]> git.ipfire.org Git - thirdparty/kernel/stable.git/commit
hv_netvsc: fix potential deadlock in netvsc_vf_setxdp()
authorSaurabh Sengar <ssengar@linux.microsoft.com>
Thu, 29 May 2025 10:18:30 +0000 (03:18 -0700)
committerJakub Kicinski <kuba@kernel.org>
Sat, 31 May 2025 02:31:25 +0000 (19:31 -0700)
commit3ec523304976648b45a3eef045e97d17122ff1b2
tree6817b670518ebbfc8f0b90d14e70b4cef63952a4
parentc1f4cb8a8d48db5c8396e1976c5b2e7922d944b9
hv_netvsc: fix potential deadlock in netvsc_vf_setxdp()

The MANA driver's probe registers netdevice via the following call chain:

mana_probe()
  register_netdev()
    register_netdevice()

register_netdevice() calls notifier callback for netvsc driver,
holding the netdev mutex via netdev_lock_ops().

Further this netvsc notifier callback end up attempting to acquire the
same lock again in dev_xdp_propagate() leading to deadlock.

netvsc_netdev_event()
  netvsc_vf_setxdp()
    dev_xdp_propagate()

This deadlock was not observed so far because net_shaper_ops was never set,
and thus the lock was effectively a no-op in this case. Fix this by using
netif_xdp_propagate() instead of dev_xdp_propagate() to avoid recursive
locking in this path.

And, since no deadlock is observed on the other path which is via
netvsc_probe, add the lock exclusivly for that path.

Also, clean up the unregistration path by removing the unnecessary call to
netvsc_vf_setxdp(), since unregister_netdevice_many_notify() already
performs this cleanup via dev_xdp_uninstall().

Fixes: 97246d6d21c2 ("net: hold netdev instance lock during ndo_bpf")
Cc: stable@vger.kernel.org
Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com>
Tested-by: Erni Sri Satya Vennela <ernis@linux.microsoft.com>
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Reviewed-by: Subbaraya Sundeep <sbhatta@marvell.com>
Link: https://patch.msgid.link/1748513910-23963-1-git-send-email-ssengar@linux.microsoft.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
drivers/net/hyperv/netvsc_bpf.c
drivers/net/hyperv/netvsc_drv.c
net/core/dev.c