xfs
[Top] [All Lists]

Re: mmap() Limitation

To: Steve Lord <lord@xxxxxxx>
Subject: Re: mmap() Limitation
From: Andi Kleen <ak@xxxxxxx>
Date: Wed, 5 Dec 2001 20:49:17 +0100
Cc: "Quang Nguyen (Ngo)" <quang.nguyen@xxxxxxxxxxxx>, "'linux-xfs@xxxxxxxxxxx'" <linux-xfs@xxxxxxxxxxx>
In-reply-to: <1007579975.22679.28.camel@xxxxxxxxxxxxxxxxxxxx>
References: <ACD4093EB009D411BC8A009027D7699660A15E@xxxxxxxxxxxxxxxxx> <1007579975.22679.28.camel@xxxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
User-agent: Mutt/1.3.22.1i
> There does not appear to be a user space interface to mmap out beyond
> a 4G boundary on ia32 linux. The mmap64 call is from irix. This looks
> like a deficiency in the linux large file support to me.

There is actually. In kernel it is called mmap2(), in glibc it should
be mmap64() or even default when compiled for largefiles. mmap2 does not
pass in a true 64bit offset, but an 32bit scaled by PAGE_SIZE, but that
works well enough as the page cache cannot represent bigger files anyways
(that is the internal interface between glibc and kernel; the user passes an 
byte 
offset to mmap64 of course) 


-andi


<Prev in Thread] Current Thread [Next in Thread>