Re: sending large buffers to wine applications

Rainer Weikusat wrote:

A 'tmpfs' is not 'a RAM disk'. It's a file system whose
contents reside only in the page cache, eventually backed by
swap space in case of memory pressure.

You can easyly prevent a file in tmpfs from being swaped: Open,
mmap and mlock it.

Yes tmpfs is not a RAM disk like the ramdisk driver provides, but
it's close enough to what the OP needs.

Wolfgang Draxinger
--
E-Mail address works, Jabber: hexarith@xxxxxxxxxx, ICQ: 134682867

.

Relevant Pages

  • Re: sending large buffers to wine applications
    ... contents reside only in the page cache, ... swap space in case of memory pressure. ... Who would have guessed that mlock actually works as it is documented ...
    (comp.os.linux.development.apps)
  • Re: is there a user mode way to flush disk cache
    ... I've got plenty of memory. ... > the disk led is on solid. ... > disk cache is nearly empty before I start. ... copies the file from tmpfs to a harddrive. ...
    (comp.os.linux.development.system)
  • Re: how to get a silent harddisk?
    ... logs and other status files are on the tmpfs)? ... > I don't know why it uses the hard disk, but if it is only reading those ... > cache, then perhaps the atime is being updated and that is causing the ... Maybe, but the behavior of squid is always the same, no matter if I put the log files, the disk cache and the rest of squid's stuff in /var on the tmpfs or not. ...
    (Debian-User)
  • RE: open(O_DIRECT) on a tmpfs?
    ... But I'd also expect any app requesting O_DIRECT in that way, ... Try to minimize cache effects of the I/O to and from this file. ... I don't see why tmpfs has to conflict with it. ...
    (Linux-Kernel)
  • Re: mounting /tmp right
    ... /tmp is a tmpfs "ramdisk" ... some other ram-based temporary storage. ... out of /tmp earlier than a system that uses little or no swap space. ...
    (alt.os.linux.suse)