A potential data race in drivers/scsi/mvumi.ko

[Pavel Andrianov]

Hi!

There is a potential data race in drivers/scsi/mvumi.ko.
Regard such case:

Thread 1					Thread 2
...
-> mvumi_reset_host_9500 -
   is called without any locking
  -> mvumi_wait_for_outstanding
    ->mvumi_start
      ->mvumi_check_handshake
        -> mvumi_handshake_event
          -> mvumi_handshake
            ->mvumi_hs_build_page

			   interrupt comes
					      -> mvumi_isr_handler
						-> mvumi_handshake
						  -> mvumi_hs_build_page

In this case the same data mhba->handshake_page is modified from two 
threads. Likely the first thread should acquire the same spinlock 
mhba->shost->host_lock as the second thread.

--
Pavel Andrianov
Linux Verification Center, ISPRAS
web: http://linuxtesting.org
e-mail: andrianov@xxxxxxxxx