[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v5 09/17] spapr_pci: add PHB unrealize
From: |
David Gibson |
Subject: |
Re: [Qemu-devel] [PATCH v5 09/17] spapr_pci: add PHB unrealize |
Date: |
Wed, 20 Feb 2019 14:26:42 +1100 |
User-agent: |
Mutt/1.10.1 (2018-07-13) |
On Tue, Feb 19, 2019 at 06:18:18PM +0100, Greg Kurz wrote:
> To support PHB hotplug we need to clean up lingering references,
> memory, child properties, etc. prior to the PHB object being
> finalized. Generally this will be called as a result of calling
> object_unparent() on the PHB object, which in turn would normally
> be called as the result of an unplug() operation.
>
> When the PHB is finalized, child objects will be unparented in
> turn, and finalized if the PHB was the only reference holder. so
> we don't bother to explicitly unparent child objects of the PHB,
> with the notable exception of DRCs. This is needed to avoid a QEMU
> crash when unplugging a PHB and resetting the machine before the
> guest could handle the event. The DRCs are removed from the QOM tree
> by pci_unregister_root_bus() and we must make sure we're not leaving
> stale aliases under the global /dr-connector path.
>
> The formula that gives the number of DMA windows is moved to an
> inline function in the hw/pci-host/spapr.h header because it
> will have other users.
>
> The unrealize function is able to cope with partially realized PHBs.
> It is hence used to implement proper rollback on the realize error
> path.
>
> Signed-off-by: Michael Roth <address@hidden>
> Signed-off-by: Greg Kurz <address@hidden>
> Reviewed-by: David Gibson <address@hidden>
Applied, thanks.
> ---
> v5: - unparent child DRCs at unrealize
> v4: - reverted to v2
> v3: - don't free LSIs at unrealize
> v2: - implement rollback with unrealize function
> ---
> ---
> hw/ppc/spapr_pci.c | 86
> +++++++++++++++++++++++++++++++++++++++++--
> include/hw/pci-host/spapr.h | 5 +++
> 2 files changed, 87 insertions(+), 4 deletions(-)
>
> diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c
> index e2bc9fec824b..ede928b0bff3 100644
> --- a/hw/ppc/spapr_pci.c
> +++ b/hw/ppc/spapr_pci.c
> @@ -1570,6 +1570,75 @@ static void spapr_pci_unplug_request(HotplugHandler
> *plug_handler,
> }
> }
>
> +static void spapr_phb_finalizefn(Object *obj)
> +{
> + sPAPRPHBState *sphb = SPAPR_PCI_HOST_BRIDGE(obj);
> +
> + g_free(sphb->dtbusname);
> + sphb->dtbusname = NULL;
> +}
> +
> +static void spapr_phb_unrealize(DeviceState *dev, Error **errp)
> +{
> + sPAPRMachineState *spapr = SPAPR_MACHINE(qdev_get_machine());
> + SysBusDevice *s = SYS_BUS_DEVICE(dev);
> + PCIHostState *phb = PCI_HOST_BRIDGE(s);
> + sPAPRPHBState *sphb = SPAPR_PCI_HOST_BRIDGE(phb);
> + sPAPRTCETable *tcet;
> + int i;
> + const unsigned windows_supported = spapr_phb_windows_supported(sphb);
> +
> + if (sphb->msi) {
> + g_hash_table_unref(sphb->msi);
> + sphb->msi = NULL;
> + }
> +
> + /*
> + * Remove IO/MMIO subregions and aliases, rest should get cleaned
> + * via PHB's unrealize->object_finalize
> + */
> + for (i = windows_supported - 1; i >= 0; i--) {
> + tcet = spapr_tce_find_by_liobn(sphb->dma_liobn[i]);
> + if (tcet) {
> + memory_region_del_subregion(&sphb->iommu_root,
> + spapr_tce_get_iommu(tcet));
> + }
> + }
> +
> + if (sphb->dr_enabled) {
> + for (i = PCI_SLOT_MAX * 8 - 1; i >= 0; i--) {
> + sPAPRDRConnector *drc = spapr_drc_by_id(TYPE_SPAPR_DRC_PCI,
> + (sphb->index << 16) | i);
> +
> + if (drc) {
> + object_unparent(OBJECT(drc));
> + }
> + }
> + }
> +
> + for (i = PCI_NUM_PINS - 1; i >= 0; i--) {
> + if (sphb->lsi_table[i].irq) {
> + spapr_irq_free(spapr, sphb->lsi_table[i].irq, 1);
> + sphb->lsi_table[i].irq = 0;
> + }
> + }
> +
> + QLIST_REMOVE(sphb, list);
> +
> + memory_region_del_subregion(&sphb->iommu_root, &sphb->msiwindow);
> +
> + address_space_destroy(&sphb->iommu_as);
> +
> + qbus_set_hotplug_handler(BUS(phb->bus), NULL, &error_abort);
> + pci_unregister_root_bus(phb->bus);
> +
> + memory_region_del_subregion(get_system_memory(), &sphb->iowindow);
> + if (sphb->mem64_win_pciaddr != (hwaddr)-1) {
> + memory_region_del_subregion(get_system_memory(), &sphb->mem64window);
> + }
> + memory_region_del_subregion(get_system_memory(), &sphb->mem32window);
> +}
> +
> static void spapr_phb_realize(DeviceState *dev, Error **errp)
> {
> /* We don't use SPAPR_MACHINE() in order to exit gracefully if the user
> @@ -1587,8 +1656,7 @@ static void spapr_phb_realize(DeviceState *dev, Error
> **errp)
> PCIBus *bus;
> uint64_t msi_window_size = 4096;
> sPAPRTCETable *tcet;
> - const unsigned windows_supported =
> - sphb->ddw_enabled ? SPAPR_PCI_DMA_MAX_WINDOWS : 1;
> + const unsigned windows_supported = spapr_phb_windows_supported(sphb);
>
> if (!spapr) {
> error_setg(errp, TYPE_SPAPR_PCI_HOST_BRIDGE " needs a pseries
> machine");
> @@ -1745,6 +1813,10 @@ static void spapr_phb_realize(DeviceState *dev, Error
> **errp)
> if (local_err) {
> error_propagate_prepend(errp, local_err,
> "can't allocate LSIs: ");
> + /*
> + * Older machines will never support PHB hotplug, ie, this
> is an
> + * init only path and QEMU will terminate. No need to
> rollback.
