From: Martin on
Yousuf Khan wrote:

> I think the fact that both an older Linux kernel and Windows worked,
> while the current Linux kernel doesn't is proof that the problem lies
> with the kernel and not the BIOS.

That logic is flawed.

Martin
From: Yousuf Khan on
ceed wrote:
> The problem may be that there's been a BIOS update and that the newer
> kernels works with the newer version of the BIOS. I had that happen on
> one of my other (Samsung Q310) laptop. The original BIOS works with
> <2.6.30 kernels while the newer BIOS works flawlessly with 2.6.30>..

Great, but shouldn't that be considered a bug? If you break something
that was working previously, then that sounds like a bug to me. And an
even more basic question is why do you need to remove older
compatibility for newer compatibility? It shouldn't be hard to have an
installable modular ACPI driver that installs for the appropriate ACPI
version.

I just ran into a flaw in Linux that was popping up an annoying startup
message. It turned out the flaw was introduced when they patched a bug
in another place.

Yousuf Khan
From: ceed on
On Mon, 04 Jan 2010 11:53:03 -0600, Yousuf Khan <bbbl67(a)yahoo.com> wrote:

> ceed wrote:
>> The problem may be that there's been a BIOS update and that the newer
>> kernels works with the newer version of the BIOS. I had that happen on
>> one of my other (Samsung Q310) laptop. The original BIOS works with
>> <2.6.30 kernels while the newer BIOS works flawlessly with 2.6.30>..
>
> Great, but shouldn't that be considered a bug? If you break something
> that was working previously, then that sounds like a bug to me. And an
> even more basic question is why do you need to remove older
> compatibility for newer compatibility?

If you have a choice you will try to support the latest bios version in
the kernel. You would of course try to be as backwards compatible as
possible. If a kernel upgrade happens to break the support for older bios
versions I would be more inclined to blame the bios vendor for the lack of
compatibility.

It shouldn't be hard to have an
> installable modular ACPI driver that installs for the appropriate ACPI
> version.

ACPI has always been problematic in Linux. Personally I think great
improvements have been made the last couple of years. I'm not sure what
you mean by "installable modular ACPI driver that installs for the
appropriate ACPI version". Who would create these drivers?

>
> I just ran into a flaw in Linux that was popping up an annoying startup
> message. It turned out the flaw was introduced when they patched a bug
> in another place.

Happens all the time. Makes Linux fun! :)

>
> Yousuf Khan


--
//ceed