From: Jean-Philippe Brucker <jean-philippe@linaro.org>
To: Eric Auger <eric.auger@redhat.com>
Cc: peter.maydell@linaro.org, mst@redhat.com, qemu-devel@nongnu.org,
peterx@redhat.com, armbru@redhat.com, qemu-arm@nongnu.org,
pbonzini@redhat.com, bbhushan2@marvell.com,
eric.auger.pro@gmail.com
Subject: Re: [PATCH v2 5/5] hw/arm/virt: Let the virtio-iommu bypass MSIs
Date: Fri, 22 May 2020 16:43:17 +0200 [thread overview]
Message-ID: <20200522144317.GB3453945@myrica> (raw)
In-Reply-To: <20200508173057.32215-6-eric.auger@redhat.com>
On Fri, May 08, 2020 at 07:30:57PM +0200, Eric Auger wrote:
> At the moment the virtio-iommu translates MSI transactions.
> This behavior is inherited from ARM SMMU. The virt machine
> code knows where the guest MSI doorbells are so we can easily
> declare those regions as VIRTIO_IOMMU_RESV_MEM_T_MSI. With that
> setting the guest will not map MSIs through the IOMMU and those
> transactions will be simply bypassed.
>
> Depending on which MSI controller is in use (ITS or GICV2M),
> we declare either:
> - the ITS interrupt translation space (ITS_base + 0x10000),
> containing the GITS_TRANSLATOR or
> - The GICV2M single frame, containing the MSI_SETSP_NS register.
>
> Signed-off-by: Eric Auger <eric.auger@redhat.com>
>
> ---
>
> v1 -> v2:
> - Test which MSI controller is instantiated
> - If GICV2M is in use, declare its doorbell as an MSI doorbell too
> ---
> include/hw/arm/virt.h | 6 ++++++
> hw/arm/virt.c | 18 ++++++++++++++++++
> 2 files changed, 24 insertions(+)
>
> diff --git a/include/hw/arm/virt.h b/include/hw/arm/virt.h
> index 6d67ace76e..ad20cb6e15 100644
> --- a/include/hw/arm/virt.h
> +++ b/include/hw/arm/virt.h
> @@ -96,6 +96,11 @@ typedef enum VirtIOMMUType {
> VIRT_IOMMU_VIRTIO,
> } VirtIOMMUType;
>
> +typedef enum VirtMSIControllerType {
> + VIRT_GICV2M,
> + VIRT_ITS,
> +} VirtMSIControllerType;
I think you need a third value for msi_controller == 0. If I instantiate a
GICv3 without ITS at the moment the V2M region gets reserved. Not a big
deal since MSIs aren't supported at all in this case, but it would be
cleaner to skip any reservation.
Thanks,
Jean
> +
> typedef enum VirtGICType {
> VIRT_GIC_VERSION_MAX,
> VIRT_GIC_VERSION_HOST,
> @@ -135,6 +140,7 @@ typedef struct {
> OnOffAuto acpi;
> VirtGICType gic_version;
> VirtIOMMUType iommu;
> + VirtMSIControllerType msi_controller;
> uint16_t virtio_iommu_bdf;
> struct arm_boot_info bootinfo;
> MemMapEntry *memmap;
> diff --git a/hw/arm/virt.c b/hw/arm/virt.c
> index 634db0cfe9..d2dd07885b 100644
> --- a/hw/arm/virt.c
> +++ b/hw/arm/virt.c
> @@ -602,6 +602,7 @@ static void create_its(VirtMachineState *vms)
> sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, vms->memmap[VIRT_GIC_ITS].base);
>
> fdt_add_its_gic_node(vms);
> + vms->msi_controller = VIRT_ITS;
> }
>
> static void create_v2m(VirtMachineState *vms)
> @@ -622,6 +623,7 @@ static void create_v2m(VirtMachineState *vms)
> }
>
> fdt_add_v2m_gic_node(vms);
> + vms->msi_controller = VIRT_GICV2M;
> }
>
> static void create_gic(VirtMachineState *vms)
> @@ -2136,8 +2138,24 @@ out:
> static void virt_machine_device_pre_plug_cb(HotplugHandler *hotplug_dev,
> DeviceState *dev, Error **errp)
> {
> + VirtMachineState *vms = VIRT_MACHINE(hotplug_dev);
> +
> if (object_dynamic_cast(OBJECT(dev), TYPE_PC_DIMM)) {
> virt_memory_pre_plug(hotplug_dev, dev, errp);
> + } else if (object_dynamic_cast(OBJECT(dev), TYPE_VIRTIO_IOMMU_PCI)) {
> + /* we declare a VIRTIO_IOMMU_RESV_MEM_T_MSI region */
> +
> + if (vms->msi_controller == VIRT_ITS) {
> + /* GITS_TRANSLATER page */
> + qdev_prop_set_uint32(dev, "len-reserved-regions", 1);
> + qdev_prop_set_string(dev, "reserved-regions[0]",
> + "0x8090000, 0x809FFFF, 1");
> + } else if (vms->msi_controller == VIRT_GICV2M) {
> + /* MSI_SETSPI_NS page */
> + qdev_prop_set_uint32(dev, "len-reserved-regions", 1);
> + qdev_prop_set_string(dev, "reserved-regions[0]",
> + "0x8020000, 0x8020FFF, 1");
> + }
> }
> }
>
> --
> 2.20.1
>
prev parent reply other threads:[~2020-05-22 14:44 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-08 17:30 [PATCH v2 0/5] VIRTIO-IOMMU probe request support and MSI bypass on ARM Eric Auger
2020-05-08 17:30 ` [PATCH v2 1/5] qdev: Introduce DEFINE_PROP_RESERVED_REGION Eric Auger
2020-05-08 17:30 ` [PATCH v2 2/5] virtio-iommu: Implement RESV_MEM probe request Eric Auger
2020-05-11 6:38 ` [EXT] " Bharat Bhushan
2020-05-11 6:55 ` Auger Eric
2020-05-11 8:42 ` Bharat Bhushan
2020-05-11 8:49 ` Auger Eric
2020-05-12 3:03 ` Bharat Bhushan
2020-05-12 3:08 ` Auger Eric
2020-05-12 3:11 ` Bharat Bhushan
2020-05-08 17:30 ` [PATCH v2 3/5] virtio-iommu: Handle reserved regions in the translation process Eric Auger
2020-05-11 21:11 ` Peter Xu
2020-05-08 17:30 ` [PATCH v2 4/5] virtio-iommu-pci: Add array of Interval properties Eric Auger
2020-05-08 17:30 ` [PATCH v2 5/5] hw/arm/virt: Let the virtio-iommu bypass MSIs Eric Auger
2020-05-22 14:43 ` Jean-Philippe Brucker [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200522144317.GB3453945@myrica \
--to=jean-philippe@linaro.org \
--cc=armbru@redhat.com \
--cc=bbhushan2@marvell.com \
--cc=eric.auger.pro@gmail.com \
--cc=eric.auger@redhat.com \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=peterx@redhat.com \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).