vmxnet3: ensure that adapter is in proper state during force_close
authorNeil Horman <nhorman@tuxdriver.com>
Fri, 12 May 2017 16:00:01 +0000 (12:00 -0400)
committerDavid S. Miller <davem@davemloft.net>
Fri, 12 May 2017 16:23:52 +0000 (12:23 -0400)
commit1c4d5f51a812a82de97beee24f48ed05c65ebda5
tree168b055b2c850758eaaf6ab8862f87577bc86d98
parent42e6cae1b35b186650f5abc7b24c20ab6986c5a0
vmxnet3: ensure that adapter is in proper state during force_close

There are several paths in vmxnet3, where settings changes cause the
adapter to be brought down and back up (vmxnet3_set_ringparam among
them).  Should part of the reset operation fail, these paths call
vmxnet3_force_close, which enables all napi instances prior to calling
dev_close (with the expectation that vmxnet3_close will then properly
disable them again).  However, vmxnet3_force_close neglects to clear
VMXNET3_STATE_BIT_QUIESCED prior to calling dev_close.  As a result
vmxnet3_quiesce_dev (called from vmxnet3_close), returns early, and
leaves all the napi instances in a enabled state while the device itself
is closed.  If a device in this state is activated again, napi_enable
will be called on already enabled napi_instances, leading to a BUG halt.

The fix is to simply enausre that the QUIESCED bit is cleared in
vmxnet3_force_close to allow quesence to be completed properly on close.

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Shrikrishna Khare <skhare@vmware.com>
CC: "VMware, Inc." <pv-drivers@vmware.com>
CC: "David S. Miller" <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/vmxnet3/vmxnet3_drv.c