Re: [Qemu-devel] [PATCH v6 09/11] megasas: change msi/msix property type

[Markus Armbruster]

Cao jin <address@hidden> writes:

> On 06/01/2016 05:14 PM, Markus Armbruster wrote:
>
>>>
>>>   static bool megasas_use_msix(MegasasState *s)
>>>   {
>>> -    return s->flags & MEGASAS_MASK_USE_MSIX;
>>> +    return s->msix == ON_OFF_AUTO_AUTO || s->msix == ON_OFF_AUTO_ON;
>>
>> s->msi != ON_OFF_AUTO_OFF, please.
>>
>>>   }
>>
>> Same correctness argument as for megasas_mask_use_msi().
>>
>
> Hi Markus,
> I Cant find function names megasas_mask_use_msi, only macro that I
> deleted. Don`t quite follow the meaning, do I need to do anything
> about it?

I meant megasas_use_msi().  Sorry for sowing confusion again :)

>
>>>
>>>   static bool megasas_is_jbod(MegasasState *s)
>>> @@ -2349,12 +2347,12 @@ static void megasas_scsi_realize(PCIDevice *dev, 
>>> Error **errp)
>>>
>>>       if (megasas_use_msi(s) &&
>>>           msi_init(dev, 0x50, 1, true, false) < 0) {
>>> -        s->flags &= ~MEGASAS_MASK_USE_MSI;
>>> +        s->msi = ON_OFF_AUTO_OFF;
>>>       }
>>>       if (megasas_use_msix(s) &&
>>>           msix_init(dev, 15, &s->mmio_io, b->mmio_bar, 0x2000,
>>>                     &s->mmio_io, b->mmio_bar, 0x3800, 0x68)) {
>>> -        s->flags &= ~MEGASAS_MASK_USE_MSIX;
>>> +        s->msix = ON_OFF_AUTO_OFF;
>>>       }
>>>       if (pci_is_express(dev)) {
>>>           pcie_endpoint_cap_init(dev, 0xa0);
>>
>> Unlike the device models we've seen in earlier patches, this one
>> overwrites its configuration to reflect actual device state.  Hmm.  I'll
>> revisit this in my review of PATCH 11.
>>
>
> I guess it is still OK?

This patch is okay.  I was just observing that this device is different.