From owner-freebsd-hackers  Wed Oct 25 05:03:21 1995
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id FAA10885
          for hackers-outgoing; Wed, 25 Oct 1995 05:03:21 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id FAA10877
          for <hackers@freebsd.org>; Wed, 25 Oct 1995 05:03:15 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id VAA00368; Wed, 25 Oct 1995 21:41:46 +1000
Date: Wed, 25 Oct 1995 21:41:46 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199510251141.VAA00368@godzilla.zeta.org.au>
To: hackers@freebsd.org, msmith@atrad.adelaide.edu.au
Subject: Re: Debugging interrupt service routines...
Sender: owner-hackers@freebsd.org
Precedence: bulk

>Can someone give me a definite 'yes' or 'no' to 
>"Is it safe to call printf() in the kernel from an interrupt handler?"

It isn't safe, but we do it anyway.  Suppose syscons is doing some
critical operation and you interrupt it and print something.  Who can
say what might happen?  Even ddb uses ordinary console output although
it has its own printf routine (it doesn't really need one - printf
itself is reentrant).  The serial console i/o routines are simpler and 
can't crash the system due to interference with ordinary serial i/o
AFAIK.

Bruce