[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v2] xen/pt: Fix a xen passthrough failure
From: |
Roger Pau Monné |
Subject: |
Re: [Qemu-devel] [PATCH v2] xen/pt: Fix a xen passthrough failure |
Date: |
Mon, 3 Dec 2018 12:20:30 +0100 |
User-agent: |
NeoMutt/20180716 |
Hello,
Thanks for the patch.
The subject should be more descriptive, "Fix a xen passthrough
failure" is too generic. How about: "allow passthrough of devices with
bogus interrupt pin" or something similar.
On Mon, Dec 03, 2018 at 12:04:38AM -0500, Zhao Yan wrote:
> For some pci device, even its PCI_INTERRUPT_PIN is not 0, it actually
> doesn't support INTx mode, so its machine irq read from host sysfs is 0.
> In that case, report PCI_INTERRUPT_PIN as 0 to guest and let passthrough
> continue.
>
> v2: fix some coding style issue
The changelog between versions should be below the '---'.
>
> Cc: Roger Pau Monné <address@hidden>
> Cc: Jan Beulich <address@hidden>
> Signed-off-by: Zhao Yan <address@hidden>
> ---
> hw/xen/xen_pt.c | 5 +++++
> hw/xen/xen_pt_config_init.c | 8 +++++++-
> 2 files changed, 12 insertions(+), 1 deletion(-)
>
> diff --git a/hw/xen/xen_pt.c b/hw/xen/xen_pt.c
> index f1f3a3727c..d601c9979c 100644
> --- a/hw/xen/xen_pt.c
> +++ b/hw/xen/xen_pt.c
> @@ -847,6 +847,11 @@ static void xen_pt_realize(PCIDevice *d, Error **errp)
> }
>
> machine_irq = s->real_device.irq;
> + if (machine_irq == 0) {
> + XEN_PT_LOG(d, "machine irq is 0\n");
I would maybe consider disabling INTX assertion here on the command
register.
> + goto out;
> + }
> +
> rc = xc_physdev_map_pirq(xen_xc, xen_domid, machine_irq, &pirq);
> if (rc < 0) {
> error_setg_errno(errp, errno, "Mapping machine irq %u to"
> diff --git a/hw/xen/xen_pt_config_init.c b/hw/xen/xen_pt_config_init.c
> index 47f9010c75..1007b6c977 100644
> --- a/hw/xen/xen_pt_config_init.c
> +++ b/hw/xen/xen_pt_config_init.c
> @@ -300,7 +300,13 @@ static int xen_pt_irqpin_reg_init(XenPCIPassthroughState
> *s,
> XenPTRegInfo *reg, uint32_t real_offset,
> uint32_t *data)
> {
> - *data = xen_pt_pci_read_intx(s);
> + if (s->real_device.irq)
> + *data = xen_pt_pci_read_intx(s);
> + else {
> + XEN_PT_LOG(&s->dev,
> + "machine irq is 0, init guest PCI_INTERRUPT_PIN to 0\n");
> + *data = 0;
The default value for the register is already zero, so you could drop
the else branch AFAICT.
Thanks, Roger.