On Mon, 2004-02-09 at 01:40, Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.3-rc1/2.6.3-rc1-mm1/
> 
> 

> 
> +highmem-equals-user-friendliness.patch
> 
>  Enhance and document the `highmem=' ia32 kernel boot option.  This also
>  gives us highmem emulation on <= 896M boxes.
> 


I'm having problems with this patch on our OSDL testing machines.  It
seems to work OK with CONFIG_NOHIGHMEM, but I get a BUG in ACPI at boot
when I have CONFIG_HIGHMEM4G set.  

ACPI: Subsystem revision 20040116
------------[ cut here ]------------
kernel BUG at arch/i386/mm/pageattr.c:110!
invalid operand: 0000 [#1]
PREEMPT SMP
CPU:    0
EIP:    0060:[<c0120434>]    Not tainted VLI
EFLAGS: 00010046
EIP is at __change_page_attr+0x14/0x220
eax: c10a0000   ebx: 00000001   ecx: 00000163   edx: 00000163
esi: c10a0000   edi: 00000000   ebp: c3fdfe80   esp: c3fdfe6c
ds: 007b   es: 007b   ss: 0068
Process swapper (pid: 1, threadinfo=c3fde000 task=c3fcf800)
Stack: c1002828 c3fff000 00000001 c10a0000 00000000 c3fdfea8 c01206ea c10a0000
       00000163 00000246 c3ff6e10 c3ff6e10 c4800ac6 c3ff6e10 c3fdff18 c3fdfec4
       c012014c c109ffd8 00000002 00000163 c3fdfef4 c4800aea c3fdfee4 c024018a
Call Trace:
 [<c01206ea>] change_page_attr+0xaa/0xe0
 [<c012014c>] iounmap+0x7c/0xa0
 [<c024018a>] acpi_tb_get_table_header+0xbb/0xbf
 [<c02409dd>] acpi_tb_get_primary_table+0x31/0x72
 [<c0240ec3>] acpi_tb_init_table_descriptor+0xa9/0x117
 [<c0240c20>] acpi_tb_get_required_tables+0x169/0x1b4
 [<c02426d6>] acpi_ut_release_mutex+0x59/0x74
 [<c0240768>] acpi_load_tables+0xd0/0x12a
 [<c051da82>] acpi_bus_init+0x33/0x111
 [<c051dbb9>] acpi_init+0x59/0xb4
 [<c050a96b>] do_initcalls+0x2b/0xa0
 [<c010712a>] init+0x5a/0x1e0
 [<c01070d0>] init+0x0/0x1e0
 [<c010a2d5>] kernel_thread_helper+0x5/0x10
                                                                                
Code: 00 a4 a8 40 c0 e9 40 ff ff ff 8d b6 00 00 00 00 8d bc 27 00 00 00 00 55 89 e5 57 56 53 83 ec 08 8b 45 08 3b 05 94 36 55 c0 72 08 <0f> 0b 6e 00 7b 24 41 c0 8b 45 08 50 e8 7b 9b 03 00 89 c2 c1 ea
 <0>Kernel panic: Attempted to kill init!


The system hangs with no output at all if I have CONFIG_HIGHMEM64G set.

If I back this patch out everything works fine with all memory
configurations.

Here is my config for 64GB memory configuration.  I have 16GB in the
box.


-- 
Mark Haverkamp <markh@osdl.org>