From owner-freebsd-hackers Sun Sep 21 10:49:06 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id KAA09892 for hackers-outgoing; Sun, 21 Sep 1997 10:49:06 -0700 (PDT) Received: from bitbox.follo.net (bitbox.follo.net [194.198.43.36]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id KAA09886 for ; Sun, 21 Sep 1997 10:49:02 -0700 (PDT) Received: (from eivind@localhost) by bitbox.follo.net (8.8.6/8.8.6) id TAA20869; Sun, 21 Sep 1997 19:48:59 +0200 (MET DST) Date: Sun, 21 Sep 1997 19:48:59 +0200 (MET DST) Message-Id: <199709211748.TAA20869@bitbox.follo.net> From: Eivind Eklund To: freebsd-hackers@freebsd.org Subject: kernel config, spl*() and interrupt masks Sender: owner-freebsd-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk With the switch from processor interrupt levels to masks, is it still useful to be able to specify the interrupt level (now mask) in the kernel config file? Doesn't this just give the user one more way to blow his foot off? For something e.g. config'ed as a tty in the config file per default, spltty() will be used in the top half of that driver anyway, and the user has no way to force it to still be blocked by this call. And while we're at it - I've had to have a device driver blocked by both spltty() and splimp() - is there a better/more correct way to do this than having the driver call INTRMASK() on both tty_imask and net_imask? This method seem like quite a nasty hack, relying on things it shouldn't - but the driver has to be in both masks, as it rely on structures in the generic parts of the kernel protected by both spltty() and splimp(). Eivind.