Re: Problems with irq mapping in qemu v5.2

[Guenter Roeck <linux@roeck-us.net>]

On 12/22/20 2:23 PM, BALATON Zoltan wrote:
> On Tue, 22 Dec 2020, BALATON Zoltan via wrote:
>> Hello,
>>
>> On Tue, 22 Dec 2020, Guenter Roeck wrote:
>>> Hi,
>>>
>>> commit 459ca8bfa41 ("pci: Assert irqnum is between 0 and bus->nirqs in
>>> pci_bus_change_irq_level") added sanity checks to the interrupt number passed
>>> to pci_bus_change_irq_level(). That makes sense, given that bus->irq_count
>>> is indexed and sized by the number of interrupts.
>>>
>>> However, as it turns out, the interrupt number passed to this function
>>> is the _mapped_ interrupt number. The result in assertion failures for various
>>> emulations.
>>>
>>> Examples (I don't know if there are others):
>>>
>>> - ppc4xx_pci_map_irq() maps the interrupt number to "slot - 1". Obviously
>>>  that isn't a good thing to do for slot 0, and indeed results in an
>>>  assertion as soon as slot 0 is initialized (presumably that is the root
>>>  bridge). Changing the mapping to "slot" doesn't help because valid slots
>>>  are 0..4, and only four interrupts are allocated.
>>
>> Is that with sam460ex? This ppc4xx_pci_map_irq appears in ppc4xx-host-bridge and hw/ppc/ppc440_pcix.c (which is used by sam460ex) has a ppc4xx-host-bridge. Other user of it may be the bamboo 440 board where this host bridge appears too. I did not see asserts with sam460ex but I've only tested it with AmigaOS which may use it in an odd way not tripping the check.
> 
> I've just remembered that for sam460ex we had this commit: 484ab3dffadc (sam460ex: Fix PCI interrupts with multiple devices) that changed that mapping for that machine so I guess you got the exception with the bamboo board then. I'm not sure though that similar fix is applicable fot that or even that this fix is correct for sam460ex but appears to work so far.
> 
Yes, it is for bamboo. I think I'll just keep the mapping I came up with;
that seems to work for me.

Thanks,
Guenter