> + */
> return;
> }
> }
> @@ -1752,7 +1824,7 @@ static void spapr_phb_realize(DeviceState *dev, Error
> **errp)
> spapr_irq_claim(spapr, irq, true, &local_err);
> if (local_err) {
> error_propagate_prepend(errp, local_err, "can't allocate LSIs:
> ");
> - return;
> + goto unrealize;
> }
>
> sphb->lsi_table[i].irq = irq;
> @@ -1772,13 +1844,17 @@ static void spapr_phb_realize(DeviceState *dev, Error
> **errp)
> if (!tcet) {
> error_setg(errp, "Creating window#%d failed for %s",
> i, sphb->dtbusname);
> - return;
> + goto unrealize;
> }
> memory_region_add_subregion(&sphb->iommu_root, 0,
> spapr_tce_get_iommu(tcet));
> }
>
> sphb->msi = g_hash_table_new_full(g_int_hash, g_int_equal, g_free,
> g_free);
> + return;
> +
> +unrealize:
> + spapr_phb_unrealize(dev, NULL);
> }
>
> static int spapr_phb_children_reset(Object *child, void *opaque)
> @@ -1977,6 +2053,7 @@ static void spapr_phb_class_init(ObjectClass *klass,
> void *data)
>
> hc->root_bus_path = spapr_phb_root_bus_path;
> dc->realize = spapr_phb_realize;
> + dc->unrealize = spapr_phb_unrealize;
> dc->props = spapr_phb_properties;
> dc->reset = spapr_phb_reset;
> dc->vmsd = &vmstate_spapr_pci;
> @@ -1992,6 +2069,7 @@ static const TypeInfo spapr_phb_info = {
> .name = TYPE_SPAPR_PCI_HOST_BRIDGE,
> .parent = TYPE_PCI_HOST_BRIDGE,
> .instance_size = sizeof(sPAPRPHBState),
> + .instance_finalize = spapr_phb_finalizefn,
> .class_init = spapr_phb_class_init,
> .interfaces = (InterfaceInfo[]) {
> { TYPE_HOTPLUG_HANDLER },
> diff --git a/include/hw/pci-host/spapr.h b/include/hw/pci-host/spapr.h
> index f6e43f48fefa..4b0443f4cfe4 100644
> --- a/include/hw/pci-host/spapr.h
> +++ b/include/hw/pci-host/spapr.h
> @@ -165,4 +165,9 @@ static inline void spapr_phb_vfio_reset(DeviceState *qdev)
>
> void spapr_phb_dma_reset(sPAPRPHBState *sphb);
>
> +static inline unsigned spapr_phb_windows_supported(sPAPRPHBState *sphb)
> +{
> + return sphb->ddw_enabled ? SPAPR_PCI_DMA_MAX_WINDOWS : 1;
> +}
> +
> #endif /* PCI_HOST_SPAPR_H */
>
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
signature.asc
Description: PGP signature
- [Qemu-devel] [PATCH v5 05/17] spapr/drc: Drop spapr_drc_attach() fdt argument, (continued)
- [Qemu-devel] [PATCH v5 05/17] spapr/drc: Drop spapr_drc_attach() fdt argument, Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 06/17] xics: Write source state to KVM at claim time, Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 02/17] spapr: Generate FDT fragment for LMBs at configure connector time, Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 04/17] spapr/pci: Generate FDT fragment at configure connector time, Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 07/17] spapr: Expose the name of the interrupt controller node, Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 09/17] spapr_pci: add PHB unrealize, Greg Kurz, 2019/02/19
- Re: [Qemu-devel] [PATCH v5 09/17] spapr_pci: add PHB unrealize,
David Gibson <=
- [Qemu-devel] [PATCH v5 08/17] spapr_irq: Expose the phandle of the interrupt controller, Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 11/17] spapr: populate PHB DRC entries for root DT node, Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 10/17] spapr: create DR connectors for PHBs, Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 13/17] spapr_pci: provide node start offset via spapr_populate_pci_dt(), Greg Kurz, 2019/02/19
- [Qemu-devel] [PATCH v5 12/17] spapr_events: add support for phb hotplug events, Greg Kurz, 2019/02/19