linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: "Laurent Lagrange" <lagrange@fr.oleane.com>
To: <linuxppc-embedded@ozlabs.org>
Subject: PPC 32bits and big RAM mapping problem
Date: Thu, 1 Dec 2005 18:47:46 +0100	[thread overview]
Message-ID: <000001c5f69f$580e6ae0$5201a8c0@GEG2400> (raw)
In-Reply-To: 


Hello,

I have a PPC board with an address width fixed to 32bits, so I can address
4GB.

   M          M        M         M          M          M           M
  0          0        0         0          0          1           e
 0          2        8         c          e          e           f
 ---------------------------------------------------------------------------
---
 | Physical |        | PCI     | Kernel   | Kernel   | User      |
Peripheral |
 | RAM      | unused | memory  | memory   | reserved | memory    | mapping
|
 |          |        |         | kmalloc  |          | vmalloc   |
|
 | ex 512MB |        | ex 1GB  | ex 512MB | 16MB     |           | ex 32MB
|
 |      2GB max      | 1GB max | max ?    |          | 464MB max |
|
 ---------------------------------------------------------------------------
---
                              (0)                   (1)         (2)

00M stand for 0x00000000
(0) kernel_base
(1) vmalloc_start
(2) vmalloc_end

The exemple above shows a classical(?) mapping with a physical RAM of 512MB.
The physical RAM can growth up to 2GB, but if the kernel maps a Kernel
memory
of 2GB, there is no more place for the other items. The 4GB are eated.

Even with only 512MB of RAM, the kernel can't fully vmalloc it for user.

I know that the kernel and drivers kmalloc in Kernel memory which has the
full
space but when the users wants more RAM, it is for vmallocations.
Now when the physical RAM increases, the User memory is reduced for vmalloc.
And I don't talk about Peripheral mappings which also reduce the User
memory.

So I'm under the impression to be cornered in my shoes.
Any idea, book, article, prediction would be welcome.

Thanks
Laurent

             reply	other threads:[~2005-12-01 17:54 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-12-01 17:47 Laurent Lagrange [this message]
  -- strict thread matches above, loose matches on Subject: below --
2005-12-01 18:43 PPC 32bits and big RAM mapping problem Rune Torgersen
2005-12-01 18:56 ` Matt Porter
2005-12-01 20:18 Rune Torgersen
2005-12-06 19:55 Rune Torgersen

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='000001c5f69f$580e6ae0$5201a8c0@GEG2400' \
    --to=lagrange@fr.oleane.com \
    --cc=linuxppc-embedded@ozlabs.org \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).