[PATCH v10 0/4] PCI: X-Gene: Add APM X-Gene v1 MSI/MSIX termination driver
Duc Dang
dhdang at apm.com
Fri Jun 5 14:11:09 PDT 2015
On Fri, Jun 5, 2015 at 2:05 PM, Bjorn Helgaas <bhelgaas at google.com> wrote:
> On Fri, May 29, 2015 at 11:24:28AM -0700, Duc Dang wrote:
>> This patch set adds MSI/MSIX termination driver support for APM X-Gene v1 SoC.
>> APM X-Gene v1 SoC supports its own implementation of MSI, which is not compliant
>> to GIC V2M specification for MSI Termination.
>>
>> There is single MSI block in X-Gene v1 SOC which serves all 5 PCIe ports.
>> This MSI block supports 2048 MSI termination ports coalesced into 16 physical
>> HW IRQ lines and shared across all 5 PCIe ports. Starting from version 5 of
>> this patch series, the total MSI vectors this driver supports is reduced to 256
>> to maintain the correct set_affinity behavior for each MSI.
>>
>> v10 changes:
>> 1. Remove inline in helper functions
>> 2. Simplify hwirq_to_cpu implementation
>>
>> v9 changes:
>> 1. Add more helpers for manipulating MSI messages
>> 2. Improve code readability
>> 3. Make PCI_XGENE_MSI selectable
>>
>> v8 changes:
>> 1. Add helper to read MSI registers
>> 2. Resample interrupt status of the same group after handling interrupts
>> in that group
>> 3. Error handling in case fail to allocate CPU mask for affinity setting
>>
>> v7 changes:
>> 1. Add more error handling cases
>> 2. Clear spurious interrupts that may happen during driver probe
>> 3. Not using free_irq for chained irqs
>> 4. Improve GIC IRQ number look up in chained handler
>>
>> v6 changes:
>> 1. Correctly allocate MSI with bitmap_find_next_zero_area
>> 2. Add notifier for CPU hotplug case
>> 3. Driver clean up
>> 4. Add more detailed information into device tree binding document
>> and move the device tree binding patch before the driver/dts patch
>>
>> v5 changes:
>> 1. Implement set_affinity for each MSI by statically allocating 2
>> MSI GIC IRQs for each X-Gene CPU core and moving MSI vectors around
>> these GIC IRQs to steer them to target CPU core. As a consequence,
>> the total MSI vectors that X-Gene v1 supports is reduced to 256.
>>
>> v4 changes:
>> 1. Remove affinity setting for each MSI
>> 2. Add description about register layout, MSI termination address
>> and data
>> 3. Correct total number of MSI vectors to 2048
>> 4. Clean up error messages
>> 5. Remove unused module code
>>
>> v3 changes:
>> 1. Implement MSI support using PCI MSI IRQ domain
>> 2. Only use msi_controller to store IRQ domain
>>
>> v2 changes:
>> 1. Use msi_controller structure
>> 2. Remove arch hooks arch_teardown_msi_irqs and arch_setup_msi_irqs
>>
>> .../devicetree/bindings/pci/xgene-pci-msi.txt | 68 +++
>> MAINTAINERS | 8 +
>> arch/arm64/boot/dts/apm/apm-storm.dtsi | 27 +
>> drivers/pci/host/Kconfig | 10 +
>> drivers/pci/host/Makefile | 1 +
>> drivers/pci/host/pci-xgene-msi.c | 595 +++++++++++++++++++++
>> drivers/pci/host/pci-xgene.c | 21 +
>> 7 files changed, 730 insertions(+)
>> create mode 100644 Documentation/devicetree/bindings/pci/xgene-pci-msi.txt
>> create mode 100644 drivers/pci/host/pci-xgene-msi.c
>
> I applied these to pci/host-xgene for v4.2, thanks!
>
> I squashed the doc, driver, and maintainer updates into a single patch with
> the following changelog:
>
> PCI: xgene: Add APM X-Gene v1 PCIe MSI/MSIX termination driver
>
> APM X-Gene v1 SoC supports its own implementation of MSI, which is not
> compliant to GIC V2M specification for MSI Termination.
>
> There is a single MSI block in X-Gene v1 SOC which serves all 5 PCIe ports.
> This MSI block supports 2048 MSI termination ports coalesced into 16
> physical HW IRQ lines and shared across all 5 PCIe ports.
>
> As there are only 16 HW IRQs to serve 2048 MSI vectors, to support
> set_affinity correctly for each MSI vectors, the 16 HW IRQs are statically
> allocated to 8 X-Gene v1 cores (2 HW IRQs for each cores). To steer MSI
> interrupt to target CPU, MSI vector is moved around these HW IRQs lines.
> With this approach, the total MSI vectors this driver supports is reduced
> to 256.
>
> [bhelgaas: squash doc, driver, maintainer update]
> Signed-off-by: Duc Dang <dhdang at apm.com>
> Signed-off-by: Tanmay Inamdar <tinamdar at apm.com>
> Signed-off-by: Bjorn Helgaas <bhelgaas at google.com>
> Reviewed-by: Marc Zyngier <marc.zyngier at arm.com>
>
Thanks, Bjorn and Marc.
Regards,
Duc Dang.
More information about the linux-arm-kernel
mailing list