Re: [PATCH 2/2] via-ide: Also emulate non 100% native mode

[Mark Cave-Ayland]

On 01/03/2020 16:42, BALATON Zoltan wrote:

>> The other part I'm not sure about is that I can't see how via_ide_set_irq() 
>> can ever
>> raise a native PCI IRQ - comparing with my experience on cmd646, should 
>> there not be
>> a pci_set_irq(d, level) at the end?
> 
> According to my tests with several guests it seems the via-ide does not seem 
> to use
> PCI interrupts as described in the previous reply, only either legacy IRQ14 
> and 15 or
> one ISA IRQ line set by a config reg in native mode (except on Pegasos2). 
> This may be
> due to how it's internally connected in the southbridge chip it's part of or 
> some
> other platform specific quirk, I'm not sure.

I think this is the key part here: how does via-ide switch between legacy and 
native
mode? For CMD646 this is done by setting a bit in PCI configuration space, and 
I'd
expect to see something similar here.

It might be that the BIOS sets legacy mode on startup, and unless the OS 
explicitly
switches to native mode then the interrupt routing remains at IRQ 14/15 (or 
whatever
value is in PCI_INTERRUPT_LINE). Is there a datasheet available for the VIA 
chip to
check this?


ATB,

Mark.