Text 7716, 72 rader
Skriven 2006-10-28 01:20:02 av Maurice Kinal (1:140/13.1)
   Kommentar till text 7715 av Paul Rogers (1:105/360.0)
Ärende: Booting
===============
Hey Paul!

Oct 27 16:55 06, Paul Rogers wrote to Maurice Kinal:

 PR> Does the kernel record somewhere what the boot device was?
 PR> Somewhere I can read it, e.g. /proc, I mean.

Used to be there but lately /sys (sysfs) has been stealing greater control.  At
the moment I am relearning much of what used to come second nature.  I've
noticed that the sysfs is more revealing then proc used to be so I am not
protesting but at this point am not exactly comfortable with what I think I
know.  Couple that with looking at new ways to do this type of stuff and I can
comfortably say at this point in time that I am not comfortable ... yet. 
Getting there though or at least planning to.

Also, although /sys is humanly readable it isn't exactly self evident what all
the stuff there really means.  I have noticed things like pciutils can yeild
different information then some of the stuff I see in /sys wrt pci.

On the plus side what works great works great without the need for firmware and
the like and I am confident I'll get there eventually.

 PR> Trinux uses syslinux and an initrd.

Quite a few do.  Myself I still use lilo and an initrd to ccomplish the same
thing.  That works across the board with everything I boot, especially on the
64-bit system where there are few choices as to working boot apps in pure
64-bit mode.

 PR> linuxrc.  It comes up and scans `dmesg | ...` looking for hard
 PR> drives & CDROM evidence, looks at /proc/filesystems, sees what
 PR> mounts where successfully.

Yeah.  I am not sure about CD's anymore.  I see that getting worse not better
especially since Sony and their competition feels that we now want something
completely different that appears to be another money grab to me.

 PR> bootstrap process order & it's not straight in my mind.

Right.  Same here which is why I am sticking to lilo for now as it has never
failed me other then when the device is questionable.

 PR> Given the fact that I insist on having a floppy option, I think
 PR> this is something you had to do for your thumb drives.

Nope.  I do them the same way I do regular hard drives.  The only problem with
booting thumb drives has been the bios but the disks themselves I have only
seen a few that couldn't boot but in those two cases it was due to purpose
abuse on my part.  I wanted to see how much abuse they could REALLY take rather
then what the manufacturer claimed.  I've been playing with these (not usb
which is more recent) for quite some time now and overall have been very happy
with them, including the usb based ones now.  Mind you the rest are more easily
made to be usb based then their so-called 'native' mode in a few cases,
compactflash being one that comes to mind.  Bootable IDE adapters for
compactflash are not that readily available and the first one I tried I had to
jump through far too many hoops to get my hands on.  Now I just use usb since
motherboards (actually bios's) are getting them right.  I still prefer regular
hard drives but the usb is looking like a great and inexpensive way to take
care of hotswapping and booting of minimal systems.  A little slow booting but
it seems solid all without the need for any specialized software as I can use
the exact same tools/apps.

Sorry if the above is a tad long winded with little real information but I just
wanted to make my point clear on this particular matter.  I honestly believe it
will all get better soon and then I may have something more informative to say
on the issue of bootability.  Working on it.  :-)

Life is good,
Maurice

--- Msged/LNX 6.2.0
 * Origin: The Pointy Stick Society XXV - Almost there (1:140/13.1)