[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device
From: |
Stefan Hajnoczi |
Subject: |
Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device |
Date: |
Tue, 29 Mar 2022 13:35:23 +0100 |
On Fri, Mar 25, 2022 at 03:19:41PM -0400, Jagannathan Raman wrote:
> Assign separate address space for each device in the remote processes.
>
> Signed-off-by: Elena Ufimtseva <elena.ufimtseva@oracle.com>
> Signed-off-by: John G Johnson <john.g.johnson@oracle.com>
> Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
> ---
> include/hw/remote/iommu.h | 18 ++++++++
> hw/remote/iommu.c | 95 +++++++++++++++++++++++++++++++++++++++
> MAINTAINERS | 2 +
> hw/remote/meson.build | 1 +
> 4 files changed, 116 insertions(+)
> create mode 100644 include/hw/remote/iommu.h
> create mode 100644 hw/remote/iommu.c
>
> diff --git a/include/hw/remote/iommu.h b/include/hw/remote/iommu.h
> new file mode 100644
> index 0000000000..8f850400f1
> --- /dev/null
> +++ b/include/hw/remote/iommu.h
> @@ -0,0 +1,18 @@
> +/**
> + * Copyright © 2022 Oracle and/or its affiliates.
> + *
> + * This work is licensed under the terms of the GNU GPL, version 2 or later.
> + * See the COPYING file in the top-level directory.
> + *
> + */
> +
> +#ifndef REMOTE_IOMMU_H
> +#define REMOTE_IOMMU_H
> +
> +#include "hw/pci/pci_bus.h"
> +
> +void remote_configure_iommu(PCIBus *pci_bus);
> +
> +void remote_iommu_del_device(PCIDevice *pci_dev);
> +
> +#endif
> diff --git a/hw/remote/iommu.c b/hw/remote/iommu.c
> new file mode 100644
> index 0000000000..13f329b45d
> --- /dev/null
> +++ b/hw/remote/iommu.c
> @@ -0,0 +1,95 @@
> +/**
> + * IOMMU for remote device
> + *
> + * Copyright © 2022 Oracle and/or its affiliates.
> + *
> + * This work is licensed under the terms of the GNU GPL, version 2 or later.
> + * See the COPYING file in the top-level directory.
> + *
> + */
> +
> +#include "qemu/osdep.h"
> +#include "qemu-common.h"
> +
> +#include "hw/remote/iommu.h"
> +#include "hw/pci/pci_bus.h"
> +#include "hw/pci/pci.h"
> +#include "exec/memory.h"
> +#include "exec/address-spaces.h"
> +#include "trace.h"
> +
> +struct RemoteIommuElem {
> + AddressSpace as;
> + MemoryRegion mr;
> +};
> +
> +struct RemoteIommuTable {
> + QemuMutex lock;
> + GHashTable *elem_by_bdf;
> +} remote_iommu_table;
> +
> +#define INT2VOIDP(i) (void *)(uintptr_t)(i)
> +
> +static AddressSpace *remote_iommu_find_add_as(PCIBus *pci_bus,
> + void *opaque, int devfn)
> +{
> + struct RemoteIommuTable *iommu_table = opaque;
> + struct RemoteIommuElem *elem = NULL;
> + int pci_bdf = PCI_BUILD_BDF(pci_bus_num(pci_bus), devfn);
> +
> + elem = g_hash_table_lookup(iommu_table->elem_by_bdf, INT2VOIDP(pci_bdf));
Why is a lock needed around g_hash_table_insert() below but no lock is
held around g_hash_table_lookup()?
Insertion isn't atomic because lookup and insert are separate operations
and they are not done under a single lock.
> +
> + if (!elem) {
> + g_autofree char *mr_name = g_strdup_printf("vfu-ram-%d", pci_bdf);
> + g_autofree char *as_name = g_strdup_printf("vfu-as-%d", pci_bdf);
> +
> + elem = g_malloc0(sizeof(struct RemoteIommuElem));
> +
> + memory_region_init(&elem->mr, NULL, mr_name, UINT64_MAX);
> + address_space_init(&elem->as, &elem->mr, as_name);
> +
> + qemu_mutex_lock(&iommu_table->lock);
> + g_hash_table_insert(iommu_table->elem_by_bdf, INT2VOIDP(pci_bdf),
> elem);
> + qemu_mutex_unlock(&iommu_table->lock);
> + }
> +
> + return &elem->as;
> +}
> +
> +static void remote_iommu_del_elem(gpointer data)
> +{
> + struct RemoteIommuElem *elem = data;
> +
> + g_assert(elem);
> +
> + memory_region_unref(&elem->mr);
> + address_space_destroy(&elem->as);
> +
> + g_free(elem);
> +}
> +
> +void remote_iommu_del_device(PCIDevice *pci_dev)
> +{
> + int pci_bdf;
> +
> + if (!remote_iommu_table.elem_by_bdf || !pci_dev) {
> + return;
> + }
> +
> + pci_bdf = PCI_BUILD_BDF(pci_bus_num(pci_get_bus(pci_dev)),
> pci_dev->devfn);
> +
> + qemu_mutex_lock(&remote_iommu_table.lock);
> + g_hash_table_remove(remote_iommu_table.elem_by_bdf, INT2VOIDP(pci_bdf));
> + qemu_mutex_unlock(&remote_iommu_table.lock);
> +}
> +
> +void remote_configure_iommu(PCIBus *pci_bus)
> +{
> + if (!remote_iommu_table.elem_by_bdf) {
> + remote_iommu_table.elem_by_bdf =
> + g_hash_table_new_full(NULL, NULL, NULL, remote_iommu_del_elem);
> + qemu_mutex_init(&remote_iommu_table.lock);
> + }
> +
> + pci_setup_iommu(pci_bus, remote_iommu_find_add_as, &remote_iommu_table);
Why is remote_iommu_table global? It could be per-PCIBus and indexed by
just devfn instead of the full BDF.
signature.asc
Description: PGP signature
- Re: [PATCH v7 08/17] vfio-user: instantiate vfio-user context, (continued)
- [PATCH v7 11/17] vfio-user: handle PCI config space accesses, Jagannathan Raman, 2022/03/25
- [PATCH v7 10/17] vfio-user: run vfio-user context, Jagannathan Raman, 2022/03/25
- [PATCH v7 07/17] vfio-user: define vfio-user-server object, Jagannathan Raman, 2022/03/25
- [PATCH v7 16/17] vfio-user: handle reset of remote device, Jagannathan Raman, 2022/03/25
- [PATCH v7 17/17] vfio-user: avocado tests for vfio-user, Jagannathan Raman, 2022/03/25
- [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Jagannathan Raman, 2022/03/25
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device,
Stefan Hajnoczi <=
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Jag Raman, 2022/03/29
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Stefan Hajnoczi, 2022/03/29
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Jag Raman, 2022/03/29
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Stefan Hajnoczi, 2022/03/30
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Peter Xu, 2022/03/30
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Stefan Hajnoczi, 2022/03/30
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Peter Xu, 2022/03/30
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Stefan Hajnoczi, 2022/03/31
- Re: [PATCH v7 12/17] vfio-user: IOMMU support for remote device, Peter Xu, 2022/03/31
[PATCH v7 13/17] vfio-user: handle DMA mappings, Jagannathan Raman, 2022/03/25