kernel/linux.git/drivers/vdpa, branch v5.15.7

vdpa_sim: avoid putting an uninitialized iova_domain

2021-12-01T08:04:55+00:00

commit bb93ce4b150dde79f58e34103cbd1fe829796649 upstream. The system will crash if we put an uninitialized iova_domain, this could happen when an error occurs before initializing the iova_domain in vdpasim_create(). BUG: kernel NULL pointer dereference, address: 0000000000000000 ... RIP: 0010:__cpuhp_state_remove_instance+0x96/0x1c0 ... Call Trace: put_iova_domain+0x29/0x220 vdpasim_free+0xd1/0x120 [vdpa_sim] vdpa_release_dev+0x21/0x40 [vdpa] device_release+0x33/0x90 kobject_release+0x63/0x160 vdpasim_create+0x127/0x2a0 [vdpa_sim] vdpasim_net_dev_add+0x7d/0xfe [vdpa_sim_net] vdpa_nl_cmd_dev_add_set_doit+0xe1/0x1a0 [vdpa] genl_family_rcv_msg_doit+0x112/0x140 genl_rcv_msg+0xdf/0x1d0 ... So we must make sure the iova_domain is already initialized before put it. In addition, we may get the following warning in this case: WARNING: ... drivers/iommu/iova.c:344 iova_cache_put+0x58/0x70 So we must make sure the iova_cache_put() is invoked only if the iova_cache_get() is already invoked. Let's fix it together. Cc: stable@vger.kernel.org Fixes: 4080fc106750 ("vdpa_sim: use iova module to allocate IOVA addresses") Signed-off-by: Longpeng Acked-by: Jason Wang Reviewed-by: Stefano Garzarella Link: https://lore.kernel.org/r/20211124015215.119-1-longpeng2@huawei.com Signed-off-by: Michael S. Tsirkin Signed-off-by: Greg Kroah-Hartman

vdpa/mlx5: Fix clearing of VIRTIO_NET_F_MAC feature bit

2021-11-18T18:16:58+00:00

[ Upstream commit ef76eb83a17e803a66b64ac95b36ae48b3d17c22 ] Cited patch in the fixes tag clears the features bit during reset. mlx5 vdpa device feature bits are static decided by device capabilities. These feature bits (including VIRTIO_NET_F_MAC) are initialized during device addition time. Clearing features bit in reset callback cleared the VIRTIO_NET_F_MAC. Due to this, MAC address provided by the device is not honored. Fix it by not clearing the static feature bits during reset. Fixes: 0686082dbf7a ("vdpa: Add reset callback in vdpa_config_ops") Signed-off-by: Parav Pandit Reviewed-by: Eli Cohen Link: https://lore.kernel.org/r/20211026175519.87795-7-parav@nvidia.com Signed-off-by: Michael S. Tsirkin Acked-by: Jason Wang Signed-off-by: Sasha Levin

vduse: Fix race condition between resetting and irq injecting

2021-10-22T10:49:14+00:00

The interrupt might be triggered after a reset since there is no synchronization between resetting and irq injecting. And it might break something if the interrupt is delayed until a new round of device initialization. Fixes: c8a6153b6c59 ("vduse: Introduce VDUSE - vDPA Device in Userspace") Signed-off-by: Xie Yongji Link: https://lore.kernel.org/r/20210929083050.88-1-xieyongji@bytedance.com Signed-off-by: Michael S. Tsirkin

vduse: Disallow injecting interrupt before DRIVER_OK is set

2021-10-22T10:49:14+00:00

The interrupt callback should not be triggered before DRIVER_OK is set. Otherwise, it might break the virtio device driver. So let's add a check to avoid the unexpected behavior. Fixes: c8a6153b6c59 ("vduse: Introduce VDUSE - vDPA Device in Userspace") Signed-off-by: Xie Yongji Link: https://lore.kernel.org/r/20210923075722.98-1-xieyongji@bytedance.com Signed-off-by: Michael S. Tsirkin

