Index
> OS Construction > Testing E820 ACPI 3.0 compatibility |
| Author |
|
|
smiddy 21 Nov 2006, 11:55
Hello,
You may or may not know that the latest ACPI specification defines INT 15h AX=E820h with extended attributes. I've yet to encounter such a machine. I am posting my code in hopes that someone here may have a machine that does. Please read the source files for instructions on usage. When you post your results please use this format: MB Make, Model, and CPU: MSI 7A266 Athlon 3000+ BIOS make and date: AMI 7/01/2003
|
|||||||||||
21 Nov 2006, 11:55 |
|
|
smiddy 21 Nov 2006, 18:51
Thanks LocoDelAssembly! It is interesting that your machine being as new as it is doesn't support the extended attributes.
It looks like you EBDA is bigger than mine too. WOW, and some of the records are out of sequence, perhaps I need to put a sorting algorithm in there, eh? |
|||
|
21 Nov 2006, 18:51 |
|
|
arafel 21 Nov 2006, 18:56
ASUSTek TUV4X, PIII EB 933MHz
Award 07/04/2002 Code: E820 - 1.00.1 Attempt to see E820 information... -smiddy E820: hSz Base Address - Memory Length - Memory Type - ExtAttr 14 : 0000000000000000 - 000000000009FC00h - Available to OS 14 : 000000000009FC00 - 0000000000000400h - Reserved Memory 14 : 00000000000F0000 - 0000000000010000h - Reserved Memory 14 : 0000000000100000 - 000000003FEEC000h - Available to OS 14 : 000000003FFEC000 - 0000000000003000h - ACPI Reclaim Memory 14 : 000000003FFEF000 - 0000000000010000h - Reserved Memory 14 : 000000003FFFF000 - 0000000000001000h - ACPI NVS Memory 14 : 00000000FFFF0000 - 0000000000010000h - Reserved Memory ------------------------------------------------------------------------------- : Total Memory : 3FF8BC00h - 1,073,265,664 bytes. ------------------------------------------------------------------------------- ASRock 775Dual-880Pro, P4 531 3.0GHz AMI 06/16/2006 Code: E820 - 1.00.1 Attempt to see E820 information... -smiddy E820: hSz Base Address - Memory Length - Memory Type - ExtAttr 14 : 0000000000000000 - 000000000009FC00h - Available to OS 14 : 000000000009FC00 - 0000000000000400h - Reserved Memory 14 : 00000000000E6000 - 000000000001A000h - Reserved Memory 14 : 0000000000100000 - 000000003FEB0000h - Available to OS 14 : 000000003FFB0000 - 0000000000010000h - ACPI Reclaim Memory 14 : 000000003FFC0000 - 0000000000030000h - ACPI NVS Memory 14 : 000000003FFF0000 - 0000000000010000h - Reserved Memory 14 : 00000000FEE00000 - 0000000000001000h - Reserved Memory 14 : 00000000FF780000 - 0000000000880000h - Reserved Memory ------------------------------------------------------------------------------- : Total Memory : 3FF4FC00h - 1,073,019,904 bytes. ------------------------------------------------------------------------------- |
|||
|
21 Nov 2006, 18:56 |
|
|
LocoDelAssembly 21 Nov 2006, 19:07
Quote: WOW, and some of the records are out of sequence, perhaps I need to put a sorting algorithm in there, eh? Hey you're right, mmmm, then if I continue with http://board.flatassembler.net/topic.php?p=43144#43144 some day now I know that I don't need to insert the records in sequence which certanly it's a much easy task  |
|||
|
21 Nov 2006, 19:07 |
|
|
smiddy 21 Nov 2006, 19:08
Thanks arafel! Your AMI BIOS is very surprising too.
