From nobody Wed Nov 27 19:27:44 2024 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; dkim=pass header.i=@intel.com; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass(p=none dis=none) header.from=intel.com ARC-Seal: i=1; a=rsa-sha256; t=1697446189; cv=none; d=zohomail.com; s=zohoarc; b=kENhsyRwbp8pU9HuFZALTKlB5AYoh5OOiAl0Dtafix2yGgT2BGIMYjvbQbLizX9kgYudMNbPAf5LCWHMajxPcqP1/ABTzkbQPrYc0RZeALkE+VKInrgY3WwysHn0NEHJMiLpCY8OGN9XQ+6g0NjEFp9QRknJIeDmnnROwAv/OEo= ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com; s=zohoarc; t=1697446189; h=Content-Transfer-Encoding:Cc:Cc:Date:Date:From:From:In-Reply-To:List-Subscribe:List-Post:List-Id:List-Archive:List-Help:List-Unsubscribe:MIME-Version:Message-ID:References:Sender:Subject:Subject:To:To:Message-Id:Reply-To; bh=gqNd0OBlz3tVsdtY8RWrGjZucwAF4owl+bJSkp2QXNE=; b=ixmWC/pqwsW32AukBThtrdBBMSXFtpiSOylo/qKjgHjbU0SwcT+PCVs01MfBYQCHuO95uowWhhoX9FCtiYXR7eLVhIZuLdT3k/E2/viz9/mLJ/YxuzbDxaCmvoIdVj5qB4Q6UJvDun8I/IXzno4saq2N4THeUqqqk5xtAl865qI= ARC-Authentication-Results: i=1; mx.zohomail.com; dkim=pass header.i=@intel.com; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass header.from= (p=none dis=none) Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 1697446189020270.2520461777349; Mon, 16 Oct 2023 01:49:49 -0700 (PDT) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1qsJHa-0002vG-9I; Mon, 16 Oct 2023 04:49:06 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1qsJHU-0002ku-32 for qemu-devel@nongnu.org; Mon, 16 Oct 2023 04:48:56 -0400 Received: from mgamail.intel.com ([192.55.52.151]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1qsJHQ-0001HS-Ru for qemu-devel@nongnu.org; Mon, 16 Oct 2023 04:48:55 -0400 Received: from orsmga007.jf.intel.com ([10.7.209.58]) by fmsmga107.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 16 Oct 2023 01:48:41 -0700 Received: from duan-server-s2600bt.bj.intel.com ([10.240.192.147]) by orsmga007-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 16 Oct 2023 01:48:37 -0700 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1697446132; x=1728982132; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=+5euWQCg0SwFy0K+UjW/imvDiJ5SW8o4bCCUOEJuxCs=; b=fkVMJnBrfox8OxS3i6n6cUDsesopx2vgydBqavVumfEz8NlxtvRq2Y5Q PyNryr+uQBJqTCtDQkh37nuL4AsgcbBb3Ubh4tuFHH0jpxYBzkh0o33Od 7UX7CM/xYQ632Rq1y5yXGsiRa8PJTr/H7VqcXChmGWzvITYretLAzG7/u iF7EGLTJZRuCutBcqLzAdWr61UB7N9xbOI0SUrmVhYETeHGFvg8/yYhad 6urnLkm+3IcsbyCkdGjTwj8SZirKkGjYG4YPsVJoQaPg8SqtJ/8xnmdFq sQnLKw56ZwRLBbxgo1vBefHG5g6pFyrhzjjuaArLjzVd30bxU0kjhC8Ew A==; X-IronPort-AV: E=McAfee;i="6600,9927,10863"; a="365737687" X-IronPort-AV: E=Sophos;i="6.03,229,1694761200"; d="scan'208";a="365737687" X-ExtLoop1: 1 X-IronPort-AV: E=McAfee;i="6600,9927,10863"; a="749223014" X-IronPort-AV: E=Sophos;i="6.03,229,1694761200"; d="scan'208";a="749223014" From: Zhenzhong Duan To: qemu-devel@nongnu.org Cc: alex.williamson@redhat.com, clg@redhat.com, jgg@nvidia.com, nicolinc@nvidia.com, joao.m.martins@oracle.com, eric.auger@redhat.com, peterx@redhat.com, jasowang@redhat.com, kevin.tian@intel.com, yi.l.liu@intel.com, yi.y.sun@intel.com, chao.p.peng@intel.com, Zhenzhong Duan Subject: [PATCH v2 19/27] vfio/iommufd: Implement the iommufd backend Date: Mon, 16 Oct 2023 16:32:15 +0800 Message-Id: <20231016083223.1519410-20-zhenzhong.duan@intel.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20231016083223.1519410-1-zhenzhong.duan@intel.com> References: <20231016083223.1519410-1-zhenzhong.duan@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: pass client-ip=192.55.52.151; envelope-from=zhenzhong.duan@intel.com; helo=mgamail.intel.com X-Spam_score_int: -43 X-Spam_score: -4.4 X-Spam_bar: ---- X-Spam_report: (-4.4 / 5.0 requ) BAYES_00=-1.9, DKIMWL_WL_HIGH=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_MED=-2.3, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: qemu-devel-bounces+importer=patchew.org@nongnu.org X-ZohoMail-DKIM: pass (identity @intel.com) X-ZM-MESSAGEID: 1697446189286100002 Content-Type: text/plain; charset="utf-8" From: Yi Liu Add the iommufd backend. The IOMMUFD container class is implemented based on the new /dev/iommu user API. This backend obviously depends on CONFIG_IOMMUFD. So far, the iommufd backend doesn't support dirty page sync yet due to missing support in the host kernel. Co-authored-by: Eric Auger Signed-off-by: Eric Auger Signed-off-by: Yi Liu Signed-off-by: Zhenzhong Duan --- include/hw/vfio/vfio-common.h | 22 ++ include/hw/vfio/vfio-container-base.h | 3 + hw/vfio/common.c | 19 +- hw/vfio/iommufd.c | 535 ++++++++++++++++++++++++++ hw/vfio/meson.build | 3 + hw/vfio/trace-events | 12 + 6 files changed, 590 insertions(+), 4 deletions(-) create mode 100644 hw/vfio/iommufd.c diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index 9f2b86581b..e72f5962ee 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -83,6 +83,23 @@ typedef struct VFIOLegacyContainer { QLIST_HEAD(, VFIOGroup) group_list; } VFIOLegacyContainer; =20 +#ifdef CONFIG_IOMMUFD +typedef struct VFIOIOASHwpt { + uint32_t hwpt_id; + QLIST_HEAD(, VFIODevice) device_list; + QLIST_ENTRY(VFIOIOASHwpt) next; +} VFIOIOASHwpt; + +typedef struct IOMMUFDBackend IOMMUFDBackend; + +typedef struct VFIOIOMMUFDContainer { + VFIOContainer bcontainer; + IOMMUFDBackend *be; + uint32_t ioas_id; + QLIST_HEAD(, VFIOIOASHwpt) hwpt_list; +} VFIOIOMMUFDContainer; +#endif + typedef struct VFIODeviceOps VFIODeviceOps; =20 typedef struct VFIODevice { @@ -110,6 +127,11 @@ typedef struct VFIODevice { OnOffAuto pre_copy_dirty_page_tracking; bool dirty_pages_supported; bool dirty_tracking; +#ifdef CONFIG_IOMMUFD + int devid; + VFIOIOASHwpt *hwpt; + IOMMUFDBackend *iommufd; +#endif } VFIODevice; =20 struct VFIODeviceOps { diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-c= ontainer-base.h index 9a5971a00a..5345986993 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -114,6 +114,9 @@ void vfio_container_init(VFIOContainer *bcontainer, void vfio_container_destroy(VFIOContainer *bcontainer); =20 #define TYPE_VFIO_IOMMU_BACKEND_LEGACY_OPS "vfio-iommu-backend-legacy-ops" +#ifdef CONFIG_IOMMUFD +#define TYPE_VFIO_IOMMU_BACKEND_IOMMUFD_OPS "vfio-iommu-backend-iommufd-op= s" +#endif #define TYPE_VFIO_IOMMU_BACKEND_OPS "vfio-iommu-backend-ops" =20 DECLARE_CLASS_CHECKERS(VFIOIOMMUBackendOpsClass, diff --git a/hw/vfio/common.c b/hw/vfio/common.c index ee2ebf4be9..6901573c32 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -1520,10 +1520,13 @@ VFIOAddressSpace *vfio_get_address_space(AddressSpa= ce *as) =20 void vfio_put_address_space(VFIOAddressSpace *space) { - if (QLIST_EMPTY(&space->containers)) { - QLIST_REMOVE(space, list); - g_free(space); + if (!QLIST_EMPTY(&space->containers)) { + return; } + + QLIST_REMOVE(space, list); + g_free(space); + if (QLIST_EMPTY(&vfio_address_spaces)) { qemu_unregister_reset(vfio_reset_handler, NULL); } @@ -1558,8 +1561,16 @@ int vfio_attach_device(char *name, VFIODevice *vbase= dev, { const VFIOIOMMUBackendOpsClass *ops; =20 - ops =3D VFIO_IOMMU_BACKEND_OPS_CLASS( +#ifdef CONFIG_IOMMUFD + if (vbasedev->iommufd) { + ops =3D VFIO_IOMMU_BACKEND_OPS_CLASS( + object_class_by_name(TYPE_VFIO_IOMMU_BACKEND_IOMMUFD_OPS= )); + } else +#endif + { + ops =3D VFIO_IOMMU_BACKEND_OPS_CLASS( object_class_by_name(TYPE_VFIO_IOMMU_BACKEND_LEGACY_OPS)= ); + } if (!ops) { error_setg(errp, "VFIO IOMMU Backend not found!"); return -ENODEV; diff --git a/hw/vfio/iommufd.c b/hw/vfio/iommufd.c new file mode 100644 index 0000000000..ee8c4620b6 --- /dev/null +++ b/hw/vfio/iommufd.c @@ -0,0 +1,535 @@ +/* + * iommufd container backend + * + * Copyright (C) 2023 Intel Corporation. + * Copyright Red Hat, Inc. 2023 + * + * Authors: Yi Liu + * Eric Auger + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + + * You should have received a copy of the GNU General Public License along + * with this program; if not, see . + */ + +#include "qemu/osdep.h" +#include +#include +#include + +#include "hw/vfio/vfio-common.h" +#include "qemu/error-report.h" +#include "trace.h" +#include "qapi/error.h" +#include "sysemu/iommufd.h" +#include "hw/qdev-core.h" +#include "sysemu/reset.h" +#include "qemu/cutils.h" +#include "qemu/chardev_open.h" + +static int iommufd_map(VFIOContainer *bcontainer, hwaddr iova, + ram_addr_t size, void *vaddr, bool readonly) +{ + VFIOIOMMUFDContainer *container =3D + container_of(bcontainer, VFIOIOMMUFDContainer, bcontainer); + + return iommufd_backend_map_dma(container->be, + container->ioas_id, + iova, size, vaddr, readonly); +} + +static int iommufd_unmap(VFIOContainer *bcontainer, + hwaddr iova, ram_addr_t size, + IOMMUTLBEntry *iotlb) +{ + VFIOIOMMUFDContainer *container =3D + container_of(bcontainer, VFIOIOMMUFDContainer, bcontainer); + + /* TODO: Handle dma_unmap_bitmap with iotlb args (migration) */ + return iommufd_backend_unmap_dma(container->be, + container->ioas_id, iova, size); +} + +static void vfio_kvm_device_add_device(VFIODevice *vbasedev) +{ + Error *err =3D NULL; + + if (vfio_kvm_device_add_fd(vbasedev->fd, &err)) { + error_report_err(err); + } +} + +static void vfio_kvm_device_del_device(VFIODevice *vbasedev) +{ + Error *err =3D NULL; + + if (vfio_kvm_device_del_fd(vbasedev->fd, &err)) { + error_report_err(err); + } +} + +static int iommufd_connect_and_bind(VFIODevice *vbasedev, Error **errp) +{ + IOMMUFDBackend *iommufd =3D vbasedev->iommufd; + struct vfio_device_bind_iommufd bind =3D { + .argsz =3D sizeof(bind), + .flags =3D 0, + }; + int ret; + + ret =3D iommufd_backend_connect(iommufd, errp); + if (ret) { + return ret; + } + + /* + * Add device to kvm-vfio to be prepared for the tracking + * in KVM. Especially for some emulated devices, it requires + * to have kvm information in the device open. + */ + vfio_kvm_device_add_device(vbasedev); + + /* Bind device to iommufd */ + bind.iommufd =3D iommufd->fd; + ret =3D ioctl(vbasedev->fd, VFIO_DEVICE_BIND_IOMMUFD, &bind); + if (ret) { + error_setg_errno(errp, errno, "error bind device fd=3D%d to iommuf= d=3D%d", + vbasedev->fd, bind.iommufd); + goto err_bind; + } + + vbasedev->devid =3D bind.out_devid; + trace_vfio_iommufd_bind_device(bind.iommufd, vbasedev->name, + vbasedev->fd, vbasedev->devid); + return ret; +err_bind: + vfio_kvm_device_del_device(vbasedev); + iommufd_backend_disconnect(iommufd); + return ret; +} + +static void iommufd_unbind_and_disconnect(VFIODevice *vbasedev) +{ + /* Unbind is automatically conducted when device fd is closed */ + vfio_kvm_device_del_device(vbasedev); + iommufd_backend_disconnect(vbasedev->iommufd); +} + +static int vfio_get_devicefd(const char *sysfs_path, Error **errp) +{ + long int ret =3D -ENOTTY; + char *path, *vfio_dev_path =3D NULL, *vfio_path =3D NULL; + DIR *dir =3D NULL; + struct dirent *dent; + gchar *contents; + struct stat st; + gsize length; + int major, minor; + dev_t vfio_devt; + + path =3D g_strdup_printf("%s/vfio-dev", sysfs_path); + if (stat(path, &st) < 0) { + error_setg_errno(errp, errno, "no such host device"); + goto out_free_path; + } + + dir =3D opendir(path); + if (!dir) { + error_setg_errno(errp, errno, "couldn't open dirrectory %s", path); + goto out_free_path; + } + + while ((dent =3D readdir(dir))) { + if (!strncmp(dent->d_name, "vfio", 4)) { + vfio_dev_path =3D g_strdup_printf("%s/%s/dev", path, dent->d_n= ame); + break; + } + } + + if (!vfio_dev_path) { + error_setg(errp, "failed to find vfio-dev/vfioX/dev"); + goto out_close_dir; + } + + if (!g_file_get_contents(vfio_dev_path, &contents, &length, NULL)) { + error_setg(errp, "failed to load \"%s\"", vfio_dev_path); + goto out_free_dev_path; + } + + if (sscanf(contents, "%d:%d", &major, &minor) !