Re: GPLv3 Position Statement

[Krzysztof Halasa <khc@pm.waw.pl>]

I generally agree with you, but...

Linus Torvalds <torvalds@osdl.org> writes:

> And it not at all uncommon to have a flash that simply cannot be upgraded 
> without opening the box. Even a lot of PC's have that: a lot (most?) PC's 
> have a flash that has a separate _hardware_ pin that says that it is 
> (possibly just partially) read-only. So in order to upgrade it, you'd 
> literally need to open the case up, set a jumper, and _then_ run the 
> program to reflash it.

I think this is history. Yes, late 486s and Pentiums (60 and 66?)
had a jumper protecting the flash. It's not true since ca. "Pentium 75+"
days - while many boards use "bootblock" chips, it's (almost?) always
unprotected (at most it just requires setting some GPIO pin(s)). The
rest of flash obviously has to be R/W to support the ESCD etc.

I think there are systems with 2 copies of the whole BIOS, and the
user selects the copy with a jumper (probably connected directly to
the most significant address line of the flash IC) - the second
copy might theoretically use a R/O bootblock but I've never checked it.

Most VGAs, disks, PCI cards etc. have flash chips with no protection
either, and I have to say I felt much better when they used (EP)ROMs.

I think almost all hardware manufacturers use a blank flash chips,
programming them "in system" with things like JTAG.
-- 
Krzysztof Halasa