Kernel error: BIOS doesn't leave a aperture memory hole

[wicking]

Hello.

Since I have a new mainboard from Gigabyte (GA-MA69GM-S2H) I always get the following error during boot process (needs 5 to 10 seconds):

Kernel 2.6.24

Checking aperture...
CPU 0: aperture @ fee000000 size 32 MB
Aperture too small (32 MB)
No AGP bridge found
Your BIOS doesn't leave a aperture memory hole
Please enable the IOMMU option in the BIOS setup
This costs you 64 MB of RAM

Kernel 2.6.25

Checking aperture...
Node 0: aperture @ 1ffe000000 size 32 MB
Aperture beyond 4GB. Ignoring.
No AGP bridge found
Your BIOS doesn't leave a aperture memory hole
Please enable the IOMMU option in the BIOS setup
This costs you 64 MB of RAM

Where or how can I activate IOMMU in the BIOS? What is it?
Or where can I change the size of the aperture memory hole?

I’ve already tried Ctrl+F1 in the BIOS to get more options, but I didn’t find any options to change that.. Or is there a possibility with kernel parameters?

[stevea]

You should first update to the latest BIOS revision for your card.
==
This is related to AGP graphics.
Are you using onboard graphics of not ?

You'll want to fuss with the BIOS settings for AGP if you are using a vid-card.
See under "Advanced", the options:
Internal Graphics Mode ("Disable" if no vid card, "Auto" otherwise)
Frame Buffer Size ((if you have vid-card then set this to some fixed size, same as card buffer)).
--

I don't have this card so I can only suggest that you try changing these graphics settings.

[leigh123linux]

You get these error messages with the 64 bit kernel in both F8 & F9 !!

[root@localhost ~]# uname -r 
2.6.25-3.fc9.x86_64
[root@localhost ~]#

output from dmseg

spurious 8259A interrupt: IRQ7.
Console: colour VGA+ 80x25
console [tty0] enabled
Checking aperture...
Node 0: aperture @ 92b0000000 size 32 MB
Aperture beyond 4GB. Ignoring.
No AGP bridge found
Memory: 1025036k/1048512k available (2649k kernel code, 23088k reserved, 1391k data, 356k init)
CPA: page pool initialized 1 of 1 pages preallocated
SLUB: Genslabs=13, HWalign=64, Order=0-1, MinObjects=4, CPUs=2, Nodes=1
Calibrating delay using timer specific routine.. 4019.79 BogoMIPS (lpj=2009897)
Security Framework initialized
SELinux:  Initializing.

[richcsst]

I get the same error and have a PCI Express system! I don't have an AGP slot, nor a way to set anything of the sort in BIOS, yet I get this same error too.

I have F8, 4GB RAM, AMD 64 6600+ X2 and a standard RPM kernel,.

Do I have to compile a custom kernel to stop this stupid error? It seems to realize there is no AGP device, yet it still goes on as if there is? WTF?

Checking aperture...
CPU 0: aperture @ eb40000000 size 32 MB
Aperture too small (32 MB)
No AGP bridge found
Your BIOS doesn't leave a aperture memory hole
Please enable the IOMMU option in the BIOS setup
This costs you 64 MB of RAM
Mapping aperture over 65536 KB of RAM @ c000000

[MSK61]

I have F9, 2GB RAM, AMD 64 4400+ X2 and the latest update kernel(2.6.25.10-86.fc9.x86_64 at the time being). I keep receiving only this message upon boot:

Aperture beyond 4GB. Ignoring.

No other notes or comments about the reason or how to resolve or suppress this message. The boot completes successfully otherwise, with no problems. This message has been appearing since the first kernel that was installed with fedora 9, and remained through all kernel updates released after installation. It didn't exist however with F7 64-bit kernel, with the same hardware configuration.

[wicking]

I have F9, 2GB RAM, AMD 64 4400+ X2 and the latest update kernel(2.6.25.10-86.fc9.x86_64 at the time being). I keep receiving only this message upon boot:

Aperture beyond 4GB. Ignoring.

