> Noone claimed they boot the system from USB.
Then BIOS is irrelevant.... > the thread drifted when some people just wanted to point out that > for your mobile use-case scenario non HD-USB booting BIOS might > be an issue. ...then BIOS is relevant. The originator of this thread asked for USB, so he presumably knows how BIOS issues pertain to PC rescue, and yet still wants USB functionality. If you want to argue about more general use cases like ours.... ...then measure importance and form an opinion. Only 3-6 major PC BIOS vendors exist, all of whom have supported USB-HD for 6 years, and USB-ZIP longer. Our opinion: legacy issues on old machines fall into a less-than-worrisome slot. The workaround is a grub boot floppy or kexec/kboot. As time marches on, legacy issues become even less a problem, while USB device capacity grows and cost shrinks. > In any case it appears to work just fine for you so all's fine - yours, The fallacy implied is that we are exceptional. What I could retort is, "too bad it doesn't work for you, since it works for everyone else." :-)) P.S. Embedded devices and field service do not work as you described. No nexus really exists between USB and embedded devices. The subject is worth a few lines to clear up confusion. For decades now (prior to JTAG, and still today) embedded systems have used serial ports. They are not always visible as DB-9 connectors on the box, but the circuitry exists. Connection can be a bare header on the embedded board. Only if the customer/device must perform serial comm is there a DB-9 port on the case. Field service involves cracking the case, plugging into the serial comm header, and flashing an upgrade. If a DB-9 port exists then that is used instead. Either way, this operation is rare, not a regular occurrence, so speed is unimportant. What matters is part count and design simplicity. USB is massively huge overkill for flash upgrades. Any company you find desiring that functionality is going broke soon. Flash is the most common boot method, not "internal storage USB-CD/stick booting" whatever is meant by that phrase. Boot flash is just memory mapped. It has nothing to do with USB. The reason flash became so popular for embedded systems is that upgrades can be performed without part exchange. The idea that companies now want to use USB flash sticks is backwards. More likely would be an EPROM. These are traditional in embedded systems. They pop straight out of their sockets, as opposed to soldered flash, which requires electronic communication to change. EPROMs are programmed with specialized EPROM programmers, as you may know. The upgrade method is to swap chips mechanically. No USB here. Recent years have seen embedded devices with network functionality using web-browser interfaces for flash upgrades. Those are your Linksys and D-Link boxes. That's Ethernet from a PC, not USB. Since they already handle Ethernet, it costs nothing extra to implement flash u/g over Ethernet. No USB here, either. PC/104 is a normal PC. Flashing the BIOS happens exactly like a normal PC. Typically you plug into the serial port, or run "flash.exe" from a boot floppy (plugging into the floppy port a spare floppy drive). Again, not USB at all. Unless you're developing a USB device, it won't enter the picture. If you are, there are dozens of vendors for USB stacks, chips, etc. But you're still not going to boot the device from USB, it will boot from memory-mapped flash ROM. The idea that USB is somehow causing embedded systems headaches is high weirdness! ----------------------------------------------------------- If you wish to unsubscribe from this mailing, send mail to [EMAIL PROTECTED] with a subject of: unsubscribe t2
