LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Boaz Harrosh <bharrosh@panasas.com>
To: Andi Kleen <ak@suse.de>
Cc: Christoph Hellwig <hch@infradead.org>,
	schwidefsky@de.ibm.com, akpm@linux-foundation.org,
	linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>
Subject: Re: [patch 2/3] CONFIG_HIGHPTE vs. sub-page page tables.
Date: Thu, 03 Jan 2008 16:01:15 +0200	[thread overview]
Message-ID: <477CEAAB.4030200@panasas.com> (raw)
In-Reply-To: <200801031412.26806.ak@suse.de>

On Thu, Jan 03 2008 at 15:12 +0200, Andi Kleen <ak@suse.de> wrote:
>> Can we please just nuke CONFIG_HIGHPTE?  There's only been a small
>> amount of 32bit machines 
> 
> It's unfortunately a larger amount :/ And for unknown reasons a lot of
> people still install 32bit kernels on new perfectly capable 64bit systems
> even if they have a lot of memory.
> 
Yes I've seen that too many times. When I comment about it, people say:
"What this 'core 2' supports AMD64 ??!!"

I think this is the distros fault.
1. Most call it AMD64 and not, i don't know what it should be called.

2. They put the first and default installation as i386 and not the x64,
   the later is only farther down or on a different link.

3. My Fedora x86_64 installation has a full blown 32bit installation inside.
   In fact, every time I yum install this or that, I get installed by default
   both the 64bit and 32bit libraries. I don't even know how to turn the 32bit
   off. This means that Distros can supply one x86 installation for both
   flavors and decide at setup time what to install. By this, follow the Kernel
   in observing that it is the same-ARCH same-INSTALL.

4. Some binary modules like Flash, media-codecs acrobat-reader etc... think
   (That snow ball effect here) that 32bit is much more common and only provide
   that. So people think that for best compatibility they should stick with 
   32bit. But this is not true. They all work perfectly here in 64bit land.
   Distros should both confirm on that publicly. And also make extensive tests 
   for 32bit compatibility on 64bit machines, even for these binary only bad
   guys.

5. Distros should jump on the 64bit bang-wagon, to stand it-self apart from
   Window 32bit-ness. It has bin proven more than once that a Huge Linux 64bit 
   machine, with lots of CPUs and memory, can be the best Windows 32bit performer 
   in existence. Both under KVM/VMWARE or Wine. 64bit Vista can never get close,
   not to speak of the "none-existent" 64bit XP.

Just my $0.02
Boaz

> I don't think removing CONFIG_HIGHPTE will be an option any time soon.
> 
> -Andi
> --


  reply	other threads:[~2008-01-03 14:02 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-11-12 14:30 [patch 0/3] page table changes schwidefsky
2007-11-12 14:30 ` [patch 1/3] add mm argument to pte/pmd/pud/pgd_free schwidefsky
2007-11-12 14:30 ` [patch 2/3] CONFIG_HIGHPTE vs. sub-page page tables schwidefsky
2008-01-02 20:44   ` Christoph Hellwig
2008-01-02 21:24     ` Geert Uytterhoeven
2008-01-02 21:28       ` Benjamin Herrenschmidt
2008-01-03 13:12     ` Andi Kleen
2008-01-03 14:01       ` Boaz Harrosh [this message]
2008-02-01 23:15   ` Andrew Morton
2008-02-03  5:37     ` Benjamin Herrenschmidt
2008-02-03  5:53       ` Andrew Morton
2008-02-03  6:46         ` Ingo Molnar
2008-02-04 10:36         ` Martin Schwidefsky
2008-02-04 10:51           ` Andrew Morton
2008-02-04 11:02             ` Russell King
2008-02-04 11:14               ` Andrew Morton
2008-02-05 14:39             ` Martin Schwidefsky
2008-02-05 18:46               ` Andrew Morton
2008-02-06  9:06                 ` Martin Schwidefsky
2008-02-06  9:09                   ` Andrew Morton
2008-02-06  9:15                     ` Ingo Molnar
2008-02-06 15:50                     ` Martin Schwidefsky
2007-11-12 14:30 ` [patch 3/3] arch_rebalance_pgtables call schwidefsky
2007-11-13 12:33   ` Nick Piggin
2007-11-14  9:26     ` Martin Schwidefsky
2007-11-14 10:06       ` Benjamin Herrenschmidt
2007-11-14 11:49         ` Martin Schwidefsky
2007-11-14 22:07           ` Benjamin Herrenschmidt
2007-11-15 17:13             ` Martin Schwidefsky

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=477CEAAB.4030200@panasas.com \
    --to=bharrosh@panasas.com \
    --cc=ak@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=benh@kernel.crashing.org \
    --cc=hch@infradead.org \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=schwidefsky@de.ibm.com \
    --subject='Re: [patch 2/3] CONFIG_HIGHPTE vs. sub-page page tables.' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).