* more than 3 GB in userspace (4G/4G patch?) for 2.6.16
@ 2006-06-13 20:30 Hanno Mueller
2006-06-13 21:30 ` H. Peter Anvin
0 siblings, 1 reply; 3+ messages in thread
From: Hanno Mueller @ 2006-06-13 20:30 UTC (permalink / raw)
To: linux-kernel
Hi,
on my 8 GB system with two 32-bit-xeons, I would like to have more than
3 gigs in userspace.
A 0.5/3.5 GB split appears to be what I need for my application.
I read about the 4G/4G patch from 2004 but couldn't find a current patch
for 2.6.16 - is there anything similar to it that I can apply on a
2.6.16 kernel?
Thanks a lot & regards,
Hanno
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: more than 3 GB in userspace (4G/4G patch?) for 2.6.16
2006-06-13 20:30 more than 3 GB in userspace (4G/4G patch?) for 2.6.16 Hanno Mueller
@ 2006-06-13 21:30 ` H. Peter Anvin
2006-06-13 21:37 ` Dave Hansen
0 siblings, 1 reply; 3+ messages in thread
From: H. Peter Anvin @ 2006-06-13 21:30 UTC (permalink / raw)
To: linux-kernel
Followup to: <448F207A.6080601@hanno.de>
By author: Hanno Mueller <sockpuppet@hanno.de>
In newsgroup: linux.dev.kernel
>
> on my 8 GB system with two 32-bit-xeons, I would like to have more than
> 3 gigs in userspace.
>
See if you can replace them with 64-bit Xeons. That would be the best option.
> A 0.5/3.5 GB split appears to be what I need for my application.
That is incompatible with PAE. The way PAE works, the split has to be
an even number of gigabytes.
-hpa
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: more than 3 GB in userspace (4G/4G patch?) for 2.6.16
2006-06-13 21:30 ` H. Peter Anvin
@ 2006-06-13 21:37 ` Dave Hansen
0 siblings, 0 replies; 3+ messages in thread
From: Dave Hansen @ 2006-06-13 21:37 UTC (permalink / raw)
To: H. Peter Anvin; +Cc: linux-kernel
On Tue, 2006-06-13 at 14:30 -0700, H. Peter Anvin wrote:
> > A 0.5/3.5 GB split appears to be what I need for my application.
>
> That is incompatible with PAE. The way PAE works, the split has to be
> an even number of gigabytes.
Well, there have been ancient patches available in the past. People pop
up now and again and ask for them. This is _horribly_ out of date, but
it at least indicates that this is possible.
The basic idea is that you never share a kernel PMD page. Each process
needs to allocate four of its own:
http://lkml.org/lkml/2003/5/9/138
-- Dave
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2006-06-13 21:38 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-06-13 20:30 more than 3 GB in userspace (4G/4G patch?) for 2.6.16 Hanno Mueller
2006-06-13 21:30 ` H. Peter Anvin
2006-06-13 21:37 ` Dave Hansen
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox