[openbeos] Re: x86 boot loader milestone
- From: "Bruno van Dooren" <bruno_van_dooren@xxxxxxxxxxx>
- To: <openbeos@xxxxxxxxxxxxx>
- Date: Thu, 22 Apr 2004 08:39:59 +0200
first of all: WAAAY cool milestone.
second: my computer is a dual PIII 1GHz with a MSI motherboard that uses a
VIA chipset.
it has raid, but no array is attached.
each time it boots, it tries to detect a raid array, and do a lot of other
BIOS stuff.
the boot process takes about 30-45 seconds to get to the bootloader.
BeOS boots in 8-10 seconds.
so even if I would be able to boot OBOS in 3 seconds, this still wouldn't
make much difference.
this is not to say it wouldn't be cool on a motherboard with a faster BIOS
startup.
kind regards,
Bruno.
----- Original Message -----
From: "Jared Eldredge" <jared@xxxxxx>
To: <openbeos@xxxxxxxxxxxxx>
Sent: Thursday, April 22, 2004 7:52 AM
Subject: [openbeos] Re: x86 boot loader milestone
> i dont get it : BeOS boots in what, 15 seconds?
> are we really trying to decrease boot time? it doesn't seem painful to
> me at all when (if ever) i have to reboot. so why the efforts?
>
> personally i think its a neat idea - but only for the sake of 'it would
> be cool if it works'. i highly doubt that you can reduce boot times
> significantly without storing the entirety of the critical OS
> components on the flash card. but hey, i'm a pessimist. if you can
> make me boot in 5 seconds i'll eat my words :)
> -jared
>
> > > This is too funny. I am working on exactly this same thing! :-)
> > >
> > > What I was thinking was that I would make a ram disk and mount home
> > > as the ram
> > > disk. And turn off VM. Certain critical things (preferences) would
> > > be
> > > copied
> > > back to the CF. So far, it does boot pretty fast. The next step was
> > > /
> > > is to
> > > investigate LinuxBIOS to see if we could boot with that. :-)
> > >
> > > Michael
> >
> > I gotta admit, when I talked with you about a month or so ago your
> > comments on what you were doing did help direct my ideas :) Here's
> > what just came to mind that I was wondering if it was feasible. It's
> > possible and rather easy now to make an image of a drive or a
> > filesystem. Is it possible, however, to create an image of RAM?
> > Given
> > that there is the AGMS RAM filesystem, I think it should be possible.
> > If so, would it be possible to boot the system until the instant
> > before
> > the UserBootscript is processed, and take an image of the current
> > memory contents and store them on the CF card? That way, when the
> > system boots, instead of re-loading things from the boot drive, etc
> > it
> > can restore the memory contents and pick up where it left off
> > (process
> > UserBootscript, etc). Of course, there would have to be a check of
> > some sort before the RAM image dump happens so that in case something
> > changes and needs reprocessing it can be done. DHCP, device
> > detection,
> > and things of that nature will have to be done after the RAM image
> > restore. Anyway, it's just a random idea I had, there may be enough
> > things that must be detected and re-configured every boot that makes
> > this system have no practical use. But some things always are loaded
> > into memory (kernel parts, app_server parts, tracker parts, etc etc),
> > and they may benefit from this.
> >
> >
> >
> > -Ben
> >
> >
> > ---------------------------------------
> > Commander Sozo
> > CommdrSozo@xxxxxxxxxxxxxxx
> >
>
>
> --
> Automated (hopefully never vulgar) fortune:
>
> Information Center, n.:
> A room staffed by professional computer people whose job it is
> to tell you why you cannot have the information you require.
>
>
Other related posts:
