The dmidecode project provides the means to learn exactly what claims your BIOS is making about your hardware. Strange as it might seem, it’s useful information, even when it’s not 100% reliable.
The dmidecode project provides the means to learn exactly what claims your BIOS is making about your hardware. Strange as it might seem, it’s useful information, even when it’s not 100% reliable.
I remember installing a 30GB harddisk in a 33Mhz 486 PC (12 MB RAM). Of course, DOS would never see 30GB due to BIOS limitations. But Slackware 4.0 ran fine
dmidecode could be useful to test pc’s for other operating systems that are more dependant on the BIOS.
“Of course, DOS would never see 30GB due to BIOS limitations.”
That’s more than just a BIOS limitation, it’s also hardcoded into MSDOS: the 504mb drive size limit.
Correct: old dos versions have a 504MB drive limit. But I always use the DOS 7.x versions (from bootable dos floppies created under Windows 98) and those newer DOS versions have no such limits. However, they still suffer from the BIOS limit.
http://www.bebits.com/app/4022
Just ported dmidecode to BeOS.
Didn’t have much to change anyway.
Opps miss read that as ‘old days’.
I almost posted a long tirate about paying through the nose for my first HD, at 30MB, and being glad that FAT12 had the forsight to support even much larger drives (all the way up to 32MB)! Wow those were the good old days.
Here at my workplace there are ~100 Fedora Core linux boxes used for grid computing; the rebuild process uses kickstart to install the OS and then yum pointed at a local repository to ensure that all the machines are consistent. We’ve created an RPM package which queries DMI to get the motherboard serial #, which then gets matched up to a hostname and IP address. On initial boot, systems come up with a dynamic address, but then reboot themselves and right away get to work without my having to visit each individual machine.
and to think I had a 10 MB HardCard…. whatever happened to those nice things? Can you just see HDs mounted on their own PCI controllers nowadays? Maybe dual slotters with fans in another PCI slot to keep em cool… ahh the good ol days…
I wonder if this was a similar problem. i tried to install linux for teh first time, and it ran very very very slowly. I eventually found it was only using 16MB of ram. I had 128 MB of ram on that machine.
I managed to track the issue down to an option in my bios that enabled a memory hole (15-16MB). I think it was needed for my sound card…or was it the modem.
Anyways. Windows 2k had no problem with detecting the memory as 128 MB.
I managed to force linux (red hat) to detect the memory as 128MB with some install option (-mem 128) or something like that. Was this likely a problem with bios reporting?
Yamin
On the IBM X440 and X445 servers the dmidecode output thinks its a Toshiba Laptop.
I wonder what a Toshiba laptop is reported as?
dmidecode blacklisting is useful for distro vendors, too. Whenever people report a machine with idiosyncratic hardware issues on mandrake (with stuff like ACPI or APIC) the standard request is to give the dmidecode output, so the motherboard can be blacklisted not to use the feature it has trouble with…
I’d never heard of this (/me hides). Nice!