ecryptfs shouldn't be running kernel_read() from the page fault handler anyway - it should pass the read off to a work queue and put the process to sleep until

ecryptfs shouldn't be running kernel_read() from the page fault handler anyway - it should pass the read off to a work queue and put the process to sleep until it completes.