BIOS Virus – 2775 days old

[Sinnercure]

My laptop hdd failed a while back in a bad way. It spontaneously set itself a hard disk platter-lock one night, leaving the drive unmountable, as I had no idea what sort of password might have been put on there.

I just got my new disk today, and before I install it, I'd like to scan my bios with a virus-scanner or else flash the bios to the latest and greatest for this chip. Problem is, I can't get the UBCD to boot up without a disk in there to run a virus scan. It gets to a line:
No disks to use; driver NOT loaded!
Kernel: allocated 34 Diskbuffers = 18088 Bytes in HMA

I can get linux liveCDs to boot up...

Anyone have any suggestions for how I might either run a virus scan on the bios without a hard disk; or how I might flash BIOS without a hard disk (or a floppy disk).

Or is it far more likely that the drive failed in a wierd way and a virus in the BIOS is not a possible culprit.

I'd rather not lose another drive that cost me a few clams.

[Sinnercure]

If it makes a difference, I was working in Linux the last time I turned off this machine without the "Enter Disk Password" greeting me on boot.

[Grogan]

You can't "run a virus scan on the bios". That's not an area that's accessible, moreover, there's no such thing as a bios virus.

It is, however, possible for some viruses to hide in the extended cmos area, but that would be relatively rare. Clearing the CMOS (e.g. remove battery or use CMOS clear jumper) takes care of that.

It is also possible for a virus to corrupt your system bios. (e.g. Win32_CIH a.k.a. Chernobyl). It writes one byte of data to the bootstrap area, rendering the computer unbootable even with a boot disk (even a Linux boot disk). This isn't itself an infection, the bios chip needs reprogramming or replacing if this occurs.

Chances are this has nothing to do with any viral activity.

[Sinnercure]

Thanks a lot for setting me straight. I went ahead and installed the new disk. All's working as expected.

My laptop's touchpad doesn't work with the 2.6 kernels. It's a separate issue from the /dev/psaux support in slack 10.2 that Blackhawk recently raised . Doesn't work from Mepis or Knoppix either. I run a lot of large file processing scripts from terminals with this machine, so I'm wondering if the native posix thread library would improve processing times significantly enough to warrant carrying around a usb mouse until I figure out the touchpad. I've seen claims that processes can be sped up by a factor of ten, which would certainly be worth it.

[Grogan]

ahh, NPTL speeds up processes that make use of threading, yes. Basically the improvements are in creating and destroying threads, signal handling, and interprocess communications (IPC). For example, mutexes can be shared between threads. Signals can now be passed between threads (and information between threads can be passed using signals as well). Things get to happen more simultaneously.

I don't think your file processing scripts will benefit from this (you may be spawning or forking processes, but that's not threading). However, the 2.6 kernel itself might give you better performance due to better i/o scheduling and the like.

Also, if you truly want to see maximal performance benefits from the 2.6 kernel, you really should build your own, configured for the correct processor, with support only for the hardware that you have and only the features you need. I'm sure you've seen the kernel compiling howto here, and Slackware is easy to build a custom kernel for.

A normal USB mouse would probably be easier to operate anyway, if it works easily for you.

I remember seeing something about some touch pads that stopped working. Something to do with how the kernel initializes a ps/2 mouse device now, where the 2.4 kernel driver didn't do that (the device is initialized by the drivers in X)

There exist touchpad drivers for X that you can try.

http://web.telia.com/~u89404340/touchpad/

[Sinnercure]

I think I did something real boneheaded.

I used fdisk to make the first 40g an ntfs partition that I'll install windows on tomorrow, and the last 40g are partitioned up for slack.

Started installing and forgot to flag an active partition... I was planning on making the windows partition active, should I quit the install and flag that now with fdisk, or would I need winxp to be properly installed on hd1 for this scheme to work?

[Grogan]

Linux (lilo) isn't going to use an active partition at all so you should be able to set the partition active later on (before installing Windows)

You REALLY would have been best to install Windows first. Boot with the Windows XP CD, create your 40 gig ntfs partition (specifying the size) and leave the rest unpartitioned. Install Windows, then boot with your Slackware CD and do the deed.

But now that you've gone this far, try it your way. You can boot with the slackware CD and do root= to jumpstart your slackware installation after Windows XP setup ruins lilo in the MBR. Specifying root= at the kernel command line will boot with the kernel from CDROM, but will run init from your slackware root filesystem. Once you get in, simply add a chainloading stanza to lilo.conf to start Windows, and run /sbin/lilo.

e.g.

other = /dev/hda1
label = Windows
table = /dev/hda