=3D 2) { + error_setg(errp, "failed to get major:minor for \"%s\"", vfio_dev_= path); + goto out_free_dev_path; + } + g_free(contents); + vfio_devt =3D makedev(major, minor); + + vfio_path =3D g_strdup_printf("/dev/vfio/devices/%s", dent->d_name); + ret =3D open_cdev(vfio_path, vfio_devt); + if (ret < 0) { + error_setg(errp, "Failed to open %s", vfio_path); + } + + trace_vfio_iommufd_get_devicefd(vfio_path, ret); + g_free(vfio_path); + +out_free_dev_path: + g_free(vfio_dev_path); +out_close_dir: + closedir(dir); +out_free_path: + if (*errp) { + error_prepend(errp, VFIO_MSG_PREFIX, path); + } + g_free(path); + + return ret; +} + +static VFIOIOASHwpt *vfio_container_get_hwpt(VFIOIOMMUFDContainer *contain= er, + uint32_t hwpt_id) +{ + VFIOIOASHwpt *hwpt; + + QLIST_FOREACH(hwpt, &container->hwpt_list, next) { + if (hwpt->hwpt_id =3D=3D hwpt_id) { + return hwpt; + } + } + + hwpt =3D g_malloc0(sizeof(*hwpt)); + + hwpt->hwpt_id =3D hwpt_id; + QLIST_INIT(&hwpt->device_list); + QLIST_INSERT_HEAD(&container->hwpt_list, hwpt, next); + + return hwpt; +} + +static void vfio_container_put_hwpt(IOMMUFDBackend *be, VFIOIOASHwpt *hwpt) +{ + QLIST_REMOVE(hwpt, next); + iommufd_backend_free_id(be->fd, hwpt->hwpt_id); + g_free(hwpt); +} + +static int __vfio_device_attach_hwpt(VFIODevice *vbasedev, uint32_t hwpt_i= d, + Error **errp) +{ + struct vfio_device_attach_iommufd_pt attach_data =3D { + .argsz =3D sizeof(attach_data), + .flags =3D 0, + .pt_id =3D hwpt_id, + }; + int ret; + + ret =3D ioctl(vbasedev->fd, VFIO_DEVICE_ATTACH_IOMMUFD_PT, &attach_dat= a); + if (ret) { + error_setg_errno(errp, errno, + "[iommufd=3D%d] error attach %s (%d) to hwpt_id= =3D%d", + vbasedev->iommufd->fd, vbasedev->name, vbasedev->= fd, + hwpt_id); + } + return ret; +} + +static int __vfio_device_detach_hwpt(VFIODevice *vbasedev, Error **errp) +{ + struct vfio_device_detach_iommufd_pt detach_data =3D { + .argsz =3D sizeof(detach_data), + .flags =3D 0, + }; + int ret; + + ret =3D ioctl(vbasedev->fd, VFIO_DEVICE_DETACH_IOMMUFD_PT, &detach_dat= a); + if (ret) { + error_setg_errno(errp, errno, "detach %s from ioas failed", + vbasedev->name); + } + return ret; +} + +static int vfio_device_attach_container(VFIODevice *vbasedev, + VFIOIOMMUFDContainer *container, + Error **errp) +{ + int ret, iommufd =3D vbasedev->iommufd->fd; + VFIOIOASHwpt *hwpt; + uint32_t hwpt_id; + Error *err =3D NULL; + + /* try to attach to an existing hwpt in this container */ + QLIST_FOREACH(hwpt, &container->hwpt_list, next) { + ret =3D __vfio_device_attach_hwpt(vbasedev, hwpt->hwpt_id, &err); + if (ret) { + const char *msg =3D error_get_pretty(err); + + trace_vfio_iommufd_fail_attach_existing_hwpt(msg); + error_free(err); + err =3D NULL; + } else { + goto found_hwpt; + } + } + + ret =3D iommufd_backend_alloc_hwpt(iommufd, vbasedev->devid, + container->ioas_id, &hwpt_id); + + if (ret) { + error_setg_errno(errp, errno, "error alloc shadow hwpt"); + return ret; + } + + /* Attach device to an hwpt within iommufd */ + ret =3D __vfio_device_attach_hwpt(vbasedev, hwpt_id, errp); + if (ret) { + iommufd_backend_free_id(iommufd, hwpt_id); + return ret; + } + + hwpt =3D vfio_container_get_hwpt(container, hwpt_id); +found_hwpt: + QLIST_INSERT_HEAD(&hwpt->device_list, vbasedev, next); + vbasedev->hwpt =3D hwpt; + + trace_vfio_iommufd_attach_device(iommufd, vbasedev->name, vbasedev->fd, + container->ioas_id, hwpt->hwpt_id); + return