Re: [PATCH 2/2] spi: pca2xx-pci: Allow MSI

[Jan Kiszka]

On 2017-01-16 10:29, Andy Shevchenko wrote:
> On Mon, 2017-01-16 at 10:06 +0100, Jan Kiszka wrote:
>> Now that the core is ready for edge-triggered interrupts, we can
>> safely
>> allow the PCI versions that provide this to enable the feature and,
>> thus, have less shared interrupts.
>>
> 
> My comments below.
> 
>> -	 if (IS_ERR(ssp->clk))
>> +	if (IS_ERR(ssp->clk))
>>  		return PTR_ERR(ssp->clk);
> 
> This doesn't belong to the patch.
>  
>> +	pci_set_master(dev);
>> +
>> +	ret = pci_alloc_irq_vectors(dev, 1, 1, PCI_IRQ_ALL_TYPES);
>> +	if (ret < 0) {
>> +		clk_unregister(ssp->clk);
>> +		return ret;
>> +	}
>> +	ssp->irq = pci_irq_vector(dev, 0);
>> +
> 
> This looks good, though I would put it closer to the initial place of
> ssp->irq assignment, i.e. before clock registering.
> 
>> +		pci_free_irq_vectors(dev);
>> +	pci_free_irq_vectors(dev);
> 
> You know my answer, right? So, please be sure that we are using
> pcim_alloc_irq_vectors().
> 
> Yes, I know there is (was?) no such API, needs to be created. Currently
> this might make a mess on ->remove().
> 

FWIW, I've an updated version of this patch already, addressing the
remarks. Just waiting for a reply on the other patch now.

Jan

-- 
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux