From: Nathan Chen <nathanc@nvidia.com> Open iommufd FD from libvirt backend without exposing these FDs to XML users, i.e. one per domain for /dev/iommu, and pass the FD to qemu command line. Set per-process memory accounting for iommufd instead of the default per-user memory accounting. Suggested-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Nathan Chen <nathanc@nvidia.com> --- src/qemu/qemu_command.c | 13 +++++++++++-- src/qemu/qemu_domain.c | 1 + src/qemu/qemu_domain.h | 2 ++ src/qemu/qemu_process.c | 43 +++++++++++++++++++++++++++++++++++++++++ 4 files changed, 57 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index db5712721e..7f96eb1a01 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -5349,9 +5349,13 @@ qemuBuildHostdevCommandLine(virCommand *cmd, static int qemuBuildIOMMUFDCommandLine(virCommand *cmd, - const virDomainDef *def) + const virDomainDef *def, + virDomainObj *vm) { size_t i; + qemuDomainObjPrivate *priv = vm->privateData; + g_autofree char *fdstr = g_strdup_printf("%d", priv->iommufd); + for (i = 0; i < def->nhostdevs; i++) { virDomainHostdevDef *hostdev = def->hostdevs[i]; @@ -5370,8 +5374,13 @@ qemuBuildIOMMUFDCommandLine(virCommand *cmd, if (subsys->u.pci.driver.iommufd != VIR_TRISTATE_BOOL_YES) continue; + virCommandPassFD(cmd, priv->iommufd, VIR_COMMAND_PASS_FD_CLOSE_PARENT); + + priv->iommufd = -1; + if (qemuMonitorCreateObjectProps(&props, "iommufd", "iommufd0", + "S:fd", fdstr, NULL) < 0) return -1; @@ -10997,7 +11006,7 @@ qemuBuildCommandLine(virDomainObj *vm, if (qemuBuildRedirdevCommandLine(cmd, def, qemuCaps) < 0) return NULL; - if (qemuBuildIOMMUFDCommandLine(cmd, def) < 0) + if (qemuBuildIOMMUFDCommandLine(cmd, def, vm) < 0) return NULL; if (qemuBuildHostdevCommandLine(cmd, def, qemuCaps) < 0) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 3366214677..8e1ebe7799 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -2042,6 +2042,7 @@ qemuDomainObjPrivateAlloc(void *opaque) priv->blockjobs = virHashNew(virObjectUnref); priv->fds = virHashNew(g_object_unref); + priv->iommufd = -1; priv->pidMonitored = -1; /* agent commands block by default, user can choose different behavior */ diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h index 88c8416aa4..3361e97315 100644 --- a/src/qemu/qemu_domain.h +++ b/src/qemu/qemu_domain.h @@ -264,6 +264,8 @@ struct _qemuDomainObjPrivate { /* named file descriptor groups associated with the VM */ GHashTable *fds; + int iommufd; + char *memoryBackingDir; }; diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 2841856454..c5b2a5fda8 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -104,6 +104,7 @@ #include "backup_conf.h" #include "storage_file_probe.h" #include "virpci.h" +#include "viriommufd.h" #include "logging/log_manager.h" #include "logging/log_protocol.h" @@ -7671,6 +7672,42 @@ qemuProcessPrepareHostBackendChardevHotplug(virDomainObj *vm, return 0; } +/** + * qemuProcessOpenIommuFd: + * @vm: domain object + * @iommuFd: returned file descriptor + * + * Opens /dev/iommu file descriptor for the VM. + * + * Returns: FD on success, -1 on failure + */ +static int +qemuProcessOpenIommuFd(virDomainObj *vm) +{ + int fd = -1; + + VIR_DEBUG("Opening IOMMU FD for domain %s", vm->def->name); + + if ((fd = open(VIR_IOMMU_DEV_PATH, O_RDWR | O_CLOEXEC)) < 0) { + if (errno == ENOENT) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("IOMMU FD support requires /dev/iommu device")); + } else { + virReportSystemError(errno, "%s", + _("cannot open /dev/iommu")); + } + return -1; + } + + if (virIOMMUFDSetRLimitMode(fd, true) < 0) { + VIR_FORCE_CLOSE(fd); + return -1; + } + + VIR_DEBUG("Opened IOMMU FD %d for domain %s", fd, vm->def->name); + return fd; +} + /** * qemuProcessOpenVfioDeviceFd: * @hostdev: host device definition @@ -7725,6 +7762,7 @@ qemuProcessOpenVfioDeviceFd(virDomainHostdevDef *hostdev) static int qemuProcessOpenVfioFds(virDomainObj *vm) { + qemuDomainObjPrivate *priv = vm->privateData; size_t i; /* Check if we have any hostdevs that need VFIO FDs */ @@ -7740,6 +7778,11 @@ qemuProcessOpenVfioFds(virDomainObj *vm) hostdevPriv->vfioDeviceFd = qemuProcessOpenVfioDeviceFd(hostdev); if (hostdevPriv->vfioDeviceFd == -1) return -1; + + /* Open IOMMU FD */ + priv->iommufd = qemuProcessOpenIommuFd(vm); + if (priv->iommufd == -1) + return -1; } } -- 2.43.0