vdpa/mlx5: Avoid executing set_vq_ready() if device is reset

2021-09-14T22:10:43+00:00

Avoid executing set_vq_ready() if the device has been reset. In such case, the features are cleared and cannot be used in conditional statements. Such reference happens is the function ctrl_vq_idx(). Fixes: 52893733f2c5 ("vdpa/mlx5: Add multiqueue support") Signed-off-by: Eli Cohen Link: https://lore.kernel.org/r/20210909063738.46970-1-elic@nvidia.com Signed-off-by: Michael S. Tsirkin Acked-by: Jason Wang

vdpa/mlx5: Clear ready indication for control VQ

2021-09-14T22:10:43+00:00

When clearing VQs ready indication for the data VQs, do the same for the control VQ. Fixes: 5262912ef3cf ("vdpa/mlx5: Add support for control VQ and MAC setting") Signed-off-by: Eli Cohen Link: https://lore.kernel.org/r/20210909063652.46880-1-elic@nvidia.com Signed-off-by: Michael S. Tsirkin Acked-by: Jason Wang

vduse: Cleanup the old kernel states after reset failure

2021-09-14T22:10:43+00:00

We should cleanup the old kernel states e.g. interrupt callback no matter whether the userspace handle the reset correctly or not since virtio-vdpa can't handle the reset failure now. Otherwise, the old state might be used after reset which might break something, e.g. the old interrupt callback might be triggered by userspace after reset, which can break the virtio device driver. Signed-off-by: Xie Yongji Link: https://lore.kernel.org/r/20210906142158.181-1-xieyongji@bytedance.com Signed-off-by: Michael S. Tsirkin Acked-by: Jason Wang

vduse: missing error code in vduse_init()

2021-09-14T22:10:42+00:00

This should return -ENOMEM if alloc_workqueue() fails. Currently it returns success. Fixes: b66219796563 ("vduse: Introduce VDUSE - vDPA Device in Userspace") Signed-off-by: Dan Carpenter Link: https://lore.kernel.org/r/20210907073223.GA18254@kili Signed-off-by: Michael S. Tsirkin Reviewed-by: Xie Yongji Acked-by: Jason Wang

Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost

2021-09-11T21:48:42+00:00

Pull virtio updates from Michael Tsirkin: - vduse driver ("vDPA Device in Userspace") supporting emulated virtio block devices - virtio-vsock support for end of record with SEQPACKET - vdpa: mac and mq support for ifcvf and mlx5 - vdpa: management netlink for ifcvf - virtio-i2c, gpio dt bindings - misc fixes and cleanups * tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (39 commits) Documentation: Add documentation for VDUSE vduse: Introduce VDUSE - vDPA Device in Userspace vduse: Implement an MMU-based software IOTLB vdpa: Support transferring virtual addressing during DMA mapping vdpa: factor out vhost_vdpa_pa_map() and vhost_vdpa_pa_unmap() vdpa: Add an opaque pointer for vdpa_config_ops.dma_map() vhost-iotlb: Add an opaque pointer for vhost IOTLB vhost-vdpa: Handle the failure of vdpa_reset() vdpa: Add reset callback in vdpa_config_ops vdpa: Fix some coding style issues file: Export receive_fd() to modules eventfd: Export eventfd_wake_count to modules iova: Export alloc_iova_fast() and free_iova_fast() virtio-blk: remove unneeded "likely" statements virtio-balloon: Use virtio_find_vqs() helper vdpa: Make use of PFN_PHYS/PFN_UP/PFN_DOWN helper macro vsock_test: update message bounds test for MSG_EOR af_vsock: rename variables in receive loop virtio/vsock: support MSG_EOR bit processing vhost/vsock: support MSG_EOR bit processing ...

Documentation: Add documentation for VDUSE

2021-09-06T11:20:58+00:00

VDUSE (vDPA Device in Userspace) is a framework to support implementing software-emulated vDPA devices in userspace. This document is intended to clarify the VDUSE design and usage. Signed-off-by: Xie Yongji Acked-by: Jason Wang Link: https://lore.kernel.org/r/20210831103634.33-14-xieyongji@bytedance.com Signed-off-by: Michael S. Tsirkin