[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 2/2] pci-assign: Add MSI affinity support
From: |
Michael S. Tsirkin |
Subject: |
Re: [Qemu-devel] [PATCH 2/2] pci-assign: Add MSI affinity support |
Date: |
Sun, 12 May 2013 14:01:36 +0300 |
On Thu, May 09, 2013 at 10:36:05AM -0600, Alex Williamson wrote:
> Track the last MSIMessage programmed so we can determine when it has
> changed and update the routing to the guest.
>
> Signed-off-by: Alex Williamson <address@hidden>
> ---
> hw/i386/kvm/pci-assign.c | 27 +++++++++++++++++++++++++++
> 1 file changed, 27 insertions(+)
>
> diff --git a/hw/i386/kvm/pci-assign.c b/hw/i386/kvm/pci-assign.c
> index 0f83a4c..e09265b 100644
> --- a/hw/i386/kvm/pci-assign.c
> +++ b/hw/i386/kvm/pci-assign.c
> @@ -134,6 +134,7 @@ typedef struct AssignedDevice {
> int msi_nr;
> struct {
> int virq;
> + MSIMessage msg;
> } *msi;
> MSIXTableEntry *msix_table;
> hwaddr msix_table_addr;
> @@ -1015,6 +1016,7 @@ static void assigned_dev_update_msi(PCIDevice *pci_dev)
> assigned_dev->msi = g_malloc(sizeof(*assigned_dev->msi));
> assigned_dev->msi_nr = 1;
> assigned_dev->msi[0].virq = virq;
> + assigned_dev->msi[0].msg = msg;
> if (kvm_device_msi_assign(kvm_state, assigned_dev->dev_id, virq) <
> 0) {
> perror("assigned_dev_update_msi: kvm_device_msi_assign");
> }
> @@ -1027,6 +1029,27 @@ static void assigned_dev_update_msi(PCIDevice *pci_dev)
> }
> }
>
> +static void assigned_dev_update_msi_msg(PCIDevice *pci_dev)
> +{
> + AssignedDevice *assigned_dev = DO_UPCAST(AssignedDevice, dev, pci_dev);
> + uint8_t ctrl_byte = pci_get_byte(pci_dev->config + pci_dev->msi_cap +
> + PCI_MSI_FLAGS);
> + MSIMessage msg;
> +
> + if (assigned_dev->assigned_irq_type != ASSIGNED_IRQ_MSI ||
> + !(ctrl_byte & PCI_MSI_FLAGS_ENABLE)) {
> + return;
> + }
> +
> + msg = msi_get_message(pci_dev, 0);
> +
> + if (msg.address != assigned_dev->msi[0].msg.address ||
> + msg.data != assigned_dev->msi[0].msg.data) {
> + kvm_irqchip_update_msi_route(kvm_state, assigned_dev->msi[0].virq,
> msg);
> + assigned_dev->msi[0].msg = msg;
> + }
> +}
> +
Okay, but why do we track the message in AssignedDevice?
To avoid extra calls to kvm_irqchip_update_msi_route?
Let's not do premature optimizations please:
just call kvm_irqchip_update_msi_route directly,
no need to track the message in AssignedDevice.
Moreover, if you can measure a performance gain
from tracking msg in AssignedDevice there's a
better way to do this: we have an outstanding patch to kvm,
which optimized the no change case away to an array
lookup. This benefits msi-x as well.
> static bool assigned_dev_msix_masked(MSIXTableEntry *entry)
> {
> return (entry->ctrl & cpu_to_le32(0x1)) != 0;
> @@ -1202,6 +1225,10 @@ static void assigned_dev_pci_write_config(PCIDevice
> *pci_dev, uint32_t address,
> if (range_covers_byte(address, len,
> pci_dev->msi_cap + PCI_MSI_FLAGS)) {
> assigned_dev_update_msi(pci_dev);
> + } else if (ranges_overlap(address, len,
> + pci_dev->msi_cap + PCI_MSI_ADDRESS_LO,
> + 10 - PCI_MSI_ADDRESS_LO)) {
> + assigned_dev_update_msi_msg(pci_dev);
> }
> }
> if (assigned_dev->cap.available & ASSIGNED_DEVICE_CAP_MSIX) {