[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH] hw/pci: prevent hotplug of devices on pcie-root-ports with w
From: |
Michael S. Tsirkin |
Subject: |
Re: [PATCH] hw/pci: prevent hotplug of devices on pcie-root-ports with wrong devfn address |
Date: |
Wed, 14 Jun 2023 16:53:53 -0400 |
On Wed, Jun 14, 2023 at 10:09:35PM +0530, Ani Sinha wrote:
>
>
> > On 14-Jun-2023, at 7:52 PM, Michael S. Tsirkin <mst@redhat.com> wrote:
> >
> > On Wed, Jun 14, 2023 at 06:01:50PM +0530, Ani Sinha wrote:
> >> PCIE root ports only allow one device on slot 0/function 0.
> >
> >
> > Why do you say this? PCI devices can be multifunction.
>
> Yeah you are right, the language needs correction. For hot plugging, I
> believe only function 0 is relevant. At least that is what it looks like from
> the code.
Adding function 0 is what makes it visible to guest.
But you can't block multifunction devices that would be a regression.
> >
> >> When hotplugging a
> >> device on a pcie root port, make sure that the device address passed is
> >> always 0x00 that represents slot 0 and function 0. Any other slot value and
> >> function value would be illegal on a root port.
> >>
> >> CC: jusual@redhat.com
> >> CC: imammedo@redhat.com
> >> Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2128929
> >> Signed-off-by: Ani Sinha <anisinha@redhat.com>
> >> ---
> >> hw/pci/pci.c | 9 +++++++++
> >> 1 file changed, 9 insertions(+)
> >>
> >> Note:
> >> I tested this with both native and acpi hotplug enabled on pcie. The
> >> check seems to work on both.
> >>
> >> (qemu) netdev_add socket,id=hostnet1,listen=:1234
> >> (qemu) device_add
> >> e1000e,netdev=hostnet1,mac=00:11:22:33:44:03,id=net1,bus=pci.6,addr=0x2.0x5
> >> Error: PCI: slot 2 function 5 is not valid for e1000e
> >> (qemu) device_add
> >> e1000e,netdev=hostnet1,mac=00:11:22:33:44:03,id=net1,bus=pci.6,addr=0x0.0
> >> (qemu) info network
> >> net1: index=0,type=nic,model=e1000e,macaddr=00:11:22:33:44:03
> >> \ hostnet1: index=0,type=socket,
> >> (qemu) device_del net1
> >> (qemu) info network
> >> hostnet1: index=0,type=socket,
> >>
> >> diff --git a/hw/pci/pci.c b/hw/pci/pci.c
> >> index bf38905b7d..bc39322dfd 100644
> >> --- a/hw/pci/pci.c
> >> +++ b/hw/pci/pci.c
> >> @@ -64,6 +64,7 @@ bool pci_available = true;
> >> static char *pcibus_get_dev_path(DeviceState *dev);
> >> static char *pcibus_get_fw_dev_path(DeviceState *dev);
> >> static void pcibus_reset(BusState *qbus);
> >> +static bool pcie_has_upstream_port(PCIDevice *dev);
> >>
> >> static Property pci_props[] = {
> >> DEFINE_PROP_PCI_DEVFN("addr", PCIDevice, devfn, -1),
> >> @@ -1189,6 +1190,14 @@ static PCIDevice *do_pci_register_device(PCIDevice
> >> *pci_dev,
> >> name);
> >>
> >> return NULL;
> >> + } else if (dev->hotplugged &&
> >> + !pci_is_vf(pci_dev) &&
> >> + pcie_has_upstream_port(pci_dev) && devfn) {
> >> + /* For an upstream PCIE port, we only support one device at slot
> >> 0 */
> >> + error_setg(errp, "PCI: slot %d function %d is not valid for %s",
> >> + PCI_SLOT(devfn), PCI_FUNC(devfn), name);
> >> + return NULL;
> >> +
> >> }
> >>
> >> pci_dev->devfn = devfn;
> >> --
> >> 2.39.1
> >