Dave Jones <[EMAIL PROTECTED]> writes: > On Mon, Feb 18, 2002 at 12:12:22PM -0700, Eric W. Biederman wrote: > > > Actually there is a dmi BIOS table that on machines in the last 3 or 4 > > years reliably reports the mothberboard name. I hadn't thought of > > that but that is definentily information we can use. > > It's ok in most cases, but I wouldn't call it reliable.
You are right reliable is a little strong but probably reliable enough for our purposes... That is it when it reports the information with other than a default value the information is correct. So using the non default values to trigger extra functionally probably works. > Eg: > Handle 0x0001 > DMI type 1, 8 bytes. > System Information Block > Vendor: > Product: > Version: > Serial Number: > Handle 0x0002 > DMI type 2, 8 bytes. > Board Information Block > Vendor: > Product: > Version: > Serial Number: > > Seems quite a few motherboard vendors can't be bothered filling > in the fields, and just leave them in the state defined in the > code from Award/AMI/whoever. I've also seen examples where all > the fields are xxxxxx 0123456789 etc.. Agreed. In my limited skimming I haven't seen the board id come back wrong. But I have seen a lot of the other fields come back nonsense. My favorite was the report of sdram in a system equipped with rimms. > On newer systems some of the ACPI tables also contain board > specific ID's iirc. In the RSDT on this Athlon board for eg, > it contains "ASUS K7V-RM". I've not seen any of these that > have been incorrect yet, but I'll not put faith into a BIOS > vendor not screwing up something this simple. > (This table is easily parsable btw, no need for AML interpretors > and other ACPI evils) Which table is easily parsable the newer ACPI stuff, or the DMI table? Eric
