From: Betov on
"rhyde(a)cs.ucr.edu" <rhyde(a)cs.ucr.edu> �crivait
news:1172699244.357060.194420(a)t69g2000cwt.googlegroups.com:

> As you, yourself, have pointed
> out, Windows is FREE.

Not that the insanities of this nerd would diserve
any answer, but, just in case some readers could
give him any credit, and believe that i really
wrote that somewhere:

I never said that Windows was free, and i don't think
that anybody, around, could have been stupid enough,
for emitting such an absurdity.


Betov.

< http://rosasm.org >




From: Wolfgang Kern on

Jim Carlock said:

> "Wolfgang Kern" stated...
> : You need one working environment as 'the tool' which allow to
> : test new OS code parts immediate (without frequent reboot).

> : My problem would be that my current OS know only keyboard,
> : mouse and old modems on USB. These RAM-Sticks are all a
> : bit different and only M$ may have enough hardware information
> : to support most of it.

> XPSP2 driver for both USB sticks (two different models, one a
> 256MB SanDisk, the other an unknown name brand) ...

> %systemroot%\system32\drivers\VolSnap.sys
> 08/03/2004 10:00 PM 52,352 volsnap.sys

Win98se doesn't even know about USB-sticks without a proper driver.
WinXP_home couldn't identify my Samsung 2GB stick (and still can't)

> I do not think either one of mine are bootable... but perhaps
> the drivers only need to get installed into the BIOS on the
> mainboard?

As Robert already mentioned:
the BIOS will be able to load and execute the first 512 bytes
of any int13-media (ROM,CD,LAN,USB,..) only.

> : So when the new OS boots from USB it must already contain
> : working USB loader/drivers/routines as the BIOS is of little help
> : here.

> USB network cards should be bootable as well, right? Back in
> 1998 or 1999 Intel promoted bootable PCI network interfaces.
> One Microsoft / Intel event I attended demonstrated a NIC that
> booted off a remote server, and the OS sent commands back
> and forth between the driveless client and ran some of the new
> terminals on the client, off the server (Terminal Server).

Yes.

> Then the other way I see it working, bootable USB memory retains
> a BIOS, like the bootable network cards, and respond as disk
> drives, just as a hard disk would. If it acts like a disk drive, it is a
> disk drive and PCI is only the funnel that moves information about.

A USB-memory cannot know about motherboard resources, as in
opposition to ATA/ATAPI, where well defined hardware standards
exists, USB-host hardware is not standardised and may vary with
the chipset on the motherboard.
So it will be impossible to merge a hardware checker in front
of the loader into 510 bytes.


I repeat the links for later use (my fast but lazy storage)

thanks Jim
__
wolfgang


http://www.digg.com/linux_unix/Create_a_bootable_USB_disk_with_DSL
http://www.damnsmalllinux.org/store/Mini_ITX_Systems/Damn_Small_Machine
http://www.damnsmalllinux.org/usb.html

> And the following link looks like pretty good reading.
http://www.bootdisk.com/pendrive.htm

> Microsoft talks about bootable USB-CDs and USB-Flash Drives
> (UFDs)...
http://www.microsoft.com/whdc/device/storage/usb-boot.mspx




From: robertwessel2 on
On Mar 1, 8:02 am, "Wolfgang Kern" <k...(a)utanet.at> wrote:
> A USB-memory cannot know about motherboard resources, as in
> opposition to ATA/ATAPI, where well defined hardware standards
> exists, USB-host hardware is not standardised and may vary with
> the chipset on the motherboard.
> So it will be impossible to merge a hardware checker in front
> of the loader into 510 bytes.


Of course that problem is no different than that of SCSI disks. No
two SCSI (or FC) HBAs present the same interface, other than what's
available via the Int 13 emulation of EFI. Somewhere before the
booting OS starts talking to the SCSI adapter directly, it needs to
load an appropriate driver (which is no different than loading an
ATAPI driver, except that the later is common enough that including it
unconditionally in the boot image is reasonable).

FWIW, USB memory devices look basically like SCSI devices. They use
the same commands, except the physical transport is USB.

From: Wolfgang Kern on

Hi Robert.

> Of course that problem is no different than that of SCSI disks. No

> two SCSI (or FC) HBAs present the same interface, other than what's
> available via the Int 13 emulation of EFI. Somewhere before the
> booting OS starts talking to the SCSI adapter directly, it needs to
> load an appropriate driver (which is no different than loading an
> ATAPI driver, except that the later is common enough that including it
> unconditionally in the boot image is reasonable).
>
> FWIW, USB memory devices look basically like SCSI devices. They use
> the same commands, except the physical transport is USB.

particulsr right. USB-host is stil unknown while USB device-behaviour
may be recognised by the BIOS (even unable to fully support it).

__
wolfgang