Re: kern/48119 (/dev/random kernel crash)

[Sergio Lopez <slp%sinrega.org@localhost>]

The following reply was made to PR kern/48119; it has been noted by GNATS.

From: Sergio Lopez <slp%sinrega.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: kern/48119 (/dev/random kernel crash)
Date: Wed, 4 Sep 2013 22:42:25 +0000

 On Tue, Sep 03, 2013 at 01:20:01PM +0000, Taylor R Campbell wrote:
 >  If someone has tested it and it works, the patch looks OK to me
 >  (remove the printf after testing before commit, of course), as a
 >  provisional measure until we split cprng_strong into two routines.
 >  When shoe-horning -1 into a size_t, I'd write an explicit cast in both
 >  cases to make it clear that we're doing something bogus, and perhaps
 >  add a comment to the effect that returning -1 is and provisional until
 >  we fix the API.
 
 You're right about the ugliness of stuffing -1 into a size_t. If this
 API change is expected to be applied in a near future, perhaps this fix
 can wait a little.
 
 Sergio.