Ollie,
The answer to your question is the latter. We're using x86's, however,
in many instances they're not wired as PCs and may or may not be using PC
chipsets. For instance, one of our designs uses a custom southbridge chip
that conflicts with typical PC compatible peripherals - if I remember
correctly, this specific ASIC also doesn't use an 8259 compatible interrupt
controller.
rpm
Ollie Lho
<[EMAIL PROTECTED] To: [EMAIL PROTECTED]
m.tw> cc: [EMAIL PROTECTED]
Sent by: Subject: Re: Linux BIOS
[EMAIL PROTECTED]
.tw
05/08/01
05:43 PM
[EMAIL PROTECTED] wrote:
>
> Hello,
>
> I'm a developer for an embedded computer manufacturer - RadiSys. I'm
> looking for a quick way to get Linux running on our embedded x86
platforms.
> Typically, we engineer the hardware platform to be "pc compatible" and
> throw a BIOS onto the board to complete its PC-ness, however, some of our
> platforms are not PCs even though they use PC chipsets (Intel BX's and.
> Serverworks). In these instances diving into the BIOS source code to
> port to a non-PC is not an effective option - the interdependencies of
the
> BIOS source code make this task essentially impossible. So, my question
> is ... is the Linux BIOS well suited for these non-PC platforms?
> Conceptually, my expectation is that the Linux BIOS would work fine on
the
> non-PC platforms and the level of port effort would be relative to how
far
> the design strayed from a PC. I would also expect that devices that
could
> not be found during the discovery and configuration phase would just be
> left alone and un-initialized. Finally, since we're done once we get
the
> ROM'ed version of Linux running, there's no need to use LOBOS to load
> another OS. Is my thinking correct? Do you foresee any problems with
> making Linux BIOS work for embedded designs as described above?
>
Does PC == x86 and non-PC == non-x86 in your question ??
or
PC == x86 PC and non-PC == x86 customized ??
Ollie
