linux-toradex.git/drivers/vhost, branch v5.17-rc1

vdpa: Protect vdpa reset with cf_mutex

2022-01-14T23:50:54+00:00

Call reset using the wrapper function vdpa_reset() to make sure the
operation is serialized with cf_mutex.

This comes to protect from the following possible scenario:

vhost_vdpa_set_status() could call the reset op. Since the call is not
protected by cf_mutex, a netlink thread calling vdpa_dev_config_fill
could get passed the VIRTIO_CONFIG_S_FEATURES_OK check in
vdpa_dev_config_fill() and end up reporting wrong features.

Fixes: 5f6e85953d8f ("vdpa: Read device configuration only if FEATURES_OK")
Signed-off-by: Eli Cohen 
Link: https://lore.kernel.org/r/20220111183400.38418-3-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin 
Reviewed-by: Si-Wei Liu
Acked-by: Jason Wang

vdpa: Avoid taking cf_mutex lock on get status

2022-01-14T23:50:54+00:00

Avoid the wrapper holding cf_mutex since it is not protecting anything.
To avoid confusion and unnecessary overhead incurred by it, remove.

Fixes: f489f27bc0ab ("vdpa: Sync calls set/get config/status with cf_mutex")
Signed-off-by: Eli Cohen 
Link: https://lore.kernel.org/r/20220111183400.38418-2-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin 
Reviewed-by: Si-Wei Liu
Acked-by: Jason Wang

vdpa: Allow to configure max data virtqueues

2022-01-14T23:50:53+00:00

Add netlink support to configure the max virtqueue pairs for a device.
At least one pair is required. The maximum is dictated by the device.

Example:
$ vdpa dev add name vdpa-a mgmtdev auxiliary/mlx5_core.sf.1 max_vqp 4

Signed-off-by: Eli Cohen 
Link: https://lore.kernel.org/r/20220105114646.577224-6-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin

vdpa: Sync calls set/get config/status with cf_mutex

2022-01-14T23:50:53+00:00

Add wrappers to get/set status and protect these operations with
cf_mutex to serialize these operations with respect to get/set config
operations.

Signed-off-by: Eli Cohen 
Link: https://lore.kernel.org/r/20220105114646.577224-4-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin

vdpa: Provide interface to read driver features

2022-01-14T23:50:53+00:00

Provide an interface to read the negotiated features. This is needed
when building the netlink message in vdpa_dev_net_config_fill().

Also fix the implementation of vdpa_dev_net_config_fill() to use the
negotiated features instead of the device features.

To make APIs clearer, make the following name changes to struct
vdpa_config_ops so they better describe their operations:

get_features -> get_device_features
set_features -> set_driver_features

Finally, add get_driver_features to return the negotiated features and
add implementation to all the upstream drivers.

Acked-by: Jason Wang 
Signed-off-by: Eli Cohen 
Link: https://lore.kernel.org/r/20220105114646.577224-2-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin

vdpa: clean up get_config_size ret value handling

2022-01-14T23:50:53+00:00

The return type of get_config_size is size_t so it makes
sense to change the type of the variable holding its result.

That said, this already got taken care of (differently, and arguably
not as well) by commit 3ed21c1451a1 ("vdpa: check that offsets are
within bounds").

The added 'c->off > size' test in that commit will be done as an
unsigned comparison on 32-bit (safe due to not being signed).

On a 64-bit platform, it will be done as a signed comparison, but in
that case the comparison will be done in 64-bit, and 'c->off' being an
u32 it will be valid thanks to the extended range (ie both values will
be positive in 64 bits).

So this was a real bug, but it was already addressed and marked for stable.

Signed-off-by: Laura Abbott 
Reported-by: Luo Likang 
Signed-off-by: Michael S. Tsirkin

vhost/test: fix memory leak of vhost virtqueues

2022-01-14T23:50:53+00:00

We need free the vqs in .release(), which are allocated in .open().

Signed-off-by: Xianting Tian 
Link: https://lore.kernel.org/r/20211228030924.3468439-1-xianting.tian@linux.alibaba.com
Signed-off-by: Michael S. Tsirkin 
Acked-by: Jason Wang

vdpa: Avoid duplicate call to vp_vdpa get_status

2022-01-14T23:50:52+00:00

It has no sense to call get_status twice, since we already have a
variable for that.

Signed-off-by: Eugenio Pérez 
Link: https://lore.kernel.org/r/20211104195833.2089796-1-eperezma@redhat.com
Signed-off-by: Michael S. Tsirkin 
Acked-by: Jason Wang 
Reviewed-by: Stefano Garzarella

vdpa: check that offsets are within bounds

2021-12-08T19:53:15+00:00

In this function "c->off" is a u32 and "size" is a long.  On 64bit systems
if "c->off" is greater than "size" then "size - c->off" is a negative and
we always return -E2BIG.  But on 32bit systems the subtraction is type
promoted to a high positive u32 value and basically any "c->len" is
accepted.

Fixes: 4c8cf31885f6 ("vhost: introduce vDPA-based backend")
Reported-by: Xie Yongji 
Signed-off-by: Dan Carpenter 
Link: https://lore.kernel.org/r/20211208103337.GA4047@kili
Signed-off-by: Michael S. Tsirkin 
Cc: stable@vger.kernel.org

vhost-vdpa: clean irqs before reseting vdpa device

2021-11-25T00:00:28+00:00

Vdpa devices should be reset after unseting irqs of virtqueues, or we
will get errors when killing qemu process:

>> pi_update_irte: failed to update PI IRTE
>> irq bypass consumer (token 0000000065102a43) unregistration fails: -22

Signed-off-by: Wu Zongyong 
Link: https://lore.kernel.org/r/a2cb60cf73be9da5c4e6399242117d8818f975ae.1636946171.git.wuzongyong@linux.alibaba.com
Signed-off-by: Michael S. Tsirkin 
Acked-by: Jason Wang