Syslinux - May 2005 - "Failed to free base memory" once again

Hi,

I got the following output booting through PXElinux. I used a netboot
floppy image for my ISA NE2000 (RTL8019) network card to download the
pxelinux.0 file from my server. I use syslinux 3.07. Here's my output:

PXELINUX 3.07 2005-01-12  Copyright (C) 1994-2005 H. Peter Anvin
########################################################
# Welcome to the Petteflet PXE booting infrastructure. #
########################################################
Loading aagje......................Ready.
Failed to free base memory, error FF01-03FF-F000FE6E
Linux version 2.6.11-gentoo-r6 (root at pluk) (gcc version 3.3.5-20050130
(Gentoo Linux 3.3.5.20050130-r1, ssp-3.3.5.20050130-1, pie-8.7.7.1)) #2
Sun May 1 13:30:57 CEST 2005
BIOS-provided physical RAM map:
 BIOS-88: 0000000000000000 - 000000000009f000 (usable)
 BIOS-88: 0000000000100000 - 0000000002800000 (usable)
40MB LOWMEM available.
DMI not present.
Allocating PCI resources starting at 02800000 (gap: 02800000:fd800000)
Built 1 zonelists
Kernel command line: BOOT_IMAGE=aagje console=ttyS0,115200n8 vga=-2
netdev=10,0x300 root=/dev/nfs rw
Initializing CPU#0
PID hash table entries: 256 (order: 8, 4096 bytes)
Using pit for high-res timesource
Console: colour VGA+ 80x50
Dentry cache hash table entries: 8192 (order: 3, 32768 bytes)
Inode-cache hash table entries: 4096 (order: 2, 16384 bytes)
Bad page state at free_hot_cold_page (in process 'swapper', page
c1000180)
flags:0x00000000 mapping:00000000 mapcount:0 count:0
Backtrace:
 [<c0102877>]
 [<c012bff4>]
 [<c012c52b>]
 [<c032a168>]
 [<c032a2cd>]
 [<c03288db>]
 [<c032071a>]
 [<c010019f>]
Trying to fix it up, but a reboot is needed
Unknown interrupt or fault at EIP 00000286 00000060 c012c22a
Memory: 37616k/40960k available (1759k kernel code, 2908k reserved, 407k
data, 128k init, 0k highmem)
Checking if this processor honours the WP bit even in supervisor mode... Ok.
SLAB: cache with size 1024 has lost its name
SLAB: cache with size 2048 has lost its name

And then the system hangs. It is a 486DX2-S with 40MB memory and no
harddisk... Aagje is the name of the kernel to load. My
pxelinux.cfg/default config file:

SERIAL  0 115200 0x303
CONSOLE 0

SAY     ########################################################
SAY     # Welcome to the Petteflet PXE booting infrastructure. #
SAY     ########################################################

DEFAULT aagje console=ttyS0,115200n8 vga=-2 netdev=10,0x300 root=/dev/nfs rw

I've tried numerous options for my kernel, all with the same results.

Scanning through the list archives, I found no precise detailed output
and requests for precise debug output, not using VMware. So is this
enough? Do you require more to fix this? Thanks in advance for your time
and effort. If you want, you can have shell access to the serial console
 of the system and write access to the config file and images...

Kind regards,

Simon de Hartog

Simon de Hartog wrote:
> Hi,
> 
> I got the following output booting through PXElinux. I used a netboot
> floppy image for my ISA NE2000 (RTL8019) network card to download the
> pxelinux.0 file from my server. I use syslinux 3.07. Here's my output:
> 
> PXELINUX 3.07 2005-01-12  Copyright (C) 1994-2005 H. Peter Anvin
> ########################################################
> # Welcome to the Petteflet PXE booting infrastructure. #
> ########################################################
> Loading aagje......................Ready.
> Failed to free base memory, error FF01-03FF-F000FE6E
> Linux version 2.6.11-gentoo-r6 (root at pluk) (gcc version 3.3.5-20050130
> (Gentoo Linux 3.3.5.20050130-r1, ssp-3.3.5.20050130-1, pie-8.7.7.1)) #2
> Sun May 1 13:30:57 CEST 2005
> BIOS-provided physical RAM map:
>  BIOS-88: 0000000000000000 - 000000000009f000 (usable)
>  BIOS-88: 0000000000100000 - 0000000002800000 (usable)
> 40MB LOWMEM available.
> DMI not present.
> Allocating PCI resources starting at 02800000 (gap: 02800000:fd800000)
> Built 1 zonelists
> Kernel command line: BOOT_IMAGE=aagje console=ttyS0,115200n8 vga=-2
> netdev=10,0x300 root=/dev/nfs rw
> Initializing CPU#0
> PID hash table entries: 256 (order: 8, 4096 bytes)
> Using pit for high-res timesource
> Console: colour VGA+ 80x50
> Dentry cache hash table entries: 8192 (order: 3, 32768 bytes)
> Inode-cache hash table entries: 4096 (order: 2, 16384 bytes)
> Bad page state at free_hot_cold_page (in process 'swapper', page
c1000180)
> flags:0x00000000 mapping:00000000 mapcount:0 count:0
> Backtrace:
>  [<c0102877>]
>  [<c012bff4>]
>  [<c012c52b>]
>  [<c032a168>]
>  [<c032a2cd>]
>  [<c03288db>]
>  [<c032071a>]
>  [<c010019f>]
> Trying to fix it up, but a reboot is needed
> Unknown interrupt or fault at EIP 00000286 00000060 c012c22a
> Memory: 37616k/40960k available (1759k kernel code, 2908k reserved, 407k
> data, 128k init, 0k highmem)
> Checking if this processor honours the WP bit even in supervisor mode...
Ok.
> SLAB: cache with size 1024 has lost its name
> SLAB: cache with size 2048 has lost its name
> 
> And then the system hangs. It is a 486DX2-S with 40MB memory and no
> harddisk... Aagje is the name of the kernel to load. My
> pxelinux.cfg/default config file:
> 
The system hang has nothing to do with "failed to free base memory", 
that error message just means you lost about 100K of memory.  For 
booting Linux, that really shouldn't make much difference (actually, in 
your case, Linux should have cleaned that up anyway.)

More likely, you're having a kernel problem independent of that.

-> One thing that does come out odd is that your BIOS-88 report and your 
"failed to free base memory" report don't match.

Try using Etherboot instead of netboot, just to try something different, 
I guess.

	-hpa