No other notes or comments about the reason or how to resolve or suppress this message. The boot completes successfully otherwise, with no problems. This message has been appearing since the first kernel that was installed with fedora 9, and remained through all kernel updates released after installation. It didn't exist however with F7 64-bit kernel, with the same hardware configuration.

I still use Fedora 8 and now the error message sounds the same as in your Fedora 9. I didn’t resolve it yet.

[jbkt23]

Add the following to your kernel parameters in grub.conf:

iommu=noaperture

Mine looks like this:

kernel /boot/vmlinuz-2.6.23.17-88.fc7 ro root=LABEL=/1 rhgb quiet iommu
=noaperture

This is an older install that I haven't upgraded yet, but your symptoms are similar to what I experienced.

[wicking]

Add the following to your kernel parameters in grub.conf:

iommu=noaperture

Mine looks like this:

kernel /boot/vmlinuz-2.6.23.17-88.fc7 ro root=LABEL=/1 rhgb quiet iommu
=noaperture

This is an older install that I haven't upgraded yet, but your symptoms are similar to what I experienced.

That was the solution. Thanks.

[MSK61]

Thanks, jbkt23. Worked like a charm. I've googled about this error for ages with no satisfactory result. Wishing one day I may similarly find a solution for the other kernel problem 'hub 1-0:1.0: unable to enumerate USB device on port 10'.

[chopperpilot99]

Can somebody tell me exactly which line to change on mine to fix this issue? Thanks.



# grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this file
# NOTICE: You have a /boot partition. This means that
# all kernel and initrd paths are relative to /boot/, eg.
# root (hd0,0)
# kernel /vmlinuz-version ro root=/dev/VolGroup00/LogVol00
# initrd /initrd-version.img
#boot=/dev/sda
default=0
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title Fedora (2.6.26.3-29.fc9.x86_64)
root (hd0,0)
kernel /vmlinuz-2.6.26.3-29.fc9.x86_64 ro root=/dev/VolGroup00/LogVol00 rhgb quiet
initrd /initrd-2.6.26.3-29.fc9.x86_64.img
title Fedora (2.6.25-14.fc9.x86_64)
root (hd0,0)
kernel /vmlinuz-2.6.25-14.fc9.x86_64 ro root=UUID=4d7d5dac-be7c-435f-8c54-32d07933340c rhgb quiet
initrd /initrd-2.6.25-14.fc9.x86_64.img

[jbkt23]

Kernel parameters are always added to the line that begins with the word "kernel". In my post above I added the parameter "iommu=noaperture" at the end of the line after the word "quiet". Parameters are separated by spaces.

[jbkt23]

Also, as it isn't clear whether you are familiar with script files the lines beginning with # are commented out which means they are not read by the grub interpreter. Lines are continuous, meaning don't hit enter at the end of your editor screen to make it wrap sooner.

I recommend that you add a # before the word "hiddenmenu", so that you have a choice of booting to the other kernel. To be safe only edit one of the lines that begins with the word "kernel", so if your edit causes it to hang you can boot to the other kernel.

I would also search the forum with the terms "booting from the rescue disk" or "booting from install disk".

Before you edit grub.conf make a copy called grub.conf.orig so if you hose it you can get back to where you were.

[chopperpilot99]

Thanks, worked great!

[JONOR]

Have you noticed the problem return in the last week Wicking ?
Since about a week ago - Sunday 22 in Mar 009 (in Britain) i have been getting this in my log messages :

(typed by hand as i haven't worked out to copy/paste messages yet so there might be minor typos!)
------------------------------------
Checking aperture...
No AGP bridge found
Node 0: aperture @ 2000000 size 32MB
Aperture pointing to e820 RAM . Ignoring.
Your BIOS doesn't leave a memory hole
Please enable the iommu option in the BIOS
This costs you 64MB of RAM
------------------------------------

Following the mentioned thread advise solves this problem but makes another.
On booting with iommu=noaperture in /etc/grub.conf on the relevent two kernal lines on my Gigabyte GA-MA78G-DS3H board (which has on-board graphics)
i miss the above messages but get a two minutes of beep starting at the log in banner. I am not quite sure what log messages to post on this yet.

[Ashish Sood]

thanks jbkt23 for a solution