[haiku-development] Re: [haiku-commits] Re: r36420 - in haiku/trunk: build/jam headers/os/drivers src/add-ons/kernel/bus_managers/acpi src/add-ons/kernel/drivers/power src/add-ons/kernel/drivers/power/acpi_embedded_controller ...

On 2010-04-23 at 19:25:46 [+0200], Jorge G. Mare <koki@xxxxxxxxxxxxx> wrote:
> Fredrik Holmqvist wrote:
> > 2010/4/23 Ingo Weinhold <ingo_weinhold@xxxxxx>:
> >   
> >>> Oh I just remembered... what about some machines that stop controlling 
> >>> the
> >>> fan speed when ACPI is enabled? Is this a myth or is this a real 
> >>> problem?
> >>>       
> >
> > I've havn't heard of anyone with that problem yet. It's a scary
> > problem. Is there anything we can/should do regarding this?
> >   
> 
> FWIW, this happens when I run ACPI-enabled Haiku on my good old AMD
> Athlon XP machine.

This means there is already one example of hardware which we would toast if 
we enable the flag in Alpha 2. Old AMD machines especially just go up in 
smoke, they (at least some older models) have no self-preservation like 
Intel CPUs have since the Pentium III.

As sad as this may be, this means we should not ship with ACPI enabled. The 
proper fix would be to have ACPI drivers which control the fan speed, but I 
see no way we have those in time for the release. Perhaps there is an easy 
way to have drivers that just put whatever hardwere in "auto" mode, if there 
is such a thing, but I have no idea how many specs there are, and it would 
again need proper testing.

Best regards,
-Stephan

Other related posts: