Hi Christoph,
On 12/7/16, 11:03 AM, "Christoph Hellwig" <hch@xxxxxxxxxxxxx> wrote:
>> static int
>> qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp)
>> {
>> #define MIN_MSIX_COUNT 2
>> #define ATIO_VECTOR 2
>> int i, ret;
>> struct qla_msix_entry *qentry;
>> scsi_qla_host_t *vha = pci_get_drvdata(ha->pdev);
>>
>> + ret = pci_alloc_irq_vectors(ha->pdev,
>> + MIN_MSIX_COUNT, ha->msix_count, PCI_IRQ_MSIX|PCI_IRQ_AFFINITY);
>
>Given that as-is the code only uses two vectors, and they are not
>for per-cpu queues using PCI_IRQ_AFFINITY is actually wrong, and
>you're better off without it. Also please fix the spacing:
>tabs for aligning continued arguments, and spaces around operators
>(although the revised version won't have an operator - so this is
>just for future reference:
>
Ack. Agree here this change should not be part of this patch.
> ret = pci_alloc_irq_vectors(ha->pdev, MIN_MSIX_COUNT,
> ha->msix_count, PCI_IRQ_MSIX);
>
>> + struct rsp_que *rsp = (struct rsp_que *)e->handle;
>
>No need for the cast here.
Ack. Will post update in new series.
>
>> + struct rsp_que *rsp = (struct rsp_que *)e->handle;
>
>Same here.
Will send new series soon.
>
��.n��������+%������w��{.n�����{������ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f