ret; +} + +static void vfio_device_detach_container(VFIODevice *vbasedev, + VFIOIOMMUFDContainer *container) +{ + VFIOIOASHwpt *hwpt =3D vbasedev->hwpt; + Error *err =3D NULL; + int ret; + + ret =3D __vfio_device_detach_hwpt(vbasedev, &err); + if (ret) { + error_report_err(err); + } + + QLIST_REMOVE(vbasedev, next); + vbasedev->hwpt =3D NULL; + if (QLIST_EMPTY(&hwpt->device_list)) { + vfio_container_put_hwpt(vbasedev->iommufd, hwpt); + } + + trace_vfio_iommufd_detach_device(container->be->fd, vbasedev->name, + container->ioas_id); +} + +static void vfio_iommufd_container_destroy(VFIOIOMMUFDContainer *container) +{ + VFIOContainer *bcontainer =3D &container->bcontainer; + + if (!QLIST_EMPTY(&container->hwpt_list)) { + return; + } + memory_listener_unregister(&bcontainer->listener); + vfio_container_destroy(bcontainer); + iommufd_backend_put_ioas(container->be, container->ioas_id); + g_free(container); +} + +static int vfio_ram_block_discard_disable(bool state) +{ + /* + * We support coordinated discarding of RAM via the RamDiscardManager. + */ + return ram_block_uncoordinated_discard_disable(state); +} + +static int iommufd_attach_device(char *name, VFIODevice *vbasedev, + AddressSpace *as, Error **errp) +{ + VFIOIOMMUBackendOpsClass *ops =3D VFIO_IOMMU_BACKEND_OPS_CLASS( + object_class_by_name(TYPE_VFIO_IOMMU_BACKEND_IOMMUFD_OPS)); + VFIOContainer *bcontainer; + VFIOIOMMUFDContainer *container; + VFIOAddressSpace *space; + struct vfio_device_info dev_info =3D { .argsz =3D sizeof(dev_info) }; + int ret, devfd; + uint32_t ioas_id; + Error *err =3D NULL; + + devfd =3D vfio_get_devicefd(vbasedev->sysfsdev, errp); + if (devfd < 0) { + return devfd; + } + vbasedev->fd =3D devfd; + + ret =3D iommufd_connect_and_bind(vbasedev, errp); + if (ret) { + goto err_connect_bind; + } + + space =3D vfio_get_address_space(as); + + /* try to attach to an existing container in this space */ + QLIST_FOREACH(bcontainer, &space->containers, next) { + container =3D container_of(bcontainer, VFIOIOMMUFDContainer, bcont= ainer); + if (bcontainer->ops !=3D ops || vbasedev->iommufd !=3D container->= be) { + continue; + } + if (vfio_device_attach_container(vbasedev, container, &err)) { + const char *msg =3D error_get_pretty(err); + + trace_vfio_iommufd_fail_attach_existing_container(msg); + error_free(err); + err =3D NULL; + } else { + ret =3D vfio_ram_block_discard_disable(true); + if (ret) { + error_setg(errp, + "Cannot set discarding of RAM broken (%d)", = ret); + goto err_discard_disable; + } + goto found_container; + } + } + + /* Need to allocate a new dedicated container */ + ret =3D iommufd_backend_get_ioas(vbasedev->iommufd, &ioas_id); + if (ret < 0) { + error_setg_errno(errp, errno, "Failed to alloc ioas"); + goto err_get_ioas; + } + + trace_vfio_iommufd_alloc_ioas(vbasedev->iommufd->fd, ioas_id); + + container =3D g_malloc0(sizeof(*container)); + container->be =3D vbasedev->iommufd; + container->ioas_id =3D ioas_id; + QLIST_INIT(&container->hwpt_list); + + bcontainer =3D &container->bcontainer; + vfio_container_init(bcontainer, space, ops); + QLIST_INSERT_HEAD(&space->containers, bcontainer, next); + + ret =3D vfio_device_attach_container(vbasedev, container, errp); + if (ret) { + goto err_attach_container; + } + + ret =3D vfio_ram_block_discard_disable(true); + if (ret) { + goto err_discard_disable; + } + + /* + * TODO: for now iommufd BE is on par with vfio iommu type1, so it's + * fine to add the whole range as window. For SPAPR, below code + * should be updated. + */ + vfio_host_win_add(bcontainer, 0, (hwaddr)-1, 4096); + bcontainer->pgsizes =3D 4096; + + bcontainer->listener =3D vfio_memory_listener; + + memory_listener_register(&bcontainer->listener, bcontainer->space->as); + + if (bcontainer->error) { + ret =3D -1; + error_propagate_prepend(errp, bcontainer->error, + "memory listener initialization failed: "); + goto err_listener_register; + } + + bcontainer->initialized =3D true; + +found_container: + ret =3D ioctl(devfd, VFIO_DEVICE_GET_INFO, &dev_info); + if (ret) { + error_setg_errno(errp, errno, "error getting device info"); + goto err_listener_register; + } + + /* + * TODO: examine RAM_BLOCK_DISCARD stuff, should we do group level + * for discarding incompatibility check as well? + */ + if (vbasedev->ram_block_discard_allowed) { + vfio_ram_block_discard_disable(false); + } + + vbasedev->group =3D 0; + vbasedev->num_irqs =3D dev_info.num_irqs; + vbasedev->num_regions =3D dev_info.num_regions; + vbasedev->flags =3D dev_info.flags; + vbasedev->reset_works =3D !!(dev_info.flags & VFIO_DEVICE_FLAGS_RESET); + vbasedev->bcontainer =3D bcontainer; + QLIST_INSERT_HEAD(&bcontainer->device_list, vbasedev, container_next); + QLIST_INSERT_HEAD(&vfio_device_list, vbasedev, global_next); + + trace_vfio_iommufd_device_info(vbasedev->name, devfd, vbasedev->num_ir= qs, + vbasedev->num_regions, vbasedev->flags); + return 0; + +err_listener_register: + vfio_ram_block_discard_disable(false); +err_discard_disable: + vfio_device_detach_container(vbasedev, container); +err_attach_container: + vfio_iommufd_container_destroy(container); +err_get_ioas: + vfio_put_address_space(space); + iommufd_unbind_and_disconnect(vbasedev); +err_connect_bind: + close(vbasedev->fd); + return ret; +} + +static void iommufd_detach_device(VFIODevice *vbasedev) +{ + VFIOContainer *bcontainer =3D vbasedev->bcontainer; + VFIOIOMMUFDContainer *container; + VFIOAddressSpace *space =3D bcontainer->space; + + QLIST_REMOVE(vbasedev, global_next); + QLIST_REMOVE(vbasedev, container_next); + vbasedev->bcontainer =3D NULL; + + if (!vbasedev->ram_block_discard_allowed) { + vfio_ram_block_discard_disable(false); + } + + container =3D container_of(bcontainer, VFIOIOMMUFDContainer, bcontaine= r); + vfio_device_detach_container(vbasedev, container); + vfio_iommufd_container_destroy(container); + vfio_put_address_space(space); + + iommufd_unbind_and_disconnect(vbasedev); + close(vbasedev->fd); +} + +static void vfio_iommu_backend_iommufd_ops_class_init(ObjectClass *oc, + void *data) { + VFIOIOMMUBackendOpsClass *ops =3D VFIO_IOMMU_BACKEND_OPS_CLASS(oc); + + ops->dma_map =3D iommufd_map; + ops->dma_unmap =3D iommufd_unmap; + ops->attach_device =3D iommufd_attach_device; + ops->detach_device =3D iommufd_detach_device; +} + +static const TypeInfo vfio_iommu_backend_iommufd_ops_type =3D { + .name =3D TYPE_VFIO_IOMMU_BACKEND_IOMMUFD_OPS, + + .parent =3D TYPE_VFIO_IOMMU_BACKEND_OPS, + .class_init =3D vfio_iommu_backend_iommufd_ops_class_init, + .abstract =3D true, +}; +static void vfio_iommu_backend_iommufd_ops_register_types(void) +{ + type_register_static(&vfio_iommu_backend_iommufd_ops_type); +} +type_init(vfio_iommu_backend_iommufd_ops_register_types); diff --git a/hw/vfio/meson.build b/hw/vfio/meson.build index eb6ce6229d..9cae2c9e21 100644 --- a/hw/vfio/meson.build +++ b/hw/vfio/meson.build @@ -7,6 +7,9 @@ vfio_ss.add(files( 'spapr.c', 'migration.c', )) +if have_iommufd + vfio_ss.add(files('iommufd.c')) +endif vfio_ss.add(when: 'CONFIG_VFIO_PCI', if_true: files( 'display.c', 'pci-quirks.c', diff --git a/hw/vfio/trace-events b/hw/vfio/trace-events index 08a1f9dfa4..9b180cf77c 100644 --- a/hw/vfio/trace-events +++ b/hw/vfio/trace-events @@ -164,3 +164,15 @@ vfio_state_pending_estimate(const char *name, uint64_t= precopy, uint64_t postcop vfio_state_pending_exact(const char *name, uint64_t precopy, uint64_t post= copy, uint64_t stopcopy_size, uint64_t precopy_init_size, uint64_t precopy_= dirty_size) " (%s) precopy 0x%"PRIx64" postcopy 0x%"PRIx64" stopcopy size 0= x%"PRIx64" precopy initial size 0x%"PRIx64" precopy dirty size 0x%"PRIx64 vfio_vmstate_change(const char *name, int running, const char *reason, con= st char *dev_state) " (%s) running %d reason %s device state %s" vfio_vmstate_change_prepare(const char *name, int running, const char *rea= son, const char *dev_state) " (%s) running %d reason %s device state %s" + +#iommufd.c + +vfio_iommufd_get_devicefd(const char *dev, int devfd) " %s (fd=3D%d)" +vfio_iommufd_bind_device(int iommufd, const char *name, int devfd, int dev= id) " [iommufd=3D%d] Successfully bound device %s (fd=3D%d): output devid= =3D%d" +vfio_iommufd_fail_attach_existing_hwpt(const char *msg) " %s" +vfio_iommufd_attach_device(int iommufd, const char *name, int devfd, int i= oasid, int hwptid) " [iommufd=3D%d] Successfully attached device %s (%d) to= ioasid=3D%d: output hwptd=3D%d" +vfio_iommufd_detach_device(int iommufd, const char *name, int ioasid) " [i= ommufd=3D%d] Detached %s from ioasid=3D%d" +vfio_iommufd_alloc_ioas(int iommufd, int ioas_id) " [iommufd=3D%d] new IOM= MUFD container with ioasid=3D%d" +vfio_iommufd_device_info(char *name, int devfd, int num_irqs, int num_regi= ons, int flags) " %s (%d) num_irqs=3D%d num_regions=3D%d flags=3D%d" +vfio_iommufd_fail_attach_existing_container(const char *msg) " %s" +vfio_iommufd_container_reset(char *name) " Successfully reset %s" --=20 2.34.1