|
|||
|
21 Nov 2006, 19:08 |
|
|
smiddy 21 Nov 2006, 19:51
LocoDelAssembly wrote:
I breifly read through the thread and that is interesting information. Thanks! I must have been elsewhere when that was posted. I suspect that the earlier version are using a combination of ways to determine RAM, like a probe or something else. Where XP is exclusively using E820. Interesting! I have developed a memory determination routine for my own OS that uses CMOS, INT 12h, INT 15h AX=E820, AX=E801, and AH=88, SMBIOS, PnP, and Probing. Both PnP and E820 return a map of sorts and should be compatible, but most times they are not. Oddly too, PnP results work under V86 mode so you can find out what RAM is there via a dos box in windows (at least in XP). |
|||
|
21 Nov 2006, 19:51 |
|
|
LocoDelAssembly 21 Nov 2006, 20:26
Thanks for the tip, I'll try this weekend to make some time to work on this.
|
|||
|
21 Nov 2006, 20:26 |
|
|
f0dder 22 Nov 2006, 00:30
ASUS A8N SLI Premium ACPI BIOS Revision 1009
10/21/2005-NF-CK804-A8NSLI-P-00 AMD64x2 4400+, 2048MB DDR400 RAM, 4x512MB blocks Code: E820 - 1.00.1 Attempt to see E820 information... -smiddy E820: hSz Base Address - Memory Length - Memory Type - ExtAttr 14 : 0000000000000000 - 000000000009D000h - Available to OS 14 : 00000000000F0000 - 0000000000010000h - Reserved Memory 14 : 00000000FEC00000 - 0000000001400000h - Reserved Memory 14 : 00000000E0000000 - 0000000010000000h - Reserved Memory 14 : 000000007FFF3000 - 000000000000D000h - ACPI Reclaim Memory 14 : 000000007FFF0000 - 0000000000003000h - ACPI NVS Memory 14 : 000000000009D000 - 0000000000003000h - Reserved Memory 14 : 0000000000100000 - 000000007FEF0000h - Available to OS ------------------------------------------------------------------------------- : Total Memory : 7FF8D000h - 2,147,012,608 bytes. ------------------------------------------------------------------------------- |
|||
|
22 Nov 2006, 00:30 |
|
|
smiddy 22 Nov 2006, 02:39
LocoDelAssembly wrote: Thanks for the tip, I'll try this weekend to make some time to work on this. When I get the chance I'll PM you the PnP source I have, which should provide a decent starting point. |
|||
|
22 Nov 2006, 02:39 |
|
|
smiddy 22 Nov 2006, 02:43
f0dder, excellent post. I am bemiffed by the randomness of the output. I have expected BIOS' with your date to be capable of 18h bytes too, but there ya go!
|
|||
|
22 Nov 2006, 02:43 |
|
|
f0dder 22 Nov 2006, 15:04
Well, I guess "reserved memory" depends on your chipset - memory mapped devices and such? Weird that these don't get mapped higher, though.
|
|||
|
22 Nov 2006, 15:04 |
|
|
vid 22 Nov 2006, 15:44
should this work under windoze, or just pure DOS?
|
|||
|
22 Nov 2006, 15:44 |
|
|
f0dder 22 Nov 2006, 15:53
Pure DOS, the E820 stuff isn't emulated by NTVDM.
|
|||
|
22 Nov 2006, 15:53 |
|
|
smiddy 22 Nov 2006, 16:38
f0dder wrote: Well, I guess "reserved memory" depends on your chipset - memory mapped devices and such? Weird that these don't get mapped higher, though. I think part of the lower memory stuff is due to compatibility with older software. |
|||
|
22 Nov 2006, 16:38 |
|
|
smiddy 22 Nov 2006, 16:42
vid wrote: should this work under windoze, or just pure DOS? f0dder wrote: Pure DOS, the E820 stuff isn't emulated by NTVDM. Yep, pure DOS, or boot it from a floppy. One person I think is trying it from a USB drive. Perhas I should include the code for Plug & Play interface so there is a comparison. Also Plug & Play interface is compatible with the Dos Box within Windows (at least in XP). |
|||
|
22 Nov 2006, 16:42 |
|
< Last Thread | Next Thread > |
Forum Rules:
|