maxdsiz=3G would allow 
1G for mappings?  The only linux app that will be
running is matlab (and I suppose any children processes
of matlab).

-- 
Steve



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20101008231210.GA90183>