[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [virtio-dev] Re: [PATCH v2 3/4] Add "Group Identifier"
From: |
Venu Busireddy |
Subject: |
Re: [Qemu-devel] [virtio-dev] Re: [PATCH v2 3/4] Add "Group Identifier" support to Red Hat PCI bridge. |
Date: |
Wed, 27 Jun 2018 22:46:16 -0500 |
User-agent: |
Mutt/1.9.2 (2017-12-15) |
On 2018-06-28 05:14:50 +0300, Michael S. Tsirkin wrote:
> On Wed, Jun 27, 2018 at 06:07:59PM -0500, Venu Busireddy wrote:
> > On 2018-06-26 23:08:12 -0500, Venu Busireddy wrote:
> > > On 2018-06-27 07:02:36 +0300, Michael S. Tsirkin wrote:
> > > > On Tue, Jun 26, 2018 at 10:49:33PM -0500, Venu Busireddy wrote:
> > > > > Add the "Vendor-Specific" capability to the Red Hat PCI bridge device
> > > > > "pci-bridge", to contain the "Group Identifier" (UUID) that will be
> > > > > used to pair a virtio device with the passthrough device attached to
> > > > > that bridge.
> > > > >
> > > > > This capability is added to the bridge iff the "uuid" option is
> > > > > specified
> > > > > for the bridge.
> > > >
> > > > I think the name should be more explicit. How about "failover-group-id"?
> > >
> > > I can change it. But don't you think it is bit long?
> > >
> > > > >
> > > > > Signed-off-by: Venu Busireddy <address@hidden>
> > > >
> > > > I'd like to also tweak the device id in this case,
> > > > to make it easier for guests to know it's a grouping bridge.
> > >
> > > Could you please recommend a name for the new ID'd definition? Something
> > > in lines of PCI_DEVICE_ID_REDHAT_<blah blah>.
> >
> > How about these names and values for the device IDs?
> >
> > PCI_DEVICE_ID_REDHAT_PCI_BRIDGE_GRP (0x0010) for pci-bridge, and
>
> You do not need the second PCI, and group is one of the
> functions of bridge anyway.
>
> PCI_DEVICE_ID_REDHAT_BRIDGE_FAILOVER?
Sounds good.
> > PCI_DEVICE_ID_REDHAT_PCIE_BRIDGE_GRP (0x0011) for pcie-downstream
> >
> > Thanks,
> >
> > Venu
>
> PCI_DEVICE_ID_REDHAT_DOWNPORT_FAILOVER?
Sounds good.
> It's becoming annoying though - we'll need one for each type :(
>
> Ideas/options:
> - use revision ID to distinguish from regular bridge
What would we use for the revision? For example, "pci-bridge" (default
one) currently uses revision 0. Should we use revision 1 for the group
bridge? If we do that, what if in the future we need to up the revision
for the default bridge?
> - use device serial # cap for the bridge when it's an express device
We could do that. But if we are defining a new Device ID for the PCI
case, we might as well do so for the PCIe case too, and be consistent?
The other advantage in defining a new ID is, 'lspci' could right away
tell what type of bridge it is (if we update PCI IDs). No need to map
the revision to a bridge type!
Unless you have strong reservations against, I think the notion of
different Device ID for the grouping bridge sounds cleaner. Do you agree?
Venu
>
> > > > > ---
> > > > > hw/pci-bridge/pci_bridge_dev.c | 8 ++++++++
> > > > > hw/pci/pci_bridge.c | 26 ++++++++++++++++++++++++++
> > > > > include/hw/pci/pcie.h | 1 +
> > > > > 3 files changed, 35 insertions(+)
> > > > >
> > > > > diff --git a/hw/pci-bridge/pci_bridge_dev.c
> > > > > b/hw/pci-bridge/pci_bridge_dev.c
> > > > > index b2d861d216..bbbc6fa1c6 100644
> > > > > --- a/hw/pci-bridge/pci_bridge_dev.c
> > > > > +++ b/hw/pci-bridge/pci_bridge_dev.c
> > > > > @@ -71,6 +71,12 @@ static void pci_bridge_dev_realize(PCIDevice *dev,
> > > > > Error **errp)
> > > > > bridge_dev->msi = ON_OFF_AUTO_OFF;
> > > > > }
> > > > >
> > > > > + err = pci_bridge_vendor_init(dev, 0, errp);
> > > > > + if (err < 0) {
> > > > > + error_append_hint(errp, "Can't init group ID, error %d\n",
> > > > > err);
> > > > > + goto vendor_cap_err;
> > > > > + }
> > > > > +
> > > > > err = slotid_cap_init(dev, 0, bridge_dev->chassis_nr, 0, errp);
