Subject: Re: Long delays during bootup in -current
To: None <tech-kern@NetBSD.org>
From: Valeriy E. Ushakov <firstname.lastname@example.org>
Date: 10/20/2006 03:13:51
[moved from port-macppc@ to tech-kern@]
On Fri, Oct 20, 2006 at 02:11:38 +0900, Izumi Tsutsui wrote:
> email@example.com wrote:
About a change to macppc that moves enabling of interrupts from
cpu_configure() to cpu_initclocks().
> > I'm not sure I like it. cpu_configure() is expected to turn the
> > interrupts on, and this patch breaks this assumption made in MI code.
> But in fact currently there is no code which assumes device interrupts
> are enabled between cpu_configure(9) and cpu_initclocks(9) in
> src/sys/kern/subr_autoconf.c and src/sys/kern/kern_clock.c.
> (only MD function around there is cpu_initclocks(9) itself)
> There are some architectures which can't disable only clock
> interrupt with the sane way (see mips CPU_INT5), but should we
> still enable all other interrupts in cpu_configure(9) only
> for the assumption?
If this is a problem for some of the arches than may be we should
reconsider the requirements of where MD code should enable the
firstname.lastname@example.org | Zu Grunde kommen
http://snark.ptc.spbu.ru/~uwe/ | Ist zu Grunde gehen