vme: Fix a possible sleep-in-atomic bug in vme_tsi148

vme: Fix a possible sleep-in-atomic bug in vme_tsi148

The driver may sleep under a spinlock.
The function call path is:
tsi148_master_write \ tsi148_master_read (acquire the spinlock)
  vme_register_error_handler
    kmalloc(GFP_KERNEL) --> may sleep

To fix it, GFP_KERNEL is replaced with GFP_ATOMIC.

This bug is found by my static analysis tool(DSAC) and checked by my code review.

Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>