[PATCH v3 1/4] PCI: X-Gene: Add the APM X-Gene v1 PCIe MSI/MSIX termination driver

[Duc Dang]

On Thu, Apr 9, 2015 at 3:39 PM, Bjorn Helgaas <bhelgaas at google.com> wrote:
> On Thu, Apr 9, 2015 at 4:52 PM, Duc Dang <dhdang at apm.com> wrote:
>> On Thu, Apr 9, 2015 at 1:11 PM, Bjorn Helgaas <bhelgaas at google.com> wrote:
>>> On Thu, Apr 09, 2015 at 10:05:03AM -0700, Duc Dang wrote:
>>>> X-Gene v1 SoC supports total 2688 MSI/MSIX vectors coalesced into
>>>> 16 HW IRQ lines.
>>>>
>>>> Signed-off-by: Duc Dang <dhdang at apm.com>
>>>> Signed-off-by: Tanmay Inamdar <tinamdar at apm.com>
>>>> ...
>>>
>>>> --- /dev/null
>>>> +++ b/drivers/pci/host/pci-xgene-msi.c
>>>> @@ -0,0 +1,407 @@
>>>> ...
>>>
>>>> +static void xgene_irq_domain_free(struct irq_domain *domain,
>>>> +                               unsigned int virq, unsigned int nr_irqs)
>>>> +{
>>>> +     struct irq_data *d = irq_domain_get_irq_data(domain, virq);
>>>> +     struct xgene_msi *msi = irq_data_get_irq_chip_data(d);
>>>> +
>>>> +     mutex_lock(&msi->bitmap_lock);
>>>> +
>>>> +     if (!test_bit(d->hwirq, msi->bitmap))
>>>> +             pr_err("trying to free unused MSI#%lu\n", d->hwirq);
>>>
>>> I'll let Marc comment on the substance of this patch, but I'd sure like to
>>> see more information in this and the other pr_errs below.  The text "failed
>>> to create parent MSI domain" in the dmesg log is not enough.  Ideally it
>>> would identify the relevant host bridge.
>>>
>> Hi Bjorn,
>>
>> X-Gene has single MSI block that serves 5 PCIe ports. so in this
>> implementation, 5 X-Gene PCIe host bridges share the same MSI domain.
>>
>> I can add X-Gene 1 MSI driver name and device tree node name into
>> these messages, but the host bridge information is not available.
>>
>> As I understand, in case of error, the upper layer MSI core and driver
>> code will check and report the error with additional device
>> information (bus number, function number, etc)?
>
> If you think the upper layers already report the error, maybe you
> could just drop the output from the X-Gene code.
>
> I just think it's pointless to print messages that are constant
> strings with no hook for the user to associate them with a device or
> event.

Hi Bjorn,

Yes, I will do so.

Regards,
Duc Dang.
>
>>>> +     else
>>>> +             clear_bit(d->hwirq, msi->bitmap);
>>>> +
>>>> +     mutex_unlock(&msi->bitmap_lock);
>>>> +
>>>> +     irq_domain_free_irqs_parent(domain, virq, nr_irqs);
>>>> +}
>>>> +
>>>> +static const struct irq_domain_ops msi_domain_ops = {
>>>> +     .alloc  = xgene_irq_domain_alloc,
>>>> +     .free   = xgene_irq_domain_free,
>>>> +};
>>>> +
>>>> +static int xgene_allocate_domains(struct xgene_msi *msi)
>>>> +{
>>>> +     msi->domain = irq_domain_add_linear(NULL, msi->settings->nr_msi_vec,
>>>> +                                         &msi_domain_ops, msi);
>>>> +
>>>> +     if (!msi->domain) {
>>>> +             pr_err("failed to create parent MSI domain\n");
>>>> +             return -ENOMEM;
>>>> +     }
>>>> +
>>>> +     msi->mchip.of_node = msi->node;
>>>> +     msi->mchip.domain = pci_msi_create_irq_domain(msi->mchip.of_node,
>>>> +                                                   &xgene_msi_domain_info,
>>>> +                                                   msi->domain);
>>>> +
>>>> +     if (!msi->mchip.domain) {
>>>> +             pr_err("failed to create MSI domain\n");
>>>> +             irq_domain_remove(msi->domain);
>>>> +             return -ENOMEM;
>>>> +     }
>>>> +
>>>> +     return 0;
>>>> +}
>>>> +
>>>> +static void xgene_free_domains(struct xgene_msi *msi)
>>>> +{
>>>> +     if (msi->mchip.domain)
>>>> +             irq_domain_remove(msi->mchip.domain);
>>>> +     if (msi->domain)
>>>> +             irq_domain_remove(msi->domain);
>>>> +}
>>>> +
>>>> +static int xgene_msi_init_allocator(struct xgene_msi *xgene_msi)
>>>> +{
>>>> +     u32 msi_irq_count = xgene_msi->settings->nr_msi_vec;
>>>> +     u32 hw_irq_count = xgene_msi->settings->nr_hw_irqs;
>>>> +     int size = BITS_TO_LONGS(msi_irq_count) * sizeof(long);
>>>> +
>>>> +     xgene_msi->bitmap = kzalloc(size, GFP_KERNEL);
>>>> +     if (!xgene_msi->bitmap)
>>>> +             return -ENOMEM;
>>>> +
>>>> +     mutex_init(&xgene_msi->bitmap_lock);
>>>> +
>>>> +     xgene_msi->msi_virqs = kcalloc(hw_irq_count, sizeof(int), GFP_KERNEL);
>>>> +     if (!xgene_msi->msi_virqs)
>>>> +             return -ENOMEM;
>>>> +
>>>> +     return 0;
>>>> +}
>>>> +
>>>> +static irqreturn_t xgene_msi_isr(int irq, void *data)
>>>> +{
>>>> +     struct xgene_msi *xgene_msi = (struct xgene_msi *) data;
>>>> +     unsigned int virq;
>>>> +     int msir_index, msir_reg, msir_val, hw_irq;
>>>> +     u32 intr_index, grp_select, msi_grp, processed = 0;
>>>> +     u32 nr_hw_irqs, irqs_per_index, index_per_group;
>>>> +
>>>> +     msi_grp = irq - xgene_msi->msi_virqs[0];
>>>> +     if (msi_grp >= xgene_msi->settings->nr_hw_irqs) {
>>>> +             pr_err("invalid msi received\n");
>>>> +             return IRQ_NONE;
>>>> +     }
>>>
>>> Bjorn
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo at vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/