Re: [PATCH v7 1/5] random: Blocking API for accessing nonblocking_pool

Am Mittwoch, 20. Mai 2015, 20:49:45 schrieb Stephan Mueller:

Hi Herbert,

quoted

This is just a convoluted way of doing an uninterruptible sleep.
Either make it uninterruptible or allow the function to return
an error.

Sorry, I overlooked the availability of wait_event. I was looking for it
initially, but missed it. I will fix it right away.

After checking a bit more, I see that an uninterruptible sleep cannot be 
canceled with cancel_work_sync. Therefore, replacing it with wait_event does 
not work.

Thus, go get an uninterruptible sleep which yet can be canceled seems to 
require wait_event_interruptible together with the check for ERESTARTSYS.

Nonetheless, I move the loop out to the DRBG code as requested.

-- 
Ciao
Stephan