From: Saurabh Sengar <ssengar@xxxxxxxxxxxxxxxxxxx> [ Upstream commit 656c5ba50b7172a0ea25dc1b37606bd51d01fe8d ] In case of invalid sub channel, release cpu lock before returning. Fixes: a949e86c0d780 ("Drivers: hv: vmbus: Resolve race between init_vp_index() and CPU hotplug") Signed-off-by: Saurabh Sengar <ssengar@xxxxxxxxxxxxxxxxxxx> Reviewed-by: Michael Kelley <mikelley@xxxxxxxxxxxxx> Link: https://lore.kernel.org/r/1654794996-13244-1-git-send-email-ssengar@xxxxxxxxxxxxxxxxxxx Signed-off-by: Wei Liu <wei.liu@xxxxxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> --- drivers/hv/channel_mgmt.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/hv/channel_mgmt.c b/drivers/hv/channel_mgmt.c index 85a2142c9384..073ff9b6f535 100644 --- a/drivers/hv/channel_mgmt.c +++ b/drivers/hv/channel_mgmt.c @@ -637,6 +637,7 @@ static void vmbus_process_offer(struct vmbus_channel *newchannel) */ if (newchannel->offermsg.offer.sub_channel_index == 0) { mutex_unlock(&vmbus_connection.channel_mutex); + cpus_read_unlock(); /* * Don't call free_channel(), because newchannel->kobj * is not initialized yet. -- 2.35.1