Coda File System

Memory mapping.

From: Alan Scheinine <scheinin_at_crs4.it>
Date: Thu, 30 Mar 2000 10:22:41 +0200 (METDST)
What memory is available for memory mapped files?
Is it:
   (1) All of the physical memory and swap space?
   (2) Just the physical memory?
   (3) The amount of space [physical [plus swap?]] not
       being used by other processes?

We run numerical simulations that use nearly all of the physical
memory -- and then some.  Does this mean the memory available to
coda is thus limited?

In previous messages the question seems highly specialized and I
am not an expert in this matter.  For example, in a message
  > Subject: Re: rdsinit death
  > Date: Mon, 31 Jan 2000 15:27:32 -0500
it was written:

   > It could be that there is not a big enough chunk of memory available in
   > the memory map around the address that we try to mmap the data segment.

   > Try to figure out where shared libraries and such are mapped. I believe
   > we use 0x20000000 as the base address for RVM data, and it sounds like
   > something else could be in your memory map about 150MB above this point
   > (around 0x25000000). On linux, /proc/<pid>/maps should give all the info
   > about which addresses are already used up.

But there are about 50 files /proc/<pid>/maps.

In a later message there are the comments:
 > > Does this mean I have to pick a custom range for my RVM Data mapping, and
 > > if so, how should I go about this?  The same way that was mentioned
 > > a week or so ago when discussing setting up RVM for a 42 Gig server,
 > > with a start address of 0x41000000 (checking, of course that this is free
 > > by checking  /proc/{pid}/maps)?

 > Yeah, 0x50000000 could be a good choice, is also used by NetBSD 1.3/14
 > and FreeBSD. RVM stores the startaddress in the header of the RVM data
 > file during initialization and will always remap to that address.

I realize that the WWW site of coda emphasizes that the software is
at beta level.  Nonetheless, asking users to study the situation as
described above does not seem to be a reasonable direction for the
development of the software.  I have installed several dozen public
domain packages and of all of them "coda" has given me the most problems.

Alan Scheinine  Email: scheinin_at_crs4.it
Received on 2000-03-30 03:24:46