> > > > > if (err) {
> > > > > goto slotid_error;
> > > > > @@ -109,6 +115,7 @@ slotid_error:
> > > > > if (shpc_present(dev)) {
> > > > > shpc_cleanup(dev, &bridge_dev->bar);
> > > > > }
> > > > > +vendor_cap_err:
> > > > > shpc_error:
> > > > > pci_bridge_exitfn(dev);
> > > > > }
> > > > > @@ -162,6 +169,7 @@ static Property pci_bridge_dev_properties[] = {
> > > > > ON_OFF_AUTO_AUTO),
> > > > > DEFINE_PROP_BIT(PCI_BRIDGE_DEV_PROP_SHPC, PCIBridgeDev, flags,
> > > > > PCI_BRIDGE_DEV_F_SHPC_REQ, true),
> > > > > + DEFINE_PROP_UUID(COMPAT_PROP_UUID, PCIDevice, uuid, false),
> > > > > DEFINE_PROP_END_OF_LIST(),
> > > > > };
> > > > >
> > > > > diff --git a/hw/pci/pci_bridge.c b/hw/pci/pci_bridge.c
> > > > > index 40a39f57cb..cb8b3dad2a 100644
> > > > > --- a/hw/pci/pci_bridge.c
> > > > > +++ b/hw/pci/pci_bridge.c
> > > > > @@ -34,12 +34,17 @@
> > > > > #include "hw/pci/pci_bus.h"
> > > > > #include "qemu/range.h"
> > > > > #include "qapi/error.h"
> > > > > +#include "qemu/uuid.h"
> > > > >
> > > > > /* PCI bridge subsystem vendor ID helper functions */
> > > > > #define PCI_SSVID_SIZEOF 8
> > > > > #define PCI_SSVID_SVID 4
> > > > > #define PCI_SSVID_SSID 6
> > > > >
> > > > > +#define PCI_VENDOR_SIZEOF 20
> > > > > +#define PCI_VENDOR_CAP_LEN_OFFSET 2
> > > > > +#define PCI_VENDOR_GROUP_ID_OFFSET 4
> > > > > +
> > > > > int pci_bridge_ssvid_init(PCIDevice *dev, uint8_t offset,
> > > > > uint16_t svid, uint16_t ssid,
> > > > > Error **errp)
> > > > > @@ -57,6 +62,27 @@ int pci_bridge_ssvid_init(PCIDevice *dev, uint8_t
> > > > > offset,
> > > > > return pos;
> > > > > }
> > > > >
> > > > > +int pci_bridge_vendor_init(PCIDevice *d, uint8_t offset, Error
> > > > > **errp)
> > > > > +{
> > > > > + int pos;
> > > > > +
> > > > > + if (qemu_uuid_is_null(&d->uuid)) {
> > > > > + return 0;
> > > > > + }
> > > > > +
> > > > > + pos = pci_add_capability(d, PCI_CAP_ID_VNDR, offset,
> > > > > PCI_VENDOR_SIZEOF,
> > > > > + errp);
> > > > > + if (pos < 0) {
> > > > > + return pos;
> > > > > + }
> > > > > +
> > > > > + pci_set_word(d->config + pos + PCI_VENDOR_CAP_LEN_OFFSET,
> > > > > + PCI_VENDOR_SIZEOF);
> > > > > + memcpy(d->config + pos + PCI_VENDOR_GROUP_ID_OFFSET, &d->uuid,
> > > > > + sizeof(QemuUUID));
> > > > > + return pos;
> > > > > +}
> > > > > +
> > > > > /* Accessor function to get parent bridge device from pci bus. */
> > > > > PCIDevice *pci_bridge_get_device(PCIBus *bus)
> > > > > {
> > > > > diff --git a/include/hw/pci/pcie.h b/include/hw/pci/pcie.h
> > > > > index b71e369703..b4189d0ce3 100644
> > > > > --- a/include/hw/pci/pcie.h
> > > > > +++ b/include/hw/pci/pcie.h
> > > > > @@ -82,6 +82,7 @@ struct PCIExpressDevice {
> > > > > };
> > > > >
> > > > > #define COMPAT_PROP_PCP "power_controller_present"
> > > > > +#define COMPAT_PROP_UUID "uuid"
> > > > >
> > > > > /* PCI express capability helper functions */
> > > > > int pcie_cap_init(PCIDevice *dev, uint8_t offset, uint8_t type,
> > > >
> > > > ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: address@hidden
> > > > For additional commands, e-mail: address@hidden
> > > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: address@hidden
> > > For additional commands, e-mail: address@hidden
> > >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: address@hidden
> For additional commands, e-mail: address@hidden
>
[Qemu-devel] [PATCH v2 4/4] Add "Group Identifier" support to Red Hat PCI Express bridge., Venu Busireddy, 2018/06/26
[Qemu-devel] [PATCH v2 virtio 1/1] Add "Group Identifier" to virtio PCI capabilities., Venu Busireddy, 2018/06/26
Re: [Qemu-devel] [PATCH v2 0/4] Use of unique identifier for pairing virtio and passthrough devices..., Michael S. Tsirkin, 2018/06/27
Re: [Qemu-devel] [PATCH v2 0/4] Use of unique identifier for pairing virtio and passthrough devices..., Roman Kagan, 2018/06/27