On Tue, Feb 18, 2025 at 6:02 AM Philippe Mathieu-Daudé <phi...@linaro.org>
wrote:
> Hi Edgar,
>
> On 4/9/24 18:15, Edgar E. Iglesias wrote:
> > From: "Edgar E. Iglesias" <edgar.igles...@amd.com>
> >
> > Add support for optionally creating a PCIe/GPEX controller.
> >
> > Signed-off-by: Edgar E. Iglesias <edgar.igles...@amd.com>
> > Reviewed-by: Stefano Stabellini <sstabell...@kernel.org>
> > ---
> > hw/xen/xen-pvh-common.c | 76 +++++++++++++++++++++++++++++++++
> > include/hw/xen/xen-pvh-common.h | 29 +++++++++++++
> > 2 files changed, 105 insertions(+)
>
>
> > +/*
> > + * We use the GPEX PCIe controller with its internal INTX PCI interrupt
> > + * swizzling. This swizzling is emulated in QEMU and routes all INTX
> > + * interrupts from endpoints down to only 4 INTX interrupts.
> > + * See include/hw/pci/pci.h : pci_swizzle()
> > + */
> > +static inline void xenpvh_gpex_init(XenPVHMachineState *s,
> > + XenPVHMachineClass *xpc,
> > + MemoryRegion *sysmem)
> > +{
> > + MemoryRegion *ecam_reg;
> > + MemoryRegion *mmio_reg;
> > + DeviceState *dev;
> > + int i;
> > +
> > + object_initialize_child(OBJECT(s), "gpex", &s->pci.gpex,
> > + TYPE_GPEX_HOST);
> > + dev = DEVICE(&s->pci.gpex);
> > + sysbus_realize_and_unref(SYS_BUS_DEVICE(dev), &error_fatal);
> > +
> > + ecam_reg = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0);
> > + memory_region_add_subregion(sysmem, s->cfg.pci_ecam.base, ecam_reg);
> > +
> > + mmio_reg = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 1);
> > +
> > + if (s->cfg.pci_mmio.size) {
> > + memory_region_init_alias(&s->pci.mmio_alias, OBJECT(dev),
> "pcie-mmio",
> > + mmio_reg,
> > + s->cfg.pci_mmio.base,
> s->cfg.pci_mmio.size);
> > + memory_region_add_subregion(sysmem, s->cfg.pci_mmio.base,
> > + &s->pci.mmio_alias);
> > + }
> > +
> > + if (s->cfg.pci_mmio_high.size) {
> > + memory_region_init_alias(&s->pci.mmio_high_alias, OBJECT(dev),
> > + "pcie-mmio-high",
> > + mmio_reg, s->cfg.pci_mmio_high.base,
> s->cfg.pci_mmio_high.size);
> > + memory_region_add_subregion(sysmem, s->cfg.pci_mmio_high.base,
> > + &s->pci.mmio_high_alias);
> > + }
> > +
> > + /*
> > + * PVH implementations with PCI enabled must provide
> set_pci_intx_irq()
> > + * and optionally an implementation of set_pci_link_route().
> > + */
> > + assert(xpc->set_pci_intx_irq);
> > +
> > + for (i = 0; i < GPEX_NUM_IRQS; i++) {
> > + qemu_irq irq = qemu_allocate_irq(xpc->set_pci_intx_irq, s, i);
> > +
> > + sysbus_connect_irq(SYS_BUS_DEVICE(dev), i, irq);
> > + gpex_set_irq_num(GPEX_HOST(dev), i, s->cfg.pci_intx_irq_base +
> i);
> > + if (xpc->set_pci_link_route) {
> > + xpc->set_pci_link_route(i, s->cfg.pci_intx_irq_base + i);
> > + }
> > + }
> > +}
>
> Some Kconfig selector seems missing here:
>
> /usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-pvh-common.c.o: in
> function `xenpvh_gpex_init':
> hw/xen/xen-pvh-common.c:174: undefined reference to `gpex_set_irq_num'
> /usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-hvm-common.c.o: in
> function `pci_dev_bus_num':
> include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-hvm-common.c.o: in
> function `cpu_ioreq_config':
> hw/xen/xen-hvm-common.c:412: undefined reference to
> `pci_host_config_read_common'
> /usr/bin/ld: hw/xen/xen-hvm-common.c:428: undefined reference to
> `pci_host_config_read_common'
> /usr/bin/ld: hw/xen/xen-hvm-common.c:438: undefined reference to
> `pci_host_config_write_common'
>
> The current 'XEN' key represents both the "accelerator" part and
> the common Xen HW, which isn't helping to follow. Anyhow, this
> snippet fixes the build issue:
>
> -- >8 --
> diff --git a/accel/Kconfig b/accel/Kconfig
> index 794e0d18d2..4263cab722 100644
> --- a/accel/Kconfig
> +++ b/accel/Kconfig
> @@ -16,4 +16,5 @@ config KVM
> config XEN
> bool
> select FSDEV_9P if VIRTFS
> + select PCI_EXPRESS_GENERIC_BRIDGE
> select XEN_BUS
> ---
>
> I'll post a patch later.
>
Sounds good, thanks Phil!
Cheers,
Edgar
>
> Regards,
>
> Phil.
>
