Norbert Kamenicky <noro <at> xmedia.sk> writes: > > > Yes, I many times removed BIOS extension chips (not needed during kernel > boot) from different cards, just to boot faster... > if it's not easy, there is always a possibility to cut chip select/ > enable (CS/CE) wire near to chip's pin.
OK > > Linux kernel runs in 32-bit protected mode, but AFAIK nearly all BIOSes > are written for 16-bit real mode and stored in 8-bit *ROMs (very slow > access). This makes it nearly unusable under linux kernel. Yes, but there is often information in the bios that is needed for certain features, or features between chips/interfaces/busses. > > If u are on doubt, linux kernel uses (or not) BIOS calls for ACPI, I2C, > sensors, SMBus, Dallas one-wire etc., please open appropriate driver in > kernel sources and check it yourself... and let me know, if u find some > driver, which uses them. I agree with all you have said. But, that's not the entire story. Often features that work on WinBlows(*) do not work on Linux, regardless of the software and drivers... Wonder why? Often it's secret or at least partially is in bios. Seemingling insignificant information in the bios, is actually critical for key features in chips/interfaces/busses... Furthermore, much of the newer bios is in an executable or other obscure form. If you look at the Datasheets for some a given chip, you can see lots of features. Newer BIOS often contain 'keys' (implemented in a variety of methods) to activate key features in chipsets and thus in WinBlows software. Take a look at the Richard Stallman rant: http://www.fsf.org/news/freebios.html I have not disagreed with what has been said. I'm just pointing out that there's much more to the bios/hidden-firmware issues than most are aware of. This is often why vendors do not want their driver code published. They are up to things that they do not want the linux community to highlight, for the masses of blissfully unaware folks..... later, James -- [email protected] mailing list
