linux-next.git/include/uapi/linux/vduse.h, branch master

vduse: Add suspend

2026-06-11T14:50:57+00:00

Implement suspend operation for vduse devices, so vhost-vdpa will offer that backend feature and userspace can effectively suspend the device. This is a must before get virtqueue indexes (base) for live migration, since the device could modify them after userland gets them. This patch does not implement resume, so VMM resets the whole device to recover from a live migration failure. Resume optimization can be implemented on top of these patches, as other vDPA devices have done in the past. Includes clag fixes by Nathan Chancellor and squashed fixes by Eugenio Pérez . Signed-off-by: Eugenio Pérez Message-ID: <20260610083452.477759-1-eperezma@redhat.com> Signed-off-by: Nathan Chancellor Message-ID: <20260610-vduse_vq_kick-fix-guard-usage-v1-1-0ce02c08006e@kernel.org> Message-ID: <20260611133806.198402-2-eperezma@redhat.com> Message-ID: <20260611133806.198402-3-eperezma@redhat.com> Signed-off-by: Michael S. Tsirkin

vduse: add F_QUEUE_READY feature

2026-06-10T06:17:01+00:00

Add the VDUSE_F_QUEUE_READY feature flag, which tells the kernel to forward vq ready state changes to the VDUSE userland instance via a new VDUSE_SET_VQ_READY message. Without this feature, the VDUSE userland instance has no way to know when a virtqueue transitions to the ready state and when the driver starts using the virtqueue. If the userland instance doesn't support VDUSE_F_QUEUE_READY, the kernel silently ignores vq ready transitions as before. If the userland instance negotiates VDUSE_F_QUEUE_READY but fails to acknowledge a VDUSE_SET_VQ_READY message, the device is marked as broken. Signed-off-by: Eugenio Pérez Message-ID: <20260415085005.584291-4-eperezma@redhat.com> Signed-off-by: Michael S. Tsirkin

vduse: add VDUSE_GET_FEATURES ioctl

2026-06-10T06:17:01+00:00

Add an ioctl to allow VDUSE instances to query the available features supported by the kernel module. Signed-off-by: Eugenio Pérez Signed-off-by: Michael S. Tsirkin Message-ID: <20260310190759.1097506-3-eperezma@redhat.com>

vduse: avoid adding implicit padding

2026-02-09T17:21:32+00:00

The vduse_iova_range_v2 and vduse_iotlb_entry_v2 structures are both defined in a way that adds implicit padding and is incompatible between i386 and x86_64 userspace because of the different structure alignment requirements. Building the header with -Wpadded shows these new warnings: vduse.h:305:1: error: padding struct size to alignment boundary with 4 bytes [-Werror=padded] vduse.h:374:1: error: padding struct size to alignment boundary with 4 bytes [-Werror=padded] Change the amount of padding in these two structures to align them to 64 bit words and avoid those problems. Since the v1 vduse_iotlb_entry already has an inconsistent size, do not attempt to reuse the structure but rather list the members indiviudally, with a fixed amount of padding. Fixes: 079212f6877e ("vduse: add vq group asid support") Signed-off-by: Arnd Bergmann Signed-off-by: Michael S. Tsirkin Message-Id: <20260202224835.559538-1-arnd@kernel.org>

vduse: add vq group asid support

2026-01-28T20:32:18+00:00

Add support for assigning Address Space Identifiers (ASIDs) to each VQ group. This enables mapping each group into a distinct memory space. The vq group to ASID association is protected by a rwlock now. But the mutex domain_lock keeps protecting the domains of all ASIDs, as some operations like the one related with the bounce buffer size still requires to lock all the ASIDs. Acked-by: Jason Wang Signed-off-by: Eugenio Pérez Signed-off-by: Michael S. Tsirkin Message-Id: <20260119143306.1818855-12-eperezma@redhat.com>

vduse: add vq group support

2026-01-28T20:32:16+00:00

This allows separate the different virtqueues in groups that shares the same address space. Asking the VDUSE device for the groups of the vq at the beginning as they're needed for the DMA API. Allocating 3 vq groups as net is the device that need the most groups: * Dataplane (guest passthrough) * CVQ * Shadowed vrings. Future versions of the series can include dynamic allocation of the groups array so VDUSE can declare more groups. Acked-by: Jason Wang Reviewed-by: Xie Yongji Signed-off-by: Eugenio Pérez Signed-off-by: Michael S. Tsirkin Message-Id: <20260119143306.1818855-4-eperezma@redhat.com>

vduse: add v1 API definition

2026-01-28T20:32:16+00:00

This allows the kernel to detect whether the userspace VDUSE device supports the VQ group and ASID features. VDUSE devices that don't set the V1 API will not receive the new messages, and vdpa device will be created with only one vq group and asid. The next patches implement the new feature incrementally, only enabling the VDUSE device to set the V1 API version by the end of the series. Acked-by: Jason Wang Reviewed-by: Xie Yongji Signed-off-by: Eugenio Pérez Signed-off-by: Michael S. Tsirkin Message-Id: <20260119143306.1818855-3-eperezma@redhat.com>

uapi: vduse: fix typo in comment

2025-09-21T21:44:20+00:00

Fix a spelling mistake in vduse.h: "regsion" → "region" in the documentation for struct vduse_iova_info. No functional change. Signed-off-by: Ashwini Sahu Message-Id: <20250908095645.610336-1-ashwini@wisig.com> Signed-off-by: Michael S. Tsirkin

vduse: relicense under GPL-2.0 OR BSD-3-Clause

2025-01-08T11:37:13+00:00

Dual-license the vduse kernel header file to dual GPL-2.0 OR BSD-3-Clause license to make it possible to ship it with DPDK (under BSD-3-Clause) for older distros. Signed-off-by: Yongji Xie Message-Id: <20241119074238.38299-1-xieyongji@bytedance.com> Signed-off-by: Michael S. Tsirkin

vduse: Support querying information of IOVA regions

2022-08-11T08:26:08+00:00

This introduces a new ioctl: VDUSE_IOTLB_GET_INFO to support querying some information of IOVA regions. Now it can be used to query whether the IOVA region supports userspace memory registration. Signed-off-by: Xie Yongji Message-Id: <20220803045523.23851-6-xieyongji@bytedance.com> Signed-off-by: Michael S. Tsirkin Acked-by: